CHAPTER3 Ur-Bantu Consonants in Pai and Swati
The aim in this section is two-fold: firstly it is an attempt to establish the position and status of Pai and Swati in relation to Ur-Bantu. Ur-Bantu consonants will be used as a point of departure to illustrate the sound-shifts in Pai and Swati consonants. This will be done in order to get a better picture of the position of Pai in relation to that of Swati. The reason why Meinhof is used and not Guthrie, is mainly because Meinhof treats both Zulu and Pedi (Northern Sotho) in detail. Swati (Tekela) is one of the two target languages in this study. Adapting Meinhofs Zulu to Swati was more convenient as both languages are familiar to me. The work done by Ziervogel (1954) is also based on Meinhof and since Ziervogel is one of the main sources of reference, it was only logical to take Meinhof as a basis for comparison.
All the information will be presented in tables, with Meinhofs Ur-Bantu forms as a starting point. The equivalent or corresponding Pai and Swati reflexes will follow after the Ur-Bantu form. In order to facilitate this issue further, a column of the postulated Proto-Southern Bantu forms of Van der Spuy (1989) will also be included (if they are listed in his work). Furthermore, the most prominent characteristics of the Ur-Bantu, Swati and Pai consonants will be listed as distinctive features to serve as additional information in compiling a comprehensive account of the status of the speech sounds of Pai.
Secondly, the aim is to compare the Swati forms with those of Pai, as opposed to the previous section where the Pai speech sounds were described and the Swati equivalents were given. The aim is therefore to view the position of Pai from a Swati perspective. All this will be done in the following manner:
- The Ur-Bantu, Pai, Swati and PSB consonants with examples in words will be listed in a table.
- A phonetic description will follow where the obvious shifts and differences will be indicated and discussed.
- These phonetic differences will then be translated into distinctive features that will represent a summary of the Ur-Bantu, Pai and Swati data. The features will be used to high-light the most obvious characteristics of the two languages. These features will be representative of works by Hyman (1975), Ladefoged (1982), Lass (1989) and Sloat (1978). As will become evident, mainly four features will be used as a basis:
- a place of articulation feature
- [±continuant] and
- [±nasal] where nasals are described.
While it is understood that the proto-consonants are only symbolic representations, these four features have been chosen and assigned to the Ur-Bantu forms because they seem to capture the essential characteristics of the Ur-Bantu forms as well as the differences between Pai and Swati.
The feature [±continuant] will be applied mainly to distinguish between fricative and non-fricative (stop consonant) sounds. A few additional features that are considered to be important in this investigation will be used to distinguish Pai and Swati phonemes and these are:
- [±aspiration] ([±asp])
- [±glottalic] ([±glott]) for ejectives
- [±delayed release] ([±del rel]) to distinguish between affricative and non-affricative sounds
- [±egressive] ([±egr]) to account for the implosive and clicks and
- [±lateral] ([±lat])
- Notes and comments on the nature of each entry will follow where it is deemed necessary.
- Finally the Swati form (reflex) with its Pai equivalent will be listed. This information will later be used for comparative purposes.
- The same abbreviations and conventions as described in the introduction to the speech sounds of Pai will be used (Chapter 2). Examples of Pai and Swati speech sounds within a text will be in bold IP A without any brackets. Phoneme slashes will only be used when the intention is to mark a sound as a phoneme and phonetic brackets to mark a sound as a phone or an allophone. Meinhof s Ur-Bantu orthography (italics) will be followed as closely as possible with the following exceptions:
- Bold italics will be used.
- The Ur-Bantu. symbol v for the voiced fricative will be replaced by bold italic
- The Ur-Bantu semi-vowel y will be replaced by bold In the IP A transcription this semi-vowel will of course be j.
Van der Spuy’s PSB forms will be represented as a starred form in bold in his own transcription.
In the final sections of this chapter where the influence on the preceding consonants of the B. close vowels is listed, only the differences will be indicated and references will be made to what has been treated before. No comments are deemed necessary since the data is self-evident and is included only for the sake of further research.
The Primary Plosives
The primary plosives will first be treated where they are followed by B. a or other open vowels in order to eliminate the possible influence of the high and close vowels.
From the above it is clear that Pai follows the typical Sotho shift of B. voiceless velar plosive to NS voiceless velar fricative (Meinhof, 1932:48), and Swati follows the Zunda shift to aspirated velar plosive. Van der Spuy (1989:35) postulates PSB *k but also includes examples of stem-medial attestations of this sound in words that contain the partially-voiced [Is] of Swati and Zulu, eg in the verbal extensions -ek- and -uk-. Meinhof (1932:83) treats this sound in the same manner and cites the voiced « g » in the extensions as one of the reflexes of B. k, but says that the others occurring in stems are exceptions, or the result of assimilation or dissimilation (1932:98-99).
