Table Of Content
- About the author(s)/editor(s)
- About the book
- This eBook can be cited
- Table of Contents
- Chapter 1 On Japanese Sluicing: Evidence for the Focus Movement & Deletion with Some Remarks on English and Polish (Naoyuki Akaso / Seichi Sugawa)
- Chapter 2 A Syntactic Solution to the Inverse-Scope Puzzle and the Contrastive VP-Focus Construction in Hungarian Based on Extraction (Gábor Alberti / Judit Farkas)
- Chapter 3 A Short Note on the Syntactic Behaviour of Demonstratives in Polish (Piotr Cegłowski)
- Chapter 4 On Nominalization: Genitives, Datives, and Elementary Predicates in Italian (Ludovico Franco)
- Chapter 5 Peeling of Subject Case: Marked Subjects in Polish (Aleksandra Gogłoza)
- Chapter 6 On Subject Experiencer Verbs in Hungarian (Réka Jurth)
- Chapter 7 Idioms: The Interplay of Domains or a No-Man’s-Land? (Joanna Kolbusz-Buda)
- Chapter 8 N Morphology and Its Interpretation: Romance -a Plurals (M. Rita Manzini / Leonardo M. Savoia)
- Chapter 9 Differentiating between the Syntactic Realisation of Complex Events and Complex Predicates of Irish (Brian Nolan)
- Chapter 10 Rounded Back Vowels Preservation in Consonantal Contexts. C-V Licensing (Leonardo M. Savoia / Benedetta Baldi)
- Chapter 11 Constraints on Multi-Noun Compounding in English: A Corpus-Based Approach (Sebastian Wasak)
- Chapter 12 Subjects or not Subjects: Polish Dative Experiencers and Adjunct Control (Sylwiusz Żychliński)
- Series index
The volume called “Studies in Formal Linguistics: Universal Patterns and Language Specific Parameters” is a collection of twelve papers whose common denominator is the search for universal principles in the natural language, on the one hand, and the exploration of parameters of variation among languages, on the other. The papers gathered in the volume cover a wide range of topics concerning syntax, phonology, morphology and semantics.
Universal principles that all languages share and parameters along which languages vary have constituted the core of syntactic theorising since the advent of the Principles and Parameters (P&P) approach of Chomsky (1981). In the Government and Binding (GB) Theory, universal principles, or principles of Universal Grammar (UG), had a rich and modular structure, and affected a number of subsystems of human language, such as the Binding Theory, the Case Theory, the Control Theory, etc. The parameters, in turn, were perceived as choices on the range of possibilities admitted by the universal principles. In other words, in the GB era, universal principles were parametrised. As a result, the language variation was encoded in universal principles, or, in other words, was “hardwired in UG” (Gallego 2011: 527). In the GB model, a single parameter setting had a wide range of consequences, commonly referred to as ‘clustering effects’ (cf. Gallego 2011: 527–528; Fábregas et al. 2015: 5–9). Since parameters were nothing more than relativised principles, Kayne (2000) observes that “the study of syntactic parameters and the study of syntactic universals go hand in hand” (Kayne 2000: vii). The main problems that GB style analyses had to face concerned the question how many settings parameters allow and how many parameters there are. Although it was commonly assumed that parameters allowed binary choices, the binary divide was not enough to capture the variation attested in different language areas, including for instance the pro-drop parameter (Fábregas et al. 2015). The number of parameters that were posited in the GB theory was high, which was undesirable from the theoretical stance of trying to obtain a restricted set of parameters in order to be able to simplify the process of language acquisition.
In the Minimalist Program (MP) of Chomsky (1993, 1995, 2000, 2001, 2007), the linguistic inquiry has centred around recasting universal principles in the light of computational efficiency considerations and interface conditions. For this reason, the main focus has been on principles of UG, not on parameters. In the MP, UG is approached from below, which is characterised by Chomsky (2007: 4) as follows: “The MP seeks to approach the problem ‘from the bottom up’: How ← 7 | 8 → little can be attributed to UG while still accounting for the variety of I-languages attained, relying on third factor principles?”. Consequently, UG, viewed from below, is highly reduced, in comparison with the UG from the GB theory. It contains the recursive operation Merge and a set of features (in particular, formal features). The third factor principles, mentioned in the quotation above, cover principles of data processing and architectural/computational-developmental constraints (Chomsky 2005), where the latter correspond to the Inclusiveness Condition, the Minimal Link Condition, the Phase Impenetrability Condition, the No Tampering Condition, and the Principle of Full Interpretation (Gallego 2011: 539; Chomsky et al. to appear). Actually, beside the third factor, Chomsky (2005: 6) also mentions Factor I and Factor II of language design, where the former refers to genetic endowment, while the latter covers external data (E-language). Language variation cannot be attributed to Factor I, the faculty of language, and neither can it be associated with Factor III, which comprises “principles of neural organization that may be more deeply grounded in physical law” (Chomsky 1965: 59). Consequently, the locus of parametrisation in the MP is to be found in the lexicon. Variation thus emerges as a result of interaction of Factor I, which comprises features necessary to build up lexical items, with Factor II, which relies on language experience. The lexical approach to parametric variation has come to be known to as Borer-Chomsky conjecture and is stated in (1) below:
(1) All parameters of variation are attributable to differences in the features of particular items (e.g. the functional heads) in the lexicon.
(Baker 2008: 156)
If the lexical view of parameters is adopted, then there are two sources of language variation in the MP, viz. (i) absence or presence of some head F, and (ii) absence or presence of some feature or set of features in the head F (Fábregas et al. 2015: 15). Still another source of variation among languages, recognized in the MP, is attributed to Externalization, i.e. Transfer to the morphophonological component or Spell-Out. Chomsky (2010) claims that the faculty of language is only optimal to meet the requirements of the Conceptual-Intentional systems, but is not designed to meet the requisites of the Sensori-Motor systems. There are various solutions adopted to externalize the structure into a physical signal, but no perfect solution is available. As a result, variation emerges, since languages make different choices as to how to solve the problem of adapting a linguistic structure to something that can be used by the organs such as the throat, the mouth, or the hands, which were not designed with a linguistic purpose. Consequently, the locus of variation is to be found in the PF-branch of grammar, perceived as an interface between syntax ← 8 | 9 → and phonology. This kind of variation is free from semantic effects, because it is attested at a point where access to semantics is no longer possible. Therefore, no language variation is to be found at the level of LF.
In Chapter 1, Naoyuki Akaso and Seichi Sugawa examine Sluicing in Japanese, English and Polish using the cartographic model of Rizzi (1997). The paper focuses primarily on Japanese, while the English and Polish data serve to provide a comparative view on the mechanisms deriving Sluicing cross-linguistically. Akaso and Sugawa argue that Sluicing in Japanese is derived by means of focus movement, followed by deletion of FinP. The main evidence adduced in favour of this line of analysis is based on one type of reason clause and one kind of negative polarity item (henceforth, NPI) from Japanese. The reason clauses analysed in the paper are treated as focus elements, placed in Spec, FocP, as in Kawamura (2008). When the reason clause functions as a sluice, Sluicing is ungrammatical, because there is only one focus position in a sentence which is occupied by the reason clause, and consequently, there is no position to which another focused element can be moved. Likewise, Sluicing is disallowed with NPIs, since in this case the NPI fills the Spec, FocP position, and thereby fails to be c-commanded by negation. Akaso and Sugawa note that the derivation of Sluicing in English involves wh-movement, not focus movement, while Polish seems to allow for both mechanisms operative in Sluicing, viz. wh- and focus movement.
In Chapter 2, Gábor Alberti and Judit Farkas investigate several instances of an extracted right-branching domain in Hungarian. The paper builds on the two prior works by Alberti (2004) and Alberti and Farkas (2017). In particular, the emphasis is laid on the inverse scope puzzle of contrastive topics, and the broadly understood phenomenon of contrastive VP-focus in Hungarian. In their analysis, Alberti and Farkas adopt Grohmann’s (2003) model with three Prolific Domains, covering the theta domain, the discourse (operator) domain, and the agreement domain. Both contrastive topics and contrastive VP-foci in Hungarian are analysed by means of remnant movement. The remnant is formed by extracting the right periphery out of the complex expression, whereby satisfying the language specific constraint on right branching structures, attested in Hungarian. The proposed analysis accounts for the puzzling word order found with contrastive topics and contrastive VP foci in Hungarian by suggesting that from the point of view of semantics, it is the complex expression that occupies a given operator position, whereas from a syntactic viewpoint, it is an appropriate remnant that figures in a corresponding syntactic layer. This is responsible for the discrepancy between the syntactic position of a given operator and the factual meaning of a sentence hosting it. ← 9 | 10 →
In Chapter 3, Piotr Cegłowski studies the syntactic behaviour and the structural position of demonstratives in Polish. The starting point for the analysis is an empirical study of acceptability of specific types of extraction, such as extraction of N-complements, extraction across a numeral, extraction related to the Left Branch Condition (LBC), and its various manifestations, including deep LBC, extraordinary LBC, and double AP LBC. The data obtained in the course of the experiment show that demonstratives do not pattern like adjectives with respect to LBC. In fact, demonstratives are much less prone to the LBC than adjectives or wh-phrases. This fact is accounted for by proposing that demonstratives are maximal projections (DemPs) generated at the edge of a functional projection, called Agr, sandwiched between D and N. The placement of DemP at the left periphery of a DP is additionally supported by the way demonstratives interact with discourse-related items such as topic and focus. The overall conclusion reached in the paper is that the syntactic behavior of demonstrative pronouns in Polish provides evidence for the existence of a DP in a language without articles, and hence supports the claim that the DP-Hypothesis constitutes a part of the UG.
In Chapter 4, Ludovico Franco analyses deverbal nominalisations in Italian that host an external and internal argument. He proposes that the process of nominalisation can be reduced to a syntactic template introducing the main participants, viz. the external and internal argument, by means of a sort of double oblique construction. Franco argues that the thematic as well as relational PPs are linked or adjoined to the eventive noun which projects to a DP. Genitives, datives and other oblique phrases involved in the process of nominalization are taken to represent part-whole relational predicates, following the work by Manzini and Savoia (2011a, b), Manzini et al. (2015) and Manzini and Franco (2016). Specifically, in this model the genitive or dative possessor acts as a whole including a part, i.e. the possessee. The part-whole relation holds between entities, as well as between entities and events or states. The analysis advanced in the paper goes against the claim that the genitive in nominalizations represents a structural case, as in Longobardi (1994) or Alexiadou (2001), inter alia. Franco also supports the idea that the genitive case marked external argument in nominalisations shows ergative properties, i.e. it does not act as a causer or agent in the event, but rather corresponds to the possessor of a property.
