Optionality and overgeneralisation patterns in second language acquisition: Where has the expletive ensconced «it»self?

by Nadia Varley (Author)
©2016 Thesis 269 Pages


This book discusses the nature of optionality in second language grammars and the indeterminacy observed in second language users’ linguistic representations. For these purposes, experimental data from 213 learners of German and 150 learners of Russian have been collected and analysed with a special focus on the acquisition of various «subjectless» and impersonal constructions as well as argument licensing. Whereas voice alternations and argument licensing are topics amply discussed in theoretical domains, their practical implementation within second language research has remained a research lacuna. This piece of work intends to fill the gap.

Table Of Contents

  • Cover
  • Title
  • Copyright
  • About the author
  • About the book
  • This eBook can be cited
  • 1. Editorial
  • 2. Editorial
  • 3. Editorial
  • Abstract
  • Acknowledgements
  • Abbreviations
  • Contents
  • 1. Introduction
  • 1.1 Addressing the object of enquiry
  • 1.1.1 The two (pivotal) “WHYs” in this book
  • 1.2 Outline
  • 2. Argument licensing and voice alternations
  • 2.1 Theoretical background
  • 2.1.1 Argument licensing, argument structure and voice alternations
  • 2.1.2 Is the subject an argument of its verb?
  • 2.2 Unaccusativity
  • 2.2.1 Unaccusative morphology and syntax: Form vs. structure
  • 2.2.2 Argument structure in passives and other unaccusatives
  • 2.2.3 So who (or what) is the subject of impersonals?
  • 2.3 Case and agreement in impersonals
  • 2.3.1 Default agreement and morphological underspecification in impersonals
  • 2.3.2 ACC licensing and CAUSE/VOICE parameterisation
  • 2.3.3. Case/DP licensing in impersonals
  • 2.4 Chapter summary
  • 3. Previous L2 research: An overview
  • 3.1 L1A vs. L2A
  • 3.2 UG or not UG?
  • 3.3 Syntactic impairment in L2A?
  • 3.3.1 The L2A of IMPPASS
  • 3.4 Underspecification and morphological deficits
  • 3.4.1 Further evidence: Underspecification in L1A
  • 3.5 L2 straddling the interfaces
  • 3.5.1 The Interface Hypothesis: an outline
  • 3.5.2 Reconciling facts and fitting them into theory: First step
  • 3.6 Chapter summary
  • 4. Parametric variation
  • 4.1 On the EPP and expletives
  • 4.2 Pro-drop
  • 4.3 Impersonals, passives, and voice alternations cross-linguistically
  • 4.3.1 Russian impersonals
  • 4.3.2 German impersonals
  • 4.3.3 L1s in the ‘L2 GE’ study
  • Slavic impersonals
  • Romance impersonals
  • Chinese
  • 4.3.4 Interim summary
  • 4.4 Parameters and clustering effects in the languages under investigation
  • 4.5 The L2A of arguments and voice alternations: Research questions and hypotheses
  • 4.6 Chapter summary
  • 5. Two studies
  • 5.1 Methodology, informants, and data elicitation
  • 5.1.1 The cloze test
  • 5.1.2 The Grammaticality Judgement Task
  • 5.1.3 Data acquisition and informants
  • 5.2 ‘L2 GE’ and ‘L2 RU’ tests: Description
  • 5.2.1 The ‘L2 GE test’
  • 5.2.2 The ‘L2 RU test’
  • 5.3 Statistical evaluation – technical details
  • 5.3.1 The Null Hypothesis
  • 5.4 Results
  • 5.4.1 ‘L2 German’ findings
  • Mixed data evaluation across levels of L2 GE proficiency
  • Parameter-based evaluation
  • 5.4.2 ‘L2 Russian’ findings
  • Data evaluation across levels of acquisition
  • 5.5 Chapter summary
  • 6. Discussion: On the complexities of L2 syntax
  • 6.1 The nature of L2 knowledge
  • 6.1.1 Argument and event structure in L2A
  • 6.1.2 Impersonal constructions in L2A: syntax, semantics, and discourse
  • 6.1.3 On zero, bound, and free pronouns: Referentiality in L2A
  • 6.1.4 The L2A of empty categories and expletives
  • 6.1.5 Case and agreement in L2A
  • 6.1.6 Interim summary
  • 6.2 Exploring and explaining optionality in interlanguage grammars
  • 6.2.1 The dynamics of linguistic knowledge: Native speakers’ vs. L2ers’ judgements
  • 6.2.2 Phases, functional domains, L1 transfer, interfaces: Which one is responsible for the variability in L2A?
  • Phases and functional domains
  • L1 transfer and positive evidence
  • Interfaces
  • 6.3 L2 morpho-syntax
  • 6.3.1 Representational deficits in light of the principle of Underspecification and linguistic defaults
  • 6.3.2 Overgeneralisation in L2A and learner defaults
  • 6.3.3 (Anti)causatives, impersonals, and (non-)canonical patterns in L2A
  • 6.4 Chapter summary
  • 7. Concluding remarks
  • References
  • Appendix A: ‘L2 GE test’
  • Appendix B: ‘L2 RU test’
  • Appendix C: Answers

