[Start] [Previous] [Next] [Contents] [References]


4. The structural characteristics of Zulu English

4.1. Introduction

In the preceding chapter I examined the methodology employed in collecting and analysing the data that forms the basis for the description of the key features of Zulu English phonology, prosody and syntax that I provide in this chapter. The aim of this overview of the structural characteristics of Zulu English is to demonstrate that it is not simply an agglomeration of learners' errors but that it forms a distinct variety of English with the necessary systematicity to form the basis for a restandardised South African English. Demonstrating systematicity, in Platt et al.'s (1984: 166-167) terms, entails showing that the distinctive (i.e. non-standard) features of a variety are:

  1. relatively frequent;
  2. grammatically rule-governed; and
  3. pragmatically rule-governed.

While my corpora of data are too small to prove the frequency of any of the syntactic features conclusively, the fact that many of them do occur more than once in a small body of data suggests they are regularly used by ZE speakers. In sections 4.2. and 4.3. I examine a range of phonological, prosodic and syntactic features of Zulu English and attempt to show that their occurrence is regular and rule-governed. One of the ways in which I do this is to show that the occurrence of these features can be accounted for in terms of substratum interference from Zulu, the effect of language universals, or else functional considerations.

Ideally one would also examine all the features to establish whether they satisfy the third criterion of systematicity, namely pragmatic rule-governedness. However, this was not feasible given that establishing the systematicity of BSAE forms just one part of my overall argument. For some of the features (e.g. aspects of prosody, 4.2.6.; resumptive pronouns in relative clauses, 4.3.6.) I have briefly suggested how they might serve the speaker's communicative needs. In section 4.4. I examine one feature, pronoun copy, in detail and argue that on the basis of my data it clearly does serve a topic-related discourse function and its use is rule-governed.<83>

4.2. Phonological and prosodic features of Zulu English

The admittedly limited literature (Bailey ms; Lanham 1984; Buthelezi 1989; Jacobs 1994) and my own observations suggest that BSAE, including Zulu English, differs significantly from mother-tongue varieties of South African English phonetically, phonologically, and prosodically. A consideration of the phonetic features of ZE is beyond the scope of this chapter although the phonetics of this variety undoubtedly contribute to its characteristic sound, to which social meaning may be attached.

In this section I rely on the descriptions of BSAE phonology and prosody that exist in the literature supplemented by my own judgements based on listening to the recorded data. The abbreviation BSAE/ZE is intended to alert the reader to the fact that although my observations are based exclusively on Zulu English, the literature refers to BSAE more generally. According to Bailey (ms) most subvarieties of BSAE appear to exhibit essentially similar phonological features. I have highlighted a few features of Zulu English (e.g. substitution of /r/ for /l/) which may be specific to this subvariety.

As I noted in section 2.2.3., it is widely accepted that phonological features of second- language varieties can be accounted for in terms of interference from the mother tongue: the phonological system of the second-language variety (including New Englishes) is viewed as the result of speakers' attempts to map the target variety onto the existing phonological system of their first language. Consequently the greatest divergences between the second-language variety and Standard English are expected to occur in those areas where the first language and English differ most, and particularly where the target language makes phonemic distinctions that the first language does not. In the case of Zulu and English the two principal areas of difference are the vowel systems and the phonotactics of the two languages. These are the two aspects of Zulu English phonology that I concentrate on.

4.2.1. The vowel systems of White South African English and Zulu contrasted

In general the vowel system of English presents most difficulties to Zulu speakers as the complex system of English vowels contrasts with the relatively simple Zulu vowel system. So-called 'Respectable' WSAE (one of three varieties of White South African English distinguished by Lass [1990: 272] following Lanham [1967]), is typical of varieties of <84> English originating in the south of England in being non-rhotic, having // before /f/ /s/ and /T/; and in distinguishing between short and long (often diphthongal) vowels. According to Lass (1990) Respectable WSAE, in addition to its 13 simple vowels (including the unstressed schwa /«/ and the vowels in DRESS /e/ and SQUARE //), has a further seven diphthongs:

1. FACE /eI/ 2. PRICE /aI/
3. CHOICE /?I/ 4. NEAR //
5. CURE // 6. GOAT /eU/
7. MOUTH /AU/


The IPA symbols used to represent the phonological system of WSAE are not intended to indicate the exact phonetic quality of the vowels in WSAE. The use of capitalised representative words provides a convenient way of referring to phonemic classes in WSAE and ZE (cf. Wells 1982). The vowel chart below (Figure. 1) represents the vowel system of Respectable WSAE. Those sounds only occurring as part of a diphthong sequence are in square brackets:

Vowel chart depicting the vowel system of Respectable WSAE.

Figure 1. The vowel inventory of Respectable WSAE

By contrast the vowel system of Zulu (Figure 2.), in common with those of the other indigenous South African languages, is considerably simpler: there are five simple vowels and, significantly, no central vowels or diphthongs. The mid vowels /E/ and /?/ have raised allophones [e] and [o] (Bailey ms).<85>

Vowel chart depicting vowel inventory of Zulu

Figure 2. The vowel inventory of Zulu

4.2.2. The vowel system of BSAE/ZE

According to Bailey (ms) Zulu English makes use of the same five vowel system as Zulu, including the allophonic variation in mid vowels. In accommodating the English vowel system to their own, Zulu speakers make a number of simplifications, some of which result in the loss of phonological distinctions.

Central vowels

As the chart in Figure 2. makes clear, the Zulu vowel inventory does not include central vowels. The central vowels found in WSAE are re-assigned to the five peripheral vowels in Zulu English. According to Bailey (ms) the following re-assignments are typical:

  1. In WSAE the vowel in BIT shows considerable allophonic variation and it is frequently realised as a high central vowel, [ö]. In ZE the vowel is assigned to the high front vowel /i> /. As a result the BEAD/BID distinction is lost.
  2. The tense central vowel in BIRD /Îù/, is assigned to the mid front vowel /E>/. As a result the BIRD/BED distinction is lost. Bailey (ms) notes that despite the phonetic similarity that exists between BIT /ö/ and BIRD /Îù/ in WSAE, Zulu English speakers make a clear distinction between the two. He suggests that this reflects the influence of spelling considerations on second-language pronunciation.
  3. The low central vowel in CUT, /Ã/, is somewhat lowered to the /a> / position, representing a minor shift in vowel quality.
  4. Zulu English speakers' re-assignment of the schwa, /«/, which is the usual realisation of unstressed vowels in WSAE, is not entirely predictable. Although /a/ <86> is the most common substitute, spelling seems to play a role in determining which vowel is used, for example, in my experience the second vowel of 'secretary' may be pronounced /E/ and the third /a/.

