Project Description

1 State of the art and preliminary work

I. State of the art

A. Functional left peripheries and clause-typing

In standard generative grammar, elements responsible for marking certain discourse-linked formal properties of clauses (clause type, finiteness) are located at the left periphery of the clause (Rizzi 1997, 2004), traditionally associated with the notion of Complementiser Phrases (CPs), and hence also referred to as the CP-domain. Overt functional elements associated with clause types can be complementisers, which are base-generated in the CP-domain (e.g. that, if), and also operators, which move from a clause-internal (thematic) position (e.g. relative operators like which), cf. Chomsky (1977, 1981). Complementisers are responsible for defining the type of the clause, while operators are contingent upon relevant (overt/covert) complementisers that attract them to their specifier positions. Since operators move as phrase-sized constituents, they may take lexical material along with them; while base-generated functional C heads cannot contain lexical material. Though the categorical distinction between (relative) operators and complementisers has been questioned by some authors (e.g. Kayne 2009, 2010a, 2010b), there seems to be ample reason to believe that it still holds (Franco 2012). The distinction is not rigid in the sense that certain operators may be reanalysed as C heads, which is a common grammaticalisation process affecting functional elements diachronically (Hopper and Traugott 1993, Heine and Kuteva 2002, van Gelderen 2004, 2009). The reanalysis of operators into complementisers follows general economy principles, as by way of base-generating an element in the CP-domain, movement from a clause-internal position to a left peripheral position can be avoided (van Gelderen 2004). On the other hand, the appearance of new operators is also common as a way of reinforcement (van Gelderen 2004). In addition to the possible co-occurrence of complementisers and operators, certain languages are known to allow multiple complementisers (as shown e.g. by Roberts 2005 for Welsh). Cartographic approaches were primarily developed for Romance languages (e.g. Rizzi 1997, 2004), where multiple complementisers are generally not allowed, though they can be observed in some present-day and historical dialects (Paoli 2007). West-Germanic languages, however, present ample empirical evidence for the existence of multiple CP layers, since a number of dialects allow the presence of multiple overt elements either historically or in their present-day forms (as pointed out by Bianchi 1999: 182; see e.g. Allen 1997 for Old and Middle English). These constructions and the rich dialectal variation they exhibit pose several questions that a rigid cartographic approach cannot handle (see Bayer and Brandner 2008 on the necessity of a flexible approach to German dialects). For instance, the distinction between a higher and a lower complementiser cannot always be described as Force and Fin (especially in cases where both seem to mark Force, e.g. als wie ‘than as’ in certain dialects of German, see Jäger 2010), and in some cases there are more than two complementisers present in one left periphery (e.g. als wie wo ‘than as where’ in comparative subclauses in Bavarian), which again contradicts the assumption that there are only two designated positions that may host such elements. In sum: while current theories recognise the internal complexity of functional left peripheries, they fail to account for the variation attested in connection with multiple heads, as well as the role of operators in clause-typing; hence further investigation is needed (see hypotheses 1–3 in section 2.2.).

B. Functional left peripheries and information structure

In addition to elements associated with clause-typing, functional left peripheries are known to be able to host other elements as well. As shown by Rizzi (1997, 2004), the CP-domain in Italian may contain focus and (iterable) topics: in his system, these are designated positions to which certain phrases move because they are equipped with features directly linked to information structural notions. The availability of such movement shows variation across languages; languages like English or German do not seem to have designated syntactic projections for topics or focus (the constituent moving to the [Spec,CP] in German V2-clauses is insensitive to information structural properties, see Frey 2005, Fanselow and Lenertová 2011). Hungarian has long been claimed to have a designated focus position (Bródy 1990, 1995; É. Kiss 1998, 2002), which is essentially at the left edge of the verbal domain (hence below the CP-domain); topics may appear higher than the focus and are lower than the CP-domain. Various authors (e.g. Jayaseelan 2001; Belletti 2001, 2004; Poletto 2006) have shown that there is a functional vP-periphery similar to the CP-domain, and some languages have been claimed to have designated projections for topic and focus in this domain (e.g. Mòcheno by Cognola 2012). Though it seems uncontroversial that functional left peripheries may host elements associated with particular information structural notions, several authors have questioned the mainstream view that information structural properties are in a one-to-one correspondence with related syntactic features; under these approaches, movement is rather due to unspecific edge features (Fanselow and Lenertová 2011) or is purely stress-driven, as claimed by Szendrői (2001) for Hungarian focus. Unlike cartographic approaches, such views do not assume there to be a rigid series of specific functional projections in either of the left peripheral domains. Furthermore, focus is not necessarily associated with movement even in languages that may have a left-peripheral position: in Italian, focus is normally realised in situ, which is in fact the only option if it is not contrastive focus but merely new information focus (Frascarelli 2000). In sum: while current theories recognise the availability of IS-related movement to functional left peripheries, they either fail to distinguish it from syntactic movement related to clause-typing, or do not consider the possibility of IS-related movement indirectly feeding the emergence of functional left peripheries; hence further investigation is needed (see hypotheses 4–5 in section 2.2.).

C. Functional left peripheries and ellipsis

Functional left peripheries are also important in terms of ellipsis; one predominant assumption concerning non-constituent ellipsis, formulated by Merchant (2001, 2004), is that ellipsis is carried out by an [E] feature located on a functional head (a C or a v head). The idea is that this feature is inserted in the syntax but it also instructs the phonological component (PF) to elide the complement of the functional head; hence not only elements in the specifier but also in the head itself will be exempt from ellipsis. The availability of the [E] feature is the property of the functional head and is therefore independent from whether there is a fully-fledged functional left periphery in the given domain or not; for instance, VP-ellipsis is claimed to be governed by an [E] feature on a functional v head, though English does not have a functional vP-periphery in the way Hungarian does. There is considerable cross-linguistic variation in terms of which functional heads may host an [E] feature; Merchant (2013) claims that there is also a lexical difference between the [E] feature that can be located on a C head and the one that can be located on a v head, and while English has both of them, German lacks the one that attaches to v heads, hence the absence of English-type VP-ellipsis in German. Moreover, the same kind of ellipsis phenomenon may be associated with different heads in different languages: for instance, sluicing in English is carried out by an [E] feature on the C head, while it is located on a lower functional head in Hungarian, as shown by van Craenenbroeck and Lipták (2006), who identify this head as the Focus head. This suggests that while the absence of a fully-fledged vP-periphery does not exclude the availability of the [E] feature there, the presence of such vP-peripheries may increase the likelihood of [E] features appearing there. In sum: while current theories assume a relation between functional heads and clausal ellipsis, they do not investigate how the [E] feature interacts with the general structure of left peripheries (clause-type markers, IS-related movement) and how this is all related to the directionality of heads; hence further investigation is needed (see hypothesis 6 in section 2.2.).

