MIT Working Papers in Linguistics #66

Proceedings of FAJL 6: Formal Approaches to Japanese Linguistics

Kazuko Yatsushiro and Uli Sauerland, 2013

for $29.95 x

Save $18.95 (31%) if bought together with Proceedings of FAJL 3: Formal Approaches to Japanese Linguistics

for $40.95 x

Save $20.95 (31%) if bought together with Proceedings of FAJL 7: Formal Approaches to Japanese Linguistics

for $45.95 x

This volume contains the proceedings of FAJL 6. It consists of the following papers:

Backward Anaphora in Japanese: Cases of Condition C Violations - Jun Abe 5

It has been well known since Polinsky and Potsdam (2002) that universally, control relations allow not only forward control, the standard one in which the controller is pronounced, but also backward control, in which the controllee is pronounced. The existence of the latter relation in some languages is taken as strong evidence for such a movement theory of OC as proposed by Hornstein (1999, 2001), since the PRO-based theory will not be able to provide any obvious means to capture backward control. This paper aims to demonstrate that such a case of backward control is attested in Japanese, though the relevant anaphoric relation does not really involve control. It thereby gives support to an A-movement approach a la Hornstein to anaphoric relations.

Judge-Dependence in Degree Constructions: Evidence from Japanese Evidentiality - Lisa Bylinina 17

The focus of this paper is the representation of subjectivity, or judge-dependence, in Japanese and in natural language in general. The empirical domain I limit myself to is degree constructions, such as positive, comparative, and 'exceed'/'too' construction. The particular question I am trying to answer is whether there is evidence for a 'judge' argument for the cases of subjectivity we come across, or maybe only for some of the cases, or maybe for none. I observe that different subjective expressions show different behaviour and at first glance seem to suggest different analyses ('two kinds of subjectivity' analysis), but careful investigation of the data leads me to a conclusion that a deciding factor for apparent argument-like 'judge' is experiential semantics of the predicate. I suggest to take this fact seriously and propose that in cases where we seem to see a 'judge' argument, we are really dealing with an Experiencer. The judge-dependence per se can be captured in a different way, uniformly for all kinds of subjective expressions. Japanese evidentials and person restrictions on judges will play a crucial role in this reasoning.

Japanese Rokuna: A lexical Alternative to an NPI Analysis - Isaac Gould 29

This paper takes a close look at the semantics of the Japanese word rokuna, which is more complex than has been previously recognized. To my knowledge, there is no detailed study on rokuna; existing mentions in the literature claim that it is an adjective meaning 'good' and is an NPI (Makino and Tsutsui 1995 [2006]; Vasishth 2001; Kataoka 2006; Miyagawa, et al. 2012). Closer examination of the facts, though, leads us to doubt both these claims. As will be discussed in Section 2, to capture the inferences associated with rokuna, it is necessary to treat rokuna as an expression that has multiple dimensions of meaning: rokuna has an ordinary semantic value of 'bad,' and it also has an alternative semantic value (Rooth 1985) of lexically specified alternatives. Rather than being an NPI, then, rokuna is exhaustified by an operator with the force of 'only'. Having proposed that rokuna co-occurs with an only-operator, in Section 3 I show that an advantage of this account is that it provides us with a way of using the semantics of the only-operator to analyze a previously unnoticed and puzzling feature of the syntactic distribution of rokuna in clefts, relative clauses, and topicalization structures.

Darou as a Deictic Context Shifter - Yurie Hara and Christopher Davis 41

This paper investigates the distribution of the Japanese sentence-final particle darou. We proceed by examining the grammaticality and interpretation of darou-sentences while varying parameters such as clause type, boundary tone, and pragmatic context, and propose that darou is both a deictic expression pointing to the speaker's beliefs as well as a context-shifter that manipulates the context in order to circumvent a possible violation of Gricean Quality.

Implicit Prosody and Acceptability: The Case of Long-Distance Scrambling in Japanese - Shinichiro Ishihara 57

This paper discusses various (apparently) syntactic phenomena observed in the so-called long-distance scrambling, and provides a unified prosodic account, based on the analysis of the clause-mate condition for multiple clefts that I have proposed (Ishihara, 2012).

