Domínguez, Hicks, and Slabakova (Reference Dominguez, Hicks and Slabakova2019) offer some arguments related to what they regard as an irreplaceable term: incomplete acquisition. As we understand it, they believe and try to contextualize why abandoning this term entails the loss of a fine-grained construct that cannot be captured with alternative terminology. They recognize that incomplete acquisition can be (mis)taken as having a pejorative connotation by people who do not understand its provenance and the true intent—a pejorative intent that, we agree, its proponents did not wish to attribute to it. At the same time, Domínguez et al. maintain that the pluses outweigh any minuses on theoretical grounds. So much so, proverbial collateral damage, however regrettable, should not provide motivation to abandon its use. We agree that if a term is truly requisite for theoretical reasons, researchers should not bend to external pressure to replace it. If all alternatives fail to equally (or better) capture the same essence, it would be irresponsible to support disuse. At most, as suggested by Domínguez et al., we should be increasingly mindful of how a given term is used and, where possible, help to educate multiple audiences on what the true intent is. We take it that Domínguez et al.’s article is, thus, intended to simultaneously deconstruct previous argumentation against the term’s usefulness, offer a defense of its theoretical need as well as explain the true aim and provenance of the term to a larger audience.
With reference to the perceived latter goal, we must recognize that the overwhelmingly vast majority of people who hear, read, and/or are told (by way of the term) that a subset of bilinguals has, virtually by default, “incomplete” grammars have no way to become (re)acquainted with the true intent and historical provenance of the term. Heritage speakers (HS), their parents and communities, language policy makers, teachers, medical practitioners, and many more are decisively not the audience of Domínguez et al.’s article. We take for granted that not a single researcher has ever intended to be derogatory or consciously evaluative by using this term, very much including ourselves in older work. A major sticking point, however, has always revolved around the inherent trickling down process to multiple stakeholders who do not have access to anything more than the words/labels we use in writing. In our view then, the present article does very little to address one—we stress one—unfortunate dimension to the use of incomplete acquisition as a descriptive term. We must pause and ask, in light of Domínguez et al.’s acknowledgment that the term can have unintended derogatory connotations, whether more neutral terms such as differential acquisition truly lack descriptive and explanatory adequacy compared to incomplete acquisition. We think not. In fact, we find such alternative terms to be approaching greater accuracy.
As cited in Domínguez et al., Kupisch and Rothman (2018, p. 573) argued that a grammar should only be considered “incomplete” if it is rogue, that is, not universally compliant and/or is otherwise unable to fulfil the remit of a natural grammatical system. For example, the grammar of the famous linguistic isolate case, Genie, after years of exposure and training as a young adult could be fairly termed incomplete given its utter lack of functional elements (Curtiss, Reference Curtiss1977). Kupisch and Rothman’s comment should be interpreted as a statement against the appropriateness of the term itself in the domain of heritage language (HL) bilingualism because such grammars are regarded as “coherent, albeit in [their] own way” (Polinsky, Reference Polinsky2008, p. 2). Domínguez et al. further write:
It seems to us that this line of criticism is based on the mistaken assumption that terms such as deficient, attrited or incomplete may be taken to represent an evaluation of the linguistic abilities of a bilingual speaker as a person. Kupisch and Rothman question the theoretical appropriateness of incomplete acquisition as a hypothesis to explain heritage speakers’ acquisition, but their criticism seems to focus on the appropriateness of the term as well as its potential offensive connotations (Kupisch & Rothman, Reference Kupisch and Rothman2018, p. 567). In our view, this criticism seems to be about beliefs and ideology (i.e., what the bilingual speaker should or should not be able to do). Yet given that generative approaches to second language acquisition and bilingualism are only interested in the grammatical representations in a speaker’s mind, they do not (and indeed, cannot) make evaluations of speakers. (p. 242)
We wish to correct the record. Kupisch and Rothman do not believe that terms like attrited or deficient are or were meant to (by those who use or used them) evaluate bilinguals as people. Used in their proper context, each could be descriptively adequate, just as “incomplete” can be if it is describing a rogue grammar. Kupisch and Rothman’s claim is merely that because the different paths and outcomes of heritage grammars are coherent grammars in their own right, describing them as incomplete measured against a somewhat arbitrary standard (a particular monolingual one) is simply conceptually and theoretically wrong. How is different incomplete? For something to be incomplete there has to be an explicit or implicit point of comparison that is complete. Are monolingual speakers from distinct dialects incomplete relative to one another? Is a Spanish speaker that does not mark gender like most standard varieties do incomplete? If so, that might mean one considers native dominant Belizean Spanish speakers to be incomplete given the assignment and agreement patterns of this dialect (Balam, Reference Balam2014). Of course, they are not. Who decides what complete is whereby all others are therefore, on a continuum, incomplete? We could not agree more with Domínguez et al.’s claim that the criticism Kupisch and Rothman offer “seems to be about beliefs and ideology (i.e., what the bilingual speaker should or should not be able to do)” (p. 242). However, Domínguez et al.’s point should be flipped around back on to their line of argumentation. The concern Kupisch and Rothman have is that using arbitrary ideas of what complete speakers of language X should be able to do to validate their completeness has no room in HL studies and thus, in and of itself, constitutes an argument against the use of incomplete acquisition. This should be especially true for generative approaches, as highlighted by Domínguez et al., precisely because they are focused on grammatical representations in the speaker’s mind. Do Domínguez et al. intend to suggest that HS representations are incompletely represented in the mind, contradicting their claimed agreement with Polinsky (2008, p. 41) that they constitute coherent grammars? If not, as we suppose is the case, we fail to see how their conclusion—that HS grammars can be incomplete as opposed to differently represented—follows.