In Chapter 5, Aleksandra Gogłoza concentrates on the syntax of marked subjects in Polish, i.e. the external arguments that are associated with case different from the nominative. In particular, Gogłoza analyses instrumental subjects in Adversity Impersonals, dative subjects in the Dative Reflexive Construction, genitive external arguments in nominalizations, and accusative subjects in numeral ← 10 | 11 → phrases. Jabłońska’s (2007) nominal functional sequence, in which different types of external arguments occupy different hierarchically ordered positions, constitutes the starting point for Gogłoza’s account. Gogłoza first mentions problems that Jabłońska’s (2007) degrees of externality give rise to. Working within Caha’s (2009) Peeling Theory of Case, in which case is decomposed into hierarchically ordered projections, Gogłoza proposes that the external argument is merged in v1 with the instrumental case on top. Once the external argument starts moving up, the instrumental case is peeled off, and the ultimate case marking of the external argument depends on how high in the structure it moves. The analysis correctly predicts that the smaller the case of an external argument is, the more subject properties it has. Conversely, the lower the external argument stays in the structure, the fewer subject properties it shows.
In Chapter 6, Réka Jurth analyses two types of Subject Experiencer (henceforth, SE) verbs in Hungarian, viz. those which have an Object Experiencer (henceforth, OE) alternate (the worry-type), and those which are intransitive, but have a transitive SE alternate (the hate-type). She observes that the two classes of SE verbs differ in their aspectual properties in that the intransitive members of the hate-class are uniformly atelic and eventive, whereas the representatives of the worry-class are more heterogeneous and comprise both telic and atelic, stative as well as eventive verbs. Moreover, intransitive verbs of the hate-type are agentive, in contradistinction to worry-verbs which are non-agentive. The two classes of SE verbs can optionally select a PP, which in the case of the worry-type acts as a causer, subject of emotion, or target of emotion, while with the hate-type verbs, the PP can never be interpreted as a causer, but always functions as a target of emotion. The two types of SE verbs also differ morphologically. The worry-type verbs pattern morphologically with anticausatives and bear some resemblance to unergatives. The hate-types predicates, in turn, show reflexive and reciprocal morphology, and are in fact unergative. Jurth offers an analysis of the two types of SE verbs in Hungarian within Reinhart’s (2001) Theta System.
In Chapter 7, Joanna Kolbusz-Buda addresses the problem of the intricate nature of idioms. Any account of this linguistic phenomenon has to take into consideration both their syntactic structure and word-like unitary nature. Idioms seem to be a borderline phenomenon, belonging to the domains of both syntax and morphology. Thus, the author delves into the linguistic behaviour of idiomatic expressions with a view to establishing their significant traits in the system of language. The author proposes that in the processing of idioms, the syntactic and morphological factors have to interact. The aim of the analysis is to establish in ← 11 | 12 → what respects and to what extent syntax and morphology need to contribute to the creation of idiomatic expressions.
Chapter 8, Rita Manzini and Leonardo Savoia’s contribution, is devoted to the structure of nominal inflections in Romance. Subject to analysis is the –a inflection in Italian varieties which realises both the plural and the feminine singular. The authors maintain that in the language varieties analysed, the –a plural externalises a nominal Class property [aggregate]. It is further proposed that in the occurrence of a-inflection, this property ‘is at the basis of the superficial syncretism between plural and singular/feminine’. The analysis of the relevant data leads the authors to the conclusion that –a, which performs the inflectional function, is in fact a lexical unit endowed with interpretive content. This holds not only for the varieties of Italian but also for the other Romance languages.
In Chapter 9, Brian Nolan takes a closer look at the Irish data, providing a functional analysis of the realisation of complex events and complex predicates in this language. The author tries to construct an account which would incorporate both top-down and bottom-up approaches to the analysed phenomena. The analysis focuses on a set of relationships between multi-verb constructions, in single and multiple clauses, and a variety of complex events. There seems to be a linkage between conceptualisation of related complex events, and their participants. Further, the article discusses the syntactic realisation of these complex constructions of contemporary Irish.
In the organisation of phonological domains a special role is performed by the nucleus which licenses the preceding consonantal position – the onset. The prosodic dependency of the onset on its licenser also has serious consequences at the melodic level. In Chapter 10, Leonardo Savoia and Benedetta Baldi focus on the language-specific manifestations of the C-V interplay. In their paper, the authors focus on the interaction between consonantal and vocalic melodic material when the latter is associated to the stressed nucleus. They address the phenomena of metaphony and [u] propagation and preservation in the pre-tonic position. The evidence comes from Southern Italian varieties. The analysis is couched within the framework of Government Phonology.
In Chapter 11 by Sebastian Wasak, the reader will find a corpus-based account of constraints on multi-noun compounding in English, with a special focus on endocentric compounds. The analysis deals with multi-noun formations, trying to determine the frequency of their usage and account for their rare occurrence in language performance. The relevant data are taken from the British National Corpus and the Corpus of Contemporary American English. The author argues ← 12 | 13 → that multi-noun structures are typically formed of lexicalised subcomponents, which accounts for their status in the morphology of English.
In Chapter 12, Sylwiusz Żychliński concentrates on Polish dative Experiencers in order to test whether they behave like subjects in this language. He examines dative Experiencers which co-occur with the nominative argument in Polish, and analyses the way they behave with respect to a number of subjecthood criteria, available in the literature, such as anaphor binding, raising, the conjunction reduction, and resumption. These tests yield negative results as to the subject status of the dative Experiencer. The diagnostic which is most thoroughly studied in relation to dative Experiencers is adjunct control in participial clauses in Polish. Żychliński emphasizes that the prescriptive rule stating that the subject of the participial clause must be identical with the subject of the main clause is in fact untenable, especially in the light of a series of acceptability surveys that he carried out. The surveys have shown that dative Experiencers can act as controllers, which might point towards their subject status. This, however, does not seem to be supported by a wider range of data, as Żychliński specifies that accusative case marked arguments can also control PRO in participial clauses. The overall conclusion reached in the paper is that dative Experiencers cannot be taken to represent syntactic subjects in Polish.
Anna Bloch-Rozmej and Anna Bondaruk
Alberti, Gábor. 2004. “Climbing for Aspect – With no Rucksack.” In Verb Clusters: A Study of Hungarian, German and Dutch. Linguistics Today 69, edited by Katalin É Kiss, and Henk van Riemsdijk, 253–289. Amsterdam/Philadelphia: John Benjamins.
Alberti, Gábor, and Judit Farkas. 2017. “Right Branching in Hungarian: Moving Remnants.” In Language Use and Linguistic Structure, Proceedings of Olomouc Linguistics Colloquium 2016, edited by Joseph Emonds, and Markéta Janebová. 209–224. Olomouc: Palackỳ University.
Alexiadou, Artemis. 2001. Functional Structure in Nominals. Nominalisation and Ergativity. Amsterdam: John Benjamins.
Baker, Mark C. 2008. “The Macroparameters in a Microparametric World.” In The Limits of Syntactic Variation, edited by Theresa Biberauer, 351–373. Amsterdam: John Benjamins.
Chomsky, Noam. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
Chomsky, Noam. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Chomsky, Noam. 1993. “A Minimalist Program for Linguistic Theory.” In The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger, edited by Ken Hale, and Samuel J. Keyser, 1–52. Cambridge, MA: MIT Press.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Chomsky, Noam. 2000. “Minimalist Inquiries: The Framework.” In Step by Step, edited by Richard Martin, David Michaels, and Juan Uriagereka, 89–155. Cambridge, MA: MIT Press.
Chomsky, Noam. 2001. “Derivation by Phase.” In Ken Hale: A Life in Language, edited by Michael Kenstowicz, 1–52. Cambridge, MA: MIT Press.
Chomsky, Noam. 2005. “Three Factors in Language Design.” Linguistic Inquiry 36: 1–22.
Chomsky, Noam. 2007. “Approaching UG from Below.” In Interfaces + Recursion = Language? Chomsky’s Minimalist and the View from Syntax-Semantics, edited by Uli Sauerland, and Hans-Martin Gärtner, 1–29. Berlin: Mouton de Gruyter.
Chomsky, Noam. 2010. “Some Simple Evo-Devo Theses: How True might they be for Language?” In The Evolution of Human Language. Biolinguistic Perspectives, edited by Richard K. Larson, 45–62. Cambridge: Cambridge University Press.
Chomsky, Noam, Ángel J. Gallego, and Dennis Ott. to appear. “Generative Grammar and the Faculty of Language: Insights, Questions, and Challenges.” In Generative Syntax: Questions, Crossroads, and Challenges. Special issue of Catalan Journal of Linguistics, edited by Àngel J. Gallego, and Dennis Ott. 2018.
Fábregas, Antonio, Jaume Mateu, and Michael Putnam. 2015. ““Parameters” in Linguistic Theory: What, Where, and How.” In Contemporary Linguistic Parameters, edited by Antonio Fábregas, Jaume Mateu, and Michael Putnam, 3–24. London: Bloomington.
Gallego, Àngel J. 2011. “Parameters.” In Oxford Handbook of Linguistic Minimalism, edited by Cedric Boeckx, 523–551. Oxford: Oxford University Press.
Grohmann, Kleanthes K. 2003. Prolific Domains: On the Anti-locality of Movement Dependencies. Amsterdam: John Benjamins.
Jabłońska. Patrycja. 2007. “Radical Decomposition and Argument Structure.” Ph. D. diss., University of Tromsø.
Kawamura, Tomoko. 2008. “Adverbial Because-clauses as Focal Elements.” Nanzan Linguistics 3(2): 103–122.
Longobardi, Guiseppe. 1994. “Reference and Proper Names.” Linguistic Inquiry 25: 609–665.
Manzini, Rita, and Leonardo M. Savoia. 2011a. Grammatical Categories. Cambridge: Cambridge University Press.
Manzini, Rita, and Leonardo M. Savoia. 2011b. “Reducing ‘Case’ to Denotational Primitives: Nominal Inflections in Albanian.” Linguistic Variation 11: 76–120.
Manzini, Rita, and Ludovico Franco. 2016. “Goal and DOM Datives.” Natural Language and Linguistic Theory 34: 197–240.
Manzini, Rita, Leonardo M. Savoia, and Ludovico Franco. 2015. “Ergative Case, Aspect and Person Splits: Two Case Studies.” Acta Linguistica Hungarica 62: 297–351.
Reinhart, Tanya. 2001. “Experiencing Derivations.” In Proceedings of Semantics and Linguistic Theory (SALT) XI, edited by Rachel Hastings, Brendan Jackson, and Zsófia Zvolenszky, 365–387. Ithaca, NY: Cornell University.
Rizzi, Luigi. 1997. “The Fine Structure of the Left Periphery.” In Elements of Grammar, edited by Liliane Haegeman, 281–337. Dordrecht: Kluwer.