II. Preliminary work

A. Functional left peripheries and clause-typing

Bacskai-Atkari (2014) examined the structure of comparative subclauses, and argued that there are certain overtness requirements on left peripheral elements (which partially also hold in the nominal domain). On the one hand, the type of the clause must be marked overtly: this role is primarily associated with the complementiser. However, in case there is an overt operator that is unambiguously associated with a clause type, then the complementiser may be zero: this is possible in certain comparative subclauses (e.g. equatives in Hungarian) and in ordinary relative clauses as well. On the other hand, comparative subclauses were shown to contain a comparative operator obligatorily: this operator moves to the specifier of a CP. It was shown that if the operator is overt, then it may take lexical material along with it; if, however, it is phonologically zero, then lexical material is not licensed; the patterns are schematised in (1) below (Op. referring to a zero):

(1) a. … than how tall Peter is. licensed (e.g. Hungarian, certain varieties of English)
b. … than how Peter is tall. licensed (e.g. Hungarian, Czech)
c. … than Op. tall Peter is. universally prohibited
d. … than Op. Peter is tall. licensed (e.g. German, Dutch, Estonian)

Cross-linguistic variation in comparatives can largely be linked to this difference. Consequently, there are phrase positions in the left periphery that are associated with certain functional features that must be overt in case lexical material is also there. The same principle was shown to be holding at the functional left edge of extended nominal expressions as well. As far as the combination of functional left peripheral elements in subordinate clauses is concerned, Bacskai-Atkari (2013a, 2012a, 2012b, 2011) and Bacskai-Atkari and Dékány (2014) argued that with the evolution of the CP-domain in Hungarian, the system of complementisers became enriched and there was an increased need for marking not only the specific type of the clause (e.g. comparative) but also mere finite embedding. Although it is possible to mark these two functions by the same element, it is also possible that a language assigns them to different elements, which occupy distinct positions. The examination of historical data in Hungarian showed that the underlying order of C + C combinations involved the pure subordination marker in the higher position and the element responsible for clause-marking in a lower one, which is exactly the opposite of what one may expect based on the traditional assumption involving ForceP above FinP (Rizzi 1997, 2004).

B. Functional left peripheries and information structure

Bacskai-Atkari (2014, 2013b, 2012c, 2012d, 2010) examined the role of contrast in comparatives expressing inequality: the comparative subclause invariably expresses contrast, which means that prototypically there it contains one overt contrastive element (a phrase-sized constituent or a verb), but it is also possible to have multiple contrasted elements. The position of the element expressing the main contrast involved in a given comparison depends on the distribution of focus in the language, and hence shows cross-linguistic variation (e.g. it appears preferably clause-finally in English and German, as opposed to Hungarian, where it appears at the functional vP-periphery). Comparatives present further evidence that there is no need for postulating designated syntactic positions for focus: the differences attested cross-linguistically correlate with the characteristics of the syntax–prosody interface in the given languages (Bacskai-Atkari 2013c). Furthermore, Bacskai-Atkari and Dékány (2014) examined the diachronic development of embedded interrogatives in Hungarian; it was found that the encoding of clause-typing and of subordination may be associated with the same element (single encoding), such as ob ‘if’ in German and ha ‘if’ in Old Hungarian, or the two functions may be split between two different projections (double encoding), such as the co-occurrence of the complementiser hogy ‘that’ with a wh-pronoun (in wh-questions) or with the interrogative marker -e (in yes-no questions) in Modern Hungarian:

(2) a. Nem tudom, (hogy) megérkezett-e Mari.
    not know-1SG  that PRT-arrived.3SG-Q Mary
    I don’t know if Mary has arrived.
  b. Nem tudom, (hogy) ki érkezett meg.
    not know-1SG  that who arrived.3SG PRT
    I don’t know who has arrived.

It was also shown that even though the interrogative marker -e in Modern Hungarian is in the verbal domain, its role is clearly that of marking the interrogative nature of the clause and is hence not directly related to the notion of focus, contrary to prevalent assumptions in the literature (e.g. van Craenenbroeck and Lipták 2008).

C. Functional left peripheries and ellipsis

Bacskai-Atkari (2014) examined the role of non-constituent clausal ellipsis in the formation of comparative subclauses in English and Hungarian. It was found that otherwise optional processes (such as sluicing or VP-ellipsis) may save constructions from being ungrammatical; this is the case e.g. in Hungarian when the operator fails to move up to the CP-domain, it can be elided from its base position only if a larger unit (including the verb) is deleted (cf. also Bacskai-Atkari and Kántor 2012, 2011). Bacskai-Atkari (2014) argued that ellipsis carried out by an [E] feature located on a functional C or v head is a linear process that operates in a strictly left-to-right fashion. This means that even though the ellipsis domain is the complement of the head equipped with the [E] feature, it is possible that not the entire ellipsis domain is actually deleted as F-marked constituents may stop linear ellipsis. In this sense, the role of contrastive elements is crucial in determining the endpoint of ellipsis, and may lead to the formation of gapping structures in languages like English, where contrastive elements remain within the ellipsis domain. Gapping effects cannot be observed in languages like Hungarian, where contrastive elements undergo fronting and hence are located above the ellipsis domain. Under this view, there is no need to postulate additional rightward movement operations that save F-marked elements from ellipsis; essentially, sluicing and various ellipsis processes affecting the verbal domain are treated in a unified way. In addition, the strict directionality of ellipsis predicts that the complement of a functional head may be elided only if it follows the head, that is, when the given projection is head-initial: this is satisfied for both the CP and the vP in English and Hungarian. However, in German the CP is head-initial but the vP is head-final; German has sluicing but no VP-ellipsis in the way it is attested in English, which can then be tied to the directionality of heads.

1.1 Project-related publications

1.1.1 Articles published by outlets with scientific quality assurance, book publications

Bacskai-Atkari, Julia (2013) Reanalysis in Hungarian Comparative Subclauses. In: Christer Platzack and Valéria Molnár (eds.) Approaches to Hungarian 13: Papers from the 2011 Lund Conference. Amsterdam: John Benjamins. 5–32.

Bacskai-Atkari, Julia (2012) Reducing Attributive Comparative Deletion. The Even Yearbook 10. 1–25.

Bacskai-Atkari, Julia (2012) The Diachronic System of the Left Periphery of Subordinate Clauses in Hungarian. In: Balázs Surányi (ed.) Proceedings of the Second Central European Conference in Linguistics for Postgraduate Students. Budapest: Pázmány Péter Catholic University. 3–23.

