Language: Rombo (E623C)

Contributor: Daisuke Shinagawa

Preliminary

Phonemic inventory

Vowel: /i, e, a, o, u/
Consonant:

• Plosive: /p, t, ʈ [ʈ~tɕ~tl], k; b, d, ɟ, g/
• Fricative: /f, s, ʃ, h; v [v~β]/
• Nasal: /m, n, ɲ, ŋ; m̩, ŋ̍/
• Liquid: /r, ɾ, l/
• Glide: /j, w/

Part A. Types of tone

1. Is your language tonal (i.e. pitch-based minimal pairs exist)?

1. yes
2. no
3. only grammatical tone is attested (lexical contrast is not expressed by tone)
4. I'm not sure

Notes

Lexical contrast of tone is illustrated by the following near-minimal pair.

• (1-1) a. m̩ɾíŋga wakwa (‘my water’) vs. b. m̩ɾuŋgú wakwa (‘my potato’) (Shinagawa & Apolinari 2015: 16)
• (1-2) a. iʈema (‘to cultivcate (inf.)’) vs. b. iʈeʈa [H] (‘to talk (inf.)’) (Shinagawa & Apolinari 2015: 25)

2. Is there a contrastive contour tone (i.e. phonological) in the tonal inventory?

1. no
2. yes, give examples of falling and/or rising tone (rare in Bantu languages, but common in Asian tonal languages)

Notes

Tonal contour can only be realised in the vowel sequence /V.V/, which is phonologically analyzed as a disyllabic unit in this language, i.e., there is no empirical evidence that this language has a phonemic contour tone.

• (2-1)

1. i.va.ko.ó.ʈa - ‘to draw them’
2. a.ká.a.mbá - ‘and s/he said...’

3. How many phonemic contrasts in tone? Provide tonal (near-)minimal pairs if possible. Count phonological toneless as one category.

1. no contrast
2. 2
3. 3
4. more than 4
5. I don't know

Notes

As shown in parameter 5, at least 4 distinctive levels of tonal height can be phonetically distinguished, i.e., in addition to [H] and [L] as a default realization of /H/ and /Ø/ respectively, upstepped H [ꜛH] and downstepped H [ꜜH] are identified.

• (3-1)

1. m̩di ʃú ꜛúiuwa - ‘This tree is falling’
2. m̩saŋgi úꜜlá - ‘That young person’

However, these two tones seem not to be phonemic as their realization can appropriately be explained by tone rules. See notes in the parameter 5 for further description.

3-1. What is the underlying tonal system of your language?

1. /H/ vs. /ø/ (H is the privative tone)
2. /L/ vs. /ø/ (L is the privative tone)
3. /H/ vs. /L/
4. /H/ vs. /L/ vs. /ø/
5. Other types?

Notes

As in many Bantu languages, only /H/ can be a target of tone rules such as tonal shift and tonal reduction. In the following examples, the high tone underlyingly assigned to the 1PL object marker dú- is realised on the next TBU when followed by the toneless verb stem -loli- (tonal shift), whereas the same high tone is deleted when followed by a high-toned verb -káb- due to the restriction of high tone sequence (tonal reduction). On the other hand, no evidence that low tone is a target of any tonal rule is found so far, i.e., there is no empirical ground of a marked /L/ tone in this language.

As in many Bantu languages, only /H/ can be a target of tone rules such as tonal shift and tonal reduction. In the following examples, the high tone underlyingly assigned to the 1PL object marker dú- is realised on the next TBU when followed by the toneless verb stem -loli- (tonal shift), whereas the same high tone is deleted when followed by a high-toned verb -káb- due to the restriction of high tone sequence (tonal reduction). On the other hand, no evidence that low tone is a target of any tonal rule is found so far, i.e., there is no empirical ground of a marked /L/ tone in this language.

4. What is the tone bearing unit (TBU)? → What are the tonal realization of a lexical CVVC sequence?

1. mora/Contour realization (either CV́VC or CVV́C) is possible
2. syllable/Only level tones (CV́V́C [CV́ːC], CVVC [CVːC]) are allowed
3. mora and syllable, both units can be a TBU Neither phonemic long vowel (including diphthongs) nor vowel sequence is observed.
4. not sure

Notes

As shown in the notes on parameter 2, this language allows phonetic contour in the vowel sequence, e.g., CVV́Cː ikoóʈa ‘to draw’, and CV́VCː akáambá ‘and s/he said...’. However, as these vowel sequences can be phonologically analyzed as a disyllabic unit, the TBU of this language seems to be regarded as a syllable rather than a mora. This interpretation can be supported by the fact that when high tone shift occurs, an underlying H is regularly realised on one syllable right from the original locus, e.g., the final H tone on the verb akáambá is lexically assigned to the root -ámb- and the high tone realises on the next syllable, not on the next mora (i.e., *akáaḿba).