In the literature the partially-voiced [ls] of Tekela and Zulu has not been sufficiently recognised as a completely different phoneme from the « g » of Meinhof or the breathy voiced (9] or [g] that supposedly derived from the disappearance of a preceding nasal. Nor has this sound, to my knowledge, been adequately described phonetically. Descriptions encountered in a number of textbooks and study guides, range from « voiced, velar plosive » to « partially voiced, velar plosive » to « voiced, velar implosive ». Therefore the classification and description of Meinhof will have to be accepted.
The ra and la often interchange in stem-medial positions and this seems to be a common variation in Pai, as in -apara and -apala. On the other hand, ra and ~a also interchange, but only in stem-initial positions while ~a and la never interchange. This possibly indicates that there are at least two distinct phonemes at work in the first two attestations above, viz /r/ and hj; phone [I) as a variant of Ir/ will therefore not be included in this description. The occurrence of the retroflex [tl in Pai may be ascribed to the presence of a preceding nasal in the proto-fonn but it will nevertheless be included in this description.
The shift in Pai is multifold: from B. voiceless alveolar plosive, to Pai
- voiced alveolar trill
- voiced retroflex fricative
- aspirated retroflex
The shift to Swati is consistent, ie to an ejective alveolar affricate that may have the variant [tf’] or [t<f>’] when the following sound is a back vowel or a labial semivowel (cf comments below, par 184.108.40.206 onwards).
From a B./Swati perspective, the differences in places of articulation in the Pai attestations (alveolar and palatal) do not seem to be the significant feature since it is the retroflexive nature of the sounds ~ and th that really distinguishes them from ordinary alveolar or post-alveolar sounds. However, from the same perspective, the features [±voice] and [±cont] in the Pai attestations certainly are significant. The phonemes /r/ and l’IJ are both voiced and continuant and stand in stark contrast to the voiceless stops of B. and Swati. While the trill Ir/ is the common reflex in NS and Tsonga and also appears in Pai, there is no apparent reason why the voiced hJ of Pai exists alongside the voiceless plosive [th] in similar environments, even if the presence of a preceding nasal is taken into consideration.
The aspirated retroflexive plosive [th] of Pai is more in line with the Zunda reflex /th/ than with the Swati affricate [ts’]. From a historical point of view it is significant that it does appear in Pai, even more so when one takes into consideration that the NS and Tsonga reflex of B. tis Ir/ (Van der Spuy, 1989:30). If the observation that Pai is more closely related to NS than to Swati or Tsonga (Ziervogel, 1954; Van Warmelo, 1935) is taken as valid, then the phoneme Ir/ should have been the reflex in Pai. Furthermore, the appearance of [th] in a common gloss such as ‘three’, may indicate an origin that is more Proto-Nguni than Proto-Sotho. Cf also the references to breathy voicing in nasal compounds, par 220.127.116.11.
As was mentioned above, the Swati [ts’] may alternate with [tf], depending on the nature of the following vowel. These affricates are sometimes also aspirated: [tsh] [tfh], but the reason why or the position where this occurs in contrast to the ejective ones, have not been established beyond any doubt. It seems that, contrary to widely accepted views by Ziervogel (1952 & 1976) and Taljaard (1991), the majority of these attestations are likely to be aspirated and not ejective.
CHAPTER 1: Aim, Background, History anti General Introtluction
1.1 Aim and Exposition
1.2 Present Situation of Pai
1.3 The Research Material
1.4 Historical Background of the Pai
1.5 General Introduction
CHAPTER 2: The Sound System of Pai
CHAPTER 3: Ur-Bantu Co11.\·01umts in Pai anti Swati
3.2 The Primary Plosives
3.3 The Primary Fricatives
3.4 The Nasals
3.5 The Voiceless Palatals
3.6 The Voiced Palatals
3.7 The Primary B. Plosives in Nasal Compounds
3.8 The Primary B. Fricatives in Nasal Compounds
3.9 The B. Palatals in Nasal Compounds
3.10 The B. Primary Plosives, Fricatives, Palatals and Nasals followed by B. i
3.11 The B. Primary Plosives, Fricatives, Palatals and Nasals followed by B. u
3.12 The B. Primary Plosives, Fricatives, Palatals and Nasals followed by B. Close Vowels f and 11
3.13 Consonants Preceding B. Semivowels
CHAPTER 4: Morphological Survey
4.2 The Noun Classes of Pai
4.3 The Nominal Derivatives
4.4 The Pronouns of Pai
4.5 Concordial Prefixes
4.6 Aspectual Prefixes
4.7 Verbal Suffixes
4.8 The Relative Construction
CHAPTER 5: Conclusion
5.2 The Voiced Plosives of Pai and Swati
5.3 The Voiced Fricatives of Pai and Swati
5.4 The Voiced Affricates of Pai and Swati
5.5 The Voiceless Plosives of Pai and Swati
5.6 The Voiceless Fricatives of Pai and Swati
5.7 The Voiceless Affricates of Pai and Swati
5.8 The Nasals of Pai and Swati
5.9 The Approximants of Pai and Swati
5.10 The Clicks of Pai and Swati
GET THE COMPLETE PROJECT
THE PAI LANGUAGE OF EASTERN MPUMALANGA AND ITS RELATION TO SWAT/