Bacskai-Atkari, Julia (2010) Parametric Variation and Comparative Deletion. The Even Yearbook 9. 1–21.

1.1.2 Other publications

Bacskai-Atkari, Julia (2013) On the Syntax–Prosody Mapping in Hungarian Comparatives. In: Péter Szigetvári (ed.) VLlxx: Papers in Linguistics Presented to László Varga on his 70th Birthday. Budapest: Tinta Könyvkiadó. 257–277.

Bacskai-Atkari, Julia (2012) English Comparatives and Parameters. In: Mária Gósy and Attila Péteri (eds.) Tanulmányok: Nyelvtudományi Doktori Iskola. [Essays: Doctoral School for Linguistic Sciences] Budapest: Eötvös Loránd University. 23–37.

2 Objectives and work programme

2.1 Anticipated total duration of the project

6 years (funding requested for 3 years)

2.2 Objectives

The main objective of the project is to provide a flexible, formal model for the syntax of functional left peripheries, which is able to link the appearance of certain functional projections to overtness requirements in general and which may therefore handle the difference among various languages in terms of whether and to what extent functional left peripheries are attested in the CP or the vP domains. As was described in section 1, most analyses of functional left peripheries rely on the cartographic approach developed primarily for Romance languages but West-Germanic languages and Hungarian present serious problems to a strict cartographic approach.

The anticipated total duration of the project is 6 years, and the overarching aim is to contest these existing cartographic approaches by providing an alternative that may be universally applicable. The aim of the first 3 years is to develop the model on the basis of West-Germanic languages (English, German and Dutch) and Hungarian, and to apply it to Romance languages; furthermore, the model will also be counter-checked against Scandinavian languages. The aim of a second three-year period would be to refine the analysis for Scandinavian, and primarily to extend it to other languages, most importantly to Slavic and Finno-Ugric languages.

The first three-year period builds on two basic assumptions. First, it is assumed that languages may have several ways of encoding certain properties, such as distinctive intonation patterns, and the hypothesis is that these interact with (purely) syntactic means of encoding, either in the sense that syntactic encoding is used in the absence of these additional ways, or that there are multiple ways operating at the same time, and hence reinforcing each other. Second, it is also assumed that even though IS-related movement and the [E] feature responsible for clausal ellipsis do not define functional left peripheries, both of these are crucial in terms of understanding the exact structure of functional left peripheries: functional left peripheries may host phrases moving via IS-related movement, and functional heads may be responsible for carrying out clausal ellipsis.

The project will be focussing on West-Germanic languages and Hungarian; cartographic approaches were first developed for Romance languages but Germanic languages and Hungarian are known to allow constructions involving multiple complementisers (with considerable differences among the individual languages and language varieties). Therefore these languages and especially the rich variation attested in connection with them presents considerable challenges to syntactic analyses of functional left peripheries, and the project aims at constructing a model that may successfully overcome these challenges. The reason why Hungarian is also taken into account is that Hungarian tends to have the same kinds of embedded clauses that are attested in Germanic languages, yet its behaviour in terms of sentential stress assignment and focus shows a markedly different pattern: hence any hypothesis set up for West-Germanic languages can easily be checked against a different language, and these differences can then be linked to more general rules that apply in the respective languages.

I. Functional left peripheries and clause-typing

The project aims at developing a formal syntactic model for functional left peripheries, based on the assumption that the primary role of left peripheries is to host elements that contain discourse-related information regarding the clause, such as information about the clause type, as well as whether the clause is embedded. These properties must be recoverable, that is, the final utterance should contain information that makes these properties perceptible. This can be achieved via various means, such as a distinctive intonation or word order, which are therefore also cues in terms of perception. Left-peripheral functional heads constitute one type of these cues:

Hypothesis 1: Left-peripheral functional heads are syntactic cues.

Hypothesis 1 is taken as an established background assumption that does not need further investigation during the project; however, it establishes the framework that serves as the basis of the model as it explicitly rejects the idea that left-peripheral functional heads would directly encode information structural properties. Functional heads such as C and v are (i) potential phase heads and (ii) potential overt markers. Phase heads are necessary for the structure to be sent to the PF and LF interfaces but this does not need to be carried out by overt heads and hence not all C/v heads are overt. On the other hand, not all C/v heads are phase heads: if there are multiple CP/vP layers, it is only the highest one that counts as a phase boundary. Overtness is related to these functional heads acting as syntactic cues, that is, as elements that mark the type of the clause; although clause-typing (and subordination) is defined by functional heads, the overt marker does not have to be a head but may be an element moving to the specifier position of that head:

Hypothesis 2: Operators moving to the left periphery are also syntactic cues.

Hypothesis 2 complements the fundamental idea expressed by hypothesis 1. Hypothesis 2 needs to be investigated inasmuch as a solid formal approach is needed for the differentiation and relation between functional heads and operators. The separation of syntactic position (that is, whether an element is a head or a specifier) and function may account for how certain clauses (e.g. relative clauses in several languages) are primarily distinguished by operators and not by C heads, without merging the category of complementisers and operators. It is also hypothesised that there is no one-to-one correspondence between syntactic cues and the properties to be encoded: an element may be equipped with several features and may conversely mark them in an unambiguous way, while it is also possible that a certain property is encoded by multiple cues (cf. also Sobin 2002). This is essentially an interaction between economy and reinforcement:

Hypothesis 3: Functional left peripheries are defined by economy and overtness requirements.

Hypothesis 3 needs to be investigated, since it goes against the standard cartographic assumptions (as described in section 1), in that the number of functional projections in a given left peripheral domain is not predefined but can be extended if there are more elements to be accommodated; the minimum number is one, since the phase must be spelt out, but the actual number may vary across languages and constructions. That is, a given language may have split left peripheries in certain constructions but not in others. On the one hand, a structure is transparent if there are overt syntactic markers, and this may be reinforced by the presence of multiple elements (e.g. the combination als wie ‘than as’ in comparatives in certain dialects of German). On the other hand, economy disfavours the presence of multiple projections and the co-presence of an overt head and overt material equipped with similar features in its specifier (a.k.a. the Doubly Filled COMP Filter). Hence if there is both an overt functional head and an operator to be accommodated, they either have to be located in the same phrase (as e.g. in Middle English relative clauses) or there are multiple projections (as e.g. in Old/Middle Hungarian relative clauses). The choice between single and multiple marking, as well as the choice between the types of multiple marking, should ideally be linked to more general properties of the language or to the lexical properties of the items involved. The size of functional left peripheries varies ultimately according to how syntactic features are distributed among overt functional elements.