Low vowels

The low vowels in the WSAE repertory are re-assigned as follows:

  1. The low back rounded vowel in the WSAE COT class is realised as /o/ in Zulu English. Thus the distinction between COT and CAUGHT is blurred.
  2. The low back vowel in the WSAE CART class is realised as /a/. The CUT/CART distinction is thus lost. In Zulu English a clear distinction is made between COT and CART despite the fact that in many varieties of WSAE the low back vowels in COT /?/ and CART // are pronounced similarly (Lass 1990). Once again spelling seems to be implicated in Zulu English vowel assignment.
  3. Although the low front vowel occurring in WSAE CAT is most frequently realised in ZE as /E/, its behaviour is somewhat unpredictable: certain words alternate between /E/ and /a/ even within the speech of one individual while other words show a preference for /a/ over /E/, for example Bailey (ms) notes that 'Africa' is invariably pronounced /afrika/ not /Efrika/.

Vowel length

Vowel length, which is non-contrastive in Zulu, does not appear to be contrastive in BSAE/ZE either (Bailey ms; Buthelezi 1989: 47-48). While vowel length is seldom the sole distinguishing feature between two WSAE phonemes, it plays a part in distinguishing the tense (long) and lax (short) vowels. The absence of distinctive vowel length therefore contributes to loss of certain distinctions in ZE:

  1. WSAE // and /ö/ [I ~ ö] -> ZE /i> / contributes to the loss of BEAD/BID distinction.
  2. WSAE // and /e/ -> ZE /E>/ results in a loss of BARED/BED distinction.
  3. WSAE /¬ù/ and /U/ -> ZE /u>/ contributes to the loss of POOL/PULL distinction.
  4. WSAE // and /?/ -> ZE /?>/ contributes to the loss of CAUGHT/COT distinction.

Bailey (ms) notes that vowels in open syllables are usually lengthened somewhat by speakers of BSAE/ZE, e.g. CAR is typically pronounced /kaù/. <87>

Table 1. summarises the re-assignment of WSAE vowels to the BSAE/ZE vowel system:

WSAE   BSAE/ZE
BEAD i i
BID I
BARED E
BED e
BIRD Îù
BAD Q a
PART
PUT Ã
COT ? ?
CAUGHT
POOL ¬ù u
PULL U
COMMA « a i E ? u

Table 1. Summary of the re-assignment of vowel phonemes in Zulu English.

Diphthongs

The Zulu vowel system does not include diphthongs, and the status of diphthongs in Zulu English is uncertain. Bailey (ms) argues that the simplest analysis is to regard diphthongs in BSAE/ZE as being realised as sequences of the five simple vowels of BSAE/ZE. I agree that this appears to be the case with the broader diphthongs (i.e. those involving relatively large tongue movements). The following vowel sequences are used:

WSAE BSAE/ZE

HEIGHT /aI/ -> /ai/

HOUSE /AU/ -> /au/

COY /?I/ -> /?i /

BEARD // -> /iE / [ie ~ ia]

MOOR // ->/ua / <88>

The narrow diphthongs (i.e. those involving relatively small tongue movements) are, in my view, replaced by simple vowels in the speech of at least some Zulu English speakers:

  1. COAT /eU/ -> /o>/ resulting in loss of COAT/COT/CAUGHT distinction.
  2. BADE /eI/ -> /e> / resulting in the loss of BADE/BED/BIRD/ distinction.

However, other speakers do not simplify these diphthongs, making use instead of the vowel sequences /ou/ and /ei/.

My observations suggest that the WSAE vowel plus schwa sequences may be 'split' by some Zulu English speakers and semi-vowels (either /w/ or /j/) inserted between them:

HOUR /aU«/ -> /awa/ POOR // -> /uwa/

LOWER /eU«/ -> /owa/ NEAR // -> /ija/

may reflect the influence of Zulu phonotactics which does not allow the juxtaposition of vowels. Where affixation would result in such juxtaposition of vowel sequences a process of 'glide formation' applies, eliminating any vowel-vowel sequences, e.g., esonto+eni becomes esontweni (Buthelezi 1989: 56).

4.2.3. Consonantal features of BSAE/ZE

In general the consonant system of Zulu English appears to differ very little phonemically from that of WSAE (Bailey ms; Buthelezi 1989). This reflects the fact that the English consonant system can for the most part be mapped onto the Zulu system (which is considerably more complex). Those English consonant phonemes for which there is no counterpart in Zulu, and where one might anticipate differences, are:

i. The voiced and voiceless dental fricatives, /D/, /T/. Substitution of the alveolar stops /t/ and /d/ according to Jacobs (1994) is one of the characteristic features of Zulu English. However, Buthelezi (1989) and Bailey (ms) claim that BSAE speakers generally do not make such substitutions and in my experience, too, substitution is not that common. Jacobs's investigation focused on mesolectal speakers of ZE and it may be that control of /T/ and /D/ is one of the characteristics distinguishing acrolectal speakers from mesolectal and basilectal speakers.

  1. The voiceless affricate /tS/. This phoneme is sometimes pronounced as /S/ by Zulu speakers, e.g. 'chance' may be pronounced, /Sa> ns/. Jacobs's (1994) investigation <89> revealed that this substitution occurred only infrequently, although Bailey (ms) claims that this occurs often in the speech of Zulu speakers. I suggest that, as in the case of dental fricatives, this is typical of basilectal and mesolectal speakers but not acrolectal speakers. In my data the only speaker who substitutes /S/ for /tS/ is the customer (C) in transcript 4 (appendix C), who is definitely not an acrolectal speaker. The voiced affricate /dZ/ occurs in Zulu and thus does not present any difficulties to speakers.
  2. The approximants /l/ and /r/. Historically these approximants were not contrastive in Zulu, although as a result of borrowing from English and Afrikaans this distinction appears now to be phonemic (Mathonsi ms). Bailey (ms) comments that most speakers of Zulu English do appear to make this distinction. I suggest that a failure to distinguish /l/ from /r/ may be indicative of a basilectal speaker.
  3. Jacobs (1994) lists devoicing of certain obstruents, particularly /b/, /d/, /k/, /dZ/ and /z/, as features of mesolectal Zulu English. Not all of these obstruents are affected in all contexts. For example /b/ is only devoiced word finally. She argues that this leads to the loss of contrasts in pairs such as 'bat' and 'bad'. I suggest, on the basis of the analysis of my data, that this distinction may in part be a matter of perception rather than reality. It is possible that the absence of vowel lengthening in ZE before final voiced consonants contributes to the perception that ZE speakers do not distinguish such pairs. Mother-tongue speakers produce the vowel in a word such as 'bat' considerably shorter than in a word such as 'bad' and they very often rely on vowel length as much as voicing of the final consonant to distinguish such pairs. In my observation devoicing of obstruents is not noticeable in acrolectal ZE. Devoicing of stops may be more typical of the English spoken by speakers of the Sotho group of languages since voiced stops are absent or infrequent in these languages (Bailey ms).

