Sage Journals: Discover world-class research

Abstract

This special issue revisits a classic topic in linguistic theory, A-bar movement, applied to developing and bilingual grammars. We claim that A-bar movement, or filler-gap dependencies, is still the quintessential linguistic phenomenon to illustrate the interaction between the biological endowment, the experience with language (past and present), and other cognitive considerations non-specific to the faculty of language (i.e. the three factors in language design discussed in contemporary Chomskyan approaches). These three factors are present in non-native and bilingual populations, in which asymmetries between grammatical knowledge and other factors are even more apparent, in such a way that we can observe the role of each of these components independently. The appearance of new data from unique populations of bilinguals and novel experimental methodologies justify the collection of articles gathered in this volume. These studies inform new and old theoretical debates about the accessibility to Universal Grammar (UG) in nonnative grammars, the relationship between the grammar and the parser, and the role of individual differences.

Keywords

A-bar movement filler-gap dependencies islands relative clauses that-trace effect wh-movement

I Introduction

Human language exhibits a number of peculiar features, such as recursion, constituency, duality of patterning, and the ‘displacement’ of semantically-related elements from one another (Lust, 2006). It is through the intense and prolonged study of these attributes of human language that we gain insights into linguistic competence, how language is acquired, and how linguistic information interacts with and conforms to real-time processing pressures. A-bar movement, also commonly known as filler-gap dependencies (see examples in Section II), are one of the most fascinating structural features of human language that have received considerable attention in the extant literature. One of the main reasons why structures exhibiting semantically-related fillers and gaps have been a point of focus for theoretical and applied linguists alike is its relevance towards key tenets of modern linguistics, such as the concept of Universal Grammar (UG) and other behavioral metrics of processing and complexity. From a generative perspective of Second Language Acquisition (GenSLA), many instances of syntactic movement perfectly illustrate a poverty-of-the-stimulus scenario in which constraints on movement (i.e. islands) are never part of the learners’ input; yet, in spite of their absence, studies have shown that learners whose native language does have overt movement in certain contexts develop clear intuitions of what is possible and what is not with respect to syntactic extractions, knowledge that cannot come from previous linguistic experience or the target input (Juffs, 2005; Juffs and Harrington, 1995; White and Juffs, 1998). These results have traditionally served as evidence of the accessibility to UG in language acquisition. Likewise, failing to recognize the limits of movement in a second language (L2) has been interpreted as partial or total inaccessibility to UG (Bley-Vroman, 1990; Hawkins and Chan, 1997; Johnson and Newport, 1991). With the introduction of new behavioral and experimental techniques, the options for (non-)convergence between native and non-native speakers multiply, as differences and similarities can be found not only in the mental representation of the structure, but also in its real-time processing. Differences in processing can be merely quantitative, such as a delayed parsing due to limited resources, or problems integrating types of knowledge (Hopp, 2007, 2009; Slabakova, 2009); alternatively, this quantitative difference can also be interpreted as the employment of different mechanisms or different parts of the brain to comprehend or produce a structure, being the Shallow Structure Hypothesis (Clahsen and Felser, 2006) the most notable proposal in non-convergent approaches. In this special volume, we explore the nature of filler-gap dependencies in bi- and multilingual individuals and what these findings reveal about the status of these grammars at the intersection of linguistic knowledge and performance biases. In some cases, proposed syntactic analyses for filler-gap dependencies make different predictions for first language (L1) or second language (L2) speakers (Kim and Goodall, 2022), or show syntactic change in heritage contexts (Putnam and Søfteland, 2022), facts that empower L2 and bilingual data to decide between competing theoretical analyses or describing syntactic change. Some of the open questions regarding the knowledge and processing of filler-gap dependencies in non-native grammars that are discussed in this collection of articles concern the role of the L1 in the acquisition of subsequent languages, the L2 learners’ sensitivity to illicit (i.e. islands, that-trace effects) configurations, the capacity of non-native speakers to use abstract syntactic information when parsing these complex sentences, the type of brain activation at gap sites, or the effect of individual differences (i.e. attentional control, working memory) when predicting or processing filler-gap dependencies.

