The rules that determine the accent of compound nouns in Tōkyō standard Japanese have long attracted attention because of their complexity. It was noticed early on (Wada, 1943) that the rules in Tōkyō and Kyōto are similar: The accent in the two dialects often agrees, the presence and the location of the accent are primarily determined by the accent of the second element, and distinctions that go back to Middle Japanese still play a role.

In other, more peripheral dialects, the rules appear to be almost opposite, as it is the first element that plays a crucial role (Hiroto & Ōhara 1953, Kobayashi, 1974, Uwano, 1997). Some rules are nevertheless shared by both dialect types: Compounds with a second element of up to 2 moras have different rules than compounds with longer second elements. A similar distinction also existed in Middle Japanese.

The fact that there are considerable differences between the modern dialects on the one hand, while at least some of the modern rules appear to have old roots, leads to questions such as: Do the rules in the different dialects go back to a common proto-system? Is it possible to reconstruct this proto-system? Was it close to Middle Japanese, or does Middle Japanese system represent a relatively late stage? How do the rules of word-tone systems (e. g. Kagoshima, Okinoshima, the Ryūkyūs) relate to the other dialects and Middle Japanese? The presenters in the panel will each address these questions from a different angle.

Compound accentuation and the historical accent classes: The case of the Hida Hagiwara dialect

Rei Fukui (University of Tokyo)

In this study the relationship between compound accentuation and the historical accent classes will be discussed by comparing data from the Hida Hagiwara and Tōkyō dialects.

It is generally known that many Japanese dialects have a set of compound accentuation rules, but these rules alone cannot explain everything. One of the exceptional cases is the phenomenon in which the location of a compound noun is determined by the historical accent classes of the second element, as was discussed by Wada (1943). There are, however, many exceptions to this rule so that it cannot be admitted in such a simplistic form. But it is also true that there are some nouns that show different accentual behavior depending on their accent class, and the reason for this is still unanswered. This study discusses this problem by presenting the data for the Hida Hagiwara dialect, one of the so-called Nairin dialects, which has a different compound accentuation rule in this respect from that of the Tōkyō dialect.

On the accent of compound nouns in the Gairin type dialects

Tatsuya Hirako (Nanzan University)

The Kitsuki (Ōita), Taisha (Shimane), Shinshiro (Aichi), and Shizukuishi (Iwate) dialects belong to the Gairin type, where accent classes 1.1 and 1.2, and accent classes 2.1 and 2.2 have merged. In this presentation, based on my own fieldwork data and a number of previous studies (Hirako and Igarashi 2014, Hirako 2018, Uwano 1997, among others), I will present the following facts concerning the accent of compound nouns in these dialects:

(1) In all of these dialects, if the first element is 2-mora or less in length, the compound accent is determined by the accent of the first element, as Uwano (1997) pointed out: If the first element has no accent, the compound noun has no accent and vice versa. However, in some dialects, if the first element is more than 3-mora in length, the compound will have an accent, whether the first element has an accent or not.

(2) There is a difference in the locus of the accent in compounds between these four dialects; in Kitsuki and Shinshiro, the accent tends to be placed on the first mora of the second element, while it tends to be placed on the second mora of the second element in Taisha and Shizukuishi. In the Kyōto dialect of the 11th-12th century too, the accent tended to be placed on the second mora of the second element, when the second element was 3-mora in length. As far as this point is concerned, the Taisha and the Shizukuishi dialects can be regarded as archaic.

Regularization of the accentual rules for noun compounds in the Japanese dialects

Elisabeth de Boer (Ruhr-Universität Bochum)

For a long time, information on the rules that determine the accent of compounds nouns in Japanese was largely limited to the more central dialects (Kyōto type and Nairin and Chūrin Tōkyō type), where accent was thought to depend on the second element of the compound only. Information on the peripheral Gairin type accent systems was incomplete but seemed to indicate that the first element decided whether a compound was accented or not. In recent years, new fieldwork has shown that in many Gairin type accent systems, compounds with unaccented first elements do in fact have accent (Hirako & Igarashi 2014, Hirako 2018). It is also well known that there are many exceptions to the rule in the central dialects that accent is determined by the second element of the compound only.

A comparison of compound accent in a number of different Tōkyō type accent systems indicates that the idea of a sharp typological divide between the rules in the central and the peripheral dialects is probably better abandoned.

In both dialect groups, accent is to a certain extent still linked to the tones of Middle Japanese, but there are many exceptions. I will argue that these exceptions are best explained if we assume that longer compounds were regularized into the tonal shape of compounds that started with an ‘even’ tone in Middle Japanese. Shorter compounds were regularized into the tonal shape of compounds that started with a ‘rising’ tone in Middle Japanese, but this regularization was less complete.