5. How many phonetic realization of tone (i.e. number of pitch levels and number of pitch contour)?

1. identical to the number of phonemic tone
2. describe additional tones that are not phonemic (For example, Xitsonga has a contour tone or a downstepped H tone in specific contexts)

Notes

As provided in the notes on parameter 3, two allotones of phonemic H are identified in this language, namely upstepped H and downstepped H. Upstepped H is typically attested at the syntactic boundary between a verb and any preverbal element. Specifically, the focus marking high tone, which is associated to the initial TBU of the verb, is realised as an upstepped high when preceded by a preverbal constituent with a lexical high tone on its final syllable. On the other hand, the downstepped high is frequently observed in the NP domain, where underlying HH sequence at the phrase final position is regularly realised as HꜜH.

Upstep creationː H→ꜛH / H] VP[_

Downstep creation (MR): H→ꜜH / nominal stem[H_]#

Part B. Functions of tone

Note: Please add examples wherever available

6. Does the language have lexical contrasts of tone?

1. Yes -> 6-1, 6-2
2. No, the language only exhibits tonal melody patterns.

6-1. How many tonal patterns are attested for bisyllabic nominal stems?

1. geometric: e.g. 2n (HH, HL, LL, LH)
2. arithmetic: e.g. n+1 (e.g. LL, LH, HL, *HH)
3. fixed: e.g. 2
4. no contrast

Notes

As shown in Table 6-1-1, all possible combinations of /H/ and /Ø/ except for /ØØ/ [LL] sequence are attested in disyllabic nominal stems.

However, in the trisyllabic stems, possible tonal patters are restricted in a way that surface realization of HH or LL sequence is basically disallowed (except for LHꜜH).

6-2 How many tonal patterns are attested for verbal roots?

1. no contrast
2. 2 patterns (H-verb vs. L-verb)
3. 3 or more (two distinct patterns of H verbs, e.g. Shambaa)

Notes

As in many Bantu languages, and in the reconstructed Proto Bantu lexicon, the contrast of lexical tone on the verb is manifested solely on the initial syllable of the stem. As shown in the Table 6-2-1, the lexical high tone is realised on the syllable next to the initial syllable, whereas a high tone assigned to the OM is realised on the initial syllable of a tone-less verb stem. It should also be noted that the tonal property of the verb seems to be at least partially relevant to the syllable structure of the stem in a way that no tone-less verb is attested in the class of -CVVC- stems.

Table 6-2-1. Tonal property of the verb stem and its syllable structure

7. Are there grammatical categories that are solely marked by tone? e.g. SM2SG. ù- vs. SM1 ú- in Xitsonga

1. Yes -> 7-1~7
2. No

7-1. Possession marking

n.a.

7-2. Person/number marker (subject marker and object marker)

Notes

The 2SG OM and the reflexive marker are segmentally identical but tonally distinctive, i.e., 2SG ku- vs. REF kú-. Thus, the other things being equal, a pair of the same transitive verb with the 2SG OM and the reflexive marker is solely differentiated by tone.

7-3. Verbal extension

n.a.

7-4. TAM marking

n.a.

7-5. Polarity marking

Notes

Main clause negation is morphologically marked by the clause final negation particle kú, which is estimated to be cognate with the class 17 locative demonstrative pronoun. In addition to the particle, tonal modification of is also relevant to the main clause negation marking. As shown in (7-5), the initial high tone that is associated with sentence affirmation (tentatively glossed as FOC in (7-5a)) is absent in the negative form and the underlying high tone assigned to kú is spread left-ward to the TAM le- (if a stem has a lexical high tone, the spread ends in the nearest tone-less TBU, i.e., the second syllable of the stem).

7-6. Focus marking (including CJ/DJ)

Notes

As mentioned in the notes on parameter 5, the focus marking clitic ní= has a lexical high tone and the segmental ni= frequently drops, resulting that the focus, more precisely saying ‘main clause affirmation’ as part of truth-value focus, is solely marked by the initial high tone. In contrast, the lack of the initial high tone on the verb defectively signals that the form is a non-main clause verb. This is illustrated in (7-6-1), where the main clause verb dúlem̩lolia is realised with the initial high tone, whereas the relative verb eledusaidja lacks the high tone.

7-7. Other functions

n.a. (so far)

Appendix

Pitch track samples of the possible nominal tone patterns of trisyllabic words