1.   Introduction

Over the past several decades there has been a bulk of generative literature on second language acquisition (henceforth L2A) (cf. Clahsen & Muysken 1986, 1989; Lardiere 2008, 2011; Meisel 1983, 1991; Montrul 1999, 2011; Slabakova 2010; Sorace 1993, 2000, 2005, 2011; White 1985, 1989, 2000, 2003, 2011a, b, to name but a few). The object of investigation ranges from discussions of second language (L2) parameter resetting, e.g. verb raising and pro-drop parameters, through first language (L1) transfer to the role interfaces play in the acquisition of L2 grammar. To the best of my knowledge, the acquisition of a particular class of impersonal subjectless constructions in both L2 German and L2 Russian1 has been left unattended so far (but see Sopata 2005 for an important exception as regards L2 German).

In order to compare the vulnerability of different interfaces (syntax-morphology and syntax-discourse in particular, and to a certain extent syntax-semantics), several types of constructions are analysed. Their attainment by second language speakers (L2ers) across four levels of acquisition is assessed. Building on two ad hoc conducted studies, I explore the rate of the L2 acquisition of arguments and voice alternations by measuring passivation and anticausativisation sensitivity, as well as the argument licensing preferences of L2ers. Some of the questions comprising the present discussion ask why language acquirers overuse expletive elements in their L2 German (1.1) or overgeneralise the anticausativisation -SJ(A) morpheme in L2 Russian (1.2):

  Gesternwurde (*es) auf dem Schiff getanzt.
  yesterday was *EXPL on  the    ship   danced
  There was dancing on the ship yesterday.’
  Intended: ‘It is easy to translate this book.’ ← 23 | 24 →

Although the Interface Hypothesis (IH) has been initially proposed in order to describe and analyse proficient end-state grammars (cf. Sorace 2011), I will show that expanding the range of the IH might be helpful to test this hypothesis against findings from several stages of developing interlanguage grammars (ILGs) in order to question the validity of the claim that L2 optionality lies within the syntax-discourse (external) component.

1.1  Addressing the object of enquiry

With the initiated empirical investigation within this book I aim to contribute to the process of fleshing out modern generative grammar while providing novel empirical data from the realm of L2A and suggesting new sources of linguistic explorations at the interfaces. I have chosen the generative approach, since, in my opinion, this framework can best account for the nature and sources of the optionality in L2A.

In the following, I outline the theoretical background and the research questions, which this book is concerned with. These are (i) argument and event structure and argument licensing, (ii) the syntax of impersonal and ‘subjectless’ constructions, and the notion of unaccusativity, and (iii) the second language acquisition thereof. This is done in order to localise the place of the constructions under investigation, with the discussion being elaborated in Chapter 6.

In this regard I deal with the category VOICE, the lowest category of T that enters into Agree relation with little v. As such, the external argument (EA) is licensed by a VOICE-v Agree operation and projected above vP (Kratzer 1996). Such a notion of VOICE (cf. Chapter 2 for a detailed discussion) is crucial for the follow-up analysis of voice alternations and argument licensing. Consider the following sentences:

  a.  John broke the window. (agentive)
  b.  The wind broke the window. (causative)
  c.  The window broke (*by the wind/*by John). (inchoative)
  d.  The window was broken (by the wind/by John). (passive)
  e.  There was dancing on the ship last night. (impersonal existential)

The examples above provide evidence of the existence of different verb classes covered by a single lexical verb “break” or – in the terminology of the present book –, by the root √break. In English, such alternations are to be found with verbs like “melt”, “read”, “sleep”, “dance”, all of which are considered members of the unergative verbal class (cf. Burzio 1986; Perlmutter 1978). Other languages ← 24 | 25 → – especially those with highly grammaticalised aspect, aktionsart, and voice morphology – may come up with different solutions with respect to these voice alternations, consider the German and Bulgarian counterparts of (1.3):

