Japanese: A Linguistic Introduction is intended to be a college-level reference book on the Japanese language that can also serve as the principal textbook in an introductory course in Japanese linguistics. It explains various linguistic phenomena organized by and in terms of analytical methods developed in the discipline of linguistics. I have endeavored to maintain breadth of scope and intellectual depth sufficient and appropriate for a college course, including discussion of why certain linguistic phenomena are interesting or important and thus continue to be investigated. The targeted readership includes undergraduate and graduate students who are interested in the Japanese language and linguistics, instructors of Japanese, and researchers who wish to survey the field of Japanese linguistics.

To reinforce the reader’s comprehension, exercises and multimedia supplementary materials are available on the book’s website: http://hasegawa.berkeley.edu/Cambridge/introduction.php. Many referenced works are also available online; their URLs are listed in the reference section of the book, although they might cease to exist at any time.

I am deeply indebted to many individuals. First and foremost, I wish to record my gratitude to my students at UC Berkeley, who for more than twenty years have provided inspiration and served as a sounding board. Without them, this book would not have materialized. I am also grateful to those who read earlier versions of the manuscript and offered critical commentary and editorial advice; they include Setsuko Arita, Dante Aurele, Kazue Hata, Yukio Hirose, Christine Jiang, Mika Kizu, Russell Lee-Goldman, Brendan Morley, Gabriel Pellikka, Dennis Ryan, Masaharu Shimada, Mitsuaki Shimojo, Eve Sweetser, Naoaki Wada, and Ikuko Yuasa.

Introduction

A clause is a linguistic unit which consists, at a minimum, of a predicate and its argument(s). It is a unit smaller than a sentence, for a sentence can consist of more than one clause. In Japanese, a predicate is a verb (verbalpredicate), an i- or na-adjective plus copula (adjectivalpredicate), or a noun plus copula (nominalpredicate). Arguments typically, but not necessarily, consist of nounphrases (NPs) such as mado ga, heya ga, kore ga in (1).

Each predicate requires a number of arguments for conveying its essential meaning. Generally, adjectival and nominal predicates, as well as some verbal predicates, require only one argument. In Japanese, this mandatory argument is marked by the postpositional particle ga, as shown in (1), and the NP is said to be in the nominative case. The term grammaticalrelation is also used to identify the grammatical function of the given NP with respect to the predicate. In (1), all of the nominative NPs hold the grammatical relation of subject to their corresponding predicates.

In striking contrast to its simple sound system, the Japanese language employs what is arguably the most convoluted writing system ever devised in human history. Sampson (1985: 173) declares,

This notoriety is ultimately due to the fact that Japanese writing evolved from that of Chinese, a language with substantially different sound and word formation systems.

Writing Japanese with kanji

In the late fourth and early fifth centuries AD, Chinese script was introduced to Japan under the tutelage of Korean scholars.1 In Chinese (as spoken and written then), each meaning unit consisted of one syllable, and was represented by a single character called a KANJI, literally ‘a character of the Han dynasty of ancient China’ (206 BC–220 AD).

Recording a language with kanji is fundamentally different from doing it with, say, the Roman alphabet, which is designed to represent speech sounds. The method of recording speech sounds is termed a PHONOGRAPHIC writing system (phono = sound). In kanji writing, by contrast, ideas, rather than sounds, are represented, so it is called an IDEOGRAPHIC writing system (ideo = idea).

Introduction

Time is an intellectual construct necessary to conceive of and to comprehend changes in the world around us. Grammatical notions of time are manifestations of the human experience of time, traditionally characterized in terms of tense, aspect, and taxis. Tense frames time linearly as a past–present–future continuum and establishes the relationship between the time of the depicted situation and the moment of speech. Aspect is defined as the assessment or characterization of the denoted situation “as it progresses or as it is distributed in time, but irrespective of the moment of speech or … of the time of another action, mentioned or implied” (Maslov 1988: 63). Taxis, which is less common than tense and aspect, is concerned with the chronological relationship between two situations: i.e. do they occur simultaneously; does one take precedence; is there a perceived sequentiality?

In the modern world, people consider tense to be indispensable for an understanding of reality. Nonetheless, a tense system is not an ontological necessity. Historically, aspect is primary, and tense secondary in Indo-European languages (Kurylowicz 1964; Bybee 1985), and creole languages are inherently aspectual (Givón 1982; Kotsinas 1989). For pre-modern people, more salient than tense per se was whether or not a certain change had occurred and whether or not the speaker could ascertain its occurrence with confidence. If the speaker was certain, the hearer would naturally interpret the change to have occurred in the past. Tense has thus emerged as secondary. Tense is abstract, intellectual, and objective; aspect, on the other hand, is definite, impressionistic, and subjective (Izui 1967: 85), involving not only the temporal contour of a situation, but a number of other factors that are not strictly temporal (Michaelis 1993: 17), e.g. modality and evidentiality (Chapter 24), resultativity (Section 9.7), transitivity (Section 7.4), and distribution of focus over sentential constituents (Section 8.9). Ancient Japanese people conceived of time quite differently than their modern Japanese counterparts as evidenced by the fact that the word toki ‘time’ in classical Japanese meant a moment or occasion appropriate to starting a certain action (Wada 1994), e.g. toki no koe [Lit. voice of the time] meaning ‘a battle cry’.

Introduction

The roles of speaker and listener constantly alternate in typical conversation. Person A is the speaker in a given moment, with B as the listener; then B takes his/her turn to speak, while A becomes the listener. Goffman (1974: 136) characterizes this state of affairs as follows.

In both English-speaking societies and Japan, possibly worldwide, the conversational norm consists of one person speaking at a time, with the speaker changing in an orderly manner. If someone starts speaking while you are talking, you are likely to be offended. By contrast, if a speaker ends his/her utterance and a dead silence follows, it might be embarrassing to those present as most people feel a need to avoid silences during conversations. These tacit rules of conversation are studied in linguistics and other academic disciplines in terms of taking the floor (i.e. the right to speak in interaction) or turn-taking. For such investigation, Sacks et al. (1974) developed an analytical framework, conversation analysis, which has been widely used for describing the orderliness and sequential patterns of verbal interaction.