4.2.4. Phonotactics of BSAE/ZE

Phonotactically English is more complex than Zulu (and the other indigenous South African languages) in that it allows consonant clusters that are not permissible in Zulu. Some speakers of Zulu English, in my observation, simplify these clusters or insert an epenthetic vowel to split the cluster and thus reduce its phonotactic complexity: <90>

  1. /strIN/ -> /steriN/
  2. /driùm/ -> /deri> m/

An epenthetic vowel is also frequently used in place of syllabic consonants: /n`/, /m`/ and /l`/, for example, 'bottle', which in Respectable WSAE is pronounced /b?tl`/ and in Zulu English would often be pronounced, /botel/.

4.2.5. Lectal variation in BSAE/ZE phonology

In section 3.3.3. I argue that a relatively prestigious, acrolectal subvariety of BSAE is most likely to form the basis for any future restandardised South African English, hence my decision to concentrate data collection on white-collar workers who are likely to make use of a relatively acrolectal variety. In the preceding discussion I have indicated where, in my view, a feature is more typical of mesolectal or even basilectal BSAE/ZE than an acrolect. It seems probable that most of the phonological features of BSAE/ZE that have been discussed could be arranged on a lectal continuum stretching from basilect to acrolect. Some features might be common to all BSAE/ZE speakers (except for hyperlectal speakers of StdSAE who fall outside the continuum), some features might be more frequently found in basilectal speakers and others in more acrolectal speakers. More research would be required to determine where exactly on this continuum the various phonological features I have highlighted fall, but in summary the following tentative generalisations seem plausible:

  1. Not making a distinction between /l/ and /r/ is typical only of basilectal speakers.
  2. The use of fricative /S/ in place of the affricate /tS/ and the substitution of /t/ and /d/ for /T/ and /D/ are features typical of basilectal and mesolectal speakers.
  3. Devoicing of obstruents is not typical of acrolectal speakers.
  4. Vowel and consonant epenthesis is not typical of acrolectal speakers.
  5. The use of vowel sequences is a more acrolectal strategy than the mono- phthongisation of diphthongs.
  6. The re-assignment of the 13 StdSAE vowels, especially the central vowels, to the five vowel system of Zulu is typical of most speakers. <91>

4.2.6. Prosodic characteristics of BSAE/ZE

Lanham (1984) identifies significant differences in the prosodic (or suprasegmental) systems of BSAE and StdSAE. These differences, which I consider below, form an important part of my account of the pragmatic role of pronoun copying. In section 4.4.2. I argue that the relatively frequent use of pronoun copying in ZE can, in part, be explained as compensating for the fact that the ZE prosodic system differs from the StdSAE one. Before examining Lanham's characterisation of the prosody of BSAE, it is necessary to briefly define a number of terms since these appear to be used in a variety of ways in the literature.

  1. Pitch refers to the perceptual frequency of voiced segments (Katamba 1989: 186) and depending on the domain to which pitch movements apply, intonation and tone can be distinguished.
  2. Intonation refers to pitch movement (or contours) that apply to an utterance as a whole (Katamba 1989: 239); while
  3. Tone refers to pitch movement as a lexical or morphemic property, as occurs in many languages (Katamba 1989: 186-187). In general, languages tend to give prominence to one or other use of pitch, giving rise to the distinction between tone languages, of which Zulu is an example, and intonational languages, to which group English belongs. All tone languages do, however, make limited intonational use of pitch as well (Katamba 1989: 239).
  4. Stress refers to the auditory prominence of a segment or word and is a combination of higher pitch, longer duration, greater intensity and, in the case of English, fuller vowel value (Sloat, Taylor & Hoard 1978: 71; Katamba 1989: 53). Stress change may be a feature of individual words (word stress or accent) or it may apply to utterances and act in concert with intonation, which Lanham (1984: 218) refers to as prominence.
  5. Prosody refers to the combination of intonation, prominence and other suprasegmental features.

As with phonology, the prosodic features of BSAE/ZE can largely be accounted for in terms of attempts by speakers to accommodate their mother-tongue patterns to those of the target language. For this reason I briefly contrast the prosodic systems of English <92> (based principally on Katamba 1989) and the African languages (based on Lanham's 1984 discussion).

Word stress

The complex system of variable word stress in English differs considerably from the rather limited role that it plays in the Nguni and Sotho languages. In English, stress contours are the property of particular words, for example, photógrapher is permissible but *photográpher is not, and the stress contour, recórd, distinguishes the verb from the noun, récord (Katamba 1989: 222; Sloat et al. 1978). English word stress also interacts with the phonological system in that unstressed vowels are frequently reduced to schwas. At a structural level word stress patterns help distinguish compound nominal structures from attribute-noun structures (e.g. bláckbird vs bláck bírd), as well as disambiguate sentences such as, 'They are cooking apples'.

By contrast, word stress in the Nguni and Sotho languages appears to be a feature only of the ideophones (interjections expressing qualities such as 'pitch blackness' or 'dead silence', Cope 1984: 92), while other lexical items have no marked stress contour (Lanham 1984: 220). The penultimate syllable of a word may be lengthened, but not stressed, when the word occurs at an utterance boundary. In the light of these differences between English and Zulu, it is not surprising that the use of stress in BSAE should differ significantly from Standard English. Lanham (1984: 221) argues that speakers of BSAE attempt to approximate the word stress patterns of StdSAE but generally BSAE stress differs from StdSAE stress in that:

  1. Only primary and secondary stress levels are distinguished, as opposed to the four level distinction in Standard English.
  2. Unstressed syllables do not occur in BSAE with the consequence that vowel reduction does not take place.

The placement of stress within lexical items quite often differs from that of StdSAE. Lanham (1984: 222) suggests that stress placement is 'echoic', meaning that it is based on recall of the speech of BSAE or WSAE models. Further research would be required to ascertain whether this is indeed an accurate assessment of BSAE stress placement or whether BSAE follows its own, perhaps more regular, rules of stress placement. <93>

In addition to stress differences, Lanham (1984: 221) claims that syllables tend to be assigned more or less equal duration in BSAE. This contrasts with StdSAE in which stressed syllables tend to occur at fairly regular intervals with the result that unstressed syllables are speeded up. It appears, therefore, that BSAE is to some extent 'syllable-timed' (like French) whereas StdSAE is 'stress-timed'. Syllable timing is a feature of many African Englishes (Schmied 1991: 64) as well as other New Englishes (Platt, Weber & Ho 1984: 136). According to Platt et al. (1984: 137) the use of syllable timing can make a New English difficult for mother-tongue speakers to understand 'because content words do not seem to "stand out" sufficiently'. Lanham similarly claims that syllable timing may affect the intelligibility of BSAE because it interferes with the assignment of utterance prominence (see below).