II. Functional left peripheries and information structure

Functional left peripheries may host elements such as topics and focus; it is hypothesised that phrases with special information structural properties behave differently from phrases that have functional operators related to clause-typing:

Hypothesis 4: Functional left peripheries may be extended but not defined by IS-related movement.

Hypothesis 4 needs to be investigated as there are conflicting views regarding the question. First, hypothesis 4 goes against cartographic approaches, according to which IS-related movement targets designated positions that are integral parts of left peripheries. Hypothesis 4 is thus in accordance with some recent alternative analyses of IS-related movement; however, the novelty underlying hypothesis 4 is that it aims at re-addressing the relation between the syntactic structure of purely functional projections in left peripheries and IS-related movement. It is assumed that CPs and vPs may project additional layers that are equipped with an unselective edge feature in the sense of Fanselow and Lenertová (2011) and the specifier of these projections may host phrases that are associated with particular information structural notions for independent reasons. It is supposed that there are no zero designated topic or focus heads (which does not exclude the possibility of overt ones in certain languages, e.g. the focus head in Gungbe, see Aboh 1999). Unlike complementisers and functional operators, these elements are not syntactic cues in terms of clause-typing. On the other hand, a given phrase may be related to both categories: for instance, wh-phrases explicitly mark the interrogative nature of the clause and they tend to receive focus interpretation as well. In languages that do not (typically) exhibit IS-related movement, this makes no difference in the marking of clause-typing; this can be observed e.g. in English, where wh-elements move regularly to the CP-domain and the marking of [+wh] is encoded in the CP-periphery. In languages that seem to have designated left peripheral positions for e.g. focus, such as in Hungarian, this may result in conflicting requirements in terms of the position of these constituent: in Hungarian interrogatives, the wh-phrase is located at the vP-periphery (traditionally identified as the FocusP, see É. Kiss 2002). Hence though syntactic markers associated with clause-typing are located in the CP-domain in Hungarian, the [+wh] feature (as opposed to the [+rel] feature) is located in the vP-periphery; this prediction seems to be borne out in embedded yes-no questions as well, where the interrogative marker is clearly below the CP-domain (cf. van Craenenbroeck and Lipták 2006). This in turn predicts that IS-related movement (which is not actually driven by IS-related features) has some bearing on whether a language has a certain (lower) functional left periphery or not:

Hypothesis 5: IS-related movement feeds the emergence of functional left peripheries.

While hypothesis 4 is relevant to the internal structure of a given left periphery, hypothesis 5 primarily concerns the relation between the CP and the vP-periphery. Hypothesis 5 needs to be investigated because it implies a cross-linguistic reconsideration of which languages have both a CP and a fully-fledged vP-periphery and to what extent (leftward) IS-related movement is attested. Under this view, the fact that a given language splits certain functions between two left peripheries, and/or has fully-fledged left peripheries in a certain domain, is related to the encoding of information structural properties in an indirect way, in that movement relevant for the syntax–prosody interface and purely syntactic movement are related in terms of economy: the encoding of certain syntactic properties may be harmonised with e.g. the requirements concerning the position of focussed constituents, thus avoiding either extra movement operations or stress shift. Again, the appearance of multiple left peripheries in itself works against economy principles, and hence is expected to appear in languages only if there is additional reason for it.

III. Functional left peripheries and ellipsis

The notion of functional left peripheries is intertwined with directionality: left peripheries arise if the relevant functional projections are head-initial, in which case there is a series of functional heads (and phrases in their specifiers) that are adjacent in the linear structure, and the complement of the lowest functional head (e.g. the TP complement of a C head or a lexical vP/VP complement of a functional v head) follows the entire periphery. This does not exclude the possibility of head-initial phrases dominating categorially non-distinct head-final phrases: only the reverse configuration is ruled out (this is the Final-over-Final Constraint, see Holmberg 2000; Biberauer, Holmberg and Roberts 2008; Biberauer, Newton and Sheehan 2009). However, functional left peripheries arise if there are multiple head-initial functional projections in the relevant domain; this is presumably related to the fact that syntactic movement is universally leftward (Kayne 1994). The project aims at investigating the difference between head-initial and head-final functional projections in terms of left peripheries, and at showing that the difference is truly one related to directionality and not to additional properties (such as extra movement operations). It is also hypothesised that while only functional left peripheral heads (that is, C and functional v heads) may host the [E] feature responsible for clausal ellipsis, there is no lexical difference between an [E] feature that attaches to a C head and one that attaches to a v head and the availability of ellipsis related to one type of functional head depends on whether the given functional projection is head-initial or not:

Hypothesis 6: Restrictions on the appearance of [E] are not related to the lexical properties of [E].

Hypothesis 6 needs to be investigated because it is contrary to more or less standard assumptions, and it aims at linking ellipsis to general, predictable properties of a given language rather than to idiosyncratic, unpredictable properties of certain items. If an [E] feature appears in a functional left periphery, then it appears as high as possible since ellipsis targets the maximal largest unit (see Merchant 2008). It is also hypothesised that the [E] feature does not contain information with respect to the endpoint of ellipsis but that it is F-marked elements that constitute the endpoint at PF, since F-marked material cannot be elided. Hence ellipsis carried out by an [E] feature on a functional head may affect the entire complement domain (e.g. in sluicing in English) or only part of it (e.g. in verb gapping constructions in English, where the F-marked object DP stops ellipsis). However, if the complement precedes the head, as in head-final functional projections, then ellipsis cannot be carried out retrospectively; this accounts for the fact that German has no VP-ellipsis in the English way since vP/VP projections are head-final (see Haider 1993), while it has sluicing in the English way, since sluicing is connected to the CP-domain in both languages. Crucially, the hypothesis that non-constituent clausal ellipsis is constrained by linear order has two predictions in terms of functional left peripheries, which the project aims at investigating in West-Germanic and Hungarian in detail. On the one hand, this kind of ellipsis should be available only for head-initial projections, the condition of which is met in (Modern) Hungarian and in most Germanic languages but not in certain domains in certain languages, such as the vP-domain in German and in Dutch. Languages that have verb-final clauses in subordinate structures but not in main clauses (e.g. V2 order in German) are therefore also predicted to have differences in terms of the elimination of the verb, since the verb may be elided if the [E] feature is on a C head but not when it is on a v head. On the other hand, the fact that certain ellipsis types are connected to different functional domains in different languages can be explained by the differences in the extent to which functional left peripheries are available in a given language. For instance, sluicing is usually connected to the CP-domain since the wh-element moves there (e.g. in English or German), while in languages that locate the same elements lower (that is, in the vP-periphery), sluicing is carried out by an [E] feature that is located lower. Again, the mechanism underlying these processes is essentially the same, and the reason behind the differences follows from more general settings connected to the left peripheries in the given languages.

