MIT Working Papers in Linguistics #62

Proceedings of the 7th Workshop on Altaic Formal Linguistics (WAFL7)

Editor: Andrew Simpson, Year:2011

This volume contains the proceedings of WAFL7 with following contributing papers:

Author: Edith Aldridge (University of Washington)
 Keywords: hentai kambun, Old Japanese, ditransitives, object shift
 Title: Hentai Kambun and Old Japanese

This paper proposes a structural syntactic analysis of word order in the Japanese 8th century hentai kambun text Kojiki. Word order in hentai kambun texts largely resembles SVO order found in contemporary Chinese and is traditionally assumed to be an imitation of Chinese. The remaining word order types which deviate from Chinese are consequently treated as mistakes in writing Chinese. In contrast to the traditional view, this paper proposes that the word order of the Kojiki can receive a straightforward and uniform analysis if it is analyzed as having been systematically derived from Japanese and not as an inferior imitation of Chinese. Specifically, I propose that the word order patterns in the Kojiki can be uniformly accounted for by assuming that all sentences are derived as Japanese. The resemblance to Chinese, as well as the deviations from both Japanese and Chinese, are the result of reordering heads to the left of their complements within vPs and PPs. All other constituents are linearized as they would be in Japanese order.

Authors: Alya Asarina and Jeremy Hartman (Massachusetts Institute of Technology)
 Keywords: genitive subjects, relative clauses, null nouns, Phase Impenetrability Condition, Uyghur
Title: Genitive Subject Licensing in Uyghur Subordinate Clauses

In this paper, we analyze embedded clauses in Uyghur, a Turkic language spoken in the Xinjiang Uyghur Autonomous Region of China. We argue that genitive case on the subjects of these clauses is licensed by agreement with a clause-external D head. We also argue that these clauses are full CPs. Putting these claims together, we show that Uyghur exhibits agreement and case-assignment over a CP boundary, a configuration that is inconsistent with Chomsky’s (1998) Phase Impenetrability Condition (PIC). The Uyghur data thus support adopting the weaker version of the PIC proposed by Chomsky (2001).

Author: Jaehoon Choi (University of Arizona)
 Keywords: Sino-Korean, compounding, disyllabicity
 Title: Disyllabicity of Sino-Korean Compounds: Evidence from Ca-compounding

This paper investigates the disyllabic requirement imposed on compounds in Sino-Korean within the framework of Distributed Morphology. I develop a device that controls syllabic length of Sino-Korean compounds, focusing on ca-compounds in particular. Basically, Sino-Korean compounds are formed via Root Merger (Zhang 2007) and conflation (Harley 2004). Building upon this, I, in the spirit of Arad (2003) and Harley (2004), propose that the disyllabic tendency of Sino-Korean compounds is due to a prosodic constraint imposed on the Sino-Korean nominal functional head n0. That is, the Sino-Korean n0 can accommodate at most two syllables. Then, an account for the ungrammaticality of trisyllabic ca-compounds becomes available. There are some consequences of this analysis. First, the derivation of quadrasyllabic ca-compounds is made possible, given that quadrasyllabic ca-compounds are coordinate phrases semantically. Second, it becomes necessary to posit two distinct nominal heads in Korean: Sino-Korean n0 and Native-Korean n0, the former of which cannot host more than two syllables, while the latter of which can. This distinction accounts for the different syllabic-length tendency between Sino-Korean and Native-Korean. Lastly, an account for non-disyllabic Sino-Korean compounds follows straightforwardly.

Authors: Marcel den Dikken and Ji Young Shim (CUNY Graduate Center) 
Keywords: feature inheritance, EPP, lexical/functional distinction 
Title: Feature Inheritance and the EPP

Assuming that T is not itself a phase head, and that only phase heads can be specified for Agree features and EPP, Chomsky (2008) proposes that T inherits its Agree features and EPP specification from C, and suggests that such features inheritance is in effect in the vP phase as well, with v activating V for the EPP. This paper endorses feature inheritance and its application to the vP phase but, taking seriously that lexical roots are not probes, rejects the idea that V can be the beneficiary of feature inheritance, instead identifying the head of the non-phase complement of v as Asp. Just as T can inherit Agree features and EPP from C, so can Asp inherit Agree features and EPP from v. But throughout, feature inheritance fails whenever a lexical root occupies the head of the complement of the phase head. It will be shown that the conjunction of feature inheritance and lexical heads’ inability to serve as probes yields the right results for both subject and object placement.