Abstract: This paper focuses on the often discussed issue of what kind of movement is involved in the derivation of Sluicing in Japanese. We claim that Sluicing in Japanese is derived by the focus movement of the remnant into the Spec of FocP, followed by the deletion of FinP. In the course of our discussion, two sets of data are provided to support our claim. First, it is observed that Sluicing is prohibited if there is another focus element, a reason clause headed by kara ‘because’ in the sluice. We argue that this is because the movement of the remnant is blocked by the existence of the KARA reason clause occupying the Spec of FocP, suggesting that it is focus movement that forces the remnant to move out of the FinP. Second, we show that NPI sika ‘only’ cannot be licensed in Sluicing. We claim that this is because the remnant with sika moves to the Spec of FocP by virtue of its focus nature, resulting in NPI being out of the c-command domain of Neg0. This suggests again that focus movement is responsible for the derivation of Sluicing in Japanese. Furthermore, based on a comparison of data from English, Japanese, and Polish, some typological implications are presented concerning what movement is involved in the derivation of Sluicing in each language. We claim that Wh-movement is involved in English Sluicing, while Focus movement is responsible for the derivation of Sluicing in Japanese. In Polish, both types of movement are available in Sluicing.
|(1)||a.||John hired someone, but I don’t know who.|
|‘John hired someone, but I don’t know who.’|
Sluicing has attracted much interest from theoretical linguists because it is recognised as an important cross-linguistic phenomenon which provides empirical grounds on which a theory of movement and ellipsis, both of which are core property of natural language, can be based.
A standard analysis of Sluicing assumes Wh-movement of the remnant to the Spec of CP, followed by the deletion of TP in the sluice (Ross 1969; Merchant 2001; among others). This is illustrated in (2).
It is quite natural to assume that Wh-movement is involved in the derivation of Sluicing in English as Wh-movement is one of the core operations available in the language. Then a question is naturally raised for wh-in-situ languages such as Japanese. Here a theoretical issue arises about whether the same kind of analysis can be extended to this type of language or whether other types of movement are involved in their derivation.
This paper provides novel sets of data to support the claim that in Sluicing in Japanese the remnant is moved out of the elided portion of the sluice by ← 18 | 19 → Focus-movement, followed by the deletion of FinP, assuming the cartographic structure of the left periphery of a clause (Rizzi 1997). Focus movement analysis of Japanese Sluicing itself is not novel: it has been proposed in several works such as Kim (1997), Hiraiwa & Ishihara (2002), and Sugawa (2008). We support this analysis by providing two sets of data in which focus movement is shown to be crucial for Sluicing in Japanese. One argument comes from the fact that the existence of the reason clause headed by the conjunction kara ‘because’ blocks the derivation of Sluicing. The other argument is built on the licensing of NPI sika ‘only’. Furthermore, based on a comparison of data from English, Japanese, and Polish, some typological implications for the type of movement involved in the derivation of Sluicing of each language are presented.
The organisation of this paper is as follows. Section 2 provides background for the following sections and presents our claim that focus movement is crucial for Sluicing in Japanese, by pointing out the problems in the previous studies. Section 3 provides two sets of data supporting our claim. One is based on the observation of the interaction of KARA reason clauses and Sluicing. The other is based on the licensing of NPI sika in the sluice. In Section 4, some typological implications for Sluicing in English, Japanese, and Polish are presented. Section 5 concludes the paper.
2. Focus Movement and Sluicing
It is well-known that Sluicing is a cross-linguistically observed phenomenon and is found in Japanese as exemplified in (3).
‘It seems that Tom bought something, but I don’t know what.’
‘It seems that Tom wrote a letter to someone, but I don’t know whether it was to Bill.’
Takahashi (1994) analyses Sluicing in Japanese in a parallel way to the analysis proposed for English Sluicing, where the sluice is derived by Wh-movement of the remnant into the Spec of CP, followed by TP deletion. The derivation of the sluice in (3a) based on this analysis is shown schematically in (4). ← 19 | 20 →
It is often pointed out, however, that several questions naturally arise regarding Takahashi’s analysis. First, it is not clear whether Wh-movement is involved in the derivation of Japanese Sluicing, since Japanese obviously lacks overt Wh-movement. If Takahashi’s analysis is on the right track, then we should assume that a wh-phrase should move to the Spec of CP in the case of Sluicing, in spite of the fact that a wh-phrase can be in situ in Japanese interrogative sentences. This analysis calls for an extra assumption that a wh-phrase must move just in the case of Sluicing. Second, not only wh-phrases but also DPs can be remnants in Sluicing in Japanese as shown in (3b). DP remnants are not acceptable on English Sluicing and this suggests that operations other than Wh-movement could be involved.
We claim, assuming a cartographic approach to the structure of CP à la Rizzi (1997), that Sluicing in Japanese is derived by Focus movement of the remnant into the Spec of FocP, followed by the deletion of FinP in the sluice. The derivation is shown schematically in (5), where XP moves out of the TP into the Spec of FocP and the deletion of FinP follows.
3. Data and proposal
In this section we provide two pieces of data to support the claim above. The data given in 3.1 supports the first part of our claim that focus movement is involved in the derivation of Sluicing in Japanese, and the data provided in Section 3.2 supports the latter part of our claim that the elided portion in the sluice is derived by FinP deletion.
3.1. KARA reason clauses as focus elements
Focus movement analysis predicts that if we set up a situation in which the focus movement is somehow blocked, the sluice cannot be derived. Example (6) illustrates this. The Spec of FocP is occupied by the focus element YP and YP blocks the movement of the remnant XP into the relevant position.
In the following argument, we show that this prediction is borne out, based on data containing KARA reason clauses. Following Kawamura’s (2008) analysis of ← 20 | 21 → because clauses in English, it is assumed that KARA reason clauses are focus elements in the Spec of FocP, blocking the focus movement of other elements.
Before getting into the details of our proposal, consider Kawamura’s (2008) proposal that because clauses in English are inherently focused. Her argument is based on the sets of data in (7) and (8).
|(7)||John didn’t call Mary [because he visited Kyoto].|
|(i)||John called Mary. Its cause was not that he visited Kyoto.|
|(ii)||John didn’t call Mary. Its cause was that he visited Kyoto.|
(Kawamura 2008, 103)
|(8)||John didn’t butter THE BAGEL.4|
|(i)||It was not the bagel that John buttered.|
|(ii)||It was the bagel that John didn’t butter.|
(Kawamura 2008, 108)
In (7) we can observe the scope interaction of the because clause and negation, resulting in ambiguity, as shown in the two possible interpretations, (7i) and (7ii). A similar scope interaction, resulting in two possible interpretations, (8i) and (8ii), is observed between the focused element and negation in (8). Based on these readings, Kawamura (2008) proposes that because clauses are interpreted as if they were inherently focused. We believe that the same is true of Japanese KARA reason clauses.
Akaso (2016) claims that Kawamura’s analysis is true of Japanese KARA reason clauses, based on Hiraiwa & Ishihara’s (henceforth, H. & I.) (2002) analysis of Cleft sentences. H. & I. (2002) argue that cleft sentences are derived from so-called NO DA sentences, a sort of focus construction. Examples of these two constructions are given in (9).
|‘It is that Taro bought a car.’|
|‘It is a car that Taro bought.’|
(9a) is a NO DA sentence and (9b) is a Cleft construction. H. & I. (2002) claim that the Cleft sentence is derived from the NO DA sentence, with multiple movement operations. According to their analysis, (9b) is derived from (9a) by focalising the object, kuruma-o. The focused element (i.e. kuruma-o) is raised to the Spec ← 21 | 22 → of FocP, as shown in (10), and then the remnant FinP is moved to the Spec of TopP, as shown in (11).
The Cleft sentence (9b) is derived by means of these two operations.
With this in mind, consider the following sentences.
|‘It is that Taro bought a car because he got a job.’|
|‘It is a car that Taro bought because he got a job.’|
(12a) sounds natural, while (12b) is quite unacceptable. This can be explained under H. & I.’s analysis as we have seen. In that analysis, the focus element is moved to the Spec of FocP. We assume that the KARA clause is located there because it is a focus element, as discussed in Akaso (2016) and Akaso and Haraguchi (2016) (henceforth, A. & H.). Once the Spec of FocP is occupied by the KARA clause, there is no slot for another element to move to this position.
If this movement is applied, the resulting sentence is ungrammatical. Thus we believe, from the above observation, that KARA clauses are connected to Foc0, as argued by Kawamura (2008) for English. ← 23 | 24 →
A. & H. (2016) offer another piece of evidence to support the claim that KARA reason clauses are connected to Foc0. Their argument is based on the Nominative/Genitive Case alternation in relative clauses in Japanese. Japanese has a peculiar Case system in which Nominative Case can be altered to Genitive Case, as shown in (14).
|‘the book that Taro read’|
This is known as Nominative/Genitive Conversion (henceforth, NGC) and this alternation is possible only within prenominal clauses, such as relative clauses. It cannot be found in a root sentence, as shown in (15).
|‘Lit. Taro read the book.’|
In (14) the adverb kino ‘yesterday’ may appear on the left of the Genitive subject. This adverb is an element of a prenominal clause, which means that the Genitive subject should be located within the clause at issue, but not in the Spec of DP.
A. & H. (2011) propose that the syntactic categories of Japanese prenominal clauses are of two types: one is Focus Phrase (FocP) for those with a Nominative subject, and the other is TP for those with a Genitive subject. Evidence for this proposal comes from the data in (16) and (17). The examples in (16) are Relative Clauses and those in (17) are Gapless Clauses.
|‘the medicine that only Taro took’|
|‘the book that only students bought’|
|‘the fact that only Taro cried’|
|‘the reason that only the Japanese walk’|
Assuming that Focus particles are licensed by the Foc0, A. & H. (2011) propose the following syntactic structures.
In the case of clauses with Nominative subjects, there is FocP between NP and TP, and Foc0 can license the Focus Particles, but this does not occur in clauses with Genitive subjects. The syntactic category of prenominal clauses with Genitive subjects is TP with no FocP above it. So, DPs with focus particles cannot be licensed in prenominal clasues with Genitive subjcts.
With this in mind, consider the following example.
|‘Taro went to the park where Ken was playing because the rain stopped.’|
Let us begin with the English translation. This English sentence yields two interpretations, both of them are given in (20).
|(20)||a.||Taro went to the park because the rain stopped, and Ken was playing there.|
|b.||Taro went to the park, where Ken was playing because the rain stopped. ← 25 | 26 →|
In (20a), the because clause modifies the main clause, giving the reason why Taro went to the park, while in (20b) the because clause modifies the relative clause, giving the reason why Ken was playing in the park.
In the case of Japanese, the Case alternation on the subject in the relative clause affects the interpretation of the sentence. The prenominal clause with a nominative subject yields two interpretations: one in which the KARA clause modifies the main clause, giving the reason why Taro went to the park and the other in which the KARA clause modifies the relative clause, giving the reason why Ken was playing in the park. This is parallel to the ambiguity found in the English sentence in (20). On the other hand, if the subject in the relative clause is marked with Genitive, the ambiguity disappears and only the former interpretation is possible.
These facts can be explained as follows: in the case where the subject in the relative clause is marked with Nominative, two Foc heads are available: one is in the CP-zone of the main clause and the other in the relative clause. In the case where the subject in the relative clause is marked with Genitive, only the Foc0 in the CP-zone of main clause is available, and there is no Foc0 in the relative clause. Thus, based on observation of the relation between the focus particles and Foc0, we may conclude that KARA clauses are licensed by Foc0 in Japanese.
