Grammatical tone in Kanise Khumi: exclusive/inclusive pronominal agreement Elissa Ikeda & Ryan Gehrmann Payap University, Chiang Mai, Thailand elissa_i@payap.ac.th, ryan_g@payap.ac.th ABSTRACT Kuki-Chin languages are known for using tonal variation to mark grammatical relationships [1]–[3]. This paper provides evidence that tone is used to mark an inclusive-exclusive distinction in pronominal subject agreement markers on Kanise Khumi verbs. An inclusive-exclusive distinction is common for Kuki-Chin independent pronouns, but not for pronominal agreement marking on the verbal complex [4], [5]. Using elicited sentences, this paper compares exclusive and inclusive agreement prefixes with respect to vowel quality (F1) and pitch (F0). When the object is 3 rd person, segmental information (i.e. vowel quality) is one feature that distinguishes exclusive from inclusive in the person- marking prefix. Pitch is another. However, in reflexive/reciprocal interpretations, pitch is the distinguishing feature between exclusive and inclusive agreement prefixes. Thus, exclusive- inclusive pronominal agreement provides an example of morphosyntactic tone in Kanise Khumi. Keywords: grammatical tone, pronominal agreement, Tibeto-Burman, Kuki-Chin, Khumi. 1. INTRODUCTION As Hyman [6] pointed out, tone can do everything that segments can do and more. In Kuki-Chin languages, tone can mark a variety of grammatical relationships such as genitive, benefactive, causative, number, negation, and nominalization to name a few. Peterson [7] argues that to his knowledge, “no other language of the area is reported to exhibit as widespread use of tone for marking morphosyntactic information as [Bangladesh] Khumi does.” This paper provides evidence that tone is used to mark an inclusive-exclusive distinction in pronominal subject agreement markers on Kanise Khumi verbs. Kanise is a Khumi variety spoken mainly in Paletwa Township, Chin State, Myanmar. Kanise Khumi distinguishes between exclusive and inclusive using both independent pronouns and agreement prefixes on the verb. For first-person reference, exclusive indicates that the speaker is included but the addressee is excluded. In contrast, an inclusive pronominal includes both the speaker and addressee. Table 1 displays the clusivity distinction in independent pronominals in Kanise Khumi. Sing. Dual Plural Exclusive kakaḭ˧ katsi ̤ि Inclusive ʔaḭ˧ ʔatsi ̤ि Table 1: First-person independent pronominals in Kanise Khumi. Kuki-Chin languages may have bound pronominal agreement prefixes on verbs that mark subject and object [4]. Among Khomic (Southwestern) Kuki- Chin languages, Kanise Khumi appears to have the most complex pronominal agreement marking system (cf. [8]–[11]). Prefixes agree with subject person (See Table 2). Subject number is marked through suffixes: dual suffix /hɔe˥/, plural suffix /u˧/, and singular -ø. Object Subject 1 excl 1 incl 2 3 1 n̩˥ aŋ˥ 2 kaŋ˥ kaŋ˥ aŋ˥ 3 or Intransitive kə˥ ka˧ n̩˥ ə˥ Reflex./Recipr. kka˧ ka˧ Table 2: Person agreement prefixes in Kanise Khumi. Three lexical tones have been observed on noun and verb roots in isolation. Tone 1 has high level pitch and modal voice quality. Tone 2 is a mid checked tone (short duration & glottalized voice quality). Tone 3 is low-falling with breathy voice quality [12], [13]. A consistent association between one tone on the inclusive prefix and a different tone on the exclusive prefix would provide evidence that tone is used to distinguish this grammatical relation [14]. However, inconsistent variation of pitch on any one of the first- person subject prefixes might indicate that 1) the prefixes are simply not specified for tone, 2) the tone of the verb root might affect the tone of the agreement prefix, or 3) that a full syllable prefix might vary freely with a reduced syllable [15], [16]. In the final case, we might expect other features of syllable reduction to correlate (i.e., neutral or mid-central vowel quality, shorter duration, gradient realizations of pitch [17]–[21]).