Authors: Kadir Gokgoz (Purdue University), Engin Arik (Okan University)
 Keywords: TİD, nonmanual markers, polar questions, borrowing, prosody, grammaticalization 
Title: Distributional and Syntactic Characteristics of Nonmanual Markers in Turkish Sign Language (TİD)

In this paper we report our findings about the distribution of certain NMMs in Turkish Sign Language (TİD) and show their relevance for the prosody and syntax of positive declarative sentences and polar questions in this language. We argue that certain NMMs are integral components of TİD grammar and in this sense they are robust in their behavior to mark syntactic relations. They are so robust that even with a case of borrowing from the surrounding majority language, i.e. Turkish, they help the grammar integrate the borrowed item into the computational system. In this regard, we will illustrate a case of interaction between these two languages and offer an analysis of grammaticalization where the behavior of NMMs is the key factor to understand the exact nature of the syntactic process.

Authors: Pavel Grashchenkov (Institute of Oriental Studies, Moscow), Vita Markman (Disney Internet Media Group, Los Angeles) 
Keywords: Turkic, ditransitives, VP decomposition, argument structure, serial verbs, differential object marking 
Title: Serial Verb Ditransitives in Turkic

Turkic complex verbal items like ‘return something’ which are created as ‘arrive’ + ‘take’ / ‘give’ are at issue. The most crucial problem for such ditransitive serial verbs is that they allow unmarked direct objects not adjacent to the transitive verbs that is out for simple transitives. We propos a structure for Turkic SVCs that follows a decompositional approach to the vP in terms of Ramchand (2008). Under this approach ‘arrive’ and ‘take’ / ‘give’ occupy different positions, but are all parts of a single V. Being in the res / proc positions, ‘arrive’ assigns theta-roles to direct objects, introducing the process and the result state of the change of possession process. ‘Take’ / ‘give’ in the init head is responsible for the accusative marking if a direct object is a DP and needs case. The analysis explains the facts and offers further empirical support for a decompositional system.

Authors: Sanae Tamura (Kyoto University), Yurie Hara (City University of Hong Kong), Youngju Kim (Hiroshima University), Hiromu Sakai (Hiroshima University) 
Keywords: causation, volitionality, nominalization, Event Semantics, abstract objects 
Title: Japanese Sentential Nominalization and Different Kinds of Causation

Natural languages have various ways to express causal relations. In this paper, we focus on the classification of causal relation in terms of volitionality (Sæbø, 1991, Degand, 2000, Sanders and Sweetser, 2009). We show that the volitionality of causal statement can be manifested in natural languages and correlates with the syntactic structures and semantic types of causal relata. In particular, we argue for a constraint on non-volitional causation which prevents a non-volitional causation from relating abstract objects. From a semantic point of view, in causal statements such as ‘A-ga B-o hikiokosu’ (A causes B), if the causal relation is non-volitional, A refers to an event. On the other hand, if the relation is volitional, A can refer to a fact. From a syntactic view point, the nominalized clause has to be an AspP in the case of non-volitional causation, whereas it can be a TP in the case of volitional causation. These syntactic and semantic properties of the two types of causal statements are derived from the analysis of clauses nominalized by proposed in Hara et al. (in press).

Author: Tomoko Ishizuka (Tama University)
 Keywords: passive voice, possessor, possessive passive, A-movement, Japanese 
Title: The Genitive Passive in Japanese - What does a modular approach tell us?