Intonation and prominence (utterance stress)

As has already been mentioned, English is an intonational language, meaning that pitch contours apply to utterances as a whole rather than individual lexical items. A more accurate term might be 'prosody-prominent language' since intonation and prominence, the two principal elements of prosody, act together. In Standard English, prosody serves a wide range of discourse functions and carries a considerable 'information load'. Certain intonational contours are associated with, and help signal, particular illocutionary acts, for instance, a low-high contour is typically associated with YES/NO-questions. Intonation and pausing can indicate syntactic structure, for example, disambiguating a sentence such as 'The old men and women were spared'. Stress is frequently used to emphasise and contrast particular words or phrases in an utterance (Katamba 1989: 243-250). Finally, utterance prominence and intonation play a very important role in establishing the discourse structure (see also section 4.4.).

By contrast Zulu is a tone language, meaning that pitch contours are primarily a lexical property serving to distinguish one word from another. Prosody in Zulu (and the other Nguni and Sotho languages) appears to carry a much smaller 'information load' than in English. According to Lanham (1984: 221) intonation exists only to a limited extent in the Southern African Bantu languages in the form of 'down-drift', which is a progressive lowering of pitch on each successive lexical item of an utterance. The junction between one downdrift and the next serves to signal utterance boundaries (Katamba 1989: 206). <94> Down-drift also plays a limited illocutionary role insofar as its absence signals an utterance as interrogative rather than declarative. There is, however, nothing in the Southern African Bantu languages akin to the use of utterance prominence or contrastive stress in English (Lanham 1984: 221) and, significantly for the argument I present in section 4.4.2., intonation and prosody do not appear to signal the pragmatic status of information as they do in English (Lanham 1984: 223).

With such marked differences between the prosodic systems of English and the substratum languages, one would expect to find some prosodic differences between White South African English and BSAE/ZE. Chick (1985: 307-308) suggests that 'the signalling load marked prosodically is less in Zulu English than in S.A. English'. According to Lanham (1984: 222) the most significant differences between BSAE and WSAE are:

  1. a tendency to 'over-focus' in BSAE, that is tonic prominence is assigned more frequently than in WSAE. The reasons for this, he argues are:
    1. Certain lexical items attract prominence because of their semantic weight.
    2. Certain intensifiers and negators attract prominence: very, many, even, never.
    3. The syllable-timed nature of BSAE means that there is pressure for the tonic segment to occur relatively early.
  2. Intonational contours are less noticeable than in WSAE and in particular low-high pitch movement at utterance boundaries is seldom used.

Implications for restandardisation

Both utterance prominence and low-high intonational junctures play a significant role in signalling pragmatic functions and thus aiding discourse structure in mother-tongue English (see section 4.4.). Their absence in BSAE/ZE could therefore be expected to have significant consequences for communication between first- and second-language speakers. Lanham (1984: 222) concludes that 'deviance in, and restricted functions for, prosodic patterns in SABE [i.e. BSAE] appear to contribute significantly to a common inability to successfully negotiate interactive discourse of any complexity with mother-tongue English users'. <95>

As Platt et al. (1984: 172, section 2.3.2.) argue, the issue of intelligibility has implications for the suitability of a variety as the basis for a local norm. They also, correctly in my view, recognise that considerations of intranational intelligibility may outweigh those of international intelligibility. With regard specifically to Lanham's conclusions, I believe, it is important to note, though, that his findings are based on precoded spoken discourse, that is, essentially dictated text. In such circumstances one might expect reduced intelligibility since the reader does not have recourse to any of the non-prosodic strategies for marking and maintaining discourse coherence. Moreover, in section 4.4. I argue that the syntactic feature of pronoun copying can be understood as just such a compensation for the absence of prosodic cues in ZE. I believe it is possible, therefore, that BSAE may have more potential for restandardisation than Lanham's account suggests.

4.3. Syntactic features of Zulu English

This section is intended to give a broad overview of some of the syntactic features of Zulu English occurring in the corpora of spoken and written data I collected. It is not possible to give a detailed account of each feature, but I have chosen to highlight some of those that I regard as the more significant ones. The primary reason to consider a feature significant, I argue, is if it involves a general structural deviation from the grammar of Standard English. Particularly interesting in this regard are features that suggest the workings of language (or acquisitional) universals or that show the influence of general functional considerations. These features are most likely to have broader implications for other aspects of the variety's syntax or even for other levels of its grammar.

By contrast, some features, such as non-standard prepositions or phrasal verb usage, involve little more than the selection of alternative lexical items and are thus perhaps relatively uninteresting from a syntactic point of view. They might nevertheless be interesting for other reasons, such as their effect on intergroup (especially second- language-first-language interactions) or their possible symbolic value as markers of the variety.

Schmied's (1991: 65-75) list of 15 syntactic features typical of African English provides a convenient starting point in considering the features of Zulu English. The examples <96> provided of each feature are not those of Schmied. Instead they are drawn from the corpora of spoken and written data I collected and thus provide evidence that each of the features does to some degree occur in ZE. Note, examples drawn from the transcripts in appendix C have references to the transcript number and turn in which they occur, e.g. (4: 92) indicates transcript 4, turn 92. The other examples are culled mainly from the written data.

1. 'Inflectional endings are not always added to the verb, but the general, regular and unmarked forms are used instead'

Intonation play_ important part in conveying the speakers attitude.

2. 'Complex tenses tend to be avoided'

I will like you to deliver because I haven't got a transport. (4: 92)

3. 'The use of the be VERB -ing constructions is extended to all verbs'

I am having a question as to why we did phonetic symbols and even wrote tests on them.

4. 'Phrasal/prepositional verbs are used differently'

The lecturer intended to take out their jackets. (meaning: 'take off their jackets')

5. 'Verb complementation (infinitives, gerunds, etc.) varies freely'

I really appreciate to get this opportunity to comment about the course.

6. 'Noun phrases are not always marked for number and case (by inflectional endings)'

I would also like to talk about my good and bad experience_.

7. 'The use of -s plural markers is overgeneralised'

I saw an Indian girl waving to me through the glasses.

8. 'Articles and other determiners tend to be omitted in front of nouns'

Intonation play __ important part in conveying the speaker's attitude.

Perhaps more common, in my experience, is an article used in a non-standard way:

I think I'm getting the cold. (2:26) <97>

9. 'Pronouns are not always distinguished by gender'

Everytime John when she giving reports everything is fine. (1:96)

10. 'Adjective forms tend to be used as adverbs'

I will also explain clear what is important in communication.

11. 'Pronouns tend to be copied as so-called resumptive pronouns'

I linked it up to Linguistics 1B which I found it very interesting at the moment.

12. 'Negative yes/no questions are confirmed by responding to the form of the question and not to the absolute "inner logic"'

Q: so you don't have problems reaching certain areas of your body?
A: yah
(meaning that he does not experience problems) (11: 43-44).