2.3 Work programme incl. proposed research methods

applicant: Bacskai-Atkari, Julia


Theoretical work will be continuous, building on already existing data and running in parallel with collecting and evaluating new data; the focus of the theoretical investigations will be shifted according to the major issues outlined in section 2.2. Phase 1 concentrates on how functional left peripheries are defined by functional heads and operators. Phase 2 investigates how these peripheries may host constituents moving via IS-related movement, either via an edge feature located on a functional head, or (for certain topics) via adjunction. Phase 3 examines how clausal ellipsis is indicative of the position of left-peripheral elements, and how it can be conditioned purely by directionality and the prohibition on eliding non-recoverable material.

Phase 1: Functional left peripheries and clause-typing

A. General considerations

The overall aim of this phase is to develop a flexible model of functional left peripheries that contains a minimal number of projections and that enables the cross-linguistic variation attested in connection with the location of clause-type markers in the CP-domain and the vP-domain. Of particular importance are complementiser combinations, as these explicitly show that there are multiple functional layers associated with clause-typing; the main issue to be investigated here is when and why certain syntactic cues are split between different functional heads. This will involve the direct comparison of the same kind of construction (e.g. comparative subclauses or embedded interrogatives) across languages (and, if applicable, dialects); since in most cases the data are already available in the relevant literature, experiments are needed only in a limited number of cases (i.e., the left periphery of comparative subclauses in selected German dialects and in Dutch). There are three major types of combinations: combinations that involve the co-presence of two complementisers that are associated with two different clause types (e.g. als ob ‘as if’ in German), combinations that involve the co-presence of a pure subordination marker and a clause-type marker (e.g. als dass ‘than that’ in German), and combinations that involve the co-presence of two complementisers with essentially the same function (e.g. als wie ‘than as’ in certain German dialects). Contrary to the cartographic approach (as given in e.g. Rizzi 1997, 2004), the goal is to establish a model that accounts for why certain languages and language varieties apply double encoding in cases where other languages have single encoding, and how the order of these elements can be predicted, given that a system based on a pre-established ForceP and FinP is not favourable. The investigation of multiple projections involves cases where there are operators moving to specifier positions, since these operators may also overtly mark the type of the clause; furthermore, given that the features of these operators are similar to those of the related complementisers, the investigation intends to clarify which present-day complementisers have most probably been reanalysed from original operators historically, and whether there are potential candidates for similar changes in present-day dialects (e.g. wo ‘where” in certain dialects of German). Since there are generally more overt left-peripheral elements in subordinate clauses, the research will chiefly focus on these. However, it is a key concern to compare all subordinate clause types to their closest main clause counterpart; for instance, interrogative clauses are expected to be different when they are embedded, due to the fact that subordination has to be marked, while the intonation pattern characteristic of main clauses is not available, which may induce differences also in terms of marking the interrogative nature of the clause. Special attention will be devoted to comparative subclauses, where both complementisers and operators tend to be available, and the operators may also carry lexical phrases. Apart from intra-peripheral issues, the investigation intends to address the problem of how various clause-type markers are distributed between the CP-periphery and the vP-periphery across languages and constructions.

B. Languages to be studied

As the combinations of complementisers (and operators) are more widespread in West-Germanic languages than in Romance, the research focuses on West-Germanic languages. In addition, Hungarian is used as a contrast: Hungarian has essentially the same types of subordinate clauses and associated left-peripheral elements, yet it is an unrelated language that is also expected to show different behaviour, given that clause-type markers may be located below the CP-domain. Special focus will be paid to English and German, where complementiser combinations have extensively been studies and hence some diachronic changes may also be taken into consideration. Apart from Standard German, selected German dialects will also be investigated (Alemannic and Hessian), especially in terms of comparative subclauses. The investigation will also consider Dutch, where multiple CP projections are widely attested; Scandinavian languages will be used to a limited extent allowing for comparison with West-Germanic.

C. Methods

Apart from widely known cases, the research also aims at testing the acceptability of certain complementiser and operator combinations that are either less frequent in usage, or are non-standard and therefore not sufficiently discussed in the literature; this includes cases showing intra-speaker and/or inter-speaker (possibly dialectal) variation. There will be three experiments involving grammaticality judgements (focussing on multiple functional elements in comparative subclauses); one in Dutch, and two in German (the German experiments directed at two specific dialect areas, Hessian and Alemannic). The experiment on Dutch will be preceded and complemented by a questionnaire directed at the basic patterns related to clause-typing, and its relation to information structure and ellipsis (similar data are already available for (Standard) English, Standard German and Hungarian). In order to gain a broader perspective and comparable results, there will be a questionnaire for Scandinavian languages as well. Furthermore, as there is considerable diachronic variation attested in German and also in Hungarian, historical data will also be consulted, by using the following corpora: Titus, Bonner Frühneuhochdeutsch-Korpus, Bochumer Korpus des Mittelhochdeutschen (German), Old Hungarian Concordance (Hungarian; note that this corpus is still being built at the Hungarian Academy of Sciences, but a large part of it is already normalised or morphologically annotated).

Phase 2: Functional left peripheries and information structure

A. General considerations

This phase aims at refining the model achieved by the previous phase, in that it considers the status and positions of other constituents that may move to functional left peripheries via IS-related movement. Since it is hypothesised that IS-related movement is driven by unspecified edge features (in the sense of Fanselow and Lenertová 2011), as opposed to specific syntactic features such as [+wh] or [+rel], the main question is where these elements may be located in a left periphery otherwise defined by designated functional elements: that is, under what circumstance may IS-related movement place elements in between designated functional heads, as opposed to placing them at either edge of the entire peripheral domain. Since there are more functional elements in embedded clauses, given that subordination has to be marked and intonation is not necessarily available in the same way as it is in main clauses, particular attention will be paid to IS-related movement in subclauses, with a comparison to main clause counterparts. Subclauses are expected to show a more intricate pattern due to the increased availability multiple functional markers: subordination markers do not appear in main clauses and certain clause-type markers are likewise not attested (or are only optionally attested) in main clauses. The research will take only finite clauses into account, since non-finite structures do not contain the same layers, which is due to reasons independent from the ones distinguishing main clauses and subclauses. The difference between main clauses and subclauses can be well observed in the case of the interrogative marker -e in Hungarian, which is obligatory in embedded clauses like in (3b) and optional in main clauses like in (3a); in addition, the embedded clause may also contain the overt complementiser hogy ‘that’, which is naturally prohibited in main clauses:

(3) a. Megjött(-e) már Mari?
    PRT-arrived.3SG-Q already Mary
    ‘Has Mary arrived yet?’
  b. Azt kérdeztem, (hogy) megjött*(-e) már Mari.
    that-ACC asked-1SG  that PRT-arrived.3SG-Q already Mary
    ‘I asked if Mary had arrived yet.’

