1 Introduction
1.1 Location
Narada is a variety of Japanese spoken in Narada village, situated at the northern edge of Hayakawa-cho town in Yamanashi Prefecture. Narada village has a rural location at the entrance to the Southern Alps of Japan. The nearest neighboring village is at least five kilometers away from Narada.
This geographical and social isolation has led to Narada’s characterization as an unexplored region. Until the 1950s, inhabitants of Narada village were slash-and-burn subsistence farmers and tended to intermarry. This environment is assumed to have led to autonomous changes in the Narada prosodic system.
1.2 Speakers
Fewer than ten Narada speakers reside in Narada village. All current Narada speakers are over 70 years old. Some speakers have migrated to other areas, but the exact number is unknown. Descendants of Narada speakers have not inherited the Narada variety of Japanese. Recent immigrants to Narada village have not become Narada speakers, placing Narada in danger of disappearing.
1.3 Classification
Tojo (1954) categorizes Narada within the Tokai-Tosan group, a subdivision of the Eastern Japanese group (see § 2.1 of Chapter 1, this volume). Despite this classification, the Narada prosodic system is markedly different from neighboring varieties. Scholars therefore refer to Narada as a language island in Japanese dialectology. Nonetheless, the Narada prosodic system is proven to have diachronically evolved from neighboring varieties (see § 2.8).
1.4 Previous Studies
Mochizuki (1951) first described Narada’s prosodic system. Based on the resemblance of surface pitch patterns, he posited that Narada was phylogenetically close to the Kinki varieties, such as Kyoto. In contrast, Hirayama (1955) and Inagaki (1957) suggested that the Narada prosodic system developed from the neighboring Yamanashi variety. Their perspective draws on the tonal class correspondences between Narada and other Japanese varieties. Subsequent studies, including Uwano (2011), corroborate this view.
As a synchronic interpretation of the prosodic system, Uwano (1975, 1976, 1984, 2012) proposed that Narada has a raising kernel (§ 2.3 of Chapter 1, this volume). This positions Narada as unlike the many other varieties with a lowering kernel. Uwano’s theory is widely accepted by scholars. Konishi et al. (2022) outlined Narada’s prosodic system, following Uwano’s theory. Moreover, in a sociolinguistic study, Konishi (2001) discussed changes in pitch patterns resulting from linguistic contact between the Narada and Yamanashi varieties.
1.5 Data
The descriptions in this chapter are based on data obtained from onsite surveys conducted from 1999 to 2024. The surveys were conducted on Narada speakers born between 1917 and 1938, who were raised in Narada village and were residents of the village at the time.
1.6 Overview of the Segmental Phonology1
Narada has five vowels, as shown in (1), and 18 consonants, as shown in (2) and Table 3.1. The consonants /ɴ, Q/ in (2) (excluded from Table 3.1) are moraic. This aspect is discussed later in this section.
(1) Vowel phonemes
/i, u, e, o, a/
(2) Consonant phonemes
/p, b, m, t, d, s, z, c, n, r, k, g, ŋ, h, w, j, ɴ, Q/
The phoneme /u/ is realized as the rounded vowel [u]. The close vowels /i/ and /u/ can be devoiced between voiceless consonants. However, devoicing is less frequent in the Narada variety than in Tokyo Japanese. The phoneme /e/ does not occur in word-initial position in traditional pronunciation, with word-initial /e/ in Tokyo Japanese corresponding to /je/ in Narada, such as in /jeki/ ‘station’, /je/ ‘picture’, and /jeza/ [jeða] ‘branch’.
Table 3.1
Consonant phoneme inventory and typical allophones. (Phonemes are indicated by slashes, whereas typical allophones are enclosed in parentheses. Symbols in square brackets are used for phonemic transcriptions if the phonetic realization may not be immediately apparent.)
|
Bilabial |
Dental |
Alveolar |
Alveolo-palatal |
Retroflex |
Palatal |
Velar |
Uvular |
Glottal |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Plosive |
/p/ |
/b/ |
/t/ |
/d/ |
(ʈ) |
(ɖ) |
/k/ |
/g/ |
||||||||||
|
Affricate |
(ts) |
(dz) |
/c/ (tɕ) |
(dʑ) |
||||||||||||||
|
Nasal |
/m/ |
/n/ |
/ŋ/ |
(ɴ) |
||||||||||||||
|
Tap |
(ɾ) |
|||||||||||||||||
|
Fricative |
(ɸ) |
( |
(ð) |
/s/ |
/z/ |
(ɕ) |
(ʑ) |
(ç) |
/h/ |
|||||||||
|
Approximant |
/w/ |
/j/ |
||||||||||||||||
|
Lateral Approximant |
/r/ [l] |
(ɭ) |
||||||||||||||||
Palatalized allophonic variations are observed in /s/, /z/, and /h/ before /i/ and /j/. In this position, they are realized as [ɕ], [ʥ ~ ʑ], and [ç], respectively. The phoneme /c/ occurs exclusively before /i/ or /j/, whereas the phoneme /t/ generally does not precede these segments, with the exception of some recent loanwords such as /tiisjatu/ [tiːɕatu] ‘T-shirt’. The phoneme /d/ is not followed by /i/ or /j/, because /di/ and /djV/ (in which “V” denotes a vowel) have historically merged into /zi/ and /zjV/, respectively. In traditional pronunciation, the phonemes /t/ and /d/ are realized as retroflexes [ʈ] and [ɖ], respectively, whereas they tend to appear as alveolars [t] and [d] in contemporary pronunciation. Similarly, the phoneme /r/ is realized as a retroflex lateral [ɭ] in traditional pronunciation but may appear as an alveolar lateral [l] or an alveolar tap [ɾ] in contemporary pronunciation.
The phonemes /t/ and /d/ before /u/ are typically not affricated (e.g., /tudumi/ [ʈuɖumi], ‘hand drum’). However, their affricate (or, in the voiced case, fricative) allophones [ʦ] and [ʣ ~ z] are also observed. When not followed by /i/ or /j/, the phonemes /s/ and /z/ exhibit variation between interdentals [
The phoneme /g/ typically occurs word-initially, such as in /goma/ [ɡoma, *ŋoma] ‘sesame’. Conversely, the phoneme /ŋ/ occurs word-medially and never word-initially, such as in /kaŋami/ [kaŋami, *kaɡami] ‘mirror’. The phoneme /g/ can also appear after a word-medial morpheme boundary, but it is not obligatory. Despite the nearly complementary distribution between [ɡ] and [ŋ], this chapter follows Uwano (2004) and addresses them as independent phonemes. This standpoint is informed by the lack of phonetic motivation for the (de)nasalization of velar plosives and the presence of minimal pairs such as /oo+garasu/ [oːɡaɭasu] ‘large glass’ versus /oo+ŋarasu/ [oːŋaɭasu] ‘large crow’ (the former from /garasu/ ‘glass’ vs. the latter from /karasu/ ‘crow’).