13. 'Question tags tend to occur in invariant form'

You can do the letter through your blue reports, isn't it? (9: 125)

14. 'The basic interrogative word order is maintained in indirect speech and questions'

I will also discuss how does the stress and intonation contribute in miscommunication between different cultural groups.

15. 'The strict English word order rules are weakened, especially for adverbs'

He's friendly always. (3: 38)

Buthelezi, (1989: 52-56) mentions eight features or constructions that she regards as having become fossilised in BSAE. Some of these are of limited significance (e.g. the use of the fixed phrase 'Thanks God') while a number of the others can be subsumed under Schmied's headings. However, three features are not mentioned by Schmied (again the examples are drawn from the data I have examined):

16. 'The construction: NP BE numerical'

They are five… we are eight there in my house. (3: 79)

17. 'The some… other construction'

Others give the stress at the first syllable and others at the second syllable.

A similar phenomenon, I have encountered, is the 'or…or' construction: <98>

see or this people were paid or not by the government.

18. 'a cheque of X rand' and other non-standard prepositions

I wish to be at home early. (3: 77)

In sections 4.3.1.- 4.3.7., below, I examine in more detail some of the features of ZE that, for the reasons outlined above, I consider significant. For each of the features I discuss, I have chosen for reasons of space to present only one or two examples in the text. Additional examples are, however, provided in appendix D. For ease of reference, I have numbered all example sentences sequentially through the remainder of this chapter from 1 to 69.

4.3.1. Non-standard verbal and tense usages

The tense-aspect systems of English and Zulu differ considerably. English has a binary tense system, marking past versus non-past. This combines with two dimensions of aspect, namely, progressive/non-progressive and perfect/non-perfect. While the future is not a fully developed tense, there are various periphrastic methods for indicating future meaning ('will', 'going to'), as well as past habitual situations ('used to'). Mood is very nearly absent although there are numerous modal auxiliaries. Zulu, by contrast, distinguishes five tenses (remote past, immediate past, present, immediate future, remote future), and six moods. In addition there are the so-called progressive and exclusive implications (Cope 1984 89-92).

It seems likely that these differences would influence the acquisition of the English tense/aspect system by Zulu speakers which, if fossilised, would result in a distinct Zulu English system. Whether this influence is the result of direct mother-tongue interference or is the result of learners resorting to more general structures and strategies in the absence of parallel first-language and second-language structures is beyond the scope of this survey of Zulu English. The following typical usages can be noted:

Extension of the Progressive

In common with many New Englishes (Platt et al. 1984: 72-73; Schmied 1991: 67), the distinction between dynamic verbs, describing actions and processes, and stative verbs, which usually describe states of affairs, is blurred in Zulu English. Syntactically, stative verbs <99> in Standard English do not usually occur in the progressive but in Zulu English verbs such as 'have', 'like', and 'know' frequently occur with -ING forms:

1. The team I am having now Thandi I must tell you I have a strong team (1: 25)

2. You find whites having their own group and Indian and blacks have their own.

3. This is surprising me because I don't see why they have to smile if they don't mean it.

The use of the progressive is extended in other ways, for instance, it may be used in place of the simple present or 'will' future:

4. They are getting a ten percent discount if they are buying. (6: 2)

5. Students in this university are coming from different cultural groups.

6. She is working as a secretary for Zulu.

The progressive also tends to be used in past reference in place of the simple, perfect and 'used to' habitual forms:

7. I was laughing that time when she was saying it. (meaning: 'I laughed when she said it') (1: 129)

8. He was speaking very fast when I was a child.

9. So my mother was always encouraging me to go to school.

10. The blacks have been learning a third grade education.

Some of these extensions of the progressive may result from an over-emphasis of the progressive aspect in the teaching of English to the detriment of the simple form, as Platt et al. suggest (1984: 73). The use of -ING with statives probably also reflects the fact that in English (unlike many other languages with a progressive aspect, see Comrie 1976:38) the progressive is often extended to stative verbs to convey the sense of a temporary state (e.g. 'I'm having a bad day, today'). The extension of the progressive may also be influenced by the absence of a progressive/non-progressive distinction in the substrate languages. Zulu, in common with other Bantu languages (Comrie 1985: 54), has a tense <100> often labelled 'progressive' although it is not semantically equivalent to the English progressive aspect (Doke 1963: 177).

Perfect tenses

The English past perfect is a relative tense, meaning that it is used to refer to an action or event occurring at a point prior to an established temporal reference point (Comrie 1985: 25). The past perfect does not necessarily imply that the event occurred a long time ago (e.g. 'I sat reading. I had just put dinner in the oven'). By contrast Zulu, in common with many Bantu languages (Comrie 1985: 83ff), has an absolute remote past tense (e.g. walala 'he went to sleep') which is used to refer to actions and events taking place in the distant past and which contrasts with the (simple) past tense (e.g. ulalile 'he went to sleep', Cope 1984: 90). There is some evidence in my data that Zulu speakers may use the English past perfect as a substitute for the Zulu remote past:

11. My father had already passed away while I was doing standard one.

12. All these changes had occurred in 1983.

In example 11 there does not seem to be any suggestion that the father's death occurred at a point prior to the writer being in standard one, as would be suggested by the use of the past perfect in Standard English. Instead the writer seems to be suggesting that both events occurred in the distant past. According to Comrie (1985: 26) substitution of the past perfect for a remote past is common among west African speakers of English. It seems that in general the perfect tenses represent an area in which there is considerable variation among BSAE speakers/writers. There are, for instance, also quite a number of examples in the written data where the writer has used a simple tense where prescriptively one might expect a perfect tense:

13. An example is used of the lecturer who taught that a child ___ said tumour when he said two more

14. After a month I got a message that I was selected for entering the University.

There are other instances in which writers have used present perfect forms where one would expect a simple tense (i.e. where there is no relevance to the present implied):

15. I've seen that they were friendly but still I was frustrated. <101>

16. But I've seen him as a person who can't respond to questions or problems effectively.

This variation is perhaps not surprising given that the perfect forms are used in Standard English in ways that are cross-linguistically relatively 'marked' (Comrie 1976: 60).

Hypotheticals

Standard English makes use of past tense verb forms to refer to hypothetical events/conditions, including those that might occur in the future. Zulu English frequently makes use of present or 'future' forms to express hypothetical states:

17. Don't you think that even with us it will be much better if we get our bonus during our birthday. (5: 7)

18. If the speaker lowered the pitch of the voice this word will be meaning the speaker is not angry.

19. I will like to improve the standard of living of my family.

The use of future tense forms to refer to hypotheticals is probably more common (i.e. less marked) in the world's languages than the use of past tense forms. This feature of ZE may therefore represent the influence of language or acquisitional universals. Ho & Platt (1993: 158) report a similar tendency to avoid using past tense forms in hypotheticals in Singaporean English.