Taken together, this collection of articles revisits and updates central topics in the literature of GenSLA, making them pertinent also to the discussions of the twenty-first century. Whereas all studies of this collection defend the position that L2 learners are able to restructure their grammars and converge with the target grammar, employing movement strategies and showing island sensitivity by not trying to pose a gap in an illicit position (Aldosari et al., 2022; Covey et al., 2022; Espírito Santo et al., 2023; Kim and Goodall, 2022), not all studies find processing similarities. For instance, Covey et al. (2022) found limited predictive processing in Chinese-speaking learners of English. Comparably, Kim and Goodall (2022) interpreted their data as the result of the L2 speakers’ difficulty with advance planning of embedded clauses. This brings us to the more recent controversy concerning the source of constraints on movement, that is, whether they have a purely grammatical nature or a cognitive one, a matter that echoes the interaction between the three factors in language acquisition (Chomsky, 2005). Similar to observations in earlier theoretical treatments of filler-gap dependencies in monolingual populations, initial explorations in L2 grammars also attempted to capture these constraints by means of purely structural (i.e. syntactic) constraints. Subsequently, research on the status of islands in monolingual grammars as well as in developing and steady-state L2 grammars also explored the phenomenon as the combination of syntactic cues interacting with other ‘third factor’ phenomena (Boxell and Felser, 2017; Clahsen and Felser, 2018; Covey, 2018; Hopp, 2015, 2016; Johnson, 2015; Johnson et al., 2016; Omaki and Schulz, 2011; Perpiñán, 2020). The articles in this special volume of Second Language Research contribute to a deeper understanding of the structural and processing (i.e. ‘third factor’ phenomena) that determine the well-formedness of filler-gap dependencies in bi-/multilingual and L2 grammars. They engage in the rigorous combination of theoretical and experimental models to explore old and new questions related to filler-gap dependencies in bi- and multilinguals. Building upon syntactic, psycholinguistic, and neurolinguistic research on these structures, and incorporating novel language combinations in different types of bilingual populations, this set of articles collectively and individually touch upon the following research questions, whose answers we summarize in this introduction:

Research question 1: How does the experimental study of the acquisition and processing of filler-gap phenomena and their proposed constraints advance our understanding about the nature of L2, bilingual, and heritage grammars?

Research question 2: How can the experimental investigation of A-bar movement in bi-/multilingual individuals and populations contribute to ongoing research on debates regarding the status of the constraints and restrictions on these filler-gap dependencies?

Research question 3: How do the three factors in language design (UG, linguistic experience, and extralinguistic ‘third factor’ principles; Chomsky, 2005) interact in non-native and bilingual grammars with respect to extraction?

II Filler-gap dependencies: Basic principles and properties

In order to better understand these debates, we will first exemplify the linguistic phenomenon in detail, and then we will go back to the open debates in developing and bilingual grammars regarding these complex structures. The literature on filler-gap dependencies is quite expansive, including formal discussion and experimental implementation. Although it is too ambitious to attempt to provide a comprehensive overview of this topic, here we wish to point out some of the core structural attributes of filler-gap dependencies. The examples in (1), taken from Chaves and Putnam (2020: 1), illustrate the core properties of filler-gap dependencies:

(1) a. The doctor_i [who]_i I think Ben said he really likes ____i has retired.

b. [Who]_i I think Ben said he really likes ____i is THAT DOCTOR.

c. It’s THAT DOCTOR_i [who]_i I think Ben said he REALLY LIKES ____i.

d. [THAT DOCTOR]_i I think Ben said he REALLY LIKES ____i.

e. [Which doctor]_i do you think Ben said that he really likes ____i?

Words that fit together in a meaningful way, i.e. are semantically-related to one another, can occur in an apparent arbitrary manner far away from one another. In all the examples in (1) above, the lexical items in brackets are interpreted as if they were realized immediately after the verb likes. The lexical item in brackets is commonly referred to as the filler, while the gap (which we indicated with the co-indexed space ___ in the examples above; it can also be symbolized with a t for trace) represents where the lexical item is (also) interpreted. There is certain debate in the literature regarding the boundedness of filler-gap dependencies. Consider the examples in (2), again adapted from Chaves and Putnam (2020: 2), that serve as a case in point, with the extraction pathway existing between the filler and the gap across three clausal boundaries in (2a), and four in (2b):

(2) a. [What]_i did you think [the students will say __ [they believe __ [the teacher was trying to do ___i]]]?

b. [What]_i do you think [the students will say __ [they believe __ [the teacher claimed __ [he was trying to do ___i]]]?

Experimental research on filler-gap dependencies over the past three decades confirms the local nature of these apparent distant extraction pathways (Hofmeister and Sag, 2010; Kluender, 1998; Phillips, 2006). The ‘correct’ gap in both (2a) and (2b) is the final one, however when parsing these sentences in their entirety, the intermediate gap positions are likely to be attempted – and subsequently rejected – during the incremental processing associated with sentence comprehension (Gibson and Warren, 2004). It is important to keep in mind that, in spite of the unbounded nature of filler-gap dependencies, some restrictions still exist. The nature of these constraints, whether they are purely structural, purely cognitive, or a mixture of these, remains hotly contested and directly addressed in Aldosari et al. (2022), Kim and Goodall (2022), and Covey et al. (2022) in this special issue. In Sections II.1 and II.2, we will describe two of the most well-known types of movement constraints: islands and the complementizer constraint (that-trace effect).

1 That-trace effects

As a proof of concept, we take a closer look at that-trace effects (also known as Complementizer constraint), illustrating what we know so far about this type of filler-gap dependency (and what we also do not know). Overly simplifying this state of affairs, there are two primary situations to which we wish to draw the reader’s attention. The first situation involves extraction patterns that are acceptable in one language while being unacceptable in another. English does not allow a subject gap position when it immediately follows an overt complementizer, such as that, that heads the embedded clause (from Chaves and Putnam, 2020: 97):

(3) a. What/who do you think ___ broke the window?

b. * What do you think that ___ broke the window?

(compare ‘Do you think that the ball broke the window?’)

c. What do you think (that) the ball broke ___?