Narada syllable structure can be generalized as follows: (C1)(G)V1(V2)(C2). The rhyme slots V1, V2, and C2 each constitute one mora. C1 represents the onset consonant, which can be any consonant except the moraic /ɴ/ or /Q/. G represents a glide, which is filled with /j/ when C1 is neither /j/ nor /w/. Alternatively, in certain loanwords, G can be filled with /w/ when C1 is /h/ (e.g. /aruhwabeQto/ [aɾuɸabetto] ‘alphabet’). V1 and V2 form a syllable nucleus. V1 is filled with any vowel, while V1V2 represents a long vowel (e.g., /ii/ or /oo/) or a diphthong (e.g., /ai/ or /oi/). The coda consonant C2 can be the moraic nasal /ɴ/ (hatsuon in Japanese linguistics) or the moraic obstruent /Q/ (sokuon in Japanese linguistics).
The place of articulation for /ɴ/ and /Q/ is dependent on context. When /ɴ/ precedes another consonant, its place of articulation assimilates to that of the subsequent consonant. When /ɴ/ precedes a vowel or semivowel, it is realized as a nasalized vowel. The place of articulation of an utterance-final /ɴ/ tends to be uvular. The place of articulation for /Q/ assimilates to the obstruent in the onset of the subsequent syllable, which forms the first part of a geminate.
C2 occurs after V1 (a short vowel) or V1V2 (a long vowel or a diphthong). Examples include /waɴ/ [waɴ] ‘soup bowl’, which consists of a heavy syllable C1V1C2, and /si.maiQ.ko/ [ɕimaikko] ‘youngest child’, in which the second syllable is a superheavy C1V1V2C2.
2 Word-Prosodic System
2.1 Classification Based on Uwano (1989, 1999)
The Narada prosodic system is classified as a multipattern system in Uwano’s (1989, 1999) typology. In a multipattern system, the number of tonal contrasts increases proportionally with the word length (see § 2.4 of Chapter 1, this volume).
2.2 Number and Nature of Tonal Contrasts
As a consequence of Narada’s multipattern system, it is challenging to determine the maximum number of tonal patterns.
Narada has a prosodic system in which tonal patterns are distinguished by the presence or absence and position of the raising kernel (Uwano 1984, 1989, 2012; Konishi et al. 2022). The tonal patterns for one- to four-mora nouns in Narada are shown in (3). “N” represents nouns in isolation, “N=ŋa” indicates that the noun is followed by the nominative particle =ŋa, and “kono N=ŋa” indicates that the noun has the nominative particle preceded by the adnominal word kono ‘this’. The notation “U”, “A1”, “A2”, “A3”, and “A4” indicate tonal patterns. “U” stands for unaccented types, while “A” stands for accented types. The number following “A” indicates the position of the accent kernel: “A1” indicates the first mora position, “A2” indicates the second mora position, and so forth. The pitch contours of utterances containing two-mora nouns in sentences are depicted in Figures 3.1a–c, including ⸢a⸣meŋa a⸢ru (‘There is a candy’), o⸢bi⸣ŋa a⸢ru (‘There is a belt’), and ⸢u⸣ma⸢ŋa⸣ iru (‘There is a horse’). The diacritics “⸢” and “⸣” indicate a pitch rise and fall, respectively, between morae (see § 2.2 of Chapter 1, this volume).
|
Type |
N (noun) |
N=ŋa (N=nom) |
kono N=ŋa |
||
|---|---|---|---|---|---|
|
(adn N=nom) |
|||||
|
(3) |
Tonal patterns of Narada |
||||
|
1- |
U |
⸢hi ‘sun’ |
⸢hi⸣ŋa |
⸢ko⸣no hiŋa |
|
|
A1 |
⸢hi ‘fire’ |
hi⸢ŋa |
⸢ko⸣no hi⸢ŋa |
||
|
2- |
U |
⸢a⸣me ‘candy’ |
⸢a⸣meŋa |
⸢ko⸣no ameŋa |
|
|
A1 |
o⸢bi ‘belt’ |
o⸢bi⸣ŋa |
⸢ko⸣no o⸢bi⸣ŋa |
||
|
A2 |
⸢u⸣ma ‘horse’ |
⸢u⸣ma⸢ŋa |
⸢ko⸣no uma⸢ŋa |
||
|
3- |
U |
⸢ku⸣ruma ‘car’ |
⸢ku⸣rumaŋa |
⸢ko⸣no kurumaŋa |
|
|
A1 |
u⸢si⸣ro ‘behind’ |
u⸢si⸣roŋa |
⸢ko⸣no u⸢si⸣roŋa |
||
|
A2 |
⸢ko⸣ko⸢ro ‘mind’ |
⸢ko⸣ko⸢ro⸣ŋa |
⸢ko⸣no koko⸢ro⸣ŋa |
||
|
A3 |
⸢ka⸣ŋami ‘mirror’ |
⸢ka⸣ŋami⸢ŋa |
⸢ko⸣no kaŋami⸢ŋa |
||
|
4- |
U |
⸢ni⸣watori ‘chicken’ |
⸢ni⸣watoriŋa |
⸢ko⸣no niwatoriŋa |
|
|
A1 |
mu⸢ra⸣saki ‘purple’ |
mu⸢ra⸣sakiŋa |
⸢ko⸣no mu⸢ra⸣sakiŋa |
||
|
A2 |
⸢a⸣sa⸢ŋa⸣o ‘morning |
⸢a⸣sa⸢ŋa⸣oŋa |
⸢ko⸣no asa⸢ŋa⸣oŋa |
||
|
glory’ |
|||||
|
A3 |
⸢ka⸣raka⸢sa ‘umbrella’ |
⸢ka⸣raka⸢sa⸣ŋa |
⸢ko⸣no karaka⸢sa⸣ŋa |
||
|
A4 |
⸢ku⸣cibiru ‘lips’ |
⸢ku⸣cibiru⸢ŋa |
⸢ko⸣no kucibiru⸢ŋa |
||
Figure 3.1
Pitch contour of (a) ⸢a⸣meŋa a⸢ru ‘There is a candy’, (b) o⸢bi⸣ŋa a⸢ru ‘There is a belt’, and (c) ⸢u⸣ma⸢ŋa⸣ iru ‘There is a horse’
Examples of two-mora nouns are presented. For ame ‘candy’, when uttered in isolation or followed by the particle =ŋa, the initial mora exhibits a high pitch, as in ⸢a⸣me and ⸢a⸣meŋa. However, when preceded by the adnominal kono ‘this’, the initial mora of kono is high-pitched, as in ⸢ko⸣no ameŋa. The pitch of the initial mora of the noun remains low unless the noun is focused. The same observation applies to uma ‘horse’. This initial high pitch can be regarded as a phrase-level tone (§ 2.6, § 4.1). However, the rises from the first mora to the second in obi ‘belt’, and from the second mora to the following particle in uma, are independent of syntactic context. A rise of this nature is absent in ame. Consequently, obi has an accent on the first mora, while uma has an accent on the second mora. (Throughout this chapter, the terms accent and accent kernel are used interchangeably.) On the other hand, the earlier example ame lacks accentuation. This interpretation extends to nouns with one, three, and four morae. Hereafter, the phonological transcriptions will represent accents as in /o⸢bi/ and /uma⸢/.
