This paper presents an analysis of what is commonly known as the “appositive genitive” in Classical Manchu. Unlike Japanese, Manchu allows the first noun phrase to appear in the definite, genitive form, as in men-i (we-GEN, “our”) hehe niyalma (woman) when expressing “we women.” As such, it is argued that the first noun phrase in such examples is the genitive subject of a subordinate clause. By analyzing the usage of adnominal forms of the copula in the corpus, I show that 1) contrary to conventional accounts, the adnominal form of the copula can be (or rather is usually likely to be) zero (i.e., phonetically empty). Furthermore, I also point out that 2) a genitive subject can be the subject of this zero copula in the same way as the subject of any other verb. On the basis of 1) and 2), we can consider the Manchu “appositive genitive” structure [NP_1i -i NP_2i] (NP: noun phrase, -i: genitive case marker, i: co-referential) to be a copula clause meaning ‘NP₁ is NP₂’ rather than a noun phrase meaning ‘NP₂ belonging to NP₁’.

The Khitan language is a dead language attested in inscriptions written in the Khitan Small Script from the 11th to the 12th centuries and is genetically related to the Mongolic languages. This paper examines the orthographic distinction of vowel length in the Khitan Small Script. Through comparison of cognates, interpretation of distribution characteristics in the texts and analysis of suffixal allomorphs and spelling alternations, this paper argues that: (1) vowels in open syllable graphemes like Vs and CVs are long (V = vowel, C = consonant including semivowel); (2) vowels in closed syllable graphemes like VCs are basically short but a series of graphemes represent a long ē plus a consonant. From a comparative linguistic viewpoint, the following two points are further argued for: (3) some long vowels in Khitan correspond to the secondary long vowels in Modern Mongolian (which resulted from vowel contraction due to the loss of intervocalic consonants); (4) other long vowels may correspond to the primary long vowels which once existed in Proto-Mongolic and Proto-Turkic.

In this paper, we discuss the Hateruma Ryukyuan accentual system. Although many previous works have claimed that Hateruma has a two-pattern accent system, some recent studies have proposed that it has a three- or four-pattern accent system.

In the first part of this paper, we address synchronic issues and argue that Hateruma has a three-pattern accent system on the basis of an analysis of F0 tracks. We also show that adjectives and compound words can have two accentual units in a single word. In the latter part of the paper, we discuss diachronic issues. Based on the so-called “classified vocabulary” (keiretsubetsu-goi) and the word-initial sound, we found that one can reliably predict a word’s accent pattern. This is true of not only Hateruma Ryukyuan but also Shiraho Ryukyuan. By comparing these two dialects, we conclude that Hateruma has retained its accent system for more than 250 years.

This paper presents a description and analysis of differential object marking (DOM) observed in Ayacucho Quechua nominalized subordinate clauses formed with -sqa and -na. This Quechua dialect allows for two patterns of object marking within nominalized subordinate clauses: -ta and -ø. Based on the results of interview surveys, this paper argues that marking with -ta is motivated by two concepts of information structure: contrastive focus and unexpectedness. The fact that Ayacucho Quechua exhibits DOM in this form has three implications for DOM with respect to typology and the theory of information structure. First, in addition to animacy, definiteness, specificity, and topicality, factors which have been reported in many languages up until now, DOM can also be triggered by the factors of contrastive focus and unexpectedness. Second, although up until now it has been said that marking of information structure is not seen in subordinate clauses, it is possible for information structure-based concepts to be marked in subordinate clauses. Third, the presence or absence of contrastiveness as an information structure property can be clearly reflected in linguistic forms.