The general concern of this study is how syntactic acceptability judgments are affected by the prosody in silent reading. According to the Implicit Prosody Hypothesis (Fodor, 1998, 2002), even in silent reading an abstract prosodic structure of the sentence is projected in the grammatical representation, and this "implicit prosody" may influence syntactic ambiguity resolution. If this hypothesis is on the right track, it would mean i) that whenever a native speaker makes a syntactic acceptability judgement of a sentence, this process always involves projection of an abstract prosodic structure of the sentence, and ii) that its influence on acceptability is expected even without an actual phonetic output of the sentence.

Complementizer Stacking, Dual Selections, and Relabeling - Toru Ishii 73

Under the symmetric Merge together with the labeling algorithms approach, Merge and labeling are independent operations. We should therefore expect that labeling can take place without Merge. This paper argues that there are cases where labeling applies independently of Merge. I will investigate complementizer stacking in Japanese, and propose that when a labeling conflict arises, relabeling, i.e. labeling without Merge, may apply as part of LF-Transfer, which accounts for prima facie puzzling syntactic and semantic selections in complementizer stacking. If the analysis to be proposed is on the right track, it presents further evidence for the symmetric Merge together with labeling algorithms approach.

Dative Subject Constructions in Child Japanese - Miwa Isobe and Reiko Okabe 85

Japanese belongs to the group of languages which allow the so-called Dative Subject Construction (DSC), and many attempts have been made to date to determine what syntactic mechanism allows this construction (e.g. Shibatani 1978, Ura 1999, Koizumi 2009, among many others). One of the characteristics of DSC is that it has a stative predicate such as a potential predicate which has no ability to assign accusative to the direct object. Moreover, the particle -ni, which marks the subject in Japanese DSC, is assumed to have several different types with various categories, including the case marker and the postposition (Sadakane and Koizumi 1995). Therefore, Japanese-speaking children have to know these properties in order to understand and produce DSC correctly. There are several acquisition studies which have reported that children have difficulty with the particle -ni (e.g. Matsuoka 1998, Murasugi and Machida 1998). But it is not yet clear when and how they acquire knowledge of DSC.

In light of the above, this study experimentally investigates whether Japanese children correctly comprehend DSC. Our experiment found that participants aged 3 to 6 were able to correctly understand -ni used in DSC and distinguish it from the benefactive use of -ni. We also conducted a transcript analysis to find out how frequently DSC is produced by adults and revealed that this construction is extremely infrequent in child-directed speech. These findings suggest that young Japanese-speaking children have adult-like knowledge of DSC, despite the fact that they are not given sufficient exposure to this construction.

Measurement by Difference Functions - Koji Kawahara 97

Genitive Object in Kansai Japanese and Slavic Languages - Kiyomi Kusumoto 109

Object NPs may appear with genitive Case in many Slavic languages. One licensing environment is under intensional verbs. Recently, Asano and Ura (2010; A&U henceforth) reported a similar phenomenon in Kansai Japanese, a dialect spoken in the western part of Japan. In addition to the well-known Case conversion called Nominative-Genitive conversion (NGC) observed in many dialects of Japanese, this dialect allows Accusative-Genitive conversion (AGC).

The goal of this paper is two-fold; (i) to give a new analysis of AGC based on a new observation regarding the so-called non-past -ta in Japanese and (ii) to explain cross-linguistic similarities and differences between Case conversion in Kansai Japanese and Slavic languages.

What Makes Root Phenomena Special? - Norio Nasu 121

Particle stranding in answers exhibits a very limited distribution. Existing analyses (Hayashi 2001, Yoshida 2004, Arita 2005, Sato and Ginsburg 2007, Sato 2012, among others) commonly argue that it occurs in sentence-initial position. This paper, on the other hand, argues against this observation on the basis of cases where a stranded particle appears in a relatively inner position. Still, this does not mean that stranding takes place unconditionally. Stranding tends to occur in a domain that is close to sentence-initial position. A question that arises is where it can occur and where it cannot. This paper attempts to identify the principles that govern the distribution of particle stranding.

Scrambling and Dependent Interpretation in Multiple Wh-Questions in Japanese - Atsushi Oho 133

This paper investigates how dependency is implemented in human language by examining properties of wh-phrases in multiple wh-questions. In the functional wh approach, there are two different types of roles in multiple wh-questions and each wh-phrase is assigned a particular role (Comorovski 1996, Dayal 1996, Hornstein 1995). This role assignment is crosslinguistically inflexible under a normal circumstance. In English, for example, a subject term must be a domain-setting wh-phrase and the other is a functionally-interpreted wh-phrase. In Japanese, though the basic role assignment is the same as in English, a fronted wh-phrase becomes a domain-setting wh-phrase. The generalization about this role assignment is given as follows:

(1) The role assignment generalization
i. The surface leftmost wh-phrase sets the domain of the function;
and ii. The other wh-phrase is functionally interpreted.