The high pitch resulting from an accent kernel is typically confined to the mora immediately after the accented mora. However, this high pitch can occasionally spread to a sequence of two or more morae, as shown in (4). The precise conditions under which this high pitch spreads are not fully understood. However, even in cases where the high pitch spreads, the particle =ŋa typically exhibits a low pitch unless there is an accent-induced rise.
|
(4) |
3- |
A1 /u⸢siro/ |
⸢ko⸣no u⸢si⸣roŋa |
⸢ko⸣no u⸢siro⸣ŋa |
|
4- |
A1 /mu⸢rasaki/ |
⸢ko⸣no mu⸢ra⸣sakiŋa |
⸢ko⸣no mu⸢rasaki⸣ŋa |
|
|
A2 /asa⸢ŋao/ |
⸢ko⸣no asa⸢ŋa⸣oŋa |
⸢ko⸣no asa⸢ŋao⸣ŋa |
As shown in (3), relatively short words with one to four morae may have n+1 tonal patterns for n-mora words. However, similar to other Japonic varieties with multipattern systems, the distributions of tonal patterns in longer words are more biased towards certain patterns. This makes it difficult to accurately determine the total number of tonal patterns.
2.3 Domain of Lexical Tone Assignment
In general, the domain of lexical tone assignment can be regarded as a lexical word. However, as will be discussed in the subsequent section, the intricate interaction between the accents of nouns and particles renders this formalization increasingly challenging.
|
Type U |
Type A1 |
Type A2 |
|||
|---|---|---|---|---|---|
|
Type |
Particle |
/ame/ ‘candy’ |
/o⸢bi/ ‘belt’ |
/uma⸢/ ‘horse’ |
|
|
(5) |
Pitch patterns with particles on two-mora nouns |
||||
|
I |
=ŋa (nom) |
⸢a⸣meŋa |
o⸢bi⸣ŋa |
⸢u⸣ma⸢ŋa |
|
|
=o (acc) |
⸢a⸣meo, ⸢a⸣mjoo |
o⸢bi⸣o, o⸢bjo⸣o |
⸢u⸣ma⸢o, ⸢u⸣ma⸢a |
||
|
=mo (add) |
⸢a⸣memo |
o⸢bi⸣mo |
⸢u⸣ma⸢mo |
||
|
=kara (abl) |
⸢a⸣mekara |
o⸢bi⸣kara |
⸢u⸣ma⸢ka⸣ra |
||
|
=Qkiri (rst) |
⸢a⸣meQkiri |
o⸢bi⸣Qkiri |
⸢u⸣maQ⸢ki⸣ri |
||
|
II |
=no (gen) |
⸢a⸣meno |
o⸢bi⸣no |
⸢u⸣mano |
|
|
III |
=made (lim) |
⸢a⸣mema⸢de |
o⸢bi⸣made |
⸢u⸣ma⸢ma⸣de |
|
|
=sikaa (rst) |
⸢a⸣mesi⸢ka⸣a |
o⸢bi⸣sikaa |
⸢u⸣ma⸢sika⸣a |
||
|
I+I |
=o=mo |
⸢a⸣meo⸢mo |
o⸢bjo⸣omo |
⸢u⸣ma⸢a⸣mo |
|
|
(acc=add) |
|||||
|
I+I |
=Qkiri=ŋa |
⸢a⸣meQkiri⸢ŋa |
o⸢bi⸣Qkiri⸢ŋa |
⸢u⸣maQ⸢ki⸣riŋa |
|
|
(rst=nom) |
|||||
|
I+II |
=Qkiri=no |
⸢a⸣meQkiri⸢no |
o⸢bi⸣Qkiri⸢no |
⸢u⸣maQ⸢ki⸣rino |
|
|
(rst=gen) |
|||||
At least three prosodic particle types follow nouns, which are shown in (5), namely, Types I, II, and III. Type I particles are unaccented, which enables the accent of the preceding lexical word to remain unchanged. This type includes the nominative particle =ŋa, as presented in (3), the accusative particle =o, the additive particle =mo, the ablative particle =kara, and the restrictive particle =Qkiri. The accusative particle typically fuses with the final vowel of the preceding noun, but the accent position is unchanged. The coda /Q/ is not realized with a high pitch; therefore, in the noun phrase umaQkiri, the accent shifts one mora to the right.
While the majority of one-mora particles belong to Type I, the genitive particle =no is exceptionally categorized as Type II. It deletes the accent on the final mora of the noun.2
Type III particles, such as =made and =sikaa, exhibit their own accent, which is realized if the lexical word is unaccented as ⸢a⸣me. However, it is not realized when the noun has an accent, such as o⸢bi and ⸢u⸣ma. The deaccenting of particles suggests that Narada has a constraint that allows, at most, one accent within the bunsetsu—a unit comprising a lexical word and any subsequent particles (see § 2.6 of Chapter 1, this volume). Thus, the bunsetsu can be interpreted as the domain of accentuation. The following text discusses the complexities of this interpretation.
When multiple particles follow a lexical word, the accent can appear within the sequence of particles, as in the cases of =o=mo, =Qkiri=ŋa, and =Qkiri=no shown in (5). Specifically, the accent is assigned to the final mora of a particle when followed by another particle. This additional accenting is observed both in constructions with an unaccented noun, such as ⸢a⸣meo⸢mo, ⸢a⸣meQkiri⸢ŋa, and ⸢a⸣meQkiri⸢no, and in constructions with an accented noun, such as in o⸢bi⸣Qkiri⸢ŋa and o⸢bi⸣Qkiri⸢no. These examples indicate that both nouns and particles carry accents, which results in two accents within the bunsetsu. This suggests that no strict constraint prohibiting multiple accents exists within the bunsetsu in Narada. However, in contrast to =made and =sikaa, the accents in particle sequences are lexically unspecified. Instead, they are assigned by morphophonological rules. Thus, the constraint on multiple accents may only apply to lexically specified accents.
Additionally, when the distance between the accent of the noun and the right boundary of the immediately following particle is minimal, the accent in the particle sequence is not manifested. Consider o⸢bjo⸣omo, ⸢u⸣maQ⸢ki⸣riŋa, and ⸢u⸣maQ⸢ki⸣rino. This deaccenting of particle sequences can be viewed as a consequence of avoiding clashes in accent and appears to operate with the bunsetsu as its domain.
Thus, although a few phenomena indicate that accentuation operates with bunsetsu as its domain, the accentual domain in Narada requires further investigation. In any case, the accent of the noun is neither deleted by subsequent particles nor is its position changed.