Now let us consider Sluicing data containing KARA reason clauses. When a KARA reason clause is put in the sluice, then Sluicing is not possible, as shown in (21).
|‘I heard that Tom criticized someone because s/he failed to get a certain prize, but I don’t know who it was because s/he failed to get the Nobel Prize.’|
We claim that the ungrammatical status of (21) can be explained if we assume that the KARA reason clause occupies the Spec of FocP. This means there is no slot for another element to move to the Spec of FocP. The second conjunct in (21) is shown schematically in (22). ← 26 | 27 →
In (22), the remnant dare-o ‘who’ cannot move to the Spec of FocP before the deletion of FinP is applied. Thus (21) is ruled out.
3.2. XP-SIKA ‘only’ NPI
The second type of supporting evidence is based on NPI (Negative Polarity Items). As shown in (23) NPIs are expressions such as any in English, which appear only in the negative environments. The NPI any cannot appear without negative licensers such as not; otherwise it is unacceptable as in (23b).
|(23)||a.||I haven’t got any books.|
|b.||* I’ve got any books.|
Japanese has several NPIs and here we use SIKA-NPI, exemplified in (24a). The phrase sika literally corresponds to only in English, but this is an NPI in Japanese, in contradiction to only. Without the negative element (-nai), it is ungrammatical, as shown in (24b).
|‘Hanako bought a comic book/comic books only.’|
|‘(intended meaning) Hanako bought a comic book/comic books only.’|
In English, an NPI appears at the object position, but not at the subject position, as shown in (25).
|(25)||a.||Tom did not meet anyone.|
|b.||* Anyone did not meet Tom.|
This is known as the subject/object asymmetry, which results for the fact that English NPIs are licensed only within the c-command domain of a negative element. Consider the clausal structure of a typical simple sentence. ← 27 | 28 →
But interestingly this asymmetry cannot be observed in Japanese. That is, SIKA-NPI can appear in either a subject position or an object position, as illustrated in (27).
|(27)||No subject/object asymmetry|
|‘Hanako bought a comic book/comic books only.’|
|‘Only Hanako bought a comic book/comic books.’|
Kishimoto (2016) explains this fact in terms of the operation of Neg-raising, by which the scope of negation could be extended to cover the whole TP.
The subject is within the c-command domain of the negative element with the Neg-raising. Assuming Kishimoto’s analysis is correct, we can predict that sluicing cannot be allowed with SIKA-NPI, and this is borne out, as shown in (29).
|‘I heard that Hakako met only someone, but Taro doesn’t know who only.’|
One might wonder if the ungrammaticality of (29) results from the deletion of the negative element, but it is not the case. Examine (30) carefully. It is what we call a Fragment (or Short) answer. The negative element is deleted in (30B), but the response is perfectly grammatical.
|‘I heard that Taro greets only someone.’|
|B:||Dare-ni-SIKA (Taro-wa t aisatsu-o siNAI no)?|
This shows that the ungrammaticality of (29) is not due to the deletion of negative element. Rather, we contend that the SIKA-NPI is out of the c-command domain of the negative expression NAI in (29). In the structure in (31), we can see the SIKA-NPI is out of the scope of NAI, as Neg-raising expands the c-command domain up to TP, but not to FocP.
|(31)||* Taro-wa [ForceP [FocP dare-ni-SIKA …[FinP [TP Hanako-ga awa-NAI-katta]-no]]-ka] siranai|
Note that when ordinary scrambling is applied to the SIKA-phrase, instead of Sluicing, the resulting sentence is grammatical, as illustrated below.
|‘Taro does not know who the only person that Hanako met is.’|
In this case, we believe that the SIKA-phrase is adjoined to TP through a clause-internal scrambling. So, the structure is as follows:
|(33)||Taro-wa [ForceP [FocP [FinP [TP dare-ni-SIKA [TP Hanako-ga awa-NAI-katta]]-no]]-ka] siranai|
We can see that the SIKA-phrase is within the c-command domain of NAI, which makes (32) grammatical, in contrast with (29).
4. Typological observations on Sluicing: English, Japanese, and Polish
This section presents a small contrastive survey on Polish sluicing. Although our understanding and knowledge about Polish is limited, it is of interest to compare Polish with English and Japanese, because Polish shares some properties with English and others with Japanese.
So far we have seen the difference between English and Japanese in terms of Sluicing. Let us introduce two terms. One is Wh-Sluicing which means the deletion follows obligatory Wh-movement. The other is Focus Sluicing, which means the deletion follows Focus movement, as observed in Japanese, which does not have the obligatory Wh-movement, since it is a wh-in-situ language. Japanese can have wh-phrases in a sentence-initial position, as in English, but this is not due to Wh-movement, but due to Scrambling.
|‘What did Taro eat?’|
This is confirmed by the fact that ordinary NPs (i.e., NPs without wh-features) can be placed in a sentence-initial position, as shown below.
|‘Taro ate sushi.’|
Thus, the seeming Wh-movement in (34b) is actually the result of the application of scrambling, an optional operation.5 As discussed in the previous sections, Japanese Sluicing consists of two operations: Focus movement and FinP deletion.
But in English, Scrambling is not allowed. Therefore, ordinary nouns (not wh-words) cannot remain as remnants in Sluicing, as shown below.
|(36)||* I don’t know (that) Tom [(she loves)]. ← 30 | 31 →|
This shows that English Sluicing consists of two operations: Wh-movement and TP-deletion.
The table in (37) is a comparison between English and Japanese Sluicing.
|(37)||English vs. Japanese|
Then what about Polish? In the literature we consulted, Polish has the English type of Wh-sluicing, as in (38).
|‘Jan will be giving someone presents but I do not know who.’|
(Grebenyova 2007, 52)
At the same time, however, in Polish, ordinary nouns can be remnants when they are associated with focus.
|‘Did you say that he will respect Maria?’|
|‘No. I said that (he will respect) JAN.’|
(Grebenyova 2007, 60)
This is the Japanese type of Focus Sluicing. That is, we may conclude from these data that Polish has both types: Wh- and Focus Sluicing.
Next consider (40), which our informant said is unacceptable.
|‘I heard that Jan will give someone presents but I do not know that who.’|
The above data show that the landing site of Focus movement is the position to the right of the complementizer ŻE, and that of Wh-Movement is to the left of ŻE. Schematically, each position is shown as follows:
|(41)||S-V [Wh-phrase / ŻE Noun] △. ← 31 | 32 →|
Leaving aside the categorial status of the deleted part by Sluicing, Polish Sluicing is the operation involved in deleting the portion following the moved element (either Wh-phrases or nouns).
Next, consider an example without Sluicing, such as (42), in which a wh-phrase appears to the right of a complementizer ŻE in embedded wh-questions:
|‘Who do you think Janek will invite?’|
(Tajsner 2015, 132)
In this case, the wh-phrase is an adjunct to TP, as mentioned in Tajsner (2015, fn 39). Therefore, (40) shows that a wh-phrase must move further to the Spec of FocP when it is qualified as a remnant of Wh-Sluicing, which requires FinP-deletion.
Although there remains the problem of what the deleted category is, we can conclude that Polish has the two types of sluicing: Wh-sluicing and Focus sluicing.
This paper provided two pieces of novel data which support our claim that Sluicing in Japanese is derived by focus movement followed by the deletion of FinP. One piece of evidence for it concerns KARA reason clauses and the other relates to SIKA-NAI NPI. Furthermore, based on a comparison of data from English, Japanese, and Polish, some typological implications were presented, and we argued that Polish has both types of Sluicing: Wh-Sluicing, as observed in English and Focus Sluicing, as in Japanese. Although our analysis in the last section is still rudimentary, we hope this will be a toehold toward a more serious contrastive research into Sluicing.
Akaso, Naoyuki, and Tomoko, Haraguchi. 2011. “On the Categorial Status of the Japanese Relative Clauses.” English Linguistics 28(1): 91–106.
Akaso, Naoyuki, and Tomoko, Haraguchi. 2016. “On Nominative in Japanese: Focus as a Case-licenser.” Poster presented at 9th Days of Swiss Linguistics, University of Geneva, Geneva, June 29–July 1.
Grebenyova, Lydia. 2007. “Sluicing in Slavic.” Journal of Slavic Linguistics 15 (1): 49–80.
Hiraiwa, Ken, and Shinchiro, Ishihara. 2002. “Missing Links: Cleft, Sluicing and ‘NO DA’ Construction in Japanese.” MIT Working Papers in Linguistics 43: 35–54.
Kawamura, Tomoko. 2008. “Adverbial because-clauses as Focal Elements.” Nanzan Linguistics Special Issue 3.2: 103–122.
Kim, Jeong-Seok. 1997. “Syntactic Focus Movement and Ellipsis: A Minimalist Approach.” PhD diss., University of Connecticut.
Kishimoto, Hideki. 2016. “Another Look at Negative Polarity Items in Japanese.” Japanese/Korean Linguistics 23: 99–118.
Merchant, Jason. 2001. The Syntax of Silence: Sluicing, Islands and the Theory of Ellipsis. Oxford: Oxford University Press.
Rizzi, Luizi. 1997. “The Fine Structure of the Left Periphery.” In Elements of Grammar: Handbook of Generative Grammar, edited by Liliane Haegeman, 281–337. Dordrecht: Kluwer.
Ross, John Robert. 1969. “Guess Who?” The Proceedings of the Chicago Linguistic Society 5: 252–286.
Sugawa, Seichi. 2008. “Ellipsis and Island Repair.” Nanzan Linguistics Special Issue 3.2: 165–183.
Tajsner, Przemysław. 2015. “On Focus Marking and Predication. Evidence from Polish with some Notes on Hausa.” Lingua Posnaniensis 57: 113–138.
Takahashi, Daiko. 1994. “Sluicing in Japanese.” Journal of East Asian Linguistics 3: 265–300.
1 This work was supported in part by JSPS KAKENHI Grant Number JP16K02785.
2 We employ the following terms to make our discussion clear.
John hired someone, but I don’t know who △.
The embedded clause, ‘who △’ in this case, in the second conjunct is called the sluice. The phrase surviving ellipsis in the sluice, ‘who’ in this case, is called the remnant. The delta △ indicates the elided portion in the sluice.
3 The abbreviations used in this paper are as follows: ACC = accusative, COMP = complementiser, DAT = dative, FIN = finiteness, FOC = focus, FUT = future, GEN = genitive, NOM = nominative, PRS = present, PST = past, Q = question marker, SG = singular, TOP = topic, 2nd = second person.
4 A focused phrase is shown in capital letters.
5 Japanese scrambling has been known for yielding surface effects (e.g. focus), but it is not necessary to appeal to scrambling/dislocation when a phrase is focalized. Stress can have the same effect without changing the word order.