The sentences in (3) illustrate the nature of a canonical that-trace effect in English: The instrument/agent argument that broke the window, which originates thematically to the embedded clause, can be extracted and appear in the matrix clause; see (3a). Example (3b) demonstrates a that-trace effect violation; in contrast to the previous example, the presence of the overt complementizer that ‘blocks’ the extraction of the instrument/agent argument. Finally, as shown in (3c), object wh-items are not subjected to that-trace effects.

The that-trace effect introduced in (3) holds in most varieties of English, however, there is a wide array of documented variation within English (Pesetsky, 1982, 2019), while other languages, such as Norwegian (Taraldsen, 1978) and Lakhota (Van Valin, 1987), appear to lack these that-trace effects altogether. Aside from the acceptability contrast of the type of complementizer (for example, embedded clauses headed by that and if lead to different levels of acceptability), one particular confound that obfuscates a purely structural/syntactic account of that-trace effects is that these effects disappear in English once we place an adverbial phrase between the complementizer that and the gap (see, amongst others, Bresnan, 1977; Culicover, 1993). Chaves and Putnam (2020: 98) dub this the ‘adverbial amelioration effect’, which is illustrated in (4):

(4) a. [Which students] did you claim that as far as you could tell ___ had read the book cover to cover?

b. [Who] did you think that after years and years of cheating death ___ finally died?

Additional evidence suggesting that non-syntactic factors, such as phonology, play a role in determining that-trace effects can be found in Nupe, a Benue-Congo language spoken in south-central Nigeria (Kandybowicz, 2006, 2009). This, of course, may also be the case (to some degree) for English, however, determining the exact nature of that-trace effects in English, Nupe, or any other language is beyond the scope of these introductory remarks. These points illustrate that in addition to purely architectural (i.e. syntactic) constraints, a number of grammatical (contextual difficulty, relevance presupposition violation, semantic–pragmatic conflicts, and phonology–syntax conflicts) and processing factors (violated expectation and working memory limitations) shape the acceptability of filler-gap dependencies in a given language (for a more elaborate discussion of these issues, see, for example, Chaves and Putnam, 2020, Section 3.2). Kim and Goodall’s (2022) article directly addresses this issue by comparing one group of L1 English speakers to two groups of L2 learners of English – namely, Spanish and Mandarin Chinese natives – and tests two competing theoretical analyses that could explain the that-trace effect: a purely syntactic one such as Antilocality (restrictions on the distance of the extraction), and the Principle of End Weight (PEW), which rely on principles of sentence processing (difficulty of having a gap at the beginning of a planning unit). The authors conclude that their study supports the PEW-based predictions, finding asymmetries in the Korean speakers in their responses in sentences with that, which showed amelioration, as opposed to sentences without that. Assuming that L2 learners have difficulty with advance planning, the presence of that provides additional time to plan the embedded clause and, hence, the presence of that facilitates the processing. However, this correlation between extraction ability and processing factors has not consistently been found for island effects (Aldosari et al., 2022; Sprouse et al., 2012).

2 Islands

Languages differ with respect to whether or not a particular type of filler-gap dependency is acceptable; however, a fundamental question is whether extraction patterns exist that are ill formed in all languages. Such universal extraction constraints are called islands (Ross, 1967) and have been the focus of intense research from myriad perspectives since their discovery. Some of the more commonly described are the following, from Aldosari et al. (2022):

(5) a. * What_i did Tom read the book [before Bill saw _ _i] Adjunct Clause

b. * What_i does Tom like [the woman who wears _ _i] Relative Clause

c. * What_i did Tom hear [the fact that Jon won __i] Complex NP

d. * What_i does Tom wonder [why the boy bought __i] Wh-Island

Some linguists interpret islands to be primarily structural in nature (Chomsky, 1973, 1986, 2001; Huang, 1982; Phillips, 2006), while, for others, they are the result of extragrammatical elements such as memory and processing constraints, as well as pragmatic and semantic considerations (Hofmeister and Sag, 2010; Hofmestier et al., 2013; Kluender, 2004; Kluender and Kutas, 1993). For the former, the ban on extraction out of islands constitutes an innate linguistic principle, part of the speaker’s genetic endowment or UG (‘first factor’ in Chomsky’s, 2005 terms); whereas for the latter, limitations on movement are due to principles not specific to the language faculty or ‘third factors’. The possibility also exists that island effects represent a combination of these competing factors.

The examples in (6), based on Phillips (2013), represent fairly common extraction islands discussed in the literature; in (6a) we observe that extraction cannot take place from a single conjunct, and in (6b) we see that extraction from an object with a definite specifier/marker is also banned. Finally, as shown in (6c), extraction from a clausal subject also results in an unacceptable structure.

(6) a. * [What]_i did [the Senate approve ___i] and [the House reject the bill]?

b. * [Who]_i did Robin believe [Simon’s news about ___i]?

c. * [What]_i did [that Ellen remembered ___i] surprise her children?

Despite the general unbounded nature of filler-gap dependencies that allow the filler to appear relatively far from the gap, constraints on extraction do exist; consider the examples in (7):

(7) a. * [Who]₁ did you give to ___₁ the canary that I brought home yesterday?

b. * [Who]₁ did you offer to ___₁ the chance to win $1,000?