As the preceding quote reveals, Domínguez et al. focused almost exclusively on Kupisch and Rothman’s and colleagues’ preoccupation with the fact that incomplete acquisition is often taken as evaluative and predicated on an underlying comparative fallacy: that the standard monolingual variety typically described in the theoretical literature, from which monolinguals also differ, albeit often less than HSs, is justified as the measure against which (in)completeness can be evaluated (e.g., Bayram et al., Reference Bayram, Rothman, Iverson, Kupisch, Miller, Puig-Mayenco and Westergaard2017; Kupisch, Reference Kupisch2013; Kupisch & Rothman, Reference Kupisch and Rothman2018; Pascual y Cabo & Rothman, Reference Pascual y Cabo and Rothman2012). In doing so, Domínguez et al. reduce our arguments in these and related papers to an inaccurate, misleading subset. Beyond the social-prudence dimension, which does concern us, we have offered several theoretical reasons why incomplete acquisition is an inaccurate and thus inappropriate label. The authors have ignored these.
Although Domínguez et al. offer argumentation for why, in their view, incomplete acquisition is a preferable term for various reasons, some of which we shed doubt on in the following text, they do so in the absence of any meaningful acknowledgment or counterargumentation to the ones we have put forth. Reducing our arguments to a terminological impasse or emotional defense of HSs serves no one, especially because the authors claim to make their case on theoretical grounds. Doing so successfully entails not only positing defensible arguments for incomplete acquisition, but also countering all previous ones in opposition.
That HS grammars are typically distinct from monolinguals of comparable age and socioeconomic status, albeit on a vast continuum, is ubiquitously documented (see Montrul, Reference Montrul2016; Polinsky, Reference Polinsky2018; for review). No one denies this. We must also bear in mind that no researcher has yet properly documented HS development, longitudinally from childhood through young adulthood. While there are practical reasons for why this is so, it means that we do not (yet) have all the data points needed to fully understand the HS developmental process. And therefore it is crucial that we maintain a clear distinction between development and outcomes, a distinction we believe is left murky by Domínguez et al. at various points. It is also imperative to recognize from the outset that incomplete acquisition is not a theory, hypothesis, or a model. Domínguez et al. write that “Kupisch and Rothman question the theoretical appropriateness of incomplete acquisition as a hypothesis to explain heritage speakers’ acquisition” (p. 242). However, Kupisch and Rothman did not question the appropriateness of incomplete acquisition as a hypothesis, simply because incomplete acquisition is not a hypothesis. It does not offer falsifiable predictions at all. It is a descriptive term and as such only more useful than potential alternatives if it were more precise.
In what remains, we question Domínguez et al.’s arguments for keeping the term incomplete acquisition. Given space limitations we can only focus on two topics: (a) failure to distinguish development versus outcome properly, and (b) the concept of incompleteness versus reanalysis on the basis of input differences available at the baseline. To begin, we will also offer a brief commentary on their choice to frame incomplete acquisition as a generative construct because we believe this to be a mistake.
WHY LIMIT THIS TO A GENERATIVE DISCUSSION?