  a.  Hans zerbrach/zerschlug    die Fensterscheibe.(agentive)
  John      broke/crashed        the window
  b.  Der Wind zerbrach die Fensterscheibe.(causative)
  the   wind broke      the window
  c.  Die    Fensterscheibe zerbrach (durch    einen    Windstoß/
  the    window             broke      through a-ACC.M wind gust/
  *durch      Hans).
  *through  John) (inchoative/anticausative)
  d.  Die Fensterscheibe wurde (vom Wind /von Hans)    zerbrochen.
  the        window      was     (by-the wind/by Hans)     broken
  e.  Gestern    nacht wurde auf      dem  Schiff  getanzt.
  yesterday night was      on        the  ship      danced
    (impersonal passive)
  a.  Ivan sčupiprozoreca.(agentive)
  Ivan PFV-broke-3SG.PSTwindow-the-M  
  b.  Vjatârât    sčupi        prozoreca.(causative)
  wind-the  PFV-broke-3SG.PST        window-the-M  
  c.  Prozorecât       se       sčupi  (ot vjatâra /*ot Ivan).
  window-the-M REFL PFV-broke-3SG.PST  by wind-the/*by John
  d.  Prozorecât      beše        sčupen  /
  window-the-M was-3SG.PST PFV-broken-PRTCPL.M/
  se      sčupi                         (ot vjatâra /?ot Ivan).
  REFL PFV-broke-3SG.PST by wind-the/ by John   (passive)
  e.  Minala-ta   nošt imaše                         tanci   na korab-a.
  last-the-F   night have-EXIST.PST      dances on ship-the-M
      (impersonal existential)

There has been much prolific research on voice alternations, unaccusativity, and (anti)causatives within different theoretical frameworks (Alexiadou et al. 2004 and the contributions in the volume; Babby & Brecht 1975; Embick 1998, 2004; Kosta 2010; Levin & Rappaport Hovav 1995; Marantz 2013; Perlmutter 1978; ← 25 | 26 → Schäfer 2008, to mention a few). In this book I deal with various kinds of voice alternations from a cross-linguistic perspective. Crucially, I have adopted the insight that verbal classes (e.g. unaccusatives, unergatives, causatives, etc.) are not associated with particular verbs stored in the lexicon, but with syntactic structures. Thus, verbs per se cannot be classified as ‘unaccusatives’ or ‘unergatives’. Rather, they are the product of unaccusative or unergative generated structures (cf. Marantz 2013 for a relevant discussion; cf. also Rosen 1984 for an important observation that unaccusativity across languages is not uniform).

I largely follow the tenets of contemporary research on argument structure (articulated in the work of Hale & Keyser 1998, 2002), which puts “emphasis on events and the connection between arguments and events, rather than on thematic roles, considered as an ordered set, or on properties of classes of verbs (where a verb is identified/distinguished by the phonological content of its stem or root)” (Marantz 2013: 165). In the present approach, argument structure is divorced from lexicon or any lexically stored information as subcategorisation frames or theta grids (Chomsky 1981 and much subsequent work). As such, my take on impersonal constructions and argument structure is essentially a morpho- syntactic one, based on the ‘syntax-prior-morphology’ approach (Halle & Marantz 1993 and much subsequent work within Distributed Morphology). Assuming that syntax is the only generative machine, I will inevitably come to discuss voice alternations and impersonal constructions as straddling the interfaces, especially with respect to the L2A thereof. Thus, the understanding how impersonals and argument structure in general work is bound to a panoply of various factors – syntactic, morphological, semantic, lexical, pragmatic, and cognitive. And this is the most demanding objective of this book – to show (at least part of) the interaction between syntax proper and the interfaces.

Of course, there is a multitude of impersonal constructions across languages. When discussing impersonals in this book, the following constructions will be addressed: (i) impersonal passives (IMPPASS) and impersonal middles (IMPMDL)/ anticausatives (for L2 German), (ii) adversity (causative) impersonals (AdvIMP), 3PL impersonals, and impersonal anticausatives (for L2 Russian).

As example (1.4e) above demonstrates, no arguments are morpho-syntactically projected in German IMPPASS constructions. This, however, does not mean that they do not involve any semantic agents/causers2. Being a passive construction, ← 26 | 27 → IMPPASS simply does not license [Spec,VOICE]. There is, nevertheless, an [+AG] feature on the VOICE/CAUSE head (cf. the preliminary structure of IMPPASS in Fig. 1.1 and see Chapter 2.2 for a closer analysis):

Figure 1.1:  Inactivated VOICE and [+AG] feature in passives


Furthermore, following Pylkkänen (1999, 2008), the distinction between syntactic subjects/agents and causers is made. Crucially, agents are formally and thematically licensed by VOICE. Thus, in the absence of VOICE we would expect no (syntactic) agents to be licensed. In contrast, causers are those arguments, which are thematically licensed in vP/VP and formally licensed by VOICE. Implementing this distinction, Pylkkänen’s ‘bundling VOICE’ parameter is applied in order to explain the curious facts about Russian adversity impersonals (AdvIMPs), in which ACC-object is licensed in the absence of any NOM argument, whereas the verb takes the non-agreeing, default form ([-AGR], homophonous with 3SG.NEUT):

  a.  Lodkuperevernulo(poryvomvetra).
  The boat was tipped over by the gust.’
  b.  Evropuzamelo(snegom).
  Europe-ACC.F swept-IMP[-AGR].PSTsnow-INSTR.M
  Europe was swept by snow.’