Generative approaches to grammar sought to develop (purely) structural constraints to account for illicit filler-gap dependencies. As a case in point, Kuno (1973: 130) proposed the Clause Non-Final Incomplete Constituent Constraint (CNFICC), which prohibits extraction out of phrases in a clause non-final position, to account for the ill-formedness of the examples in (8). Counterexamples to the CNFICC do in fact exist – as is the case for many extraction constraints defined solely according to structural configurations, as shown by Jackendoff and Culicover (1972), Hukari and Levine (1991), Fodor (1992), Haegeman et al. (2013), and others; see (8).

(8) a. [Who]₁ can I talk to ___₁ about my depression?

b. [What]₁ do astronauts like to take pictures of ___₁ from space?

Research focusing on the nature of islands continues to dominate the landscape. For example, regarding the CNFICC examples discussed above, Chaves and Putnam (2020: 54) make the case that this particular island condition may be due to the misparsing of sentence in situations where the preposition is (or is not) perceived as being stranded in relation to its sentence-medial position. Although this treatment may hold for this particular island, it is unclear if this analysis would hold for all islands. Relatedly, this special issue also contains two articles devoted to the placement and parsing of prepositions in movement constructions (Espírito Santo et al., 2023; Perpiñán and Cardinaletti, 2022), with both sets of authors claiming that, despite defective morphology (missing Pied-Piping or even the obligatory preposition altogether), these structures involve movement and adhere to structural constraints. In a related sense, the contribution of Putnam and Søfteland (2022) is worth mentioning in this context as an example of restructuring wh-morphosyntactic elements encoding modality in the remaining North American Norwegian speakers. This study is further evidence that early bilinguals (heritage speakers) can be agents of structural linguistic change, as argued extensively in Montrul (2023), also in the domain of filler-gap dependencies.

III Scope and content of this volume

Returning to the research questions from the introduction, here we provide a brief summary of the key findings and contributions of these articles and how they enrich and expand our current understanding of filler-gap dependency research in developing and bilingual grammars.

Research question 1 focuses on the nature of L2, bilingual, and heritage grammars, and how the experimental study of the acquisition and processing of filler-gap dependencies and their proposed constraints advance our general understanding about these grammars. In their review of studies of second language acquisition (SLA) that investigated island constraints, Belikova and White (2009) concluded that changes in theoretical accounts have directly affected the debate on the nature of interlanguage grammars since they render it impossible to distinguish between which L2 behavior originated from L1 transfer and which from UG accessibility. Belikova and White’s review confirmed that, assuming that islands are regulated by universal computational principles, island constraints have become a moot point for the debates on differences between L1 and L2 grammars. Following this line of reasoning, the results of the studies in this volume demonstrate that L2 learners’ grammars are indeed sensitive to movement constraints (islands, that-trace effects) as much as L1 speakers, and that, if we find differences, these are in the realm of processing capabilities or morphological expression.

To illustrate, in their study on event related potentials (ERPs) on English wh-dependency resolution by Mandarin Chinese speakers, Covey et al. (2022) found that, whereas L2 learners demonstrated island sensitivity (also in an additional off-line grammaticality judgment task), their processing was different from that of native speakers. They also found distinct brain responses for L2 learners (P600 effects) as compared to native speakers (N400 effects) while reading for comprehension; attentional control (a number Stroop task) taken as an individual measure that could capture variability in predictive processing modulated these results. The native speakers’ N400 effects were interpreted as a result of encountering unpredicted lexical material when resolving a filler-gap dependency, i.e. native speakers do not predict gaps inside islands, even in grammatical ones. This avoidance of positing a gap in the presence of an island domain modestly correlates with attentional control resources. Following Sprouse et al.’s (2012) rationale, the authors interpreted these individual results as evidence of grammatically guided dependencies. In their results, L2 learners did not attempt to pose a gap inside the island. In other words, there was no thematic link between the filler and the gap in this position, against what the Shallow Structure Hypothesis (Clahsen and Felser, 2006; Felser, 2019) would suggest. Thus, even though the learners’ brain activity was qualitatively different from that of the native speakers, L2 learners showed difficult syntactic integration only in licit contexts. In addition, the authors did not find a strong correlation between individual measurements of attentional control and processing, but they did find that learners with better Stroop scores may have experienced greater syntactic integration efforts, a result that they interpret as counterevidence for a processing account of islands. Covey et al.’s (2022) findings also included the lack of an N400 in L2 learners, which was interpreted as limited L2 processing, in line with Grüter et al.’s (2017) Reduced Ability to Generate Expectation (RAGE) Hypothesis.