This paper analyzes genitive passives (a.k.a. possessive passives) in Japanese, adapting Chris Collins' (2005) smuggling analysis of English passives. The approach taken here is a modular one, which assumes that the phenomenon of passivization is brought about by the lexical properties of the passive morpheme -rare (i.e. smuggling the internal argument past the external argument by attracting the VP shell to its specifier) interacting with the independently motivated properties of the language (i.e. EPP of T, availability of possessor raising). The analysis predicts that properties of the genitive no-phrase should carry over to the genitive passive wherein the derived subject was originally merged as a genitive no-phrase. This modular approach leads to a new way to look at genitive passives, which allows us to subsume a subset of gapless passives (a.k.a. indirect or adversative passives)—passives seemingly lacking an active source—under genitive passives. Consequently, it significantly reduces the number of Japanese gapless passives lacking an active source, making inroads into a true unification of Japanese passives as a raising phenomenon.

Author: Koji Kawahara (Fuji Women’s University)
 Keywords: gradable adjectives, comparatives, Japanese, English 
Title: The Semantics of Japanese Gradable Adjectives

Comparing objects according to some standard is one of the fundamental aspects of human cognitive ability. Natural language typically reflects this ability by building comparative constructions. The goal of this paper is to explore the cross-linguistic difference between English and Japanese with respect to semantic behavior in comparisons. I investigate whether two types of comparatives, which must be differentiated from morphological and semantic perspectives in English, need to be assumed in Japanese or not. I will show that based on the semantic diagnostics for telling one type from the other the two different types of comparatives need not be hypothesized in Japanese. I will also point out that Japanese comparatives are based on degree comparisons. Finally, I address implicational differences that are caused by several classes of standard markers, showing that the implicational differences must be analyzed in terms of a pragmatic perspective.

Authors: Shin-Sook Kim and Peter Sells (SOAS, University of London)
 Keywords: Negation, scope, NPIs, contrastive phrases, Immediate Scope Constraint, reconstruction, Korean, Japanese 
Title: Reconstruction and Scope of Negation in Korean (and Japanese)

In this paper we discuss examples from Korean in which there are two different kinds of phrases – negative polarity items (NPIs) and contrastive phrases – which have to enter into a licensing relationship with negation at the same time. We show that the surface order of phrases corresponds quite closely to the scope interpretation, due to the availability of scrambling within the language. In order to enter into licensing relationships with the relevant phrases, the scope of negation is quite variable. We briefly show that Japanese has the same properties, and it appears that Turkish does too. Although the surface order of phrases largely corresponds to their scope relations, some scrambled contrastive phrases may reconstruct to their trace position, in order to be licensed by negation. In fact, this “reconstruction” is possible for all non-subjects (both arguments and adjuncts), but is never possible for subjects. In the last section of the paper we briefly consider the implications of this subject/non-subject asymmetry for the mapping between syntax and semantics (LF).

Author: Maki Kishida, University of Maryland
 Keywords: reflexive anaphors, parametric variation, Japanese
 Title: Parametric Variation in Reflexive Classification: Evidence from Japanese

This paper discusses the classification of reflexive anaphors in languages. Many languages have more than one form of anaphors. We propose that types of reflexive anaphors in such languages are classified based on their reflexivity (semantic) differences as ‘Pure reflexive anaphors’ and ‘Near reflexive anaphors,’ as proposed in Lidz (1996, 2001a,b). Further, we claim that there is a parametric variation in classification of reflexives into the two types among languages. Our proposal is that the parametric variation depends on the ways of the two types of reflexivity marking on predicates. In languages in which semantic reflexivity marking and lexical reflexivity marking occur independently (e.g. Dutch, Kannada, Malayalam, Norwegian etc.), morphologically simplex anaphors are Pure reflexive anaphors and morphologically complex anaphors are Near reflexive anaphors. On the other hand, in languages in which the two types of reflexivity marking occur simultaneously (Japanese, Russian, Turkish, Spanish etc.), affixal reflexives are Pure reflexive anaphors and non-affixal reflexives are Near reflexive anaphors. This parametric variation analysis of reflexive classification sheds a new light on the typological research of reflexivity and coreference in generative grammar.