Particular attention will be paid to the position of focus and the evolution of functional left peripheries: in languages where the focus of the sentence is associated with a particular left peripheral position (e.g. Hungarian), the role of IS-related movement is crucial and the model has to capture the difference between languages where IS-related movement is fully optional and ones where the lack of IS-related movement (the movement of the focussed element) results in an infelicitous structure. Furthermore, fully-fledged functional vP-peripheries are expected to appear if IS-related movement canonically targets that domain of the sentence, which is expected to be the case if sentential stress is also assigned there. The investigation will particularly be concerned with comparative structures, which obligatorily contain at least one contrastive element, the position of which can be understood with respect to other potential contrastive elements, non-contrastive GIVEN elements, and the left-peripheral functional projections themselves.

B. Languages to be studied

Building on the results of the previous phase, the main focus here again will be on West-Germanic languages and Hungarian. Of particular importance is German, which is known to have V2 in main clauses, and a main clause CP is hence able to host various kinds of phrases. Movement to the left periphery shows differences in subclauses and main clauses in German; the same differences are not expected to appear in Scandinavian languages, where V2 order is widespread in subclauses as well (see Kroch et al. 2000). The investigation will concentrate only on issues directly related to functional elements in the left periphery and does not address general questions of word order in these languages. In addition to West-Germanic languages, there will be special attention devoted to Hungarian, which seems to have a designated position for focus in the left periphery, and where topics seem to be located in between the CP-periphery and the vP-periphery.

C. Methods

Apart from relying on well-established assumptions concerning information structural properties of the various languages, this phase will make use of the questionnaire data provided in Phase 1 regarding the position of elements moving via IS-related movement to functional left peripheries. In particular, the position of contrastive and non-contrastive elements in Dutch and Hungarian comparative subclauses has to be clarified; grammaticality judgements are already available for Hungarian, and the Dutch data will be provided by Phase 1 (especially the questionnaire, but partly also the experiment). Since the main aim is not to provide an (alternative) analysis for the encoding of information structure in any of the selected languages in particular, but to link IS-related movement to functional left peripheries, the investigation will necessarily rely on general methods of language comparison, and will rather be concerned with how and why the same type of clauses may show variation among languages and language varieties.

Phase 3: Functional left peripheries and ellipsis

A. General considerations

The aim of the last phase is to implement the findings concerning the model of functional left peripheries in the investigation of clausal ellipsis processes, and to provide evidence that clausal ellipsis is tied to the existence of a functional left periphery but does not require the availability of IS-related movement to these peripheries. Under this view, instances of clausal ellipsis are essentially governed by the same [E] feature on a functional head, either a C head or a functional v head, and gapping effects result from independent properties of a given language, in that gapping effects are expected to appear in languages where contrastive elements tend to appear clause-finally. It is also assumed that ellipsis is a strictly directional process, which proceeds at the PF interface linearly from the left to the right; in this way, only head-initial functional projections may be responsible for ellipsis, since the elimination of the complement is not possible in a head-final projection, where the head follows the ellipsis domain. The availability of V2 orders is also expected to play a role in the derivation of the final structures, in that constructions that have verb movement out of a head-final vP-domain may show gapping effects, while ones that do not apply verb movement do not. There will be particular attention paid to the derivation of comparative subclauses, since a typical comparative subclause is reduced to a single contrastive phrase and the rest of the clause is elided, as in (4a), hence ellipsis plays a crucial role in their derivation, which can further be investigated in the case of multiple contrastive elements, as in (4b):

(4) a. Mary bought more dogs than Peter bought x-many dogs.
  b. Mary bought more dogs than Peter bought x-many cats.

The hypothesis is that all functional C and v heads are able to host the [E] feature responsible for ellipsis, hence the [E] feature is not a lexical property of certain functional heads and is not linked to special positions either; rather, the unavailability of ellipsis in certain cases is expected to be the result of more general requirements, such as the prohibition on eliding F-marked material.

B. Languages to be studied

As this phase builds on the results of the previous two phases, the focus will be on West-Germanic languages and Hungarian. As far as West-Germanic languages are concerned, English and German will be of particular interest, and though others will also be considered, especially in terms of V2 effects, the scope of the present research cannot include a detailed investigation of all the languages in question. The reason why English and German are in focus is that ellipsis processes have widely been discussed in connection with these languages, and the present research project aims at reconsidering these findings in terms of strict directionality. German is especially important since the vP-domain contains head-final projections, while CPs are head-initial, and hence the effect of directionality can be studies within one language. Finally, Hungarian is of special interest since in this language the focussed constituent moves to the vP-periphery and is hence located higher than the [E] feature, which is expected to result in the lack of gapping effects.

C. Methods

Since the chief concern of this phase is to examine the acceptability of remnant structures produced by clausal ellipsis, the theoretical conclusions will primarily be drawn upon grammaticality judgements (data are already available in the literature for English and German; the Dutch data come from the literature and from questionnaires. However, testing is inevitable in the case of Hungarian, where only certain clausal ellipsis phenomena have been investigated (e.g. sluicing by van Craenenbroeck and Lipták 2006). This will involve one experiment (grammaticality judgement test). The focus will be on ellipsis in subordinate clauses, with the aim of testing whether all left-peripheral functional v heads are able to host the [E] feature, and to what extent locating the [E] feature as high as possible is a non-violable constraint that rules out the convergence of non-contrastive elements right above the ellipsis domain.


The 3 years of the project correspond to the 3 major phases described above (12 months each), which can be further divided into smaller periods, which will be devoted to the examination of more specific issues within the individual phases.