Aldosari et al. (2022) also attempts to correlate individual differences, such as Working Memory, with island sensitivity in Najdi Arabic learners of English, another language which does not display overt wh-movement. This investigation had two main objectives: (1) to investigate whether L2 learners are sensitive to island constraints, and (2) to elucidate the source of island effects, which contributes to the debate introduced in the general research question 2. With an acceptability judgment task (AJT) that tested four types of islands (adjunct, subject, complex NP, and whether islands), all sentences introduced with a complex wh-filler phrase (e.g. which keys) and preceded by a short background context, the two main conditions tested for each type of sentence were the length of extraction (short, from a matrix clause, or long, from an embedded clause) and island vs. non-island configuration. The only ungrammatical configuration was extraction from an embedded clause forming an island. Results for both, native and L2 speakers indicated that longer (embedded) wh-dependencies were rated lower than those with shorter (matrix) extractions; and sentences with islands were rated lower than non-island sentences. These two main factors interacted in such a way that the effect of the island structure was greater in sentences with a long extraction (i.e. the illicit island), indicating a strong sensitivity to these structures. Additionally, there were low acceptability ratings only for the ungrammatical island violation, compared to the other three grammatical conditions; at the same time, and replicating previous results (Johnson and Newport, 1991; Martohardjono, 1993; Pham et al., 2020), these findings were greater for strong islands (adjunct and subject islands) than for weak islands, particularly for whether islands (complex NPs island behaved like strong islands). Working memory measurements, in contrast, did not correlate with the AJT scores in either type of speaker, a result that Aldosari et al. (2022) interpreted as strong evidence for the grammatical nature of island effects in both, natives and L2 learners.

The results from the previous study are somewhat at odds with those found by Kim and Goodall (2022), whose study suggests a non-syntactic approach to extractions, in this case investigating the that-trace phenomenon. They employed an AJT with long-distance English wh-questions manipulating extraction site (embedded subject vs. embedded object) and presence or absence of that to test L2 English with Korean (a wh-in-situ language) and Spanish (a wh-movement language without the that-trace effect) native speakers. The results indicated a significant interaction of extraction site and presence of that in English native speakers, since they only rejected subject extraction with that, the ungrammatical structure, whereas L2 learners only presented a main effect for extraction site and no interaction (subject extractions were generally rated significantly worse than the rest of the structures). Additionally, L1 Korean speakers also showed a main effect for that. The authors argued that, if a syntactic account (Antilocality) were responsible for the that-trace effect, native and bilingual/L2 speakers would behave similarly. Instead, L2 learners experienced difficulties with all instances of subject extraction independently of the presence of that, a fact that the authors interpret as evidence for the PEW-based account. This processing burden is more apparent in L2 learners, thus explaining why this extraction asymmetry with no that effect is only present in L2 learners’ grammars. Furthermore, for Korean speakers, the presence of that even produces an amelioration of acceptability, which is explained in terms of predictive planning: Korean speakers make use of that to facilitate the processing of already difficult embedded clauses; this complementizer would allow extra time for planning in precisely the speakers that are not used to planning long-distance filler-gap dependencies.

Another group of L2 speakers whose language does not instantiate overt wh-movement is Chinese. Espírito Santo et al.’s (2023) study investigated the acquisition of European Portuguese (EP) relative clauses, in particular prepositional relative clauses, which in EP instantiate wh-movement with preposition pied-piping (the preposition moves along with the relative pronoun). Additionally, oral EP also allows the omission of the obligatory preposition, a strategy termed as Null-Prep (also investigated in Perpiñán and Cardinaletti, 2022) and resumptive pronouns in colloquial EP contexts. Chinese (Mandarin and Cantonese), on the other hand, only allows prepositional relative clauses with resumptive pronouns. With an oral production task and two self-paced reading AJTs, Espírito Santo et al. investigated (1) whether Chinese learners of EP would be able to acquire the morphological and syntactic properties of EP prepositional relative clauses, and (2) whether they would obey constraints on wh-movement with island configurations. Their results indicated that these L2 learners are able to restructure their grammar as they produce and/or accept pied-piping and Null-Prep in similar percentages to those found from native speakers. In addition, Chinese L2 learners do not transfer the resumptive pronoun strategy from their L1, not even at intermediate stages of acquisition. The data from island conditions further suggest an account of a restructured interlanguage grammar with wh-movement, as well as the inability to rescue island configurations with resumptive pronouns. This effect was kept constant across all groups, native and non-native, and questions the traditional view that resumptive pronouns can ameliorate island effects. These findings open a new interpretation of resumptive pronouns’ role, such as a discourse–pragmatics function related to comprehension (Beltrama and Xiang, 2016) and not exclusively to syntax (Polinsky et al., 2013). This study provides evidence of convergence between L1 and L2 grammars in production and acceptability ratings in a quite infrequent and variable structure such as prepositional relative clauses in EP, a result that the authors take as a clear indicator that non-native speakers are able to fully access UG and restructure their grammars, even in cases of poverty-of-stimulus.