Like Domínguez et al., our work is informed by generative linguistics. Our views on HS grammars are no exception. While our research within HL bilingualism is considered generative as a result of who we are, the concepts and constructs in much of our work on HSs are no more generative than they could be considered usage based when used by a different set of authors. That is, terms such as incomplete acquisition are not generative. We take it that Domínguez et al. frame their discussion within generative approaches to bilingualism given the concepts they appeal to in support of their view—debates on accessibility to Universal Grammar, delays in applying formal linguistic principles in child acquisition, and notions such as truncation of hierarchical syntactic structure in child development. However, we see this decision as a missed opportunity.
Indeed, the term incomplete acquisition is omnipresent in all subfields of applied linguistics that work with HSs and the debate regarding its terminological adequacy is universally applicable—no pun intended—irrespective of paradigm. We understand that Domínguez et al. want to engage this discussion in terms that we, as generative acquisitionists, first and foremost understand. That said, given that generative HS research is a fraction of the overarching field of HL bilingualism, it seems that their discussion could have been equally good and potentially more effective if the larger field of applied linguistics had been addressed. Something we, as generative scholars, have enjoyed about working within HL bilingualism studies is that paradigmatic walls in this domain of bilingualism seem less important than the shared desire to understand the processes and challenges pertaining to HL acquisition. We submit that this is for the better. The discussion in Domínguez et al. is applicable outside the world of generative approaches and we hope that it sparks discussion across the whole of HL bilingualism research circles. It is precisely because this term, referring to the same set of observable phenomena, is shared across all paradigms that its defense cannot be reduced to generative conceptualizations.
DISTINGUISHING DEVELOPMENT VERSUS OUTCOME
It is puzzling to us that Domínguez et al. use terminology from the field of child developmental acquisition as evidence, using analogy, in favor of their argumentation for incomplete acquisition. Consider the following quote:
Our main point here is that research has shown that children’s grammars are often not adultlike in certain respects: their grammars lack certain adult grammatical features, they map input onto syntactic structure incorrectly, or they interpret sentences differently, for example. Since the target features or properties are available in the input, one must assume that children must have had an opportunity to acquire them but somehow failed to do so. In this respect, child grammars are characterised as being “deficient” or “incomplete” and while grammatical deficiency or incompleteness is typically only temporary, it persists throughout most of the acquisition process. This characterisation accounts for the names of some of the proposed explanations for the observed patterns of acquisition (our emphases), e.g., Truncation Model (Rizzi, 1993/Reference Rizzi1994); A-chain deficit (Borer & Wexler, Reference Borer, Wexler, Roeper and Williams1987, Reference Borer and Wexler1992; Principle B delay (Chien & Wexler, Reference Chien and Wexler1990; Grodzinsky & Reinhart, Reference Grodzinsky and Reinhart1993), missing subjects and parameter missetting (Hyams, Reference Hyams, de Villiers and Roeper2011), among many others. (p. 6)
To begin with, it is never wise to predicate, even part of, an argument on precedence alone. Therefore, it is of no consequence that in a related, yet different field similar terms are used to describe developmental stages where the growing grammar of a child is distinguished from its target (the adult input model). More importantly, however, in all the cases cited previously, the terms used describe a point in a developmental process where the grammar is being constructed, not an endpoint of ultimate attainment as in the case of incomplete acquisition for HSs. In other words, these theoretical models address the developmental problem of language acquisition (Hyams, Reference Hyams1988) that sits alongside the logical problem (Chomsky, Reference Chomsky1965). Indeed, acquisitionists must explain why child acquisition is quick but not instantaneous and why it passes through stages that reflect nontargetlike hypothesis testing. For example, the Truncation Model seeks to explain why children at early stages of development do not produce in adultlike manner elements that would require knowledge of functional categories above particular spots in the functional tree architecture, like negation and tense. Parameter missetting is also an attempt to explain systematicity in child grammar. The very point of these terms, however, is to argue that the child grammar, albeit different from its adult target model (and eventual endpoint), is a proper grammar with its own complete structure.
In light of the preceding, we do not agree that the labels used for the developing child language grammars cited by Domínguez et al. make reference to some type of incompleteness. Terms like those in the preceding quote refer to development. Yet incomplete acquisition, at present, can only refer to ultimate attainment. There are indeed many studies that examine HS development during childhood (see e.g., Silva Corválan, Reference Silva-Corvalán2014, as well as many others that have studied HS without using the label [see Kupisch, Reference Kupisch2013, p. 206 for discussion]). However, the vast majority of studies in HL acquisition that make claims about and reference to incomplete acquisition are synchronic snapshots of the end result of the acquisition process in (young) adulthood.