In this paper, I will propose an interpretive dependency condition for multiple wh-questions as in (2) and argue that the role assignment generalization is a natural consequence of this required dependency condition.

(2) Multiple Wh-Question Dependency Condition (MWDC)
A scope of a wh-phrase x can be interpreted as parasitic on a scope of the other wh-phrase y only if x's deleted copy of operator is dependent on y's deleted copy of operator.

A crucial claim is that it is an interpretive restriction induced by the syntax of dependency that results in the inflexible role assignment. This claim is different from what has been said in the literature in that each role is not an inherent property of wh-phrase, but it is a byproduct of the required dependency condition.

Interference and Subcategorization Information: A case of Pre-Verbal NPs in Japanese - Hajime Ono, Miki Obata, Noriaki Yusa 145

Study of cue-based memory retrieval is getting more attention these days because it enables us to make clear the connection between the model of syntactic structure building and the domain general working memory system. It should also be clear that investigating the exact mechanisms of encoding and retrieval constitutes a very important empirical problem in the field. Below, we will briefly review some empirical findings regarding "similarity-based interference effects." We will review one study about semantic interference, and another about syntactic interference (Van Dyke, 2007, Van Dyke & McElree, 2006, Van Dyke & Lewis, 2003). Then we will proceed to discuss a phenomenon called pre- verbal attachment in Japanese (Kamide & Mitchell, 1999, Miyamoto, 2002). Incremental processing based on the sequence of NPs is one of the key phenomena in Japanese sentence processing, but we would like to further elaborate the implications of the incremental structural commitment from the viewpoint of cue-based memory retrieval. Specifically, we will explore a case of clause boundary insertion triggered by the sequence of a topic-marked NP and a nominative NP. We argue that this sequence of NPs leads the parser to encode the topic-marked NP as the subject of the verb that takes a CP as a complement. We will examine this claim in a self-paced reading experiment using the similarity-based interference effect as an index. We will show that an interference effect is observed when the parser reads the verb that takes a CP as a complement such as yooboo-suru 'demand' as the embedded verb. We will also discuss some further implications of the results obtained from the experiment.

Violable and Inviolable OCP Effects on Linguistic Changes: Evidence from Verbal Inflections - Shin-Ichiro Sano 157

This paper presents an empirical examination of the effects of Obligatory Contour Principle (OCP, Leben 1973; Goldsmith 1976; McCarthy 1986) in diachronic contexts, using a large-scale spontaneous speech corpus of Japanese. Specifically, I focus on two kinds of ongoing morphophonological changes in Japanese verbal inflectional paradigm, sa-Insertion and ra-Deletion, and show how the progress of these changes is affected by the OCP.

Precision and Manners of Measurements: The Case of Japanese Minimizers - Osamu Sawada 169

The Subject-Object Asymmetry Regarding Japanese Disjunctive ka in Child Japanese - Hiroyuki Shimada 181

It has been observed that interpretations of disjunction in downward entailment contexts (e.g., in negative sentences) show cross-linguistic variations (Szabolcsi 2002, Notley et al. 2011). For example, in English, the disjunction operator "or" receives a conjunctive interpretation in a negative sentence when it is c-commanded by negation. In contrast, in Japanese, the disjunction operator "ka" does not receive a conjunctive interpretation but a disjunctive interpretation in a simple negative sentence. In child languages, however, Goro and Akiba (2004) / Goro (2007) and Komine (2012) report that Japanese children, unlike adults, initially assign a conjunctive interpretation to the disjunction operator "ka" in a negative sentence and in a comparative sentence.

In this study, I experimentally show that Japanese-speaking children assign a disjunctive interpretation to the subject with the disjunction operator "ka" in a negative sentence, while they assign a conjunctive interpretation to the scrambled object with "ka" in a negative sentence. This observation indicates that children's interpretation of "ka" in a negative sentence is determined by abstract hierarchical structure (i.e., c-command relation), and the scrambled object undergoes "reconstruction," an abstract operation, even in a child language (Murasugi and Kawamura 2005, Goro 2007, Sano 2007; Leddon and Lidz 2006).