2.4 Tone-Bearing Units
The minimal tone-bearing unit (TBU) is the mora. The mora also serves as the primary “counting unit” (McCawley 1978) for measuring phonological length or distance (see § 2.7 of Chapter 1, this volume). Narada exhibits a lexical contrast between accents on the first and second mora of a heavy syllable, as noted by Uwano (2019). Example (6) illustrates this accentual contrast between the first and second morae, forming the heavy syllables /C(G)VV/ and /CVC/.3 In examples (6a1) and (6a2), which share the structure /Cjoo.ri/ (where “C” denotes a consonant and a dot indicates a syllable boundary), (6a1) has the accent on the first mora, whereas (6a2) has it on the second mora of the heavy syllable /Cjoo/. Similarly, examples (6b1) and (6b2), which share the structure /Ci.rii/, have an accent on the first and second morae of the heavy syllable /rii/, respectively. These examples demonstrate that both the first and second morae, which form a syllable with a long vowel, can carry the accent. This observation also holds for words with a heavy syllable containing a moraic nasal. Consider examples (6c1) and (6c2), which have the structure /Caɴ.bo/, and (6d1) and (6d2), which have the structure /Co.Ciɴ/. In both cases, the first and second morae of the heavy syllable can bear an accent.
|
(6) |
a1. |
/rjo⸢ori/ |
rjo⸢o⸣ri=ŋa ‘cooking=nom’ |
|
a2. |
/zjoo⸢ri/ |
zjoo⸢ri⸣=ŋa ‘sandals=nom’ |
|
|
b1. |
/biri⸢i/ |
biri⸢i⸣=ŋa ‘intestine=nom’ |
|
|
b2. |
/hirii⸢/ |
⸢hi⸣rii=⸢ŋa ‘lunch=nom’ |
|
|
c1. |
/na⸢ɴbo/ |
na⸢ɴ⸣bo ‘how much’ |
|
|
c2. |
/taɴ⸢bo/ |
taɴ⸢bo⸣=ŋa ‘rice filed=nom’ |
|
|
d1. |
/kobi⸢ɴ/ |
kobi⸢ɴ⸣=ŋa ‘forehead=nom’ |
|
|
d2. |
/joniɴ⸢/ |
⸢jo⸣niɴ=⸢ŋa ‘four people=nom’ |
These lexical contrasts cannot be observed in heavy syllables that end with the moraic obstruent /Q/. The accent cannot fall on the mora preceding /Q/, because /Q/ is not realized with a high pitch, although /Q/ can bear an accent such as in naQ⸢pa⸣=ŋa ‘vegetable leaf=nom’.
The accent of loanwords typically falls on the antepenultimate mora, which is the third mora from the end of a word, as illustrated in (7). However, early loanwords frequently deviate from this pattern, as indicated by the symbol † in (7). As /Q/ does not bear a high pitch, instances in which /Q/ is the penultimate mora and carries an accent also adhere to this rule. These patterns indicate that the mora is, indeed, the counting unit.
|
(7) |
3- |
/ba⸢taa/ ‘butter’, /bi⸢iru/ ‘beer’, /da⸢ɴsu/ ‘dance’, /ke⸢eki/ ‘cake’, |
|
/ku⸢rasu/ ‘class’, /piano/ (U)† ‘piano’ |
||
|
4- |
/apa⸢ato/ ‘apartment’, /bita⸢miɴ/ ‘vitamin’, /kjara⸢meru/ ‘caramel’, |
|
|
/see⸢taa/ ‘jumper’, /arubamu/ (U)† ‘album’, /sa⸢idaa/† ‘soda pop’, |
||
|
/teeburu/ (U)† ‘table’ |
||
|
5- |
/jooŋu⸢ruto/ ‘yoghurt’, /bisukeQ⸢to/ ‘biscuit’, /haamonika/ (U)† ‘harmonica’ |
|
|
6- |
/daijamo⸢ɴdo/ ‘diamond’, /aruhwabeQ⸢to/ ‘alphabet’, /oriɴpiQ⸢ku/ ‘Olympics’ |
|
|
† Exceptions to the prosodic rule of loanwords |
||
2.5 Directionality
Directionality is the direction in which the position of word-level prominence is calculated (see § 2.8 of Chapter 1, this volume). Narada directionality is right-to-left, as indicated by the productive tonal patterns of loanwords, verbs, and adjectives.
The productive tonal pattern of loanwords typically has the accent falling on the antepenultimate mora (see § 2.4 of this chapter). Verbs and adjectives are classified into unaccented and accented types. The regular accent location for the latter is the penultimate mora. This classification is outlined in § 3.1. These rules suggest that Narada directionality is right-to-left, meaning that the accent location is calculated from the right edge of the domain.
2.6 Obligatoriness and Culminativity
Obligatoriness requires at least one prominence within the word domain, whereas culminativity requires at most one prominence within the word domain (Hyman 2006; see § 2.5 of Chapter 1, this volume). The Narada prosodic structure does not satisfy the obligatoriness requirement. However, it appears to satisfy the culminativity requirement. Marginal cases are discussed below. The three-mora noun examples from § 2.2 (3) are presented again in (8).
|
N (noun) |
N=ŋa (N=nom) |
kono N=ŋa |
|||
|---|---|---|---|---|---|
|
(adn N=nom) |
|||||
|
(8) |
Tonal patterns of three-mora nouns in Narada. |
||||
|
U |
/kuruma/ |
⸢ku⸣ruma ‘car’ |
⸢ku⸣rumaŋa |
⸢ko⸣no kurumaŋa |
|
|
A1 |
/u⸢siro/ |
u⸢si⸣ro ‘behind’ |
u⸢si⸣roŋa |
⸢ko⸣no u⸢si⸣roŋa |
|
|
A2 |
/koko⸢ro/ |
⸢ko⸣ko⸢ro ‘mind’ |
⸢ko⸣ko⸢ro⸣ŋa |
⸢ko⸣no koko⸢ro⸣ŋa |
|
|
A3 |
/kaŋami⸢/ |
⸢ka⸣ŋami ‘mirror’ |
⸢ka⸣ŋami⸢ŋa |
⸢ko⸣no kaŋami⸢ŋa |
|
Narada’s three-mora nouns exhibit the four patterns shown in (8). These can be re-examined considering obligatoriness and culminativity.
A phrase with the noun kuruma usually shows a fall in pitch from the first to the second mora. However, this does not occur when preceded by the adnominal kono. This indicates that this HL pitch is a phrase-level tone rather than a word-level tone. Therefore, the noun kuruma can be interpreted as unaccented. Consequently, Narada does not satisfy obligatoriness.
Phrases with the noun kokoro can exhibit a fall in pitch from the first to the second mora and a rising pitch from the second to the third mora. However, the fall in pitch does not occur when preceded by the adnominal kono. This suggests a phrase-level tone. The rising pitch from the second to the third mora is then the only word-level tone. Similarly, for words like kaŋami=ŋa in (8) or other words with two instances of prominence, the fall in pitch from the first to the second mora can be interpreted as a phrase-level tone.
When considering culminativity, there is a marginal case involving the accents in particle sequences discussed in § 2.3. When multiple particles follow a lexical word, accents can appear within the particle sequences. This is seen in examples like =o=mo, =Qkiri=ŋa, and =Qkiri=no. If the lexical word preceding the particle sequence is lexically accented, the accents of both the lexical word and the particle sequences can be realized under certain conditions. Consider o⸢bi⸣Qkiri⸢ŋa and o⸢bi⸣Qkiri⸢no. This results in two prominences within the bunsetsu. As discussed in § 2.3, the accent in particle sequences is not lexically specified but is determined by a morphophonological rule. However, this volume operationally defines word-level tones as those assigned within the bunsetsu. The accent in the particle sequences should then be regarded as a word-level tone. Therefore, when considering this marginal case, Narada fails to meet the culminativity requirement.