Importantly, the instrumental argument in (1.6) is optional. Despite the translation, this argument is not to be regarded as a by-modifier (the ‘doer’) as in the case of “The letter was written by John”, where John is the implied agent. This modifier is generated below CAUSE (adjoined to vP), as will be shown in §2.3.3.

Beyond doubt, the discussion of impersonals and their computation may shed light on the long-standing puzzle where the generation of external arguments actually takes place, as well as what theoretical bearings this has on the discussion of the universality of semantico-syntactic features (e.g. [+AG]/[+CAUS]), and their accessibility in L2A. ← 27 | 28 →

1.1.1  The two (pivotal) “WHYs” in this book

In this section, the motivation behind the choice of the research topics underlying the present book is explained:

•  Why voice alternations, (impersonal) passives and (un)accusatives?

Within the generative framework, passive constructions have received considerable attention (Baker et al. 1989; Jaeggli 1986; Perlmutter & Postal 1977, to name but a very few). The interest in passives is not a random one, since from a sample of 373 languages, approximately 40% (N=162) have grammaticalised passive constructions (Siewierska 2011). In descriptive terms, the ‘passive’ stands for a configuration of certain properties, such as: (i) non-finite (passive/participle) verbal morphology usually in combination with some sort of auxiliary verb (Haspelmath 1990), (ii) demoted subject leading to suppressed agentive thematic role, (iii) promoted object, leading to NP movement into the ‘subject position’ (ACC→NOM in accordance with Burzio’s Generalisation, Burzio 1986), (iv) possibility of adjoining agentive ‘by’-phrase modifiers, (v) subject control of secondary predicate agreement, and (vi) unaccusatives’ (general) inability to build passives.

Arguably, impersonals (IMPs) and impersonal passives have offered a broad research avenue for typologists, morpho-syntacticians, and semanticists. To say the least, both structural and interpretative potential of IMPs has been largely exploited in the literature from different theoretical perspectives (e.g. Abraham & Leiss 2006; Blevins 2003; Bowers 2002; Cabredo Hofherr 1999; Carnie & Harley 2005; Fischer 2010; Lavine 2005; Mohr 2004; Schäfer 2008; Sigurðsson & Egerland 2009; Sopata 2005, a.m.o.). In Chapter 2, both the morpho-syntactic properties and the derivation of IMPs, passives, anticausatives, and subjectless constructions in the languages under investigation are drawn on. Assuming that external arguments are generated in the VOICE projection above vP and then related to the event can elegantly explain the different kinds of impersonals, passives, and middles across languages.

•  Why L2A?


ISBN (Hardcover)
Publication date
2015 (September)
Interlanguage grammars Syntax and interfaces Generative Grammar
Frankfurt am Main, Berlin, Bern, Bruxelles, New York, Oxford, Wien, 2015. 269 pp., 17 tables, 58 graphs

Biographical notes

Nadia Varley (Author)

Nadia Varley received her PhD in Germanic Linguistics at the University of Wuppertal. Her research focuses on syntax, interface phenomena and second language acquisition. She is also interested in the empirical verification of the theoretical claims within the framework of Generative Grammar.


Title: Optionality and overgeneralisation patterns in second language acquisition: Where has the expletive ensconced «it»self?
book preview page numper 1
book preview page numper 2
book preview page numper 3
book preview page numper 4
book preview page numper 5
book preview page numper 6
book preview page numper 7
book preview page numper 8
book preview page numper 9
book preview page numper 10
book preview page numper 11
book preview page numper 12
book preview page numper 13
book preview page numper 14
book preview page numper 15
book preview page numper 16
book preview page numper 17
book preview page numper 18
book preview page numper 19
book preview page numper 20
book preview page numper 21
book preview page numper 22
book preview page numper 23
book preview page numper 24
book preview page numper 25
book preview page numper 26
book preview page numper 27
book preview page numper 28
book preview page numper 29
book preview page numper 30
book preview page numper 31
book preview page numper 32
book preview page numper 33
book preview page numper 34
book preview page numper 35
book preview page numper 36
book preview page numper 37
book preview page numper 38
book preview page numper 39
book preview page numper 40
272 pages