Non-standard verbal complementation

Non-standard verbal complements, including the 'discuss about' example listed by Buthelezi, occur quite frequently in ZE:

20. Furthermore she discussed about the present debate.

21. I will start by discussing about low level of bilinguality.

22. She explained to me about the structure of my degree.

23. I asked him to explain for me that part.

It is possible that some of these usages are frequent enough to be considered stable features of ZE, but a considerably larger corpus would be required to confirm this. There is one usage that does occur a number of times in my data: 'make… to'. It appears from <102> the data that in ZE 'make' is usually followed by a 'to' infinitive rather than a bare infinitive, even in the speech of those whose English is very close to standard:

24. What makes them to stop that product if there are people who do come to that shop and buy them. (11: 217)

25. So what will we… made you to come and buy. (4: 9)

The use of a 'to' infinitive also occurs quite frequently in the written data:

26. This make cultural group to understand what she/he meaning.

27. That make the meaning to be different than other countries.

28. ELS makes the second language students to be able to adapt themselves to the university.

The failure of writers to edit 'make… to' sequences out of their writing (given that this would be quite easy) suggests that they are not conscious that this feature of their English differs from StdSAE. The use of a 'to' infinitive is clearly a simplification as relatively few verbs permit a bare infinite (others are 'let' and 'help') and those that do are often paralleled by verbs with similar meaning that do not, e.g., 'force', 'cause', 'aid', 'permit' (Celce-Murcia & Larsen-Freeman 1983: 480) . The sentences below suggest that the use of 'make… to' may be part of a wider preference for 'to' infinitive complements over other types of complements:

29. Do you think they'll be interested to take it out? (5: 27)

30. You don't end up to do the job of refiling. (9: 40)

31. So do you mind to wait? (4: 95)

32. This approach is very significant since it could prevent to have more than eight languages to be learned.

33. English must stop to be the only dominant language in the country.

4.3.2. Pro-drop: subject omitted

34. This week __ is playing with Orlando Pirates (3: 34)

The occasional omission of subjects in ZE can be explained in terms of the pro-drop parameter which refers to a cluster of associated properties found in certain languages, such as Zulu, including: <103>

  1. relatively free inversion of constituents;
  2. the absence of pleonastic (i.e. semantically empty, non-referential) subjects (Cook 1993: 163), which is the subject of the following section; and
  3. the optional omission of overt subjects (Lumsden 1988: 27).

Whether a language allows pro-drop or not seems to be related to its inflectional complexity and, in particular, its agreement system: overt subjects are redundant in languages such as Zulu which have elaborate subject-verb agreement (Haegeman 1991: 416). Language universals may play a role in ZE pro-drop since investigations into first-language acquisition (Hyams 1986 cited in Cook 1993: 166) have suggested that pro-drop is the unmarked or default setting of the parameter. However, research into second-language acquisition has been inconclusive as to whether first-language transfer or language universals are most significant (Cook 1993: 167-170).

4.3.3. Existential THERE

Another feature, cited by Buthelezi as typical of BSAE, which relates to the pro-drop parameter, is the use of constructions such as:

35. we were nine (Buthelezi 1989: 52)

This construction is used in preference to the existential THERE construction: 'There were nine of us'. Similarly in the following sentences the existential THERE seems to be being avoided:

36. Miscommunication happens because they are many students from different cultural groups.

37. At home we are a family of seven.

'There' is required in standard English as it is a non-pro-drop language and the subject position has to be filled, by a 'dummy' subject if no other is available. As with the omission of subject NPs, this may reflect the influence of language universals: If the pro-drop is indeed unmarked the avoidance of 'dummy' subjects such as 'there' by second-language speakers is what one would expect. <104>

4.3.4. Embedded questions

Inversion of NP-AUX in embedded questions is a well recognised feature of South African Black English and many other learner, second-language and even native varieties of English (Adendorff 1990). Sentences such as the following are noticeable in the written as well as the spoken data:

38. Nobody will come to me and ask what is the score (3: 9)

39. Unfortunately we don't know where is she at the moment.

40. Let me just work out your instalment and tell you how much is your instalment with finance charges (4: 95)

Inversion of NP-AUX occurs quite often in the written data as well suggesting that as with 'make…to' this feature is not recognised as non-standard by many speakers and writers:

41. She asked me why didn't I tell her I am busy.

42. This is surprising me because I don't see a reason why do they have to smile if their smile is not genuine.

Evidence from second-language acquisition research suggests that the tendency to invert subject and auxiliary in embedded clauses is particularly resistant to change in second-language varieties. According to Pienemann & Johnston (1987 cited in Cook 1993: 100) learning not to invert NP-AUX in embedded clauses depends on overcoming the general principle that permutations in embedded clauses are to be avoided. This, they argue, only occurs in the speech of the most proficient second-language speakers.

In more general terms, Adendorff (1990: 16) points out that inversion in embedded clauses is psychologically motivated in that it represents a simplification since learners need only learn one rule: NP-AUX inversion occurs in all question clauses. Adendorff (1990: 17) also argues that inversion is functionally motivated in terms of topic maintenance considerations. I shall return to the question of pragmatic motivation for certain ZE features when examining pronoun copying in section 4.4.

4.3.5. Prepositions

Buthelezi includes among her list of syntactic features of BSAE the use of the construction 'a cheque of X rand' (1989: 55-56). Although Buthelezi singles out this <105> particular construction, in my view it forms part of a broader pattern of non-standard prepositional use. Examples that occur in the data I examined include:

43. she still needs some help for that (10: 27)

44. just to create an awareness to them (10: 91)

45. you're wrong by doing this (10: 115)

46. as far as I understand a person is built up in three parts spiritual, emotional physical (10: 135)

Non-standard prepositional use is also frequent in the written data:

47. The language would have its roots from the variety used in South Africa.

48. Different ways of speaking are used to different people.

49.Those who belong to the same group with him…

50. They made friends to each other

The use of non-standard prepositions is reported to be very widespread in second-language varieties (Celce-Murcia & Larsen-Freeman 1983: 250). This can be accounted for by the fact that prepositional use is highly language specific. In the case of Zulu English this is compounded by the fact that Zulu makes use of inflections to signal certain relations expressed by means of prepositions in English.

4.3.6. Relative clauses

Mesthrie (1992: 3) reports 19 relativisation strategies as occurring in his corpus of spoken South African Indian English. I took the most frequently occurring of these strategies as my starting point for coding the relative constructions occurring in my corpus of written Zulu English. I noted each instance of a relative construction and then categorised it according to the following relativisation strategies:

51. Parataxis: He thought of that guy as a racist and liked to make fun of other people. This does not strictly involve relativisation but paratactic constructions are found in pidgins, creoles and second-language varieties.

52. 0 Relative Pronoun: When the student ___ is used in one language is mixed with other students… one loses confidence. <106>

53. Resumptive pronoun: My mother is a housewife whom she always dreamed about his [father's] son holding a better profession.

