Tonal and Non-Tonal Intonation in Shekgalagari

Abstract

The study of intonation in a (fully) tonal language presents both a challenge and an opportunity. The challenge is to see how a language which exploits F0 mainly for the purpose of lexical and grammatical contrasts succeeds in encoding the functions often expressed by means of intonation in non-tonal languages. As is well-known, word-level tonal distinctions can be quite rich and can be subject to considerable manipulation in the phrasal phonology. Particularly when tone systems are complex in these ways, the question is how there can be much room left for intonation to modify or add pitch specifications without obscuring the word-level tonal contrasts. The goal of this paper is to examine how Shekgalagari, a Bantu language of Botswana which exploits F0 mainly for the purpose of lexical and grammatical contrasts, succeeds in encoding the functions often expressed by means of intonational phonology in non-tonal languages. Whereas other languages with lexical tone adopt different strategies for incorporating intonational pitch features, we demonstrate that most of the intonational marking is non-tonal in Shekgalagari, thereby raising the question of what is a possible intonational system.

APPENDIX: Monosyllabic Words and PLL

As was seen in (24), declarative PLL does not apply when the last word of the utterance is monosyllabic, a property which is thus far limited to Shekgalagari among the Bantu languages for which we have information. Here we consider a few more facts in order to determine how this fact might be account for.

First, it should be noted that monosyllabic words are very limited in Shekgalagari. Among the ones we have identified by independent criteria are the following:

(46)	a.	monosyllabic verbs in the imperative: k-á ‘mention!’, gy-á ‘eat!’
	b.	demonstratives: ∫ é ‘this (one)’ (cl. 7), ӡé ‘these’ (cl. 10) etc.
	c.	adverbs: ӡ wá ‘in this way’, thέ ‘really’
	d.	the preposition qá ‘with’, which, however, cannot occur finally

As seen, all of the above monosyllabic words have /H/ tone. Monosyllabic imperatives have a bisyllabic variant, which can occur with Emph PLL: i :-k-á ‘mention!’, i :-gy-á  ‘eat!’. Similarly, although monosyllabic words block declarative PLL, Emph PLL may assign length to the final vowel of the preceding word:

(47)	a.	a-bal-a qá ∫e	∼	a-bal-a qá: ∫e	‘he has just counted with this’
	b.	a-rí-bál-a ӡwá	∼	a-rí-bál-a: ӡwá	‘he has just counted them like this’

(In (47a) /qá + ∫é/ ‘with this’ becomes qá ∫ e by a rule discussed by Crane (2008, 2009a, b).)

But should utterance-level intonation, here PLL, be allowed to have access to word boundaries? If yes, monosyllabic words can block PLL by virtue of not having a penultimate syllable. If no, an alternative is needed to avoid direct reference to word boundaries.

While we suspect that intonation can know where the word boundaries are, if it were necessary to exclude them from intonational implementation, the following metrical solution would work:

(48)	a.	construct a trochaic foot over the last two syllables of each word
	b.	in case the last word is monosyllabic, the trochaic foot will have only one syllable
	c.	declarative PLL specifically targets the nucleus of the penultimate syllable (vowel or syllabic nasal) of the last foot of an utterance or pause-marked intonational phrase (IP)
	d.	Emph essentially encliticizes an IP-final monosyllabic word in which case PLL is free to target the nucleus of the penultimate syllable across the word boundary.