Abstract: The paper overviews several instances of the scenario of an extracted right-branching domain in Hungarian. It shows, among others, how to account for the inverse-scope puzzle of the contrastive topic by attributing the puzzling word order to remnant movement, practically eliminating the puzzle this way. It also suggests remnant-movement based analyses of data that Kenesei (1998, 2002) discusses under the umbrella term Hungarian contrastive VP-focus. With these aims, it serves as the third member of a series of papers in which we plan to scrutinize syntactic cases in Hungarian puzzling in the following sense: there is a phrase in the sentence occupying a certain operator position, however, its predictable pragmasemantic contribution is not in harmony with the factual meaning of the given sentence. Instead, the factual meaning is such as if a more complex expression occupied the operator position in question. The straightforward solution which we attempt to elaborate in each case is always based on the assumption that in a semantic sense it is indeed the complex expression that occupies the given operator position while in a syntactic sense it is its appropriate remnant that can be found in the corresponding syntactic layer. The “appropriate remnant” is formed by extracting the right periphery of the complex expression, satisfying in this way language-specific realizations of universal constraints on right branching.
Keywords: inverse scope, VP-focus, remnant movement, Hungarian generative syntax
Our point of departure, as in Alberti and Farkas (2017), is the wide-spread opinion that “Hungarian is a more or less regular head-final language below the level of the (tensed) sentence, that is, in its NPs, APs, PPs, etc.” (Kenesei 2014: 225). The source ← 35 | 36 → of this opinion is Szabolcsi and Laczkó’s (1992: 189–190) argumentation against the existence of the postnominal complement zone in Hungarian noun phrases, on the basis of a constituency test resting upon focus constructions. Alberti et al. (2015), however, point out the inadequacy of the focus test as a constituency test, based on the property of the Hungarian focus that it cannot host right-branching phrases by any means, and they propose a contrastive-topic-based constituency test, exploiting the fact that the specifier of this layer readily tolerates right branching. In the light of this, Hungarian does not prove to be head-final (or proves to be head-final only “statistically”).1
This paper, as a continuation of Alberti and Farkas (2017), is about the other side of the coin. There are (indeed) several syntactic positions in Hungarian which do not tolerate right branching (and which makes Hungarian appear a head-final language). Nevertheless, even such a position, marked as (Spec, αP) in Figure 1 below, can host a right branching constituent, βP, at the cost of extracting the right-branching part, γP, in order to provide βP with the pragmaticosemantic contribution peculiar to the operator hosted in α. As shown in Figure 1, we should make the relevant syntactic scenario more precise in at least two respects. The raised phrase βP can be regarded as right branching not only relative to its lexical head β1 (whose representation in the figure is irrelevant at this point) but also relative to (some of) its functional heads. It is β2 that refers to the head whose complement is extracted, and hence this head is presented in the figure due to its immediate relevance (NB: β is the highest functional head, so the raised phrase of the lexical head β1 is practically the projection βP). The coincidence between β2 with the lexical head β1 is obviously not excluded. The constituent εP is the one which hosts the extracted γP. ← 36 | 37 →
Note in passing that the linear order of the relevant heads β2, β1, and α and the phonetic content of the extracted γP can also be accounted for by Ramchand’s (2014) approach, which discards the need to resort to unmotivated word order movements. Instead, it is argued that there is a specific linearization algorithm of the base structure: one can simply specify where a morpheme spells out by means of a diacritic @ rather than resort to syntactic movement.
It is at this point that some general remarks are worth making on the (unfavorably highly model-specific) decision of the syntactic category of εP, in order for us to be able to concentrate on the more relevant nodes αP and βP in Sections 2–4. In earlier syntactic models of the Hungarian sentence such as that of É. Kiss (2002: 61, 120), in which no morphosyntactic positions (i.e. abstract agreement layers) are considered and the VP is assumed to be flat (but a rich system of topic, quantifier, focus and aspectual layers is applied to account for the different word-order permutations and intonational variants occurring in Hungarian), it is straightforward to assume εP on the right periphery to be identical to this VP itself, with γP occupying an extra position as a sister of V (i.e. as a daughter of V’), at least as a default.
If a fully hierarchized (bifurcating) Grohmann-style (2003) model is assumed with three Prolific Domains, the default solution to the problem of εP is as follows. (i) As there is ab ovo no reason to assume that the extracted component γP as such gets a (new) thematic role or operator function, εP is not assumed to belong either to the thematic domain (Θ) or the operator domain (Ω). (ii) As the order of postverbal arguments depends on their phonetic weight rather than their thematic roles (see É. Kiss (2009) on the role of the Behaghel Law in Hungarian), ← 37 | 38 → the Φ domain must be made responsible for the order of postverbal dependents. Furthermore, γP – depending on its position in the given word order – should be placed somewhere in the specifier of a special extracted-component-hosting εP layer within this domain, containing layers such as Asp(ectual)P, T(ense)P, AgrSP, and AgrOP.2 ← 38 | 39 →
All in all, in what follows εP will not be discussed any more. The operator domain is assumed to contain layers headed by the five operators listed in Table 1. They can be identified on the basis of the system of the five types of pragmaticosemantic content given in the table as follows. If the reference r of a noun phrase is associated with a particular operator character in an utterance, by referring to r, a whole set of its pragmatic alternatives is evoked as background knowledge shared by the interlocutors. Such alternatives are thus not referred to explicitly, but only implicitly. Due to the given operator, some logical claim is predicated of the implicit referents.
|∃||Q∃: also-quantifier||CTop: contrastive topic|
|∀||Q∀: each-quantifier||Foc: (contrastive) focus|
|TopP: (non-contrastive) topic|
In all five examples shown in Table 2, the set of implicit referents consists of persons who can be regarded in a given context as alternatives to a person who is called Lilla. They all together form the relevant set. Suppose the implicit participants are Anna, Bea and Cili; so the relevant set now consists of four people. The corresponding sentence with an also-quantifier then provides the additional semantic information – in addition to the “explicit content” that Lilla came here, which is true in all the five variants – that what holds for Lilla also holds for (✓) at least one (∃) implicit participant. The additional information due to the contrastive topic is that what holds for Lilla does not hold for (¬) at least one (∃) implicit participant. The contribution of focus is captured in the table as follows: what holds for Lilla is a piece of information that uniformly (∀) does not hold for (¬) the implicit participants. The each-quantifier realizes the fourth logical possibility in the following sense: everyone is referred to implicitly (since the general expression mindenki “everyone” can have no other function in the given context than evoking what is termed above the relevant set), and hence the corresponding sentence can be interpreted as claiming that the information “someone came here” holds truly (✓) for each implicit participant (∀). As for the fifth operator, the ← 39 | 40 → non-contrastive topic, it can be placed in the system just sketched as an operator realizing the logical alternative of providing no information on the implicit participants. The translations illustrate these (context-based) semantic contributions.
As for the formal cues of these operators, relative to the basic variant with a topic, the two types of quantifier can be recognized by means of characteristic elements such as is ‘also’ and the morpheme mind- ‘each’. The two contrastive operators can be recognized relying on peculiar intonational and word-order phenomena. The contrastively topicalized element bears a special rising and then falling intonation (^) and is followed by a short pause (#). The focused element bears a strong FOCUS STRESS and seems to substitute for the preverb compared to the neutral word order. Note that we follow Brody (1990) in analyzing the placement of the verb in a focus construction in terms of head movement of the verb to the head of a Foc(us) functional projection that hosts the operator in its specifier.
We are going to overview several instances of the scenario of an extracted right-branching domain in Hungarian, sketched in Figure 1, starting in Section 2 with the types discussed in Alberti and Farkas (2017). Section 3 shows how to account for the inverse-scope puzzle of contrastive topics by attributing the puzzling word order to remnant movement, practically eliminating the puzzle this way. Section 4 provides remnant-movement based analyses of data that Kenesei (1998: 233–240) discusses under the umbrella term Hungarian contrastive VP-focus. Section 5 is a short summary. ← 40 | 41 →
2. A few examples of extracted right-branching domains in Hungarian from Alberti and Farkas (2017) as a point of departure for extending the given type of analysis
Our analysis is based on the assumption that aspect is expressed in Hungarian by raising a typically (right) branching argument selected by the verb to serve as its aspectualizer into a position, (Spec, AspP), left-adjacent to the surface position of the verb stem, in which (right) branching is not tolerated. It is to this tension between the opposite requirements that Alberti (2004) attributes the emergence of the Hungarian-specific climbing-preverb structures, presented in (1a-b) (see É. Kiss and van Riemsdijk (2004) and the rich underlying literature therein).
In (1a), the argument with the structure [AdvP up + sublative case-marked DP] is raised into (Spec, AspP) at the cost of divorcing from its right branching component, the sublative case-marked DP (see footnote 1). The argument for considering the sequence fel a fára to be a constituent is that it can serve as a possible short answer to such questions as ‘Where are you climbing?’ In (1b), akar ‘want’ “uses” the infinitival phrase (which is its argument referring to the object of demand) as its aspectualizer. As the InfP is perfective, its structure reflects that of its finite counterpart (1a), so the adverbial head ‘up’ in (Spec, AspInfP) constitutes its left periphery, with the InfP qualifying as the right branching part. Table 3 provides the relevant syntactic details of both analyses on the basis of Figure 1 in Section 1, together with the analysis of another type, presented in (1c).
|‘I (am going to) climb up the tree.’|
|‘I want to climb up the tree.’|
|‘And I WILL climb up the tree.’ [as a continuation of (1b)]|
Example (1c) illustrates a special indeed-construction in which the role of the matrix αP in Figure 1 is played by an also-quantifier layer. It can be used only as a continuation of a text in which what is claimed in the given sentence to take or have taken place has been “promised” as a plan or a prediction; (1c), for instance, can serve as a continuation of the expression of intent in (1b). Thus, the particle is ‘also’ (see Table 1 in Section 1 concerning the logical interpretation of this operator) confirms the plan held by the presupposition behind the fact expressed by the shaded word string in the last row of Table 3. Here the shaded strings of words happen to constitute an AspP, but VPs, FocPs and NegPs can also appear in the indeed-construction. The is particle is inserted in the strings immediately after the highest layer (β2P), triggering the extraction of the phonetic material in their complement (γP).