54. Near Standard: Blacks which have had their high school education in whites schools are not having this problem.

55. Standard: This country is inhabited by different cultured people who speak different languages.

The most noticeable feature of the relative clauses occurring in my corpus of written ZE is that the vast majority are entirely standard: of the 139 instances of relativisation I recorded, 118 (84,5 percent) were standard. No non-standard strategy accounted for more than 3,4 percent of the sentences. Two qualifications to these figures need to made: first, the corpus of data is small and, secondly, the students may to some degree have been able to model their relative constructions on those occurring in the readings they consulted. Despite these reservations, I believe that the use of relative constructions in written Zulu English is quite close to standard English. In the corpus of spoken Zulu, relative constructions occur infrequently. I would expect a larger corpus of spoken BSAE to reveal more non-standard constructions since production pressures are that much greater in the spoken medium.

As regards the non-standard forms that do occur, the use of parataxis appears to be a clear case of simplification, while the use of resumptive pronouns (e.g. 'That is the mani whoi Bongani saw himi') may also be viewed as functionally motivated. Resumptive pronouns help maintain the canonical SVO structure of the relativised clause and allow for easier interpretation (Keenan & Comrie 1977: 92).

4.3.7. Pronoun copying

Pronoun copying refers to the occurrence in sentences such as 'The parents, they supposed to pay ten rands', of an NP followed by a pronoun with the same referent. The use of the term 'copy pronoun' reflects the fact that the pronoun appears to be a copy of the full NP in terms of person and number (Platt et al. 1984: 110; Mesthrie 1992: 79; Bickerton 1981: 34).

I exhaustively searched the spoken and written corpora for instances of copy pronouns in subject and object positions. The preponderance of spoken examples suggests that the <106> use of copy pronouns is primarily an oral strategy. In addition, subject NPs were copied far more frequently than object NPs. The following examples, drawn from the transcribed data, suggest the type of sentences involved. The first four examples are taken from the speech of B. (appendix C., transcript 10), a community worker and university graduate, who made particularly frequent use of pronoun copying but whose speech is in most other respects close to standard (the subject NP is italicised and the co-referential copy pronoun is underlined):

56. I think maybe even the lady who is in charge of that she need more training. (10: 2)

57. This ladies some they've got the know-how. (10: 75)

58. Y'know creative teachers they make something themselves. (10: 102)

59. What I'm trying to do now y'know these women who are domestic workers we are still trying to bring them together and do something themselves. (object copy pronoun) (10: 144)

The following examples are drawn from a number of other speakers of Zulu English:

60. The other ones they don't like this thing. (3: 89)

61. Everytime John when she giving reports everything is fine. (1: 96)

62. You know some people they buy things like this. (6: 36)

63. casuals they not hearing any stories about that. (9: 119)

Pronoun copying is usually analysed as involving a so-called 'dislocation construction' in which the copy pronoun occupies the usual position of the subject NP and the logical subject NP occupies a pre-subject position (cf. Radford 1988: 530). On the assumption that ZE pronoun copying does involve dislocation, I argue in section 4.4. that it serves to signal a range of discourse information.

Evidence from other varieties of African English and New Englishes generally suggests that pronoun copying is a fairly widespread phenomenon among non-native varieties of English. Schmied (1991: 73), Bokamba (1992: 131), Bamgbose (1992: 155) all mention pronoun copying as being typical of African English. Considering New Englishes more generally, Platt et al. (1984: 119-121), state that 'a very common feature in many New Englishes is what is sometimes called pronoun copying'. In addition to examples from <108> East African English, they provide examples from Hong Kong, Bangladesh, Jamaica, Fiji and Singapore.

Syntactic interference as a source of pronoun copying

The evidence of the widespread occurrence of pronoun copying in other varieties of English raises doubts about whether mother-tongue interference alone can provide a complete explanation of ZE pronoun copying. However, Thomason & Kaufman (1988: 58) argue that the existence of a parallel structure in the first language means that the possibility of interference in the second-language should not be overlooked, even if the feature in question is common.

There are two potential analogues in Zulu (and the other languages in the southern Bantu group) and thus two potential interference sources. The first of these, and the one usually mentioned (e.g. Bokamba 1992: 131), is the extensive subject-verb agreement of Bantu languages. In Zulu, the substratum language of the data I am considering, the noun class of the subject is obligatorily reflected by a verbal concord marker (underlined):

64. aba-ntu ba-thenga impahla (Cope 1984: 97)

NC people SC buy goods NC = noun class marker, 'people buy goods' SC = subject concord marker

According to Bokamba, the use of a copy pronoun in effect replaces the subject concord marker. The second, and in my opinion closer, analogue in Zulu occurs in sentences such as the following:

65. abafana bona baphumile

boys they left

The use of a so-called absolute pronoun (underlined) following the logical subject of a Zulu sentence results in a structure remarkably similar to Zulu English pronoun-copying.

It is quite possible that subject-verb agreement and absolute pronoun structures work together to reinforce pronoun-copying in Zulu English. Ultimately, however, I believe that an explanation in terms of syntactic interference alone is unsatisfactory for a number of reasons. In the first place, it does not provide any reason for why Zulu speakers should transfer this particular feature of their mother-tongue syntax and not others. In the <109> second place, it does not explain why this feature should frequently be subject to fossilisation in the speech of even very proficient speakers such as B. (appendix C, transcript 10). Finally, notwithstanding Thomason & Kaufman's (1988) position, it ignores the widespread occurrence of pronoun copying in other New Englishes.

4.4. The function and source of ZE pronoun copying

In this section I examine Zulu English pronoun copying in a little more detail with two related aims in mind. The first is to show that, in addition to being relatively frequent and syntactically rule-governed (as section 4.3.7. showed), pronoun copying is also pragmatically rule-governed, in other words it satisfies all three criteria of systematicity identified by Platt et al. (1984, see section 4.1.). In section 4.4.1. I argue that Zulu English pronoun copying serves a topic-related discourse function. In the light of this, the second aim of this section is to briefly reconsider alternative explanations of the source of pronoun copying that, I believe, go some way to addressing the shortcomings that I identified above of a syntactic interference account.

4.4.1. The pragmatic role and rule-governedness of Zulu English pronoun copying

Pronoun copying does occur in mother-tongue varieties of English, even in the speech of standard English speakers, and its use appears to be essentially pragmatically determined. The literature (Givón 1976: 153; Keenan & Schieffelin 1976: 381; Siewierska 1991: 150; Finegan 1994) suggests that dislocation (i.e. pronoun copying) is a marked strategy that mother-tongue speakers most often use to signal either that they are re-introducing an old topic that has lapsed or else to signal a very definite shift in topic. The reason that dislocation is marked, is that topic development in mother-tongue English is primarily signalled intonationally. A new or re-activated topic is generally indicated by a low-high pitch juncture between utterances (Lanham 1984: 222; Brown, Currie & Kenworthy 1980: 26).