Period Issues to be investigated
Phase 1, m. 1–3 ● clause-typing in English and German
● complementiser combinations in German – synchronically and diachronically
● instances of reanalysis from operator into complementiser in English and German
Phase 1, m. 4–6 ● multiple CP projections in German and Dutch – a special focus on comparatives
● experiments: 2 German (Hessian, Alemannic),1 Dutch
● questionnaires (Dutch, Scandinavian languages)
Phase 1, m. 7–9 ● functional left peripheries in West-Germanic – main clauses and subclauses
● comparison with clause-typing in Scandinavian
● developing a flexible minimalist model
Phase 1, m. 10–12 ● functional left peripheries in Hungarian synchronically and diachronically
● the split between the CP-periphery and the vP-periphery
● a comparison between West-Germanic and Hungarian – a cross-linguistic model
Phase 2, m. 1–3 ● the typical positions of topics and foci in English and German
● instances of IS-related movement
● differences between main clauses and subordinate clauses
Phase 2, m. 4–6 ● comparison with the typical positions of topics and foci in Scandinavian
● IS-related movement to left peripheries in West-Germanic
● the position of topics and foci in functional left peripheries
Phase 2, m. 7–9 ● a flexible model of functional left peripheries instances of IS-related movement
● edge features versus specific syntactic features
● a comparison with Hungarian – main and subordinate clauses
Phase 2, m. 10–12 ● functional left peripheries in Hungarian –sentential stress and movement
● the role of the topic field between the two left peripheries
● functional left peripheries and IS-related movement
Phase 3, m. 1–3 ● ellipsis in English and German
● gapping as VP-ellipsis; the similarities of VP-ellipsis and sluicing
● modelling ellipsis as a strictly directional process
Phase 3, m. 4–6 ● ellipsis and the vP-periphery cross-linguistically
● developing a unified approach to ellipsis in the CP-domain and in the vP-domain
● experiment: Hungarian (ellipsis in embedded contexts)
Phase 3, m. 7–9 ● ellipsis and the role of prosody – ellipsis in comparatives in West-Germanic
● outlook: other Germanic languages (based on the questionnaires from Phase 1)
● the position of contrastive elements in Hungarian versus West-Germanic
Phase 3, m. 10–12 ● ellipsis and directionality
● head-final versus head-initial projections and ellipsis
● main clauses vs. subordinate clauses and ellipsis – a special focus on German

3 Bibliography

Aboh, Enoch Oladé (1999) From the Syntax of Gungbe to the Grammar of Gbe. PhD dissertation. Geneva: University of Geneva.

Allen, Cynthia (1977) Topics in Diachronic English Syntax. PhD dissertation. Amherst, MA: University of Massachusetts.

Bacskai-Atkari, Julia (2010) Parametric Variation and Comparative Deletion. The Even Yearbook 9. 1–21.

Bacskai-Atkari, Julia (2011) A komparatív operátor esete a mondatbevezetővel: Szintaktikai változások a magyar hasonlító mellékmondatokban. [The Comparative Operator’s Affair with the Complementiser: Syntactic Changes in Hungarian Comparative Subclauses] In: Katalin É. Kiss and Attila Hegedűs (eds.) Nyelvelmélet és diakrónia [Linguistic Theory and Diachrony]. Budapest & Piliscsaba: Szent István Társulat. 103–119.

Bacskai-Atkari, Julia (2012a) The Diachronic System of the Left Periphery of Subordinate Clauses in Hungarian. In: Balázs Surányi (ed.) Proceedings of the Second Central European Conference in Linguistics for Postgraduate Students. Budapest: Pázmány Péter Catholic University. 3–23.

Bacskai-Atkari, Julia (2012b) A magyar összetett kötőszók történetéhez. [On the History of Hungarian Complex Complementisers] In: Andrea Parapatics (ed.) FÉLÚTON 7: A 7. Félúton konferencia (2011) kiadványa. [FÉLÚTON 7: The Issue of the 7th Félúton Conference (2011)] Budapest: ELTE Nyelvtudományi Doktori Iskola. 1–14.

Bacskai-Atkari, Julia (2012c) Reducing Attributive Comparative Deletion. The Even Yearbook 10. 1-25.

Bacskai-Atkari, Julia (2012d) English Comparatives and Parameters. In: Mária Gósy and Attila Péteri (eds.) Tanulmányok: Nyelvtudományi Doktori Iskola. [Essays: Doctoral School for Linguistic Sciences] Budapest: Eötvös Loránd University. 23–37.

Bacskai-Atkari, Julia (2014) The Syntax of Comparative Constructions: Operators, Ellipsis Phenomena and Functional Left Peripheries. Potsdam: Universitätsverlag Potsdam.

Bacskai-Atkari, Julia (2013a) Reanalysis in Hungarian Comparative Subclauses. In: Christer Platzack and Valéria Molnár (eds.) Approaches to Hungarian 13: Papers from the 2011 Lund Conference. Amsterdam: John Benjamins. 5–32.

Bacskai-Atkari, Julia (2013b) Az Attributív Komparatív Törlésről [On Attributive Comparative Deletion]. In: Zsuzsanna Gécseg (ed.) LingDok 12: Nyelvészdoktoranduszok dolgozatai [LingDok 12: Papers Written by PhD Students of Linguistics]. Szeged: Szegedi Tudományegyetem Nyelvtudományi Doktori Iskola. 9–33.

Bacskai-Atkari, Julia (2013c) On the Syntax–Prosody Mapping in Hungarian Comparatives. In: Péter Szigetvári (ed.) VLlxx: Papers in Linguistics Presented to László Varga on his 70th Birthday. Budapest: Tinta Könyvkiadó. 257–277.

Bacskai-Atkari, Julia and Éva Dékány (2014) From Non-finite to Finite Subordination: The History of Embedded Clauses. In: Katalin É. Kiss (ed.) The Evolution of Functional Left Peripheries in Hungarian Syntax. Oxford: Oxford University Press. 148–223.

Bacskai-Atkari, Julia and Gergely Kántor (2012) Deletion in Hungarian, Finnish and Estonian Comparatives. Finno-Ugric Languages and Linguistics 1:1–2. 44–66.

Bacskai-Atkari, Julia and Gergely Kántor (2011) Elliptical Comparatives in Finnish, Estonian, and Hungarian. In: Sándor Csúcs et al. (eds.) Congressus XI. Internationalis Fenno-Ugristarum Pars VI: Dissertationes symposiorum ad linguisticam. Piliscsaba: Reguly Társaság. 234–241.

Bayer, Josef and Ellen Brandner (2008) On Wh-Head-Movement and the Doubly-Filled-Comp Filter. In: Charles B. Chang and Hannah J. Haynie (eds.) Proceedings of the 26th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project. 87–95.

Belletti, Adriana (2001) Inversion as Focalization. In: Aafke C. J. Hulk and Jean-Yves Pollock (eds.) Subject Inversion in Romance and the Theory of Universal Grammar. Oxford: Oxford University Press. 60–90.

Belletti, Adriana (2004) Aspects of the Low IP Area. In: Luigi Rizzi (ed.) The Structure of CP and IP: The Cartography of Syntactic Structures, Volume 2. Oxford: Oxford University Press. 16–51.

Bianchi, Valentina (1999) Consequences of Antisymmetry: Headed Relative Clauses. Berlin: Walter de Gruyter.