Perpiñán and Cardinaletti (2022) also investigated oblique movement in Spanish prepositional relative clauses, interrogatives, and sluicing, focusing on the Null-Prep structure, a strategy recurrently found in L2 grammars. Null-Prep has been analysed as a ‘wild’ element that exists ‘outside’ of UG (Klein, 1995), or a non-movement (binding) derivation (Dekydtspotter et al., 1998). The authors propose instead that Null-Prep appears as a result of movement, i.e. as a byproduct of the complexity of the derivation. They investigated the phenomenon in L2 Spanish within a group of Moroccan Arabic speakers, which is a language that instantiates resumptive pronouns and also pied-piping in formal contexts, and an English-speaking group, with English resorting to preposition stranding, as well as pied-piping in formal contexts. Neither standard Spanish, Moroccan Arabic, nor English accept Null-Prep. The results of two oral production tasks indicated that Null-Prep is produced in relative clauses to a greater extent than transfer structures (resumptives for the Arabic-speaking learners; preposition stranding for the English-speaking learners). Furthermore, the Null-Prep strategy alternated with movement (pied-piping), however, not with binding strategies. The results of a self-paced grammaticality judgment task further showed differential acceptance of Null-Prep across constructions and participants, being moderately accepted in relative clauses, then in sluicing, and to a lesser extent in wh-questions. Perpiñán and Cardinaletti concluded that Null-Prep follows a gradient scale in accordance with a complexity hierarchy of syntactic derivation. To summarize, the more complex the structure is, the more frequent Null-Prep will be, and the more pied-piping is accepted, the less Null-Prep is accepted, since these two structures are in competition. Null-Prep, then, appears as a derivationally less complex structure than pied-piping, as it requires fewer merging operations (P + DP is spared). The authors also argue that Null-Prep is a natural phenomenon in languages that display pied-piping, although more persistent in developing grammars not linked to any particular language.

The final contribution is based on corpus data extracted from the Corpus of American Nordic Speech (CANS). Putnam and Søfteland (2022) analysed wh-infinitives (indirect questions) in North American Norwegian (NAmNo), a moribund heritage variety of Norwegian spoken predominantly in the Upper Midwest of the USA. Speakers of NAmNo produce wh-infinitives, equivalent to those found in English, e.g. ‘I don’t know what to do’, which is ungrammatical in homeland Norwegian. Putnam and Søfteland make the case that wh-infinitives have become a viable option in the grammar of remaining NAmNo speakers. In their theoretical analysis, Putnam and Søfteland show that the restructuring of the NAmNo grammar to license such wh-infinitives requires the encoding of modality on the wh-element in wh-infinitives (rather than on a modal verb, as is the case in homeland Norwegian). Crucially, these speakers can also use the ‘original’ Norwegian way of expressing overt modality, i.e. by producing an overt modal verb marked with tense, which, in turn, allows the appearance of an overt subject in the embedded clause. These results touch up on research question 3 associated with the contributions of this special issue.

Collectively, these articles draw upon empirical evidence from acceptability judgment studies, elicited production, experiments involving behavioral methods including electrophysiological techniques such as ERPs, and data extracted from corpora. One of the broader implications of the articles in this issue is that they collectively argue for the theoretical position that some form of mental representation guides and directs parsing procedures. Factors – such as extraction length, embeddedness, or the complexity of the derivation – make grammatically more costly structures; however, this is not always automatically reflected in longer reaction times in processing. This burden can be relieved by dropping a preposition and sparing a step in the derivation, expressing modality covertly, or preferring a that at the beginning of a unit to have extra time for planning. These creative solutions are found more frequently in L2 learners and bilingual populations in environments when more than one grammatical structure competes, and their experiences with the languages may have a stronger impact. In this collection of articles we do not observe individual differences strongly correlating with the knowledge or the processing of filler-gap dependencies, a fact that emphasizes their universal nature. We contemplate the possibility that the absence of a correlation between measurements of individual cognitive abilities and acceptability ratings does not invalidate a hybrid approach to extractions in which both, purely grammatical (first factor) and extragrammatical considerations (third factor), intervene. In summary, this collection of articles touches on critical linguistic topics in non-native and bilingual populations, which advance our general knowledge about the human language capacity and enlightens the field for its relevance in combination with other cognitive sciences such as psychology, neurology, and anthropology.

Footnotes

Acknowledgements

We are highly appreciative of helpful comments and suggestions as well as interesting questions from Silvina Montrul and Roumyana Slabakova on an earlier version of this article. The usual disclaimers apply.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sílvia Perpiñán

References

Aldosari

Covey

Gabriele

(2022) Examining the source of island effects in native speakers and second language learners of English. Second Language Research. Advance online publication. http://doi.org/10.1177/02676583221099243

Belikova

White

(2009) Evidence for the fundamental difference hypothesis or not? Island constraints revisited. Studies in Second Language Acquisition 31: 199–23.

Beltrama

Xiang

(2016) Unacceptable but comprehensible: The facilitation effect of resumptive pronouns. Glossa: A Journal of General Linguistics 1: 29.

Bley-Vroman

(1990) The evolving context of the fundamental difference hypothesis. Studies in Second Language Acquisition 31: 178–98.

Boxell

Felser

(2017) Sensitivity to parasitic gaps inside subject islands in native and non-native sentence processing. Bilingualism: Language and Cognition 20: 494–511.

Bresnan

(1977) Variables in the theory of transformations. In: Akmajian

Culicover

Miller

(eds) Linguistic theory and psychological reality. Cambridge, MA: MIT Press, pp. 1–59.

