This paper aims to describe the system of reflexive pronouns in the Aragusuku dialect of Miyako Ryukyuan (henceforth Aragusuku). Aragusuku has three ways to express ‘oneself’, una, duu, and nara, all of which are called reflexive pronouns in this study. The focus of this study is on the identification of the functions of these forms, analyzing in which situation one over the other forms can be (or must be) used.

It is a common feature of Southern Ryukyuan that one dialect has more than one reflexive pronoun as in the case of Aragusuku. Much research has pointed out that one crucial factor is person: the choice of the reflexive forms depends on the person of the referent in question. For example, in Ishigaki, duu is used for first and second person pronouns while nara is only used for third person pronoun (Miyara 1995). However, paying attention to person alone cannot necessarily lead to the accurate description of the reflexive pronominal system of many dialects, including Aragusuku.

In addition to person (B in Figure 1), this study introduces two additional factors to the analysis of the reflexive pronominal system of Aragusuku: distributivity (A in Figure 1) and case (C in Figure 1).

As is summarized in the table, duu is the default (unmarked) choice, for it can be used under any circumstances. By contrast, una and nara are marked choices with their own restrictions. The form una can be used only if it refers to a plural referent with a distributive reading. On the other hand, nara is used only if it refers to a third-person referent whose case must be one of the following: nominative, accusative, genitive, dative, allative and ablative.

The contributions of the present study are two-folds. First, it has introduced the two new factors for the analysis of reflexive pronouns in Ryukyuan languages, which may be helpful not only for Aragusuku but for other Southern Ryukyuan languages that have more than one reflexive pronouns. Second, the factor pertaining to case (C in Figure 1) has theoretical and typological implications. It is observed that cases which can co-occur with nara follows the “Case Hierarchy” (Blake 2001, NOM > ACC (/ ERG) > GEN > DAT > LOC > ABL / INST > COMP > others), a typological hierarchy which is originally introduced to make predictions about the possible size (inventory) of cases in a given language.

In Present-day Japanese, the conjugation (y)oo, used at the end of subordinate clauses, expresses only an inference and cannot express the speaker's will. In contrast, in Late Middle and Early Modern Japanese, there are examples in which (y)oo at the end of ga, hodoni, kara, keredo, and si clauses appears to express the speaker's will.

This paper shows that these examples of (y)oo contain this meaning of will, based on a comparison with the conclusive form of the verb and an examination of materials from different periods. From this discussion, we distinguish the following three stages in the history of (y)oo.

(I) In Late Middle Japanese, (y)oo was used at the end of subordinate clauses to mark irrealis mood, including will and inference.

(II) In Early Modern Japanese, (y)oo ceased to mark irrealis mood, while still possibly expressing will.

(III) In Modern Japanese, (y)oo no longer expressed will at the end of subordinate clauses; will could thus only be expressed in the main clause ending.