The example series in (2) evokes (see Alberti and Farkas 2017: Section 4) how we can focus a noun phrase which is so complex that, relative to the lexical noun head, it has both right- and left-branching parts – by raising its appropriate remnant into (Spec, FocP). This position, in contrast to (Spec, TopP), does not tolerate right branching from the lexical head (cf. (6a-a’)), but it tolerates left branching (2b), at least as an option in addition to another option according to which the remnant in (Spec, FocP) only consists of the dative case-marked possessor on the left periphery of the nominal expression (2c).
|(2)||A complex noun phrase in (Spec,FocP)|
|‘It is Móricz’s rhyme about the cows that he is going to recite (of several literary works).’|
Table 4 provides the relevant technical details. What is relevant in this section (see the last row of Table 4) is that the dative case-marked possessor is assumed to be hosted in an operator layer built upon the DP layer on the left periphery of the noun phrase, on which even further operator layers can be based in the spirit of the clausal-DP hypothesis (see (9b-c)). This hypothesis (Grohmann 2003: 200) argues that “essentially all types of properties found in the clause can also be found in the nominal layer” (see also Farkas et al. 2017). ← 42 | 43 →
3. The puzzle of inverse-scope taking contrastive topics
In Hungarian, the scope order of preverbal (non-in-situ) constituents corresponds to their surface order (3a-a’); as shown by the proposed paraphrase of (3b), however, quantifiers in (Spec,CTopP) give the impression of having inverse scope, apparently violating this generalization (É. Kiss 2002: 25). The solution to this scope-inversion puzzle in Hungarian proposed by Gyuris (2009: 150) rests upon the following, unexplained, observation: “only those Hungarian sentences containing a contrastive topic are well-formed that have well-formed counterparts with the contrastive topic expression in postverbal position.”
|(3)||Illustration of the inverse-scope puzzle in Hungarian|
|‘It holds only for few students that the given student has read each novel.’|
|‘It holds for each novel that few students have read it.’|
|Meaning: practically the same as that given in (3a), and not that given in (3a’)|
Our remnant-movement-based techinique straightforwardly provides a solution to the scope-inversion puzzle concerning (3b), with the following point of departure: a certain part of the aforementioned “well-formed counterpart”, which is (3a) in the particular case, should undergo remnant movement. First of all, however, a tiny detail concerning the precise postverbal position of the quantifier should be clarified. As exemplified in (iii) in footnote 2, there is a perfect word-order variant with the preverb preceding the quantifier, but there is also a slightly dispreferred variant with the quantifier separating the verb stem from the preverb (NB: it is exactly this separation that the dispreferred status can be attributed to, due to not only their semantic togetherness but also a phonetic demand that requires the post-focal unstressed zone to be as long as possible). The latter variant is the one in ← 43 | 44 → which the quantifier is spelt out in the Grohmannian operator domain ΩΔ (above AspP), which makes this word order unambiguous. The former variant, in which the quantifier can be regarded as spelt out in the agreement or in the thematic domain, is ambiguous (depending on slight differences in stress pattern), since its covered Ω-position can be either that shown in (3a) or that shown in (3a’). As (3b) with the given stress pattern is unambiguous, its “well-formed counterpart” is the word-order variant of (3a) with the quantifier in ΩΔ, namely, in (Spec, Q∀P). We will argue that it is exactly this Q∀P projection, or rather its non-right-branching remnant, that is moved to the specifier of a CTopP-layer built upon FocP. Before concluding the analysis, however, let us consider the data whose analysis has suggested the idea of accounting for narrow-scope taking contrastive topics by means of remnant movement.
The data in question contain noun phrases as complex as those presented in (2) with the exception that these are headed by deverbal nominals. If a quantifier belongs to the head of a deverbal nominal construction as its argument, ambiguity may emerge due to a potential (“external-scope”) reading according to which the quantifier is understood as belonging to the finite/matrix verb (4a), in addition to the primary (“internal-scope”) reading (4a’).
|(4)||External/internal-scope taking argument of a nominal head|
|‘It holds for both colleagues that I am against the option according to which the given colleague would be sent away [neither of them should be sent away].’|
|‘As for the option according to which both colleagues would be sent away, I am against that [but I am not against sending away either of them].’|
|Meaning: as in (4a).|
|Meaning: as in (4a’).|
Farkas et al. (2017) attribute the possibility of external-operator interpretation to a universal rule concerning the percolation of arbitrary operator features. This is the rule that Horvath (1997: 548) bases her theory on wh-feature percolation in certain Hungarian interrogative subordinate constructions (Horvath 1997: 547–557). Kenesei also applies essentially the same universal rule (Kenesei 1998: 223–225), which he provides as a minimalist reformulation (Chomsky 1995) of a ← 44 | 45 → rule by Höhle (1982) and Selkirk (1984), to certain focus constructions in Hungarian. We thus apply the rule to (some kind of) universal quantifier feature, an each-feature. Determining components of the rule are that (i) the original position of the percolating feature should be an argument, and not an adjunct (Horvath 1997: 540–546, Kenesei 1998: 228), and (ii) it ceases to constitute an operator of the given kind (Horvath 1997: 549–550). As for formal details, while the quantifier determiner prefix mind- ‘each’ is morphologically attached to an element in the depth of a noun phrase, the pragmasemantic contribution of the each-feature counts as if it were attached to the noun head of the given noun phrase.
Thus, in (4a) the object of the matrix verb counts as the quantifier of the matrix verb in the clause-level information structure, while in (4a’) the phonetically marked contrastive topic position of the same object makes it unambiguous that the quantifier status belongs to the possessor within the complex internal structure of the given object. As for the internal information-structural function of the possessor in (4a), it ceases to constitute a quantifier within its matrix noun phrase due to the feature percolation but, as it remains foregrounded within the deverbal nominal construction, it should be considered to be a (non-contrastive) topic of the nominal head. The surprising fact comes now: if the complex noun phrase occupies its operator position on the left periphery of the sentence as a remnant, throwing off its right-branching periphery – which is not obligatory in the case of a universal quantifier or a contrastive topic but an option available (4b-b’) – the corresponding interpretations do not change at all. Table 5 provides the technical details. An interesting element is that the possessor, independently of the splitting, retains its aforementioned internal topic/quantifier function (NB: the non-contrastive topic function is the result of the percolation of the each-feature).
What has been called surprising above is surprising only at first glance, since we have attributed no pragmasemantic contribution to remnant formation so far. The lack of any change in meaning is thus what should be predicted on the basis of the analyses presented in Section 2. In other words, it is all the same with respect to ← 45 | 46 → both the external and the internal semantic factors whether a complex deverbal nominal construction is intact (4a-a’) or split (4b-b’).
Observe that the same analysis can also be applied to finite constructions. Relative to the variant in (5a), the one with a contrastive topic on the left periphery in (5a’) seems to have an inverse-scope taking quantifier. What takes scope over the focus, however, is not the accusative case-marked argument of the verb but the remnant of the phonetic form of the proposition describing the option that ‘both colleagues have been sent away from x, where x is a project’. This is the Q∀P projection appearing in the complement of the wide-scope taking focus in the sentence variant presented in (5a”), which is the Gyuris-style “well-formed counterpart” of (5a’). At this point, however, there emerges a question. Does contrastively topicalizing a proposition supply no pragmasemantic contribution? As can be seen below, different translations can be associated with the variants (5a’-a”) – but, what is crucial, with a coinciding F>Q scope hierarchy between the elative case-marked and the accusative case-marked argument of the verb. Anyway, the two readings practically coincide: what is explicitly indicated in (5a’) by the contrastive topicalization can be regarded as an implication in the “well-formed counterpart” in (5a”), since a wide-scope taking focus over a proposition P indicates or implies that alternatives of P hold for the elements of the complement subset of the relevant set in the background of the focus. Hence, the option expressed by P can safely be claimed to have alternatives in (5a”), too.
|(5)||External/internal-scope taking arguments of a nominal head|
|‘It holds for both colleagues that he has been sent away only from one of the projects [i.e., both could remain in one of the projects].’|
|‘As for the option according to which both colleagues have been sent away from the given project, that holds only for one of the projects [hence, another option holds for the other project].’|
|‘It holds only for one of the projects that both colleagues have been sent away from it.’|
|‘As for the option of having read each novel, that holds for few students [hence, other options hold for the majority of students].’|
Table 6 provides the crucial elements of our remnant-movement-based analysis concerning (5a’), together with the same analysis concerning (3b), repeated here as (5b), but with a translation exactly expresses the process of contrastively topicalizing the proposition associated with the Q∀P projection. Note that our ← 46 | 47 → approach in which we derive (5b) (i.e., (3b)) from the marked variant of (3a) is corroborated exactly by the marked status of the variant in question: although at the cost of blurring the scope hierarchy, we avoid the awkward stress pattern (and the characteristic contrastive-topic stress pattern serves as a sufficient warning that the calculation of the scope hierarchy requires care).
This approach resembles Baron Münchhaussen’s method of pulling himself out of a mire by his own hair (see the picture): a high operator-layer (αP) of the finite verb (β1) offers its specifier to one of its other operator-layers, a much lower positioned one (β2P), which moves from under a mediate operator-layer λP (here λP=FocP), taking scope over β2P, at the cost of throwing off a fourth projection of the finite verb (γP) and producing the illussion that β2P takes scope over δP. This is, however, only an illusion, indeed, since only the phonetic material of the β2P-layer appears in (Spec,αP), as the representative of the propositional content of the entire β2P projection including its whole complement on the right periphery of the given sentence. The benefit is a sentence type with a usual operator order, and it seems that the inverse-scope relation is indicated safely by the peculiar contrastive-topic stress.
4. The Hungarian contrastive VP-focus
This section claims that four observations on Hungarian contrastive VP-foci reported by Kenesei (1998: 233–240) without any formal explanation can be explained on the basis of the following straightforward hypothesis: what is going on is indeed nothing else but the focusing of a higher VP-projection by Baron ← 47 | 48 → Münchhaussen’s method discussed above, with the purpose of referring to an option typically in contrast with one of its alternatives.
In (6a), for instance, the particular word order suggests an analysis in which the option of reading out Hamlet to Mari is genuinely expressed by a TopP (with the drama foregrounded), just like the option of taking apart a car into tiny pieces in (6a’) (with the car foregrounded). Then these two options are compared by focusing their phonetic forms, or rather the appropriate remnants of these phonetic forms. As presented in Table 7, AspPs are assumed to be extracted in order for the TopPs to make it possible to occupy (Spec,FocP) via depriving them of the right-branching AspPs. Note that the AspPs in question lack the phonetic form of the verb since the Hungarian verb in the contrastive-focus construction has the distinguished mission of occupying the Foc position (see Section 1). In (6b-b’) a similar comparison between options is carried out by focusing only one option (6b) while applying an also-quantifier to the other (6b’). Note that the matrix quantifier-construction is such that it does not exploit the phonetic form of the verb, so the verb appears in the extracted material γP. It is also important to note that the extracted material can be recognized on the basis of the special stress pattern typical of the matrix operator α (and it is needless to assume that parts of the extracted material should occupy any potential extraordinary focus positions on the right periphery because only βP as a whole is associated with the contrastive-focus interpretation, cf. Kenesei 2002: 305, rule (23)).
Kenesei’s (1998) first observation to be accounted for is illustrated by the minimal pair presented in (7): it is dispreferred for βP to start with the verb stem, that is, it is dispreferred for the remnant to consist of the single verb stem (7b).
|(7)||Can a verb constitute the material of the remnant?|
|‘What Peter was doing was run around rather than read out Hamlet to Mari.’|
|Intended meaning: (7a)|
The only syntactic analysis, demonstrated in Table 7 above, provides the straightforward explanation of the dispreferred (or, according to several speakers, definitely unacceptable) word order discussed. The problem is with the double role of the verb stem: its semantic content belongs to the “option” to be expressed by βP, on the one hand, but the verb stem is the phonetic pillar that the Hungarian ← 49 | 50 → focus construction rests upon, on the other. Given its latter function, it should be unstressed, in contrast to its former function.