Author: Hideki Kishimoto, Kobe University 
Keywords: coordinate structure, mo, topic, modal projection, ATB movement
 Title: Topicalization and Coordination in Japanese

In Japanese, the particle mo can be used to construct coordinate structures. It is shown in this paper that in the type of coordinate construction where the particle mo is added to the verbs of the conjoined clauses, TPs are coordinated when subjects are included in the conjuncts. We argue that an island effect observed in when a topic appears inside the TP-conjoined structure accrues from the topic’s undergoing LF movement into ModP projected above TP. The data regarding mo-coordination illustrate that wa-marked topics must be placed in ModP for their formal licensing at LF, even if they occur in clause-medial position superficially.

Author: Seongyeon Ko, Cornell University
 Keywords: Mongolic languages, vowel harmony, vowel shift, contrastive hierarchy
 Title: Vowel contrast and vowel harmony shift in the Mongolic languages

It had long been assumed that the Mongolic languages including Proto-Mongolic have a palatal harmony system, until Svantesson (1985) presented acoustic evidence that Khalkha and other Mongolian dialects have a ‘pharyngeal’, not a palatal, harmony system. Faced with this discrepancy between the modern RTR(=[Retracted Tongue Root]) systems and the assumed-to-be pre-modern palatal systems, Svantesson (1985) proposes a vowel shift hypothesis which holds that the basis of vowel harmony has shifted from palatal to RTR contrast (except for Kalmyk/Oirat which retains the old palatal contrast). In this paper, I challenge this idea by a careful examination of the synchrony and diachrony of the Mongolic vowel systems within the framework of the contrastive hierarchy theory (Dresher 2009). As a result of this line of research, I propose a reverse shift from RTR to palatal harmony. This reverse shift meets the various criteria in historical linguistics such as directionality, ‘majority wins’, and economy (Campbell 2004). It is consistent with typological expectations as well, taking other Altaic languages such as Tungusic and Korean into consideration. This supports the idea that RTR was the original harmonic contrast in Altaic (Vaux 2009).

Author: Jaklin Kornfilt (Syracuse University and University of Stuttgart)
 Keywords: Sentential Subject Constraint (SSC), CED, Turkish, freezing effects, topicalization, Left Dislocation, extraction, sub-extraction 
Title: The Sentential Subject Constraint /CED as a Left-Dislocation Constraint in Turkish

Turkish seems to freely allow SSC/CED-violations in the formation of relative clauses, although it otherwise obeys certain island constraints. However, in the standard variety of Turkish, only one of two potentially possible RC-constructions targeting a constituent within a sentential subject is well-formed: the agreement-less, special “subject participle“ must be used, while the regular indicative nominalization form followed by subject agreement can’t. I propose that this is due to freezing the left-dislocated sentential subject, an adjunct of the main clause, in construction with the pro-subject in that main clause: The Agr-morphology of the indicative nominalization licenses and enforces a pro-subject, and therefore a LD-construction. In the well-formed, agreement-less construction, the sentential subject is topicalized, but not dislocated. No freezing effects are encountered on the topicalized sentential subject, despite its CED-nature and despite its movement to Spec,TopP, due to a parameter that I propose, whose marked positive value is set by a language’s being topic-oriented, and thus offering movement from lower CED-positions to topic positions as a strategy of last resort, thus treating information-structural specifier positions as though they were theta-positions. Evidence is offered for the relatively high position of the sentential subjects involved, and against subextractions proceeding from a low position.