2.7 Prosodic Compound Rules
The prosodic compound rules for nouns are right-dominant, whereas those for verbs are left-dominant (as discussed in § 3.2).
The examples in (9) present instances of compound nouns with three- and four-mora-final elements.4 The initial elements play no role in the accentuation of compound nouns. In cases in which the final element is not A2, the accent falls on the first mora of the final element. When the final element is A2, the accent falls on the second mora of the final element. The accentuation of the final element of compound nouns then influences the accentuation of the compound noun, which demonstrates the correct dominance of prosodic compound rules. The rules described feature a few exceptions, such as in kaza+ŋuruma, natu+jasumi, and mici+aɴnai.
|
Initial morpheme |
Final morpheme |
Compound noun |
|
|---|---|---|---|
|
(9) |
Prosodic patterns of compound nouns |
||
|
A1 /ni⸢/ ‘luggage’ |
U /kuruma/ ‘wheel, car’ |
A2 /ni+ŋu⸢ruma/ ‘cart’ |
|
|
A1 /ha⸢/ ‘tooth’ |
A2 /ha+ŋu⸢ruma/ ‘cog-wheel’ |
||
|
A1 /u⸢ba/ ‘wet nurse’ |
A3 /uba+ŋu⸢ruma/ ‘pram, baby buggy’ |
||
|
U /kaze/ ‘wind’ |
A5† /kaza+ŋuruma⸢/ ‘windmill’ |
||
|
A1 /te⸢/ ‘hand’ |
A1 /rjo⸢ori/ ‘cooking’ |
A2 /te+rjo⸢ori/ ‘home cooking’ |
|
|
A1 /na⸢be/ ‘pot’ |
A3 /nabe+rjo⸢ori/ ‘hot pot’ |
||
|
U /sakana/ ‘fish’ |
A4 /sakana+rjo⸢ori/ ‘seafood dish’ |
||
|
A2 /tama⸢ŋo/ ‘egg’ |
A4 /tamaŋo+rjo⸢ori/ ‘egg dish’ |
||
|
U /huraɴsu/ ‘France’ |
A5 /huraɴsu+rjo⸢ori/ ‘French cooking’ |
||
|
U /goma/ ‘sesame’ |
A2 /abu⸢ra/ ‘oil’ |
A4 /goma+abu⸢ra/ ‘sesame oil’ |
|
|
U /daidu/ ‘soybean’ |
A5 /daidu+abu⸢ra/ ‘soybean oil’ |
||
|
A1 /ki⸢kai/ ‘machine’ |
A5 /kikai+abu⸢ra/ ‘machine oil’ |
||
|
A2 /nata⸢ne/ ‘rapeseed’ |
A5 /natane+abu⸢ra/ ‘rapeseed oil’ |
||
|
U /teɴpura/ ‘tempura’ |
A6 /teɴpura+abu⸢ra/ ‘tempura oil’ |
||
|
A1 /a⸢me/ ‘rain’ |
A3 /otoko⸢/ ‘man’ |
A3 /ame+o⸢toko/ ‘man who brings rain’ |
|
|
A2 /iro⸢/ ‘color’ |
A3 /iro+o⸢toko/ ‘good-looking man’ |
||
|
A3 /gozjuu⸢/ ‘fifty’ |
A4 /gozjuu+o⸢toko/ ‘fifty-year-old man’ |
||
|
A2 /oo⸢kami/ ‘wolf’ |
A5 /ookami+o⸢toko/ ‘werewolf’ |
||
|
A1 /ha⸢ru/ ‘spring’ |
A3 /jasumi⸢/ ‘recess, holiday’ |
A3 /haru+ja⸢sumi/ ‘spring vacation’ |
|
|
A1 /na⸢ka/ ‘mid’ |
A3 /naka+ja⸢sumi/ ‘recess’ |
||
|
A2 /na⸢tu/ ‘summer’ |
A3, A5† /natu+ja⸢sumi, natu+jasumi⸢/ ‘summer vacation’ |
||
|
A4 /sjooŋatu⸢/ ‘January’ |
A5 /sjooŋatu+ja⸢sumi/ ‘New Year’s vacation’ |
||
|
A3 /otoko⸢/ ‘man’ |
U /tomodaci/ ‘friend’ |
A4 /otoko+to⸢modaci/ ‘male friend’ |
|
|
A3 /oɴna⸢/ ‘woman’ |
A4 /oɴna+to⸢modaci/ ‘male friend’ |
||
|
A3 /otoko⸢/ ‘man’ |
A1 /kjo⸢odai/ ‘sibling’ |
A4 /otoko+kjo⸢odai/ ‘brothers’ |
|
|
A1 /i⸢kuɴniɴ/ ‘how many people’ |
A5 /ikuniɴ+kjo⸢odai/ ‘how many siblings’ |
||
|
A3 /saɴni⸢ɴ/ ‘three people’ |
A5 /saɴniɴ+kjo⸢odai/ ‘three siblings’ |
||
|
A1 /ke⸢e/ ‘light (motor vehicle)’ |
A2 /zido⸢osja/ ‘motor vehicle’ |
A4 /kee+zido⸢osja/ ‘light motor vehicle’ |
|
|
U /koŋata/ ‘small size’ |
A5 /koŋata+zido⸢osja/ ‘compact car’ |
||
|
A4 /sakenomi⸢/ ‘drinker’ |
A3 /ziɴzi⸢i/ ‘elderly man’ |
A5 /sakenomi+zi⸢ɴzii/ ‘bibacious elderly man’ |
|
|
U /mici/ ‘road, way’ |
A4 /aɴnai⸢/ ‘guide’ |
A3, A6† /mici+a⸢ɴnai, mici+aɴnai⸢/ ‘directions’ |
|
|
U /kaɴkoo/ ‘sightseeing’ |
A5 /kaɴkoo+a⸢ɴnai/ ‘tourist information’ |
||
|
† Exceptions to the prosodic rule of compound nouns |
|||
2.8 Merger and Split of Historic Tonal Classes
The tonal class correspondences for one- and two-mora nouns between Middle Japanese (MJ) (see § 2.10 of Chapter 1, this volume) and Narada is presented in Table 3.2.