Chaves

Putnam

(2020) Unbounded dependency constructions: Theoretical and experimental perspectives. Oxford: Oxford University Press.

Chomsky

(1973) Conditions on transformations. In: Anderson

Kiparsky

(eds) A festschrift for Morris Halle. New York: Holt, Rinehart, and Winston, pp. 232–86.

Chomsky

(1986) Barriers. Cambridge, MA: MIT Press.

10.

Chomsky

(2001) Derivation by phase. In: Kenstowicz

(ed.) Ken Hale: A life in language. Cambridge, MA: MIT Press, pp. 1–50.

11.

Chomsky

(2005) Three factors in language design. Linguistic Inquiry 36: 1–22.

12.

Clahsen

Felser

(2006) Grammatical processing in language learners. Applied Psycholinguistics 27: 3–42.

13.

Clahsen

Felser

(2018) Some notes on the shallow structure hypothesis. Studies in Second Language Acquisition 40: 693–706.

14.

Covey

(2018) An ERP investigation of individual differences in the processing of wh-dependencies by native and non-native speakers. Unpublished PhD thesis, University of Kansas, Lawrence, KS, USA.

15.

Covey

Fiorentino

Gabriele

(2022) Island sensitivity in L2 learners: Evidence from acceptability judgments and event-related potentials. Second Language Research. Advance online publication. http://doi.org/10.1177/02676583221116039

16.

Culicover

(1993) Evidence against the ECP accounts of the that-t effect. Linguistic Inquiry 24: 557–61.

17.

Dekydtspotter

Sprouse

Anderson

(1998) Interlanguage A-bar dependencies: Binding construals, null prepositions and Universal Grammar. Second Language Research 14: 341–58.

18.

Espírito Santo

Alexandre

Perpiñán

(2023) The role of resumption in the acquisition of European Portuguese prepositional relative clauses by Chinese learners. Second Language Research. Advance online publication. http://doi.org/10.1177/02676583221137715

19.

Felser

(2019) Structure-sensitive constraints in non-native sentence processing. Journal of the European Second Language Association 3: 12–22.

20.

Fodor

(1992) Islands, learnability, and the lexicon. In: Goodluck

Rochemont

(eds) Island constraints: Theory, acquisition, and processing. Dordrecht: Kluwer, pp. 163–223.

21.

Gibson

Warren

(2004) Reading-time evidence for intermediate linguistic structure in long-distance dependencies. Syntax 7: 55–78.

22.

Grüter

Rohde

Schafer

(2017) Coreferences and discourse coherence in L2: The roles of grammatical aspect and reference form. Linguistic Approaches to Bilingualism 7: 199–229.

23.

Haegeman

Jiménez-Fernández

Radford

(2013) Deconstructing the subject condition: Cumulative constraint violation and tolerance thresholds. The Linguistic Review 31: 73–150.

24.

Hawkins

Chan

(1997) The partial availability of universal grammar in second language acquisition: The ‘failed functional features hypothesis’. Second Language Research 13: 187–226.

25.

Hofmeister

Sag

(2010) Cognitive constraints and island effects. Language 86: 366–415.

26.

Hofmeister

Casasanto

Sag

(2013) Islands in the grammar? Standards of evidence. In: Sprouse

Hornstein

(eds) Experimental syntax and island effects. Cambridge, UK: Cambridge University Press, pp.42–63. https://doi.org/10.1017/CBO9781139035309.004

27.

Hopp

(2007) Ultimate attainment at the interfaces in second language acquisition: Grammar and processing. Groningen: Grodil Press.

28.

Hopp

(2009) The syntax–discourse interface in near-native L2 acquisition: Off-line and on-line performance. Bilingualism: Language and Cognition 12: 463–83.

29.

Hopp

(2015) Differential learning of prediction. Linguistic Approaches to Bilingualism 5: 476–81.

30.

Hopp

(2016) The timing of lexical and syntactic processes in L2 sentence comprehension. Applied Psycholinguistics 37: 1253–80.

31.

Huang

CTJ

(1982) Move wh in a language without wh movement. The Linguistic Review 1: 369–416.

32.

Hukari

Levine

(1991) On the disunity of unbounded dependency constructions. Natural Language and Linguistic Theory 9: 97–144.

33.

Jackendoff

Culicover

(1972) A reconsideration of dative movement. Foundations of Language 7: 397–412.

34.

Johnson

(2015) Individual differences in predictive processing: Evidence from subject filled-gap effects in native and non-native speakers of English. Unpublished PhD thesis, University of Kansas, Lawrence, KS, USA.

35.

Johnson

Fiorentino

Gabriele

(2016) Syntactic constraints and individual differences in native and non-native processing of wh-movement. Frontiers in Psychology 7: 549.

36.

Johnson

Newport

(1991) Critical period effects on universal properties of language: The status of subjacency in the acquisition of a second language. Cognition 39: 215–18.

37.

Juffs

(2005) The influence of the first language on the processing of wh-movement in English as a second language. Second Language Research 21: 121–51.

38.

Juffs

Harrington

(1995) Parsing effects in L2 sentence processing: Subject and object asymmetries in wh-extraction. Studies in Second Language Acquisition 17: 483–516.