Author: Emi Mukai, University of Southern California 
Keywords: Japanese, floating numerals, bound variable construal
 Title: Bound Variable Construal with a 'Discontinuous' Binder

A combination of a floating numeral and its host NP can be taken as a binder. We call such a binder a 'discontinuous' binder. The main issue of this paper is how the Bound Variable Construal with a 'discontinuous' binder should be constrained. I claim, contrary to Hoji & Ishii 2004, that it is the positions of the host NP and the bindee (i.e., arguments) at LF, not that of floating numerals, that are crucial for the availability of the construal in question. I present evidence in support of the claim by investigating what I call the 'non-local' cases, where the subject precedes the object, and the object precedes the floating numeral whose host NP is the subject as in NPSubj (=host NP) NPObj floating-numeral V-T. I also discuss Miyagawa & Arikawa's 2007 account of the (un)acceptable status of the 'non-local' cases. I show that their proposal that the surface Subject-Object-Verb (SOV) order can correspond to two distinct types of LF representations makes wrong predictions and conclude that the SOV order necessarily corresponds to the LF representation in which S asymmetrically c-commands O, as has been widely accepted in the field.

Author: Junya Nomura, MIT 
Keyword: adpostions, comparatives, Japanese, locatives 
Title: Adpositional Comparatives in Japanese

Japanese locative constructions can classify into two types according to what kind of argument the adpostion takes. The first type, which I call the no-locative, takes the genitives marked DP, while the other one (the yori-locative) takes the DP followed by yori. Little attention has been paid to the differences between them. I observe that there are at least four differences between them. First, the truth condition of the no-locative is more restricted than that of the yori-locative. Second, when the quantifier phrases are an argument of the adpositions, the no-locatives become ambiguous, while the yori-locatives remain unambiguous. Third, some adpostions can be used in the no-locatives, but cannot be used in the yori-locatives. Lastly, the modifiers tyoudo “just” and sugu “right” are compatible with the no-locatives, but they are incompatible with the yori-locatives. I analyze that the no-locatives are “ordinary” locatives like English in front of and under, while the yori-locatives involve “comparatives.” It is shown that this analysis explains the four differences straightforwardly.

Author: Umut Özge, Middle East Technical University
 Keywords: indefinite noun phrases, discourse linking, specificity, Turkish, accusative case 
Title: Turkish Indefinites and Accusative Marking

This paper addresses the issue of the effects of overt accusative (Acc) vs. zero () marking on the interpretation of indefinite direct objects in Turkish. Previously, the overt accusative case has been associated with certain semantic and pragmatic categories such as discourse-linking (Nilsson 1985; Enç 1991; Zidani-Eroğlu 1997), “specificity” (von Heusinger 2002; von Heusinger and Kornfilt 2005), presuppositionality (Kennelly 1997; Kelepir 2001). Challenging these interpretations on the basis of data from intensional constructions, we show that Acc-marked indefinites can freely interact with intensional operators and the notion of D-linking does not uniformly apply to Acc-marked indefinites, but is highly contingent on contextual factors. We argue that interpreting Acc-indefinites as generalized Skolem terms (Farkas 1997; Steedman 2010) that are susceptible to implicit contextual domain restriction (Bende-Farkas and Kamp 2001;Schwarzschild 2002) accounts for the data presented in the paper.

Authors: Nilüfer Gültekin Şener and Serkan Şener, University of Connecticut
 Keywords: Null subject, indexical, de se pronoun, Turkish, Uyghur 
Title: Null Subjects and Indexicality in Turkish and Uyghur

Contra Kaplan’s (1977) conjecture, indexical shift is a well-noted phenomenon. In this paper, we focus on the interpretation of the null 1st person pronominal subject in Turkish, and show that it differs from the overtly realized pronoun ben/I with respect to its semantic value. Specifically, null 1st person pronoun has a shifted reading as opposed its overt counterpart, which cannot shift. We account for this observation by proposing that Turkish has a de se pronoun, which is always null. We further compare Turkish with the typologically related language Uyghur, and tentatively suggest that Uyghur also has a de se pronoun. The de se pronoun in Uyghur differs from its Turkish counterpart in two respects; it can be overt, and it must be specified for a specific case marker, namely Nominative.

Authors: Ryosuke Shibagaki (SOAS) and Dolgor Guntsetseg (University of Stuttgart)
 Keywords: Mongolian, Resultatives, clausal structure, complement/adjunct, -tAl converb
 Title: Resultatives in Mongolian: between Japanese and Korean?