Table 3.2
Correspondences in tonal classes between MJ and Narada
|
MJ tonal classes |
Narada tonal classes |
|---|---|
|
1.1 |
Type U |
|
1.2 |
Type U |
|
1.3 |
Type A1 |
|
2.1 |
Type U |
|
2.2 |
Type A2 |
|
2.3 |
Type A2 |
|
2.4 |
Type A1 |
|
2.5 |
Type A1 |
Despite significant differences in the types of accent kernels and surface patterns between Narada and Tokyo, Narada generally aligns with Tokyo in its correspondence with the MJ tonal classes. Interestingly, Narada shows a more consistent tonal correspondence with Yamanashi than with Tokyo (Konishi 2001). Examples are provided in (10). Most Class 2.2 and Class 2.3 nouns exhibit A2 across the three varieties. However, the Class 2.2 noun nasi ‘pear’ uniquely shows U in Narada and Yamanashi, and the Class 2.3 noun kumo ‘cloud’ uniquely displays A1 in Tokyo. Class 3.5 nouns generally show A2 in all three varieties. The Class 3.5 noun namida ‘tear’ is an exception, displaying A1 in Tokyo but A2 in Narada and Yamanashi. Additionally, for modern-era words like kisja ‘train’ and jakuba ‘town hall’, Narada’s tonal correspondence aligns more closely with Yamanashi than Tokyo.
|
Narada |
Yamanashi |
Tokyo |
||||
|---|---|---|---|---|---|---|
|
(10) |
Tonal patterns in Narada, Yamanashi, and Tokyo. |
|||||
|
⸢na⸣si=ŋa ‘pear=nom’ |
U |
na⸢si=ŋa |
U |
na⸢si⸣=ŋa |
A2 |
|
|
ki⸢sja⸣=ŋa ‘train=nom’ |
A1 |
⸢ki⸣sja=ŋa |
A1 |
ki⸢sja⸣=ŋa |
A2 |
|
|
⸢ku⸣mo⸢=ŋa ‘cloud=nom’ |
A2 |
ku⸢mo⸣=ŋa |
A2 |
⸢ku⸣mo=ŋa |
A1 |
|
|
ja⸢ku⸣ba=ŋa ‘town hall=nom’ |
A1 |
⸢ja⸣kuba=ŋa |
A1 |
ja⸢kuba⸣=ŋa |
A3 |
|
|
⸢na⸣mi⸢da⸣=ŋa ‘tear=nom’ |
A2 |
na⸢mi⸣da=ŋa |
A2 |
⸢na⸣mida=ŋa |
A1 |
|
These tonal correspondences led Inagaki (1957) and Uwano (2011) to conclude that the Narada’s prosodic system developed from changes in the Yamanashi system. Although Kindaichi (1977: 166–167) classified Narada as a “third type of accent similar to the Keihan type”, it should rather be considered a variant of the Churin type.
3 Detailed Description of the Prosodic System
3.1 Prosodic Patterns of Verbs and Adjectives
The number of tonal contrasts in verbs and adjectives in Narada is generally less than that in nouns except for compound words, as discussed in § 3.2. This aspect is similar to other Japanese varieties of multipattern systems. Although verbs and adjectives undergo various inflections, this discussion solely focuses on the patterns of non-past (basic) and past forms. Word length is indicated by the number of morae in the non-past form.
Verbs in Narada are classified into two inflectional classes based on the stem-final segment, namely, those with a vowel-ending stem (V-stem) and a consonant-ending stem (C-stem). Example (11) provides instances of the non-past form ‑(r)u and the past form ‑too. Verbs are fundamentally divided into the unaccented (U) and accented (A) types. This division is based on the tonal pattern of the ‑(r)u form regardless of length and inflectional class. Type U verbs have an unaccented ‑(r)u form and an accented ‑too form with the accent on the penultimate mora. Type A verbs have an accent on the penultimate mora in the ‑(r)u form. In the ‑too form, the accent falls on the antepenultimate mora, if it is three morae long, such as mi⸢to⸣o in (11), and falls on the fourth mora from the end if it is four morae or longer. A notable exception is that some three-mora C-stem verbs carry an accent on the first mora in the ‑(r)u and ‑too forms. Only a few verbs such as arik‑ and hair‑ ‘enter’ belong to this A’ category.
|
Class |
Type |
Stem |
-(r)u |
-too |
||
|---|---|---|---|---|---|---|
|
(11) |
Tonal patterns of verbs in Narada. |
|||||
|
2- |
V |
U |
ne- ‘lie down’ |
⸢ne⸣ru |
⸢ne⸣to⸢o |
|
|
A |
mi- ‘see’ |
mi⸢ru |
mi⸢to⸣o |
|||
|
C |
U |
kik- ‘hear’ |
⸢ki⸣ku |
⸢ki⸣ito⸢o |
||
|
A |
kak- ‘write’ |
ka⸢ku |
ka⸢i⸣too |
|||
|
3- |
V |
U |
maŋe- ‘bend’ |
⸢ma⸣ŋeru |
⸢ma⸣ŋeto⸢o |
|
|
A |
tate- ‘stand, build’ |
⸢ta⸣te⸢ru |
ta⸢te⸣too |
|||
|
C |
U |
tomar- ‘stop’ |
⸢to⸣maru |
⸢to⸣maQto⸢o |
||
|
A |
ojoŋ- ‘swim’ |
⸢o⸣jo⸢ŋu |
⸢o⸣jo⸢i⸣doo |
|||
|
A’ |
arik- ‘walk’ |
a⸢ri⸣ku |
a⸢ri⸣Qtoo |
|||
|
4- |
V |
U |
narabe- ‘line up’ |
⸢na⸣raberu |
⸢na⸣rabeto⸢o |
|
|
A |
kazoi- ‘count’ |
⸢ka⸣zoi⸢ru |
⸢ka⸣zo⸢i⸣too |
|||
|
C |
U |
hiroŋar- ‘spread’ |
⸢hi⸣roŋaru |
⸢hi⸣roŋaQto⸢o |
||
|
A |
kawakas- ‘dry’ |
⸢ka⸣waka⸢su |
⸢ka⸣aka⸢i⸣too |
|||
Adjectives in Narada belong to a single inflectional class. Example (12) presents instances of the non-past form ‑i and the past form ‑kaQtoo. Similar to verbs, adjectives are also divided into unaccented (U) and accented (A) types according to their non-past tonal pattern. Only three two-mora adjectives exist, namely jo‑ “good”, na‑ “not exist”, and ko‑ “strong in color or taste”, which are of the accented type. Type U adjectives carry an accent on the antepenultimate mora in the past form. Type A adjectives have an accent on the penultimate mora in the ‑i form. In the past form, the accent typically falls on the first mora for two- and three-mora adjectives. For four-mora adjectives, the accent typically falls on the second mora. However, as demonstrated by uresi‑ in (12), the accent shifts one mora to the right when the vowel of the high-tone-bearing mora is devoiced, as in [uɭeɕi̥kattoː].
|
Type |
Stem |
-i |
-kaQtoo |
||
|---|---|---|---|---|---|
|
(12) |
Tonal patterns of adjectives in Narada |
||||
|
2- |
A |
jo- ‘good’ |
jo⸢i |
jo⸢ka⸣Qtoo |
|
|
3- |
U |
aka- ‘red’ |
⸢a⸣kai |
⸢a⸣kakaQ⸢to⸣o |
|
|
A |
siro- ‘white’ |
siro⸢i |
si⸢ro⸣kaQtoo |
||
|
4- |
U |
akari- ‘bright’ |
⸢a⸣karii |
⸢a⸣karikaQ⸢to⸣o |
|
|
A |
urusa- ‘noisy’ |
⸢u⸣rusa⸢i |
⸢u⸣ru⸢sa⸣kaQtoo |
||
|
uresi- ‘glad’ |
⸢u⸣resi⸢i |
⸢u⸣resi⸢ka⸣Qtoo |
|||
3.2 Prosodic Rules for Compound Verbs
The prosodic rules for compound verbs are left-dominant but right-dominant for compound nouns (see § 2.7).