Kenesei’s second and third observations pertain to the choice of the topic within βP (see Table 7 above, again).
|(8)||Different types of topic in the remnant|
|‘What Peter was doing was read out Hamlet in the garden rather than swim.’|
|Intended meaning: (7a)|
|In’d meaning: ‘What Peter was doing was read out Hamlet aloud rather than swim.’|
Here the radical differences in acceptability between (8a-c) can simply be attributed to the usual differences with respect to topicalizability. An argument (8a) is a better (non-contrastive) topic than an adjunct (8b), and only referential adjuncts can serve as non-contrastive topics; non-referential adjuncts (8c) are not suitable for this function. Observe that this argumentation intensively exploits the assumption that βP is a TopP and should be focused as such (see Table 7 above).
Kenesei’s fourth observation pertains to idioms. As shown in (9), idioms are not ab ovo excluded from the “VP-focus” construction under discussion, presumably due to the fact that it is the idiom as a whole that is focused, even if its meaningless parts wear focus stresses. There are, however, radical differences:
Again, our explanation is based on the internal structure and word order of βP, which is summarized in Table 7 above. Idioms also have a tripartite Grohmannian cycle construction. That is, certain idiom parts can, or must, or must not, serve as particular operators (formally, while lacking the corresponding pragmasemantic contribution). ‘The wet blanket’, for instance, cannot be spelt out as either the topic or the focus of the corresponding expression (9a). Its such spelling out obligatorily triggers the literal meaning with a real wet blanket. On the idiomatic reading, ‘the wet blanket’ should appear postverbally. However, the position of the inessive case-marked phrase in the idiom presented in (9b) and the accusative case-marked one in (9b”) is (Spec,FocP) while that of the accusative case-marked one in (9b’) is (Spec,AspP). These three conclusions are based on their stress patterns under normal circumstances; nevertheless, it is not easy to differentiate the occupation of (Spec,AspP) from that of (Spec,FocP) (or sometimes from that of (Spec,TopP)) without the separate interpretation of the given noun phrases (not at our disposal on the idiomatic readings). Fortunately, our analysis exploits only their “normal” word order, and it is sure that they can appear preverbally within the corresponding idioms. And that is why they are suitable for playing the role of the remnant in the “VP-focus” construction as construed in our approach. Thus, even the strange case in which the remnant of a FocP occupies the specifier of a higher FocP (9b,b”) can happen.
In a series of papers, with this being the third one following Alberti (2004) and Alberti and Farkas (2017), we plan to scrutinize syntactic cases in Hungarian puzzling in the following sense: there is a phrase in the sentence occupying a certain operator position, however, its predictable pragmasemantic contribution is not in harmony with the factual meaning of the given sentence. Instead, the factual meaning is such as if a more complex expression occupied the operator position in question. The straightforward solution which we attempt to elaborate in each case is always based on the assumption that in a semantic sense it is indeed the ← 51 | 52 → complex expression that occupies the given operator position while in a syntactic sense it is its appropriate remnant that can be found in the corresponding syntactic layer. The “appropriate remnant” is formed by extracting the right periphery of the complex expression, satisfying in this way language-specific constraints on right branching (Hinterhölzl 2010).
After a brief overview of some basic cases of the afore-mentioned problematic syntactic scenario, we showed a way to account for the inverse-scope puzzle of contrastive topics (Gyuris 2009), practically via the elimination of the problem. In addition, we provided remnant-movement based analyses of data discussed by Kenesei (1998: 233–240) under the umbrella term Hungarian contrastive VP-focus.
It is left to future research to prepare a thorough overview of all the cases of “inverse-scope taking contrastive topics” in Hungarian, which could be based on Gyuris’s (2009) book, and to comprehensively elaborate on the topic of mirror-focus constructions (Alberti and Medve 2000).
Alberti, Gábor. 2004. “Climbing for Aspect – with no Rucksack.” In Verb Clusters: A study of Hungarian, German and Dutch. Linguistics Today 69, edited by Katalin É. Kiss, and Henk van Riemsdijk, 253–289. Amsterdam/Philadelphia: John Benjamins.
Alberti, Gábor, and Judit Farkas. 2017. “Right Branching in Hungarian: Moving Remnants.” In Language Use and Linguistic Structure, Proceedings of the Olomouc Linguistics Colloquium 2016, edited by Joseph Emmonds, and Markéta Janebová, 209–224. Olomouc: Palacký University.
Alberti, Gábor, Judit Farkas, and Veronika Szabó. 2015. “Arguments for Arguments in the Complement Zone of the Hungarian Nominal Head.” In Approaches to Hungarian 14, edited by Katalin É. Kiss, Balázs Surányi, and Éva Dékány, 3–36. Amsterdam: John Benjamins.
Alberti, Gábor, and Anna Medve. 2000. “Focus Constructions and the Scope–Inversion Puzzle in Hungarian.” Approaches to Hungarian 7, 93–118. Szeged: JATEPress.
Brody, Michael. 1990. “Some Remarks on the Focus Field in Hungarian.” In UCL Working Papers in Linguistics, edited by John Harris, 2: 201–226. London: University College London.
Chomsky, Noam. 1995. The Minimalist Program. Cambridge, MA: MIT Press.
Farkas, Judit, Veronika Szabó, and Gábor Alberti. 2017. “Information-structurally (Un)ambiguous Nominal Constructions in Hungarian.” Manuscript. Department of Linguistics, University of Pécs. ← 52 | 53 →
Grohmann, Kleanthes K. 2003. Prolific Domains: On the Anti-Locality of Movement Dependencies. Linguistik Aktuell 66. Amsterdam: John Benjamins.
Gyuris, Beáta. 2009. The Semantics and Pragmatics of the Contrastive Topic in Hungarian. Budapest: Lexica.
Hinterhölzl, Roland. 2010. “Collapsing the Head Final Filter and the Head Complement Parameter.” Working Papers in Linguistics 20: 35–65. Venice: University of Venice.
Höhle, Tilman. 1982. “Explikationen for ‘normale Betonung’ und ‘normale Vorstellung’.” In Satzglieder in Deutschen, edited by Werner Abraham, 75–154. Tübingen: Gunther Narr.
Horvath, Julia. 1997. “The status of ‘wh-expletives’ and the partial wh-movement construction in Hungarian.” Natural Language and Linguistic Theory 15: 509–572.
Kenesei, István. 1998. “Argumentumszerkezet és VP-fókusz [Argument Structure and VP Focus].” In A mai magyar nyelv leírásának újabb módszerei [New Methods in the Description of Today’s Hungarian Language] III, edited by László Büky, and Márta Maleczki, 223–242. Szeged: JATEPress.
Kenesei, István. 2002. “A fókusz megjelenítése a szintaxisban és a fonológiában [Representing Focus in Syntax and in Phonology].” In A mai magyar nyelv leírásának újabb módszerei [New Methods in the Description of Contemporary Hungarian Language] László Büky, and Márta Maleczki (eds.) V, 289–310. Szeged: SzTE.
Kenesei, István. 2014. “On a multifunctional derivational affix.” Word Structure 7.2: 214–239.
Kiss, Katalin É. 2002. The Syntax of Hungarian. Cambridge: Cambridge University Press.
Kiss, Katalin É. 2009. “Is Free Postverbal Order in Hungarian a Syntactic or a PF Phenomenon?” In The Sound Pattern of Syntax, edited by Nomi Erteschik-Shir, and Lisa Rochman, 53–71. Oxford and New York: Oxford University Press.
Kiss, Katalin É, and Henk van Riemsdijk, eds. 2004. Verb Clusters: A Study of Hungarian, German and Dutch. Linguistics Today 69. Amsterdam/Philadelphia: John Benjamins.
Piñón, Christopher. 1995. “Around the Progressive in Hungarian.” In Approaches to Hungarian 5, edited by István Kenesei, 153–190. Szeged: JATEPress.
Ramchand, Gillian. 2014. “Deriving variable linearization. A commentary on Simpson and Syed (2013).” Natural Language and Linguistic Theory 32: 263–282.
Surányi, Balázs. 2009. “Verbal particles inside and outside vP.” Acta Linguistica Hungarica 56: 201–249.
Szabolcsi, Anna, and Tibor Laczkó. 1992. “A főnévi csoport szerkezete [The Structure of Noun Phrase].” Strukturális magyar nyelvtan 1. Mondattan [A Structuralist Hungarian Grammar I. Syntax], edited by Ferenc Kiefer, 179–298. Budapest: Akadémiai Kiadó.
* We are grateful to NKFIH 120073 for their financial support. This paper is based on Alberti and Farkas (2017).
1 The following abbreviations are used in the glosses:
(i) case suffixes: ACC(usative), DAT(ive), DEL(ative), ELA(tive), INE(essive), INS(trumental), NOM(inative), SUB(lative);
(ii) other suffixes on nouns: PL(ural), POSS (possessedness suffix);
(iii) affixes on verbs: PAST, 1SG/…/3PL (agreement suffixes);
(iv) derivational suffixes: INF(initive).
2 The thematic domain (Θ) of a hierarchized Hungarian clause structure is analysed in Surányi’s (2009: 234, 237, 238) sophisticated model as follows. Besides the customary VP layer (“containing oblique, goal and theme arguments, as well as internal stative locatives”) and vP layer (“hosting the external argument subjects, and probably also dominating source and orientation of trajectory adverbials”), we need a position for preverbs and other verbal modifiers “below the base position of those elements that cannot “incorporate” [into the verb] and above the base position of those that can.” The given layer is termed PredP by Surányi, because the (phrasal) verbal modifier and the verb form a complex predicate. However, we term this thematic layer θOblP, given the following typical relation between the preverb and an oblique argument: if a preverb has a compositional meaning contribution, it characterizes the relation between the kind of movement described by the VP and a Goal, Source or Location described by the given oblique argument. Sentences (i-ii) can serve as a sketchy illustration of this syntactic model. The base positions from bottom up are as follows: the accusative case-marked theme is base-generated in the VP layer, then the adverbial phrase (rá) a megcímzett képeslapra ‘(onto) the addressed postcard.SUB’ is in (Spec, PredP), and the nominative case-marked subject is hosted in (Spec, vP). An argument for considering the adverbial phrase as a (later split) constituent is the fact that both versions can serve as short answers to questions such as ‘Where did Ili put a stamp?’ Another relevant point of the structure is the analysis of the phonological unit containing the verb (with one stressed syllable, which is the first syllable, as it always is in Hungarian). Surányi (2009: 226–229) accounts for this unit by assuming the following specifier–head configuration: the surface position of the verb is the T(ense) head, and the verbal modifier, which is the accusative case-marked bare noun phrase in (i) and an adverbial phrase in (ii), occupies (Spec, T). At this point, however, we prefer the somewhat different earlier solutions (Piñón 1995, Alberti 2004), based on the assumption that such verbal modifiers are in (Spec, AspP); thus we assume a separate aspectual layer over TP. The three word-order variants presented in (iii), all acceptable, illustrate that an accusative-case marked quantifier, for instance, can be spelled out in all three Grohmann-domains, indicating, from left to right, its operator position, its thematic position, and its Φ-position on the basis of Behaghel’s Law (É. Kiss 2009).