We offer an account of the syntactic and semantic properties of Mongolian resultatives. Mongolian uses the -tAl converb to express the resultative interpretation. Although Mongolian “resultatives” have the true resultative interpretation and aspectual property, they seem to take an adjunct structure rather than a complement one. We will show the evidence that the Mongolian resultative phrases form an embedded TP clause, where the object-oriented resultative phrases adjoin to V’ and the subject-oriented resultative phrases to vP. Washio (2002) noted that the Mongolian -tAl converb construction looks very much on the surface like the Korean -key resultative, but not the Japanese complement type resultative. Our investigation of Mongolian indeed lines up in important ways with the study of Korean by Sells (1998) and Shim & den Dikken (2007), in that Korean resultatives are adjuncts rather than complements.

Author: Deniz Tat (University of Arizona)
 Keywords: relative clauses, adjectives, Turkic, bir, LCA
 Title: APs as reduced relatives: The case of bir in (some) varieties of Turkic

In this paper, I show that, except for a small set of non-predicative adjectives, Turkish APs are always formed by indirect adjectival modification that has an underlying clausal structure. Kayne's (1994) LCA can account for this derivation as long as we assume that the subject of the underlying clause is always a fully specified one. APs derived in this manner lead to multiple determiners. It seems that languages with indirect modification have different ways of benefiting from the multiple copies of definites or indefinites by phonologically realizing them only if that augments meaning.

Author: Reiko Vermeulen (Ghent University)
 Keywords: contrastive topic, contrastive focus, Japanese, Korean, interface, syntax
 Title: Contrastive Focus and Contrastive Topic at the Syntax-Information Structure Interface

This paper provides a uniform account of the syntactic distribution of contrastive focus and contrastive topic in Japanese and Korean. Systematic comparison reveals that contrastive foci show the same distribution in the two languages, but contrastive topics do not. The account is based on two core ideas: (i) the notions [topic], [contrast] and [focus] can be targeted by mapping rules at the interface between syntax and information structure; and (ii) contrastive topic and contrastive focus are composites: the former comprises of [contrast] and [topic], and the latter, [contrast] and [focus]. The two languages have the same mapping rule for [contrast] and the same rule for [topic], but no relevant rule for [focus]. Contrastive foci are therefore subject to the rule for [contrast], and hence behave alike. On the other hand, Korean contrastive topics are subject to the rule for [contrast], showing a distribution like contrastive foci, while those in Japanese are subject to the rule for [topic], displaying properties in common with non-contrastive topics. The proposal makes several correct predictions regarding the word order between contrastive focus and contrastive topic, highlighting the advantages of the current proposal over the standard accounts of contrastive topics for both languages.

Author: Noriko Yoshimura (University of Shizuoka) 
Keywords: double o construction, accusative Case marker, major object, multiple Spec
 Title: New Perspectives on Double O Constructions in Japanese

This paper presents new perspectives on the nature of double o constructions (DOC) in Modern Japanese by considering both diachronic and synchronic data pertinent to the issue, and in particular, we argue that the DOC involves a structure of [VP [SpecDP-o] [V’DP-o V], based on Miyagawa’s (2010) major subject hypothesis together with Hoji’s (2005) and Takubo’s (2007) major object analysis for Japanese. In so doing, we first show that the DOC existed and functioned as an adjunct in Old Japanese. We then demonstrate that the DOC is permitted in yielding the topic/focus reading in Kumamoto Japanese because it uses ba as the accusative Case particle, an allophone of the topic marker ha. A detailed discussion of these diachronic and synchronic facts from MJ, OJ, and KJ leads to the adoption of the major object hypothesis for the o-marked adverbial DP in the DOC construction. By assuming that the o assigned to the major object is a morphological case, it is concluded that the DOC construction is possible in Japanese because being a non-forced agreement language, Japanese has a morphological case marking system as well as structural Case assignment (Kuroda 1978), permitting multiple occurrences of Spec (Kuroda 1988).