Examples of compound verbs are given in (13). The prosodic information in the final elements does not influence the accentuation of compound verbs. Instead, the accentedness of the initial elements determines the accentuation. These observations indicate the left dominance of the prosodic compound rules for verbs. Specifically, if the initial element is unaccented (i.e., Groups U + U and U + A in (13)), then the accent falls on the penultimate mora of the entire compound verb. Conversely, if the initial element is accented (i.e., Groups A + U and A + A in (13)), then the accent falls within the initial element on its penultimate mora if it is polymoraic. However, if it is monomoraic, then the accent falls on the compound-final mora. These instances are marked with A+ in the right column of (13). Exceptions are cases in which the initial element is short. A few compound verbs exhibit the accent on the penultimate mora of the entire compound verb, even if the initial element is accented, which is marked with A† in (13).
|
Group |
Initial morpheme |
Final morpheme |
Compound verb |
||||
|---|---|---|---|---|---|---|---|
|
(13) |
Prosodic patterns of compound verbs |
||||||
|
U + U |
U |
/hik-u/ ‘pull’ |
U |
/nuk-u/ ‘unplug, extract’ |
A |
/hiki+nu⸢k-u/ ‘pull out’ |
|
|
U |
/jar-u/ ‘do’ |
A |
/jari-nu⸢k-u/ ‘carry out to completion’ |
||||
|
U |
/tatakaw-u/ ‘fight’ |
A |
/tatakai+nu⸢ku/ ‘fight to the end’ |
||||
|
A + U |
A |
/ki⸢r-u/ ‘cut’ |
A+ |
/ki⸢ri+nuk-u/ ‘cut out’ |
|||
|
A |
/ka⸢t-u/ ‘win’ |
A+ |
/ka⸢ci+nuk-u/ ‘win through’ |
||||
|
A |
/mamo⸢r-u/ ‘defend’ |
A+ |
/mamo⸢ri+nuk-u/ ‘defend to the end’ |
||||
|
A |
/kurusi⸢m-u/ ‘suffer’ |
A+ |
/kurisi⸢mi+nuku/ ‘suffer to the end’ |
||||
|
A |
/kaɴŋai⸢-ru/ ‘think’ |
A+ |
/kaɴŋa⸢i+nuku/ ‘think thoroughly’ |
||||
|
A |
/mi⸢-ru/ ‘see’ |
A† |
/mi+nu⸢k-u/ ‘see through’ |
||||
|
A |
/iki⸢-ru/ ‘live’ |
A† |
/iki+nu⸢ku/ ‘survive’ |
||||
|
U + A |
U |
/ki-ru/ ‘wear’ |
A |
/da⸢s-u/ ‘put out’ |
A |
/ki+da⸢s-u/ ‘wear’ |
|
|
U |
/ur-u/ ‘sell (vt)’ |
A |
/uri+da⸢s-u/ ‘put on the market’ |
||||
|
U |
/ure-ru/ ‘sell (vi)’ |
A |
/ure+da⸢s-u/ ‘begin to sell’ |
||||
|
A + A |
A |
/de⸢-ru/ ‘go out’ |
A+ |
/de⸢+das-u/ ‘begin to appear’ |
|||
|
A |
/ha⸢k-u/ ‘vomit’ |
A+ |
/ha⸢ki+das-u/ ‘spit out’ |
||||
|
A |
/niŋe⸢-ru/ ‘escape’ |
A+ |
/ni⸢ŋe+das-u/ ‘run away’ |
||||
|
A |
/ku⸢-ru/ ‘come’ |
A† |
/ki+da⸢s-u/ ‘begin to come’ |
||||
|
† Exceptions to the prosodic rule of compound verbs |
|||||||
4 Phrase-Level Prosody
4.1 Phrase-Initial Tone
As discussed in § 2.2 and § 2.6, words positioned at the beginning of a prosodic phrase typically exhibit a high pitch on the first mora unless an accent is present on that mora. However, this phrase-level H-tone is not always clearly observed, even in phrase-initial words. Although the precise conditions and influencing factors are not fully described, it is plausible that the distance between the phrase-initial mora and the accented mora plays a significant role in the realization of the phrase-initial H tone.
Figures 3.2 and 3.3 show the pitch contours of sentences (14a) and (14b), respectively. The noun /mimi⸢/ ‘ear’ in (14a) and the noun /warubi⸢/ ‘bracken’ in (14b) bear accents on the second and third morae. In both sentences, the initial syntactic phrases, /mimi⸢=ŋa/ and /warubi⸢=ŋa/, are in focus. In (14a), the proximity between the phrase-initial mora and the accented mora results in a minimal excursion of the phrase-initial H, rendering it nearly imperceptible. Conversely, in (14b), the greater distance between the phrase-initial mora and the accented mora allows for a more pronounced excursion of the phrase-initial H, making it notably more prominent than in (14a).
(14)
a.
mimi⸢=ŋa
tu⸣beta-i
ear=nom
cold-npst
‘My ears are cold.’
b.
⸢wa⸣rubi⸢=ŋa⸣
to⸢re⸣-too=⸢jo
bracken=nom
get-pst=sfp
‘I got some bracken.’
4.2 Interrogative Intonation
Although Narada possesses an interrogative sentence-final particle =ka, it is possible to form interrogative sentences without it. Significantly, interrogatives lacking the marker =ka do not exhibit marked intonation contours at the end of the utterance. Rising or falling boundary pitch movements (BPM s; Venditti et al. 2008: 471) are absent in interrogatives in Narada.
Examples of the accented verb /mi⸢-ru/ ‘watch-npst’ are provided in (15), and examples of the unaccented verb /ne-ru/ ‘go.to.bed-npst’ are shown in (16). In each case, (a) represents wh-interrogatives (WHQ), (b) yes/no interrogatives (YNQ), and (c) declaratives. The rise in the final mora ru in (15) is due to the word’s accent, not the rising BPM. In all instances of (16), a rising BPM on the utterance-final mora ru is ungrammatical.
(15)
a.
o⸢i⸣sjaa
i⸢ku⸣-zi=ni
mi⸢-ru
2.top
what-time=dat
watch-npst
‘What time do you watch it?’
b.
o⸢i⸣sjaa
zju⸢u⸣-zi=ni
mi⸢-ru
2.top
ten-time=dat
watch-npst
‘Do you watch it at 10:00?’
c.
⸢o⸣rjaa
zju⸢u⸣-zi=ni
mi⸢-ru
1.top
ten-time=dat
watch-npst
‘I watch it at 10:00.’
(16)
a.
o⸢i⸣sjaa
i⸢ku⸣-zi=ni
ne-ru
2.top
what-time=dat
go to bed-npst
‘What time do you go to bed?’
b.
o⸢i⸣sjaa
zju⸢u⸣-zi=ni
ne-ru
2.top
ten-time=dat
go to bed-npst
‘Do you go to bed at 10:00?’
c.
⸢o⸣rjaa
zju⸢u⸣-zi=ni
ne-ru
1.top
ten-time=dat
go to bed-npst
‘I go to bed at 10:00.’