On the Long-Distance Movement of Subject in Japanese - Koji Shimamura 191

In this paper, we will investigate the properties of Long-Distance (LD) movement of subjects (SUBJ) in Japanese. LD movement of SUBJ can be bifurcated into scrambling (an instance of A'-movement) and hyper-raising (an instance of A-movement). A typical instantiation of the latter is the Raising-to-Object (RtO). As for scrambling, Saito (1985) argues that this operation cannot apply to SUBJ, either clause-internal or long-distance.1 Regarding the RtO, a lot of works have been dedicated to this construction (Kuno, 1976, Ura, 1994, Sakai, 1998, Hiraiwa, 2001, Tanaka, 2002, Ogawa, 2007, Obata, 2010, Takeuchi, 2010 inter alia). The consensus of those works is that the RtO as hyper-raising is movement out of CP. Then, if this state of affairs is on the right track, something like (1) can be stated.

(1) LD movement of SUBJ in Japanese must be hyper-raising but not scrambling.

Since (1) is just a mere observation, we must reduce (1) to some independent mechanism of the grammar, which is the main concern of this paper. However, we have to pin (1) down in that LD scrambling of SUBJ is really unavailable. Otherwise, our attempt would end up aiming at nothing. Oku (1998), for examples, claims that the prima facie unavailability of it is ascribed to some sort of processing problem. Yamashita (2012) seems (at least to me) to follow the same intuition as Oku has, but he formalizes it in syntactic terms like features. Crucial here is that they do not exclude LD scrambling of SUBJ from the grammar; if certain factors are met, such movement is in principle implementable. I nevertheless show in section 2 that it is impossible. My argument is based on one famous diagnostic for LD scrambling in Japanese, termed radical reconstruction effect (Saito, 1989). There, we observe that SUBJ, albeit apparently scrambled, never reconstructs. This is striking if there is any movement involved in scrambling SUBJ.

A Note on Parallelism for Elliptic Arguments - Daiko Takahashi 203

The purpose of this article is to provide a new piece of evidence for the hypothesis that null arguments in languages like Japanese can arise through ellipsis. The evidence comes from the observation that they are subject to the same parallelism constraint as other ellipsis processes such VP-ellipsis and NP-ellipsis.

Some Asymmetries in Japanese N'-Deletion and their Theoretical Implications - Kensuke Takita and Nobu Goto 215

The main empirical focus of this paper is the case of N'-deletion taking place within PPs. As far as we can tell, not many cases have been examined in the literature, but their grammaticality indicate that N'-deletion seems to be possible within PPs. Providing more controlled examples, however, we show that there are certain environments where PP-internal N'-deletion is blocked. Then, we propose an analysis that can accommodate this novel observation, discussing some theoretical implications.

Raising out of V+tate Phrases - Hidekazu Tanaka, Peter Sells, and Mika Kizu 227

Toward a Better Understanding of Japanese Scramblings: On the Absence, Emergence, and Disappearance of Minimality-like Crossing-over constraint effects - Hideaki Yamashita 239

The main aim of this paper is to discuss the implications of the patterns of crossing-over constraints effects which shows minimality-like effects, which Japanese scrambling in general exhibits. By demonstrating the absence, emergence, and disappearance of minimality-like crossing-over constraints effects are found not only with long-distance scrambling of objects but also of subjects, I conclude that scrambling of subjects is in principle possible, contra Saito's (1985) classic and widespread claim that (Nominative) subjects cannot undergo (long-distance) scrambling in Japanese.

Structure of &P and Contextual Allomorphy - Yusuke Yoda 251

There are three widely discussed aspects of the TE-form in Japanese. The first is semantic. Namely, when two or more sentences are connected with te, the semantic interpretation between the connected sentences varies, just as in the English participle constructions. Another aspect is on the morphology of TE-form which has been investigated for a long time from both theoretical and descriptive perspectives. The last aspect is syntactic one, that is, the structural relation between the non-terminated te-clause(s) and the terminal clause.

The aim of this paper is to give an adequate explanation of the second and third aspects, drawing from evidence from the syntax-morphology interface within the framework of Distributed Morphology (Halle and Marantz (1993), Embick and Noyer (2006), inter alia). The goal of this paper is to illustrate the following; (i) both te and to are coordinators in Japanese and they are an instance of allomorphy conditioned by the category left-adjacent to them at PF; (ii) Despite the fact that Japanese is regarded as an instance of absolute head-final languages, the structure of Coordination in Japanese is head-initial structure.