Biberauer, Theresa, Anders Holmberg and Ian Roberts (2008) Disharmonic Word-order Systems and the Final-over-Final-Constraint (FOFC). In: Antonietta Bisetto and Francisco Barbieri (eds.) Proceedings of XXXIII Incontro di Grammatica Generativa. Bologna: Dipartimento di Lingue e Letterature Straniere Moderne. 86–105.

Biberauer, Theresa, Glenda Newton and Michelle Sheehan (2009) Limiting Synchronic and Diachronic Variation and Change: The Final-over-Final Constraint. Language and Linguistics 10.4. 701–743.

Bródy, Michael (1990) Some Remarks on the Focus Field in Hungarian. UCL Working Papers in Linguistics 2. 201–225.

Bródy, Michael (1995) Focus and Checking Theory. In: István Kenesei (ed.) Approaches to Hungarian 5: Levels and Structures. Szeged: JATE. 31–43.

Chomsky, Noam (1977) On WH-Movement. In: Peter Culicover et al. (eds.) Formal Syntax. New York: Academic Press. 71–132.

Chomsky, Noam (1981) Lectures on Government and Binding: The Pisa Lectures. Dordrecht: Foris.

Cognola, Federica (2012) The Mixed OV/VO Syntax of Mòcheno Main Clauses: On the Interaction between High and Low Left Periphery. Available at: (Last accessed: 20 February 2015).

Craenenbroeck, Jeroen van and Anikó Lipták (2006) The Crosslinguistic Syntax of Sluicing: Evidence from Hungarian Relatives. Syntax 9.3. 248–274.

Craenenbroeck, Jeroen van and Anikó Lipták (2008) On the Interaction between Verb Movement and Ellipsis: New Evidence from Hungarian. In: Charles B. Chang and Hannah J. Haynie (eds.) Proceedings of the 26th West Coast Conference on Formal Linguistics. Somerville, MA: Cascadilla Proceedings Project. 138–146.

É. Kiss, Katalin (1998) Identificational Focus versus Information Focus. Language 74. 245–273.

É. Kiss, Katalin (2002) The Syntax of Hungarian. Cambridge: Cambridge University Press.

Fanselow, Gisbert and Denisa Lenertová (2011) Left Peripheral Focus: Mismatches between Syntax and Information Structure. Natural Language & Linguistic Theory 29.1. 169–209.

Franco, Ludovico (2012) Complementizers are not (Demonstrative) Pronouns and viceversa. Available at: (Last accessed: 20 February 2015).

Frascarelli, Mara (2000) The Syntax–Phonology Interface in Focus and Topic Constructions in Italian. Dordrecht: Kluwer.

Frey, Werner (2005) Zur Syntax der linken Peripherie im Deutschen. In: Franz Josef d’Avis (ed.) Deutsche Syntax: Empirie und Theorie. Göteborg: Acta Universitatis Gothoburgensis. 147–171.

Gelderen, Elly van (2004) Grammaticalization as Economy. Amsterdam: John Benjamins.

Gelderen, Elly van (2009) Renewal in the Left Periphery: Economy and the Complementiser Layer. Transactions of the Philological Society 107:2. 131–195.

Haider, Hubert (1993) Deutsche Syntax – generativ: Vorstudien zur Theorie einer projektiven Grammatik. Tübingen: Gunter Narr Verlag.

Heine, Bernd and Tania Kuteva (2002) World Lexicon of Grammaticalization. Cambridge: Cambridge University Press.

Holmberg, Anders (2000) Deriving OV Order in Finnish. In: Peter Svenonius (ed.) The Derivation of VO and OV. Oxford: Oxford University Press. 123–152.

Hopper, Mike and Elizabeth Traugott (1993) Grammaticalization. Cambridge: Cambridge University Press.

Jayaseelan, Karattuparambil (2001) IP-internal Topic and Focus Phrases. Studia Linguistica 55.1. 39–75.

Jäger, Agnes (2010) Der Komparativzyklus und die Position der Vergleichspartikeln. Linguistische Berichte 224. 467–493.

Kayne, Richard (1994) The Antisymmetry of Syntax. Cambridge, MA: MIT.

Kayne, Richard (2009) Antisymmetry and the Lexicon. Linguistic Variation Yearbook 8. 1–31.

Kayne, Richard (2010a) Some Preliminary Comparative Remarks on French and Italian Definite Articles. In: Richard Kayne (ed.) Comparison and Contrasts. Oxford: Oxford University Press. 3–28.

Kayne, Richard (2010b) Why Isn’t This a Complementizer? In: Richard Kayne (ed.) Comparison and Contrasts. Oxford: Oxford University Press. 190–227.

Kroch, Anthony, Ann Taylor and Donald Ringe (2000) The Middle English Verb-second Constraint: A Case Study in Language Contact and Language Change. In: Susan C. Herring et al. (eds.) Textual Parameters in Older Languages. Amsterdam: John Benjamins. 353–392.

Merchant, Jason (2001) The Syntax of Silence: Sluicing, Islands, and the Theory of Ellipsis. Oxford: Oxford University Press.

Merchant, Jason (2004) Fragments and Ellipsis. Linguistics and Philosophy 27: 661–738.

Merchant, Jason (2008) Variable Island Repair under Ellipsis. In: Kyle Johnson (ed.) Topics in Ellipsis. Cambridge: Cambridge University Press. 132–153.

Merchant, Jason (2013) Voice and Ellipsis. Linguistic Inquiry 44. 77–108.

Paoli, Sandra (2007) The Fine Structure of the Left Periphery: COMPs and Subjects. Evidence from Romance. Lingua 117.6. 1057–1079.

Poletto, Cecilia (2006) Parallel Phases: A Study of the High and Low Left Periphery of Old Italian. In: Mara Frascarelli (ed.) Phases of Interpretation. Berlin: Mouton de Gruyter. 261–294.

Rizzi, Luigi (1997) The Fine Structure of the Left Periphery. In: Liliane Haegeman (ed.) Elements of Grammar. Dordrecht: Kluwer. 281–337.

Rizzi, Luigi (2004) Locality in the Left Periphery. In: Adriana Belletti (ed.) Structures and Beyond: The Cartography of Syntactic Structures, Volume 3. Oxford: Oxford University Press. 223–251.

Roberts, Ian (2005) Principles and Parameters in a VSO Language. Oxford: Oxford University Press.

Sobin, Nicholas (2002) The Comp-trace Effect, the Adverb Effect and Minimal CP. Journal of Linguistics 38. 527–560.

Szendrői, Kriszta (2001) Focus and the Phonology–Syntax Interface. PhD dissertation. London: University College London.