39.

Kandybowicz

(2006) Compe-trace effects explained away. In: Baumer

Montero

Scanlon

(eds) Proceedings of the 25th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Press, pp. 220–28.

40.

Kandybowicz

(2009) Comp-trace effects and the syntax–phonology interface. In: Bunting

Desai

Peachey

Straughn

Tomková

(eds) Proceedings of CLS 42. Chicago, IL: Chicago Linguistics Society, pp. 103–15.

41.

Kim

Goodall

(2022) The source of the that-trace effect: New evidence from L2 English. Second Language Research. Advance online publication. http://doi.org/10.1177/02676583221104604

42.

Klein

(1995) Evidence for a ‘wild’ L2 grammar: When PPs rear their empty heads. Applied Linguistics 16: 87–117.

43.

Kluender

(1998) On the distinction between strong and weak islands: A processing perspective. In: Culicover

McNally

(eds) The limits of syntax. Leiden: Brill, pp. 241–79.

44.

Kluender

(2004) Are subject islands subject to a processing account? In: Schmeiser

Chand

Kelleher

Rodriguez

(eds) Proceedings of WCCFL 29. Somerville, MA: Cascadilla Press, pp. 101–25.

45.

Kluender

Kutas

(1993) Bridging the gap: Evidence from ERPs on the processing of unbound dependencies. Journal of Cognitive Neuroscience 5: 196–214.

46.

Kuno

(1973) Constraints on internal clauses and sentential subjects. Linguistic Inquiry 4: 363–86.

47.

Lust

(2006) Child language: Acquisition and growth. Cambridge: Cambridge University Press.

48.

Martohardjono

(1993) Wh-movement in the acquisition of a second language: A crosslinguistic study of three languages with and without overt movement. Unpublished PhD thesis, Cornell University, Ithaca, NY, USA.

49.

Montrul

(2023) Native speakers, interrupted: Differential object marking and language change in heritage languages. Cambridge: Cambridge University Press.

50.

Omaki

Schulz

(2011) Filler-gap dependencies and island constraints in second-language sentence processing. Studies in Second Language Acquisition 33: 563–88.

51.

Perpiñán

(2020) Wh-movement, islands and resumption in L1 and L2 Spanish: Is (un)grammaticality the relevant property? Frontiers in Psychology 11: 395.

52.

Perpiñán

Cardinaletti

(2022) Null-Prep as a systematic interlanguage phenomenon: Evidence from relative clauses, interrogatives, and sluicing constructions. Second Language Research. Advance online publication. http://doi.org/10.1177/02676583221132198

53.

Pesetsky

(1982) Paths and categories. Unpublished PhD thesis, Massachusetts Institute of Technology, Cambridge, MA, USA.

54.

Pesetsky

(2019) Complementizer-trace effects. In: Everaert

van Riemsdijk

(eds) A companion to syntax. 2nd edition. Oxford: Blackwell, pp. 1–34.

55.

Pham

Covey

Gabriele

Aldosari

Fiorentino

(2020) Investigating the relationship between individual differences and island sensitivity. Glossa: A Journal of General Linguistics 5: 94.

56.

Phillips

(2006) The real-time status of island phenomena. Language 82: 795–823.

57.

Phillips

(2013) On the nature of island constraints I: Language processing and reductionist accounts. In: Sprouse

Hornstein

(eds) Experiment syntax and island effects. Cambridge: Cambridge University Press, pp. 64–108.

58.

Polinsky

Clemens

Morgan

Xiang

Heestand

(2013) Resumption in English. In: Sprouse

Hornstein

(eds) Experimental syntax and island effects. Cambridge: Cambridge University Press, pp. 341–59.

59.

Putnam

Søfteland

(2022) When covert modality sneaks into your grammar: Wh-infinitives in American Norwegian. Second Language Research. Advance online publication. https://doi.org/10.1177/02676583221103741

60.

Ross

(1967) Constraints on variables in syntax. Unpublished PhD dissertation, Massachusetts Institute of Technology, Cambridge, MA, USA.

61.

Slabakova

(2009) L2 fundamentals. Studies in Second Language Acquisition 31: 155–73.

62.

Sprouse

Wagers

Phillips

(2012) A test of the relation between working memory capacity and syntactic island effects. Language 88: 82–123.

63.

Taraldsen

(1978) The scope of wh movement in Norwegian. Linguistic Inquiry 9: 623–40.

64.

Van Valin

(1987) The role of government in the grammar of head-marking languages. International Journal of American Linguistics 53: 371–79.

65.

White

Juffs

(1998) Constraints on wh-movement in two different contexts of non-native language acquisition: Competence and processing. In: Flynn

Marohardjono

O’Neil

(eds) The generative study of second language acquisition. New York: Lawrence Erlbaum, pp. 111–19.

Filler-gap dependencies in bi- and multilingual grammars: Findings,challenges,and unknowns

Abstract

Keywords

I Introduction

II Filler-gap dependencies: Basic principles and properties

1 That-trace effects

2 Islands

III Scope and content of this volume

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

ORCID iD

References