Laura J. Downing, ZAS

The defining hallmark of the process of reduplication is that the Base and the reduplicated portion of the word should be as similar as possible. As a result, one would expect the tonal information of the base to be copied as faithfully and regularly as the segmental information. However, work on reduplication in various African languages shows a wide variety of correlations between the tone of the reduplicant and the tone of the Base. In some languages, like Yoruba, reduplicants can have their own morphological tone and so do not copy the tone of the Base. As shown in (1), the reduplicative prefix of deverbal nouns in Yoruba always has a high tone, no matter what the tone of the base verb is:

(1) Yoruba deverbal noun reduplication (from Pulleyblank 1988, p 265, fig (61))

At the other extreme are Bantu languages like Chichewa (at least the dialect called ‘Chichewa-Al’ in Hyman & Mtenje 1999), in which tone is regularly copied in reduplicating verb stems (and nouns (Myers & Carleton 1996), as shown in (2):

(2) Chichewa verbal reduplication (from Hyman & Mtenje 1999, p 108, fig (34); the tone patterns given abstract away from some tone processes for clarity of exposition)

However, one also finds numerous Bantu languages like Kinande (Mutaka & Hyman 1990) and Runyankore (Poletto 1998), where the tone is not copied. As shown by the Runyankore data in (3), in these languages, only one high tone is realized within the reduplicant-base complex, and it is realized on the same syllable as in the non-reduplicated form:

(3) Runyankore verbal reduplication (Poletto 1998, p 168, fig (5.4))

A final complication is found in a number of Southern Bantu languages, like some dialects of Chichewa (Chichewa-Sam in Hyman & Mtenje 1999; the dialect analyzed in Myers & Carleton 1996), Isixhosa (Cassimjee 1994) SiSwati (Downing 1994) and Shona (Hewitt 1992). In these languages, the realization of the tone in the reduplicant-Base complex depends on the length of the Base. As shown in (4a), in Chichewa, the tone is copied in bases of 3 syllables or longer, but not copied in 1-2 syllable bases. As shown in (4b), in SiSwati, tone is realized only within the Base verb stem in bases of 3 syllables or longer, but on the preceding reduplicant with 1-2 syllable bases:

(4)

(a) Chichewa tone realization in verbal reduplication (Myers & Carleton 1996, p 49, fig (25))

(b) SiSwati tone realization in verbal reduplication (Downing 1994, p 90, fig (20) and field notes)

Some previous work on tone in reduplication, like Mutaka & Hyman (1990) and Walsh (1992) propose purely phonological solutions to this problem. Tone should reduplicate if it is lexically associated to the base or associated segmentally (as proposed to prosodically). However, as Myers & Carleton (1996) argue, this approach cannot explain why tone copies in reduplicating verb stems in some Bantu languages but not in others when the representation of tone is considered to be identical in these languages. Other work (Myers & Carleton 1996, Hewitt 1992, Hyman & Mtenje 1999) has proposed a morphological explanation. Tone copies when the reduplicant and the base are in independent morphological domains for tone association, but does not when they are in the same domain. However, this approach cannot explain languages like Yoruba where reduplicants are assigned their own tone nor can it straightforwardly explain why shorter bases are more likely to be parsed into a single domain with their reduplicants than longer bases are.

In this paper, I will sketch analyses of the cases illustrated in (1) – (4) above. I will propose that whether or not tone copies depends both on morphological factors, like the grammatical status of tone in the language and the morpho-prosodic parse of the reduplicative complex, and on phonological factors, like the OCP and prosodic constraints on tonal domains. The intended contributions of the paper are twofold. First, I briefly survey the problems found with tone in reduplication. Secondly, I propose that a variety of interacting factors – not just one either phonological or morphological factor – explains in a particular language why tone is (or is not) copied in reduplication.

Cassimjee, Farida. 1994. Isixhosa tonology: An Optimal Domains Theory Analysis. ms., University of Illinois at Urbana-Champaign.

Downing, Laura J. 1994. SiSwati verbal reduplication and the theory of Generalized Alignment. Proceedings of NELS 24, 81-95.

Hewitt, Mark S. 1992. Vertical maximization and metrical theory. PhD dissertation, Brandeis University.

Hyman, Larry M. & Al Mtenje. 1999. Prosodic Morphology and tone: the case of Chichewa. In René Kager, Harry van der Hulst and Wim Zonneveld, eds. The Prosody-Morphology Interface. Cambridge: Cambridge University Press, 90-133.

Myers, Scott & Troi Carleton. 1996. Tonal transfer in Chichewa. Phonology 13, 39-72.

Mutaka, Ngessimo & Larry M. Hyman. 1990 Syllables and morpheme integrity in Kinande reduplication. Phonology 7, 73-119.

Poletto, Robert E. 1998. Topics in Runyankore Phonology. PhD dissertation, The Ohio State University.

Pulleyblank, Douglas. 1988. Vocalic underspecification in Yoruba. Linguistic Inquiry 19, 233-270.

Walsh, Laura. 1992. Tone in reduplication. Proceedings of CLS 28, 543-553.