Figures 3.4 and 3.5 show the pitch contours of (15b) and (15c), respectively. In both sentences, the pitch rise due to the accent on the focused /zju⸢u⸣-zi=ni/ ‘ten-time=dat’ exhibits an expanded pitch range. The rise from the accent on the verb /mi⸢-ru/ ‘watch-npst’ has a suppressed range. Neither shows any BPM. These observations suggest that the yes/no interrogative in (15b) and declarative in (15c) are neither morphosyntactically distinguished nor distinguished by boundary prosodic patterns. Instead, the distinction may be pragmatically supplemented by factors such as the person of the subject5 or other prosodic features, such as the pitch range of the focus phrase.
Figure 3.4
Pitch contour of yes/no interrogative (15b)
Acknowledgements
This research was supported by JSPS KAKENHI Grant Numbers JP25H00472, JP23H00007, JP23H00634, JP21H04351, JP20H00015, JP17K02777; and the NINJAL collaborative research projects “Research on the Conservation of Endangered Languages” and “Evidence-based Study on the Intonational Diversity of Japanese and Ryukyuan”. I am deeply indebted to the consultants of Narada, Kiichi Fukasawa, Fukuo Fukasawa, Hanae Fukasawa, Isao Fukasawa, and Minoru Fukasawa, for their patient collaboration.
The content of this section is based on Konishi et al. (2022: § 2.1, § 2.2) with a few modifications.
The genitive particle =no in Tokyo Japanese behaves similarly. Tokyo Japanese has a word-prosodic system with a lowering kernel (see § 2.3 of Chapter 1, this volume).
In several examples in (6), such as in biri⸢i⸣=ŋa, the phrase-initial H tone is not realized. See § 4.1.
Generalizing words with final elements of one or two morae is challenging due to their numerous exceptions; thus, their description is omitted.
In (15a, b) and (16a, b), the researcher presented the speaker with sentences in Standard Japanese that omitted the subject, such as nan-zi=ni/{mi-ru/ne-ru}=no and zjuu-zi=ni/{mi-ru/ne-ru}=no. The speaker then translated these into Narada, spontaneously supplying a second-person subject. In cases where morphological and prosodic means to distinguish between YNQ s and declaratives are lacking, the supplied subject may aid in interpreting the sentence. However, this requires verification through examples in discourse data and other investigations.
References
Hirayama, Teruo (1955) Gengotō no onchō taikei seiritsu to sono kaishaku [The formation and interpretation of the prosodic system in a language island]. Kokugo to Kokubungaku 32(12): 44–57.
Hyman, Larry M. (2006) Word-prosodic typology. Phonology 23(2): 225–257.
Inagaki, Masayuki (1957) Narada hōgen no akusento [The accent system of the Narada dialect]. In Masayuki Inagaki, Shigeo Shimizu, and Masashi Fukasawa (eds.) Narada no hōgen [The Narada dialect], 33–65. Kofu: Yamanashi Minzoku no Kai.
Kindaichi, Haruhiko (1977) Akusento no bunpu to hensen [The distribution and changes of accent systems]. In Iwanami kōza Nihongo 11: hōgen [Iwanami series: Japanese language, vol. 11: Dialects], 129–180. Tokyo: Iwanami Shoten.
Konishi, Izumi (2001) Narada hōgen akusento no genzai: Tōkyō-shiki akusento no shūtoku heiyō to jūrai akusento no henka [The current accent of the Narada dialect: Acquisition and co-usage of the Tokyo-type accent system and changes in the traditional accent system]. Jinbun Gakuho 320: 25–40.
Konishi, Izumi, Yosuke Miki and Noriko Yoshida (2022) Narada [Narada]. In Kenan Celik, Nobuko Kibe, Yosuke Igarashi, Hayato Aoi and Hajime Oshima (eds.) Nihon no shōmetsu kiki gengo hōgen no bunpō kijutsu [Grammatical descriptions of endangered languages and dialects of Japan], 77–150. Tokyo: NINJAL.
McCawley, James D. (1978) What is a tone language? In Victoria A. Fromkin (ed.) Tone: A linguistic survey, 113–131. New York: Academic Press.
Mochizuki, Nobuhiko (1951) Yamanashi-ken Narada buraku no akusento no kenkyū: tōzai akusento to hikaku shite [A study of the accent system in Narada village, Yamanashi prefecture: comparison with the accent systems in Eastern and Western Japanese]. In Kishio Terakawa, Haruhiko Kindaichi, and Masayuki Inagaki (eds.) Kokugo akusento ronsō [Studies on Japanese accentuation], 451–484. Tokyo: Hosei Daigaku Shuppan-kyoku.
Tojo, Misao (1954) Nihon hōgengaku [Japanese dialectology]. Tokyo: Yoshikawa Kobunkan.
Uwano, Zendo (1975) Akusento-so no benbetsu tokuchō [Distinctive features of prosodemes]. Gengo no Kagaku 6: 23–84.
Uwano, Zendo (1976) Narada akusento-so no shozoku goi [Lexical items of the tonal classes of Narada]. Bunkei Ronso: Bungaku-hen 11: 1–32.
Uwano, Zendo (1984) Niigata-ken Murakami hōgen no akusento [The accent of the Murakami dialect in Niigata Prefecture]. In Kindaichi Haruhiko Koki Kinen Ronshu Henshu Iinkai (ed.) Kindaichi Haruhiko hakase koki kinen ronbunshū 2: Gengogaku hen [Festschrift for the 70th birthday of Dr. Haruhiko Kindaichi 2: Linguistics], 347–390. Tokyo: Sanseido.
Uwano, Zendo (1989) Nihongo no akusento [Accent in Japanese]. In Miyoko Sugito (ed.) Kōza Nihongo to Nihongo kyōiku 2: Nihongo no onsei on’in jō [Series on Japanese language and Japanese language education 2: Japanese phonetics and phonology volume 1], 178–205. Tokyo: Meiji Shoin.
Uwano, Zendo (1999) Classification of Japanese accent systems. In Shigeki Kaji (ed.) Proceedings of symposium cross-linguistic studies of tonal phenomena: tonogenesis, typology, and related topics, 151–178. Tokyo: ILCAA.
Uwano, Zendo (2004) On no kōzō [The structure of speech sounds]. In Kiyozo Kazama, Zendo Uwano, Kazuto Matsumura, and Ken Machida (eds.) Gengogaku [Linguistics], 195–250. Tokyo: Tokyo Daigaku Shuppankai.
Uwano, Zendo (2011) “Agekaku” no yurai: Narada akusento no seiritsu katei [The origin of the “raising kernel”: The formation process of the accent system of Narada]. In Rikiji Sakazume (ed.) Gengo henka no bunseki to riron [The analysis and theory of language change], 599–614. Tokyo: Ohu.
Uwano, Zendo (2012) Three types of accent kernels in Japanese. Lingua 122(13): 1415–1440.
Uwano, Zendo (2019) Tokushuhaku no shomondai [Some issues related to moraic phonemes]. On’in Kenkyu 20: 129–140.
Venditti, Jennifer J., Kikuo Maekawa, and Mary E. Beckman (2008) Prominence marking in the Japanese intonation system. In Shigeru Miyagawa and Mamoru Saito (eds.) Handbook of Japanese linguistics, 458–514. Oxford: Oxford University Press.