We are equally perplexed about Domínguez et al.’s appeal to precedence of potentially distasteful terminology in the domain of adult L2 acquisition. Here again, we must highlight that the concepts they cite are largely relevant in the sense of developing L2 interlanguage grammars. For others, it relates to accessibility to Universal Grammar in adulthood, which is irrelevant in the case of HSs because they are native child acquirers of their HL. The Failed Functional Features hypothesis (FFFH) refers to a potential critical period for accessing primitive features not instantiated in a speaker’s L1, not to the speaker’s grammar. Local and global impairment also refer to access to (sub)parts of Universal Grammar, not a learner’s grammar. Because the relevant authors in Domínguez et al. have argued against proposals such as FFFH and local impairment of L2 grammars in their own work, we are surprised to see them draw analogies with the term incomplete acquisition. Theories such as FFFH offer testable hypotheses that make falsifiable predictions, whereas incomplete acquisition is a term that serves as a descriptor of an end result. Thus, the terminological analogies drawn by Domínguez et al. are simply not at the same level, resulting in an (un)intentional strawman. If anything, the common theme revealed by the terms in child and adult L2 acquisition that Domínguez et al. appeal to underscores that, as a field, we have a long tradition of ignoring comparative fallacies (e.g., Bley-Vroman, Reference Bley-Vroman1983). We have an opportunity to reflect on this in general to avoid it in HL studies and invite others to do the same. We should definitively not, however, use precedence from distinct contexts to legitimize the continuation of insensitive and inaccurate terms or perpetuate the “monolingual-centric” comparative approach in HL bilingualism.
INCOMPLETENESS VERSUS REANALYSIS
We are especially concerned with the fact that reanalysis, according to Domínguez et al., would fall under the umbrella of incompleteness. Domínguez et al. write:
[G]rammar of a heritage speaker could be incomplete due to the acquisition conditions which are specific to that speaker, regardless of whether it resembles the grammars of their parents or not. Our assumption is that incomplete acquisition can arise when input conditions change during the course of acquisition (and, crucially, before a mature and stable grammar has fully developed) and that change impacts upon the overall pattern of acquisition and resulting grammar. (p. 242)
We do not disagree that this may happen. In essence, this is what Putnam and Sánchez (Reference Putnam and Sánchez2013) predict in their model of reanalysis of the heritage grammar in relation to input/intake over time. The preceding quote is also reminiscent of the arguments put forth by Pascual y Cabo and Rothman (Reference Pascual y Cabo and Rothman2012), supported by strong empirical data in Pascual y Cabo (Reference Pascual y Cabo2013, Reference Pascual y Cabo2018). The latter has shown how input differences laden with previous generational attrition that HSs are confronted with cause changes to the ensuing HL systems at the level of representation (whether matching the L1 attrition of the parents or regularizing innovations in their input to new grammatical representations distinct from their parents). If reanalysis results in systems that are different from a previous generation or other native speakers that grow up in monolingual environments, what about the innovative system speaks to its incompleteness? Can it not be labeled better to reflect the reality of a distinct environment peppered with differences to the typical monolingual one?
In the case of creole genesis (e.g., Lefebvre, Reference Lefebvre1998) or theories postulating a link between child acquisition and diachronic change (e.g., Lightfoot & Westergaard, Reference Lightfoot and Westergaard2007), reanalysis has always been considered a naturally occurring phenomenon. In the case of HSs, who have an innate push to regularize ensuring the end result is a comprehensive grammar, reanalysis takes place to accommodate and compensate for input quality/quantity issues. Therefore, in our view, reanalysis and incompleteness cannot go hand in hand. A reanalyzed grammar is different from the standard target, yes. But it is not an incomplete version of it.
FINAL STATEMENT
Science is nothing without opportunity for discussion and debate. While there is a matter of terminology at stake here, this is not a mere terminological debate. We thank Domínguez et al. for highlighting the need for this discussion, which in turn provided opportunity to clarify points we have raised in previous papers. We share the very same goals even if we do not agree on all the facts. As such, we look forward to the immediate future where we move beyond terminological debates and work toward the common goals we share in a way that projects empathy to our object of study and accuracy to the concepts we seek to understand.