ZE pronoun copying also appears to serve primarily a topic-related function. However, the higher frequency of pronoun copying in ZE compared to mother-tongue varieties suggests that it is a less marked construction, serving a more general function than in standard English. An examination of the data exerpted below shows that the main <110> function of pronoun copying is simply to signal a change in topic referent, rather than being restricted to re-introducing topics as in standard English.

66. Extract from transcript 10, starting at turn 25 (Information about the participants and context is provided in section 3.3.4.). Conventions: Significant new referents are underlined. Co-indexation shows the subsequent development of these referent NPs as topics. Dislocated structures are italicised. Indistinct words are transcribed as xxx.

B.: if a person feel like sewing she can xxx everything xxx and she can go straight
    to the church every Wednesday afternoon at two to four o' clock.
J.: okay [THEME CHANGE] so if had to ask you just er..... to give me three or four
    needs that maybe we can get involved in what would you say
B.: at the moment erm we are still looking for for sewing machines1
J.:                                                                right
B.:                                                                       even can
    .. one1 or two it's okay
J.:                          alright
B.:                                  because this1 would be for the group
J.:                                                                       right
B.: and er the lady2 who is up here at the creche. she still needs some help for that 
    I think she2 might come next year here to TREE or college
J.:                                                          yes.
B.:                                                               and the problem
    is maybe sometime.. y'know the parents3 they supposed to pay ten rand
R.:                                                                       yes
B.:                                                                           and it's
    okay that's super but there are people4 who are not working up there which means
    that those children belong to the parents who are not working they are not accepted
    in the creche…

The extract begins just before J. introduces a new theme [THEME CHANGE], namely, the needs of the community. In answer to J.'s question, B. introduces a new referent, 'sewing machines'. This serves as the topic of her next two turns. In turn 9, B. introduces an entirely new topic referent, 'the lady who is up here the creche', making use of a dislocated phrase with a copy pronoun. The next dislocated phrase with a copy pronoun, 'the parents', not only introduces a new topic referent but also marks a rephrasal after a false start. The dislocated phrase is also accompanied by a marked rise in volume, all of which suggests an emphatic establishment of topic.

Pronoun copying in the other transcripts lends support to this view of the function of dislocation. For example, in the extract below the manager introduces a new topic referent, 'the casuals' as a dislocated phrase with a pronoun copy: <111>

67. Extract from transcript 9, starting at turn 117.

MANAGER: yah and these cabinets1 Y. what is going to happen there
STAFF:   aren't these1 going to the new shop
MANAGER: so leave them1 here until we go to the new shop there's still a lot of filing2
         that needs going on here because it2 will be done in the new shop ... casuals3 they not hearing any stories about that2.

Another clue to the role of dislocated phrases with copy pronouns is the fact that in eight of the 18 instances in B.'s speech (Transcript 10) the dislocated phrase is preceded by 'y'know', which is itself a prevalent feature of Zulu English and which frequently serves to establish a topic (Li & Thompson 1976: 484). This use of 'y'know' or 'y'see' is also noticeable in some of the other speakers:

68. Extract from transcript 11, starting at turn 45.

SHADO: and you have no problems in manoeuvring anything you really like to do?
CALLER 2: ... no no y'see like myself I'm big y'know but I manage myself
              y'know as long as you dress neatly.

The data thus provides evidence that pronoun copying (dislocation) plays a pragmatic role in signalling topic development and that it is pragmatically as well as syntactically rule-governed and of relatively frequent occurrence (as shown in section 4.3.7.). One can, therefore, conclude that this feature of Zulu English displays considerable systematicity, in Platt et al.'s (1984) terms.

4.4.2. Possible sources of Zulu English pronoun copying

In addition to providing evidence of the systematicity of pronoun copying, the data also suggests that the reason that pronoun copying occurs relatively frequently in Zulu English (compared with StdSAE) is because it serves as the unmarked (usual) way of signalling topic development for ZE speakers. This does not, however, explain why the ZE pattern of usage differs from StdSAE in the first place. In the following sections I, first, consider whether the observed pattern of pronoun copying in Zulu English reflects a compensatory development at the syntactic level for changes in Zulu English at the prosodic level, and I then briefly consider two alternative, more general explanations.

Functional compensation

As I note in section 4.2.6., Zulu English prosody differs significantly from that of StdSAE. Lanham (1984: 223) goes so far as to state that 'prosodic distinctions and their placement in the BSAE of our subjects do not serve discourse functions'. The absence of <112> prosodic signalling of topic (and other pragmatic categories such as focus) can perhaps be thought of as creating a 'gap' that syntactic structures, such as pronoun copying, fill. It thus seems possible that the use of pronoun copying in Zulu English to signal the pragmatic category of topic may be viewed as compensating for changes at a prosodic level.

This preference for syntactic signalling of pragmatic information may reflect indirect influence from the mother tongue. In Zulu, prosody as a whole appears to play a relatively minor role in signalling pragmatic information (Lanham 1984: 221) and in its place greater use is made of morphological and syntactic strategies (Cope 1984: 85-97). Dislocation appears to be regularly used to signal topic development in Zulu (and, according to Demuth [1987: 93-94], also Sotho).

Although this explanation still relies to some extent on the notion of mother-tongue interference, it is in my view preferable to the explanations of direct syntactic interference that I presented in section 4.3.7. In the first place, the explanation does not depend on the possibility of mother-tongue interference at a syntactic level which, as I noted in section 2.2.3., is a contentious issue. Instead, pronoun copying is understood to be a compensatory development resulting from changes at a phonological/prosodic level where the notion of mother-tongue interference is generally accepted. In the second place, reference to pronoun copying filling what would otherwise be a functional 'gap' in Zulu English provides some explanation for why Zulu speakers should make use of this feature and not others. In the third place reference to the function that pronoun copying fulfils also provides a clue as to why this feature should be prone to fossilisation in the speech of even very proficient speakers: one would expect features which serve speakers' communicative needs to be among those most likely to be retained.

Language universals and communicative stress

The evidence of pronoun copying in other New Englishes suggests that its use may be a response to something more general than the influence of the mother tongue, even if indirectly. There are at least two possible explanations for this evidence:

  1. Syntactic signalling of discourse information (including topic development) may in some sense be a less marked option than the use of prosodic cues. There may thus <113> be a tendency for non-native varieties of English to make use of strategies such as pronoun copying in preference to intonation. This possibility is suggested by amongst other things the fact that pronoun copying is also common in creoles which, according to Bickerton (1981), are the languages that most closely reflect the workings of the putative linguistic bioprogram (see section 2.2.3.). Additional, albeit indirect, evidence is suggested by the fact that most languages allow to some extent for pragmatic relations to be expressed syntactically but relatively few do so intonationally (Creider 1