|‘Finally, Ili put a stamp (or more stamps) on the addressed postcard.’|
|‘Finally, Ili put a stamp on the addressed postcard.’|
|‘Ili put 〈both ornamented stamps〉 on the postcard.’ (〈m2db〉: mindkét díszes bélyeget ‘both ornamented stamp.ACC’ in Hungarian)|
Abstract: Following the well-grounded assumption that phrases are endocentric, the absence of exponents of a given category in a language is taken to signify that the language lacks the particular type of phrasal projections in its inventory. To the extent that this claim is correct, languages without articles (the flagship category of determiners) are often considered to belong to the “NP-camp” (Fukui 1986, 1988; Trenkic 2004; Bošković 2005, 2007, 2009, 2012a, 2012b, 2013, 2014a, 2014b, 2015a, 2015b; Despić 2011, 2013, among others). In this respect, demonstrative pronouns constitute a puzzling issue. Their syntactic location and categorial classification has long been on the linguistic agenda (Longobardi 1994; Zlatić 1997; Migdalski 2003; Bošković 1994, 2005, 2007, 2009; Bašić 2004; Trenkic 2004; Pereltsvaig 2007; Rutkowski 2007; Norris 2014, among many others).
This article examines the selected empirical facts concerning the syntactic behaviour of demonstrative pronouns in Polish, in particular the fact that they are not easily extractable, i.e. there is a marked contrast between the acceptability of extractions involving demonstratives and other nominal constituents (e.g. wh-, prenominal adjectives). A closer look at the relevant statistics obtained in the course of an empirical acceptability study of the specific extraction types substantiates the observation that demonstratives exert the characteristics of phrases (DemP) and are located in the Agr domain of N. By virtue of their specific location, they interact in some interesting ways with other nominal constituents, including fronted topics. At a more general level, the relevant interactions and word order configurations give reasonable grounds to assume that (languages like) Polish, despite appearances (i.e. lack of articles), nevertheless feature a fairly elaborate nominal structure with a rich left periphery, possibly including DP. ← 55 | 56 →
Keywords: demonstratives, numerals, Left Branch Extraction (LBE), Antilocality, Phase Impenetrability Condition (PIC), determiners
The aim of this article is to present two interesting empirical facts concerning the syntactic conduct of demonstrative pronouns in Polish, i.e. their relative reluctance to undergo Left Branch Extraction (LBE), as well as their interaction with other, discourse-related movements in the nominal left periphery. On the basis of these observations, it will be argued that although demonstrative pronouns (Demonstrative Phrases under this account) may technically not belong to the category D, their syntactic behaviour provides circumstantial evidence for the existence of DP or an equivalent functional projection placed high up in the left periphery of the nominal. In a more general sense, an assumption along these lines provides further support for the DP-Hypothesis (Abney 1987) as part of UG. Thus, Section 2 offers a succinct overview of the selected cross-linguistic arguments relating to the syntactic classification of demonstratives. Section 3 outlines the gist of the alleged correlation between LBE and the presence/absence of determiners (Uriagereka 1988; Zlatić 1997; Bošković 2005, among others). Section 4 reports on the results of the survey into various extraction types from nominals in Polish, focusing specifically on demonstratives. Section 5 raises the question of the treatment of demonstratives in the syntactic hierarchy as heads or phrases and examines their interaction with the topicalised N-complements in the left periphery of the nominal. Section 6 sums up the observations and concludes the discussion.
2. Demonstratives are (not?) exponents of D – selected arguments
Cross-linguistically, demonstratives display different syntactic characteristics as to their function, position in the nominal structure and interactions with other NP-constituents. These seem, to a large extent, determined by their potential syntactic classification as (a subgroup of) determiners. While the latter claim is generally uncontested in languages with overt articles like English, the issue is far from clear-cut whenever overt articles are missing. For instance, as noted by Fukui (1986, 1988), English and Japanese seem to ascribe different functions to the demonstrative pronoun, which is manifested through their syntactic behaviour. Thus, while English obligatorily places them at the left edge of the nominal, ← 56 | 57 → the strict demonstrative pronoun – modifier order need not be maintained in the latter, witness the difference between (1a) and (1b) (Shin-Sook 1991: 731).
|‘Professor Yamada’s that/the lecture’|
Fukui further argues that functional categories such as C, I or D do not exist in Japanese, and thus lexical categories (N, V, A, and P) may project the X’ level and get reiterated. Another interesting observation concerns the function of demonstratives with respect to possessive and indefinite pronouns. In English, due to their complementary distribution with articles, the three are classified as the exponents of the same grammatical category of determiners. This, however, is not the case cross-linguistically, witness the examples below from, respectively, Hungarian, Javanese and Italian (Progovac 1998: 166).
Turning to Slavic languages, given that most of them (excluding Bulgarian and Macedonian) do not feature articles in their inventory of lexical items, demonstratives, possessives and indefinites are (by analogy to English) often considered to be determiners. However, Corver (1990: 332) raises a number of serious arguments against this classification. As he observes, they are morphologically similar to adjectives, can appear in predicative positions and need not precede other nominal constituents (see also Fukui 1986; Shin-Sook 1991). The three observations are presented for Polish (3) and Czech (Corver 1990: 332–333) in (4), respectively.
|‘this good student’|
|‘this good student’|
|‘This pen is mine / new.’|
|‘this good student’|
Another interesting argument concerns NP/DP-internal movement facts. In particular, Progovac (1998) notes a peculiar configurational discrepancy between nominal expressions with intensifying prenominal adjectives and regular nouns [ADJ–N], and those with pronouns [PRON–ADJ], which is also observed for English (5), see Abney (1987). Similar discrepancy is observed for Polish by Rutkowski (2007).1 Relevant examples are provided in (6) (Serbo-Croatian, Progovac 1998: 168) and (7) (Polish, Rutkowski 2007: 85).
|(5)||a.||I want to talk to someone famous / *famous someone.|
|‘Chomsky/he himself slept during that lecture.’|
|he||alone [himself]/||alone [himself]||he||slept||on||that||lecture|
|‘Chomsky/he himself slept during that lecture.’|
Following Abney’s (1987) original reasoning, Progovac (1998) assumes a fairly elaborate sequence of functional projections which the pronoun ‘passes through’ before it eventually reaches the DP (Spec and head position for possessive and personal pronouns, respectively). In that sense, the facts observed in (6) and (7) seem to bear resemblance to the well-known N-D movement facts from Swedish or Italian (see Delsing 1993; Longobardi 1994).
Turning to demonstratives, a correlation along the lines of (6) and (7) may also be observed in (8).
|‘That man was old’|
|‘That man was old’|
However, Zlatić (1998: 3) offers an alternative interpretation of these facts. Specifically, she discards the analysis of the N-DEM word order in terms of N-D movement (see Longobardi 1994), since such an analysis could not be extended to, for instance, Serbo-Croatian, in which (whenever three constituents are present, i.e. DEM, A, N), the resulting word order is either A-N-DEM or DEM-N-A. Although Zlatić (1998) expressly excludes N-DEM-A, judgements are not uniform (see Bašić (2004) for Serbian; Caruso (2011, 2013) for Croatian).
3. The relevance of the Left Branch Extraction
Another interesting point that may shed some light on the status of demonstratives concerns the correlation between the presence/absence of determiners and Left Branch Extraction (LBE, see Ross 1967; Uriagereka 1988; Zlatić 1998; Bošković 2005, 2007, 2009, among others). Examples in (9) illustrate the illicit extraction of, respectively, a determiner, an adjective and a wh-.
|(9)||a.||* Thosei he took [ti apples]|
|b.||* Interestingi he read [ti books]|
|c.||* Whichi did he find [ti key]? ← 59 | 60 →|
Among Slavic languages, only Bulgarian and Macedonian follow the pattern of English in that they disallow LBE, relevant examples from Bulgarian (Bošković 2005: 3) provided below.
|(10)||a.||* Kakvai||prodade||Petko||[ti kola]?|
|‘What kind of car did Petko sell?’|
|b.||[Kakva kola]i prodade Petko ti?|
However, analogous examples are apparently acceptable in other Slavic languages, including Polish (11).
|(11)||a.||(?) Tamtei||zabrał||[ti jabłka]|
|‘He took those apples.’|
|‘He read interesting books.’|
|‘Whose key did he find?’|
Thus, Uriagiereka (1988) observes that LBE is possible only in languages without overt articles. Corver (1992) captures this observation in his GB-based account drawing on the concept of the Empty Category Principle. In particular, he takes the difference between (9) and (10) on the one hand, and (11) on the other to instantiate the difference between languages featuring functional (DP) and lexical (NP) nominal projections, respectively. The proposal seems to ‘make the right cut’ for languages like English, Bulgarian or Macedonian. Additionally, as Boeckx (2003: 40) points out, Corver’s (1992) account also predicts the difference between (12a) and (12b), examples from Chamorro (data from Chung 1998: 282–286, ex. 9a, 14c).
|‘Whose child had you scolded?’|
|b.||* Hayi||un-li’i’||[i||gä’- ña||ga’lagu t°]|
|‘Whose dog did you see?’|
The presence of the overt article i in (12b) renders the sentence ungrammatical, i.e. the article apparently blocks the wh- extraction from within the nominal. On the other hand, Hungarian has got overt determiners and still allows extraction, provided it affects elements located at the edge of the nominal phrase, the position marked with Dative (Szabolcsi 1983, examples from Gavruseva 2000: 744, ex. 1a-c). ← 60 | 61 →
|c.||*(a)||kiNOM/||ki-nekDAT||ismer-te-ték||[tDAT a (*tNOM) vendég-e-ø]|
|‘Whose guest did you know?’|
The minimalist derivation imposes a number of serious constraints on the course of the syntactic derivation and dispenses with the relation Government, which effectively renders ECP-based approaches like Corver’s (1992) no longer applicable. In turn, Bošković’s phase-based account of the LBE (Bošković 2005, 2012a,b, 2013) is based on a neat interplay of the Phase Impenetrability Condition and the Antilocality Condition (Bošković 1994; Saito & Murasugi 1995, Grohmann 2000/2003; Ticio 2003, among others), both provided below.
- ISBN (PDF)
- ISBN (ePUB)
- ISBN (MOBI)
- ISBN (Book)
- Publication date
- 2019 (January)
- Universal principles Parameters Structure Syntax Phonology Morphology Semantics
- Berlin, Bern, Bruxelles, New York, Oxford, Warszawa, Wien, 2018. 253 pp., 2 fig. col., 10 fig. b/w, 20 tables,