You have
Access
Domínguez, Hicks, and Slabakova (Reference Dominguez, Hicks and Slabakova2019) offer some arguments related to what they regard as an irreplaceable term: incomplete acquisition. As we understand it, they believe and try to contextualize why abandoning this term entails the loss of a fine-grained construct that cannot be captured with alternative terminology. They recognize that incomplete acquisition can be (mis)taken as having a pejorative connotation by people who do not understand its provenance and the true intent—a pejorative intent that, we agree, its proponents did not wish to attribute to it. At the same time, Domínguez et al. maintain that the pluses outweigh any minuses on theoretical grounds. So much so, proverbial collateral damage, however regrettable, should not provide motivation to abandon its use. We agree that if a term is truly requisite for theoretical reasons, researchers should not bend to external pressure to replace it. If all alternatives fail to equally (or better) capture the same essence, it would be irresponsible to support disuse. At most, as suggested by Domínguez et al., we should be increasingly mindful of how a given term is used and, where possible, help to educate multiple audiences on what the true intent is. We take it that Domínguez et al.’s article is, thus, intended to simultaneously deconstruct previous argumentation against the term’s usefulness, offer a defense of its theoretical need as well as explain the true aim and provenance of the term to a larger audience.
With reference to the perceived latter goal, we must recognize that the overwhelmingly vast majority of people who hear, read, and/or are told (by way of the term) that a subset of bilinguals has, virtually by default, “incomplete” grammars have no way to become (re)acquainted with the true intent and historical provenance of the term. Heritage speakers (HS), their parents and communities, language policy makers, teachers, medical practitioners, and many more are decisively not the audience of Domínguez et al.’s article. We take for granted that not a single researcher has ever intended to be derogatory or consciously evaluative by using this term, very much including ourselves in older work. A major sticking point, however, has always revolved around the inherent trickling down process to multiple stakeholders who do not have access to anything more than the words/labels we use in writing. In our view then, the present article does very little to address one—we stress one—unfortunate dimension to the use of incomplete acquisition as a descriptive term. We must pause and ask, in light of Domínguez et al.’s acknowledgment that the term can have unintended derogatory connotations, whether more neutral terms such as differential acquisition truly lack descriptive and explanatory adequacy compared to incomplete acquisition. We think not. In fact, we find such alternative terms to be approaching greater accuracy.
As cited in Domínguez et al., Kupisch and Rothman (2018, p. 573) argued that a grammar should only be considered “incomplete” if it is rogue, that is, not universally compliant and/or is otherwise unable to fulfil the remit of a natural grammatical system. For example, the grammar of the famous linguistic isolate case, Genie, after years of exposure and training as a young adult could be fairly termed incomplete given its utter lack of functional elements (Curtiss, Reference Curtiss1977). Kupisch and Rothman’s comment should be interpreted as a statement against the appropriateness of the term itself in the domain of heritage language (HL) bilingualism because such grammars are regarded as “coherent, albeit in [their] own way” (Polinsky, Reference Polinsky2008, p. 2). Domínguez et al. further write:
It seems to us that this line of criticism is based on the mistaken assumption that terms such as deficient, attrited or incomplete may be taken to represent an evaluation of the linguistic abilities of a bilingual speaker as a person. Kupisch and Rothman question the theoretical appropriateness of incomplete acquisition as a hypothesis to explain heritage speakers’ acquisition, but their criticism seems to focus on the appropriateness of the term as well as its potential offensive connotations (Kupisch & Rothman, Reference Kupisch and Rothman2018, p. 567). In our view, this criticism seems to be about beliefs and ideology (i.e., what the bilingual speaker should or should not be able to do). Yet given that generative approaches to second language acquisition and bilingualism are only interested in the grammatical representations in a speaker’s mind, they do not (and indeed, cannot) make evaluations of speakers. (p. 242)
We wish to correct the record. Kupisch and Rothman do not believe that terms like attrited or deficient are or were meant to (by those who use or used them) evaluate bilinguals as people. Used in their proper context, each could be descriptively adequate, just as “incomplete” can be if it is describing a rogue grammar. Kupisch and Rothman’s claim is merely that because the different paths and outcomes of heritage grammars are coherent grammars in their own right, describing them as incomplete measured against a somewhat arbitrary standard (a particular monolingual one) is simply conceptually and theoretically wrong. How is different incomplete? For something to be incomplete there has to be an explicit or implicit point of comparison that is complete. Are monolingual speakers from distinct dialects incomplete relative to one another? Is a Spanish speaker that does not mark gender like most standard varieties do incomplete? If so, that might mean one considers native dominant Belizean Spanish speakers to be incomplete given the assignment and agreement patterns of this dialect (Balam, Reference Balam2014). Of course, they are not. Who decides what complete is whereby all others are therefore, on a continuum, incomplete? We could not agree more with Domínguez et al.’s claim that the criticism Kupisch and Rothman offer “seems to be about beliefs and ideology (i.e., what the bilingual speaker should or should not be able to do)” (p. 242). However, Domínguez et al.’s point should be flipped around back on to their line of argumentation. The concern Kupisch and Rothman have is that using arbitrary ideas of what complete speakers of language X should be able to do to validate their completeness has no room in HL studies and thus, in and of itself, constitutes an argument against the use of incomplete acquisition. This should be especially true for generative approaches, as highlighted by Domínguez et al., precisely because they are focused on grammatical representations in the speaker’s mind. Do Domínguez et al. intend to suggest that HS representations are incompletely represented in the mind, contradicting their claimed agreement with Polinsky (2008, p. 41) that they constitute coherent grammars? If not, as we suppose is the case, we fail to see how their conclusion—that HS grammars can be incomplete as opposed to differently represented—follows.
As the preceding quote reveals, Domínguez et al. focused almost exclusively on Kupisch and Rothman’s and colleagues’ preoccupation with the fact that incomplete acquisition is often taken as evaluative and predicated on an underlying comparative fallacy: that the standard monolingual variety typically described in the theoretical literature, from which monolinguals also differ, albeit often less than HSs, is justified as the measure against which (in)completeness can be evaluated (e.g., Bayram et al., Reference Bayram, Rothman, Iverson, Kupisch, Miller, Puig-Mayenco and Westergaard2017; Kupisch, Reference Kupisch2013; Kupisch & Rothman, Reference Kupisch and Rothman2018; Pascual y Cabo & Rothman, Reference Pascual y Cabo and Rothman2012). In doing so, Domínguez et al. reduce our arguments in these and related papers to an inaccurate, misleading subset. Beyond the social-prudence dimension, which does concern us, we have offered several theoretical reasons why incomplete acquisition is an inaccurate and thus inappropriate label. The authors have ignored these.
Although Domínguez et al. offer argumentation for why, in their view, incomplete acquisition is a preferable term for various reasons, some of which we shed doubt on in the following text, they do so in the absence of any meaningful acknowledgment or counterargumentation to the ones we have put forth. Reducing our arguments to a terminological impasse or emotional defense of HSs serves no one, especially because the authors claim to make their case on theoretical grounds. Doing so successfully entails not only positing defensible arguments for incomplete acquisition, but also countering all previous ones in opposition.
That HS grammars are typically distinct from monolinguals of comparable age and socioeconomic status, albeit on a vast continuum, is ubiquitously documented (see Montrul, Reference Montrul2016; Polinsky, Reference Polinsky2018; for review). No one denies this. We must also bear in mind that no researcher has yet properly documented HS development, longitudinally from childhood through young adulthood. While there are practical reasons for why this is so, it means that we do not (yet) have all the data points needed to fully understand the HS developmental process. And therefore it is crucial that we maintain a clear distinction between development and outcomes, a distinction we believe is left murky by Domínguez et al. at various points. It is also imperative to recognize from the outset that incomplete acquisition is not a theory, hypothesis, or a model. Domínguez et al. write that “Kupisch and Rothman question the theoretical appropriateness of incomplete acquisition as a hypothesis to explain heritage speakers’ acquisition” (p. 242). However, Kupisch and Rothman did not question the appropriateness of incomplete acquisition as a hypothesis, simply because incomplete acquisition is not a hypothesis. It does not offer falsifiable predictions at all. It is a descriptive term and as such only more useful than potential alternatives if it were more precise.
In what remains, we question Domínguez et al.’s arguments for keeping the term incomplete acquisition. Given space limitations we can only focus on two topics: (a) failure to distinguish development versus outcome properly, and (b) the concept of incompleteness versus reanalysis on the basis of input differences available at the baseline. To begin, we will also offer a brief commentary on their choice to frame incomplete acquisition as a generative construct because we believe this to be a mistake.
WHY LIMIT THIS TO A GENERATIVE DISCUSSION?
Like Domínguez et al., our work is informed by generative linguistics. Our views on HS grammars are no exception. While our research within HL bilingualism is considered generative as a result of who we are, the concepts and constructs in much of our work on HSs are no more generative than they could be considered usage based when used by a different set of authors. That is, terms such as incomplete acquisition are not generative. We take it that Domínguez et al. frame their discussion within generative approaches to bilingualism given the concepts they appeal to in support of their view—debates on accessibility to Universal Grammar, delays in applying formal linguistic principles in child acquisition, and notions such as truncation of hierarchical syntactic structure in child development. However, we see this decision as a missed opportunity.
Indeed, the term incomplete acquisition is omnipresent in all subfields of applied linguistics that work with HSs and the debate regarding its terminological adequacy is universally applicable—no pun intended—irrespective of paradigm. We understand that Domínguez et al. want to engage this discussion in terms that we, as generative acquisitionists, first and foremost understand. That said, given that generative HS research is a fraction of the overarching field of HL bilingualism, it seems that their discussion could have been equally good and potentially more effective if the larger field of applied linguistics had been addressed. Something we, as generative scholars, have enjoyed about working within HL bilingualism studies is that paradigmatic walls in this domain of bilingualism seem less important than the shared desire to understand the processes and challenges pertaining to HL acquisition. We submit that this is for the better. The discussion in Domínguez et al. is applicable outside the world of generative approaches and we hope that it sparks discussion across the whole of HL bilingualism research circles. It is precisely because this term, referring to the same set of observable phenomena, is shared across all paradigms that its defense cannot be reduced to generative conceptualizations.
DISTINGUISHING DEVELOPMENT VERSUS OUTCOME
It is puzzling to us that Domínguez et al. use terminology from the field of child developmental acquisition as evidence, using analogy, in favor of their argumentation for incomplete acquisition. Consider the following quote:
Our main point here is that research has shown that children’s grammars are often not adultlike in certain respects: their grammars lack certain adult grammatical features, they map input onto syntactic structure incorrectly, or they interpret sentences differently, for example. Since the target features or properties are available in the input, one must assume that children must have had an opportunity to acquire them but somehow failed to do so. In this respect, child grammars are characterised as being “deficient” or “incomplete” and while grammatical deficiency or incompleteness is typically only temporary, it persists throughout most of the acquisition process. This characterisation accounts for the names of some of the proposed explanations for the observed patterns of acquisition (our emphases), e.g., Truncation Model (Rizzi, 1993/Reference Rizzi1994); A-chain deficit (Borer & Wexler, Reference Borer, Wexler, Roeper and Williams1987, Reference Borer and Wexler1992; Principle B delay (Chien & Wexler, Reference Chien and Wexler1990; Grodzinsky & Reinhart, Reference Grodzinsky and Reinhart1993), missing subjects and parameter missetting (Hyams, Reference Hyams, de Villiers and Roeper2011), among many others. (p. 6)
To begin with, it is never wise to predicate, even part of, an argument on precedence alone. Therefore, it is of no consequence that in a related, yet different field similar terms are used to describe developmental stages where the growing grammar of a child is distinguished from its target (the adult input model). More importantly, however, in all the cases cited previously, the terms used describe a point in a developmental process where the grammar is being constructed, not an endpoint of ultimate attainment as in the case of incomplete acquisition for HSs. In other words, these theoretical models address the developmental problem of language acquisition (Hyams, Reference Hyams1988) that sits alongside the logical problem (Chomsky, Reference Chomsky1965). Indeed, acquisitionists must explain why child acquisition is quick but not instantaneous and why it passes through stages that reflect nontargetlike hypothesis testing. For example, the Truncation Model seeks to explain why children at early stages of development do not produce in adultlike manner elements that would require knowledge of functional categories above particular spots in the functional tree architecture, like negation and tense. Parameter missetting is also an attempt to explain systematicity in child grammar. The very point of these terms, however, is to argue that the child grammar, albeit different from its adult target model (and eventual endpoint), is a proper grammar with its own complete structure.
In light of the preceding, we do not agree that the labels used for the developing child language grammars cited by Domínguez et al. make reference to some type of incompleteness. Terms like those in the preceding quote refer to development. Yet incomplete acquisition, at present, can only refer to ultimate attainment. There are indeed many studies that examine HS development during childhood (see e.g., Silva Corválan, Reference Silva-Corvalán2014, as well as many others that have studied HS without using the label [see Kupisch, Reference Kupisch2013, p. 206 for discussion]). However, the vast majority of studies in HL acquisition that make claims about and reference to incomplete acquisition are synchronic snapshots of the end result of the acquisition process in (young) adulthood.
We are equally perplexed about Domínguez et al.’s appeal to precedence of potentially distasteful terminology in the domain of adult L2 acquisition. Here again, we must highlight that the concepts they cite are largely relevant in the sense of developing L2 interlanguage grammars. For others, it relates to accessibility to Universal Grammar in adulthood, which is irrelevant in the case of HSs because they are native child acquirers of their HL. The Failed Functional Features hypothesis (FFFH) refers to a potential critical period for accessing primitive features not instantiated in a speaker’s L1, not to the speaker’s grammar. Local and global impairment also refer to access to (sub)parts of Universal Grammar, not a learner’s grammar. Because the relevant authors in Domínguez et al. have argued against proposals such as FFFH and local impairment of L2 grammars in their own work, we are surprised to see them draw analogies with the term incomplete acquisition. Theories such as FFFH offer testable hypotheses that make falsifiable predictions, whereas incomplete acquisition is a term that serves as a descriptor of an end result. Thus, the terminological analogies drawn by Domínguez et al. are simply not at the same level, resulting in an (un)intentional strawman. If anything, the common theme revealed by the terms in child and adult L2 acquisition that Domínguez et al. appeal to underscores that, as a field, we have a long tradition of ignoring comparative fallacies (e.g., Bley-Vroman, Reference Bley-Vroman1983). We have an opportunity to reflect on this in general to avoid it in HL studies and invite others to do the same. We should definitively not, however, use precedence from distinct contexts to legitimize the continuation of insensitive and inaccurate terms or perpetuate the “monolingual-centric” comparative approach in HL bilingualism.
INCOMPLETENESS VERSUS REANALYSIS
We are especially concerned with the fact that reanalysis, according to Domínguez et al., would fall under the umbrella of incompleteness. Domínguez et al. write:
[G]rammar of a heritage speaker could be incomplete due to the acquisition conditions which are specific to that speaker, regardless of whether it resembles the grammars of their parents or not. Our assumption is that incomplete acquisition can arise when input conditions change during the course of acquisition (and, crucially, before a mature and stable grammar has fully developed) and that change impacts upon the overall pattern of acquisition and resulting grammar. (p. 242)
We do not disagree that this may happen. In essence, this is what Putnam and Sánchez (Reference Putnam and Sánchez2013) predict in their model of reanalysis of the heritage grammar in relation to input/intake over time. The preceding quote is also reminiscent of the arguments put forth by Pascual y Cabo and Rothman (Reference Pascual y Cabo and Rothman2012), supported by strong empirical data in Pascual y Cabo (Reference Pascual y Cabo2013, Reference Pascual y Cabo2018). The latter has shown how input differences laden with previous generational attrition that HSs are confronted with cause changes to the ensuing HL systems at the level of representation (whether matching the L1 attrition of the parents or regularizing innovations in their input to new grammatical representations distinct from their parents). If reanalysis results in systems that are different from a previous generation or other native speakers that grow up in monolingual environments, what about the innovative system speaks to its incompleteness? Can it not be labeled better to reflect the reality of a distinct environment peppered with differences to the typical monolingual one?
In the case of creole genesis (e.g., Lefebvre, Reference Lefebvre1998) or theories postulating a link between child acquisition and diachronic change (e.g., Lightfoot & Westergaard, Reference Lightfoot and Westergaard2007), reanalysis has always been considered a naturally occurring phenomenon. In the case of HSs, who have an innate push to regularize ensuring the end result is a comprehensive grammar, reanalysis takes place to accommodate and compensate for input quality/quantity issues. Therefore, in our view, reanalysis and incompleteness cannot go hand in hand. A reanalyzed grammar is different from the standard target, yes. But it is not an incomplete version of it.
FINAL STATEMENT
Science is nothing without opportunity for discussion and debate. While there is a matter of terminology at stake here, this is not a mere terminological debate. We thank Domínguez et al. for highlighting the need for this discussion, which in turn provided opportunity to clarify points we have raised in previous papers. We share the very same goals even if we do not agree on all the facts. As such, we look forward to the immediate future where we move beyond terminological debates and work toward the common goals we share in a way that projects empathy to our object of study and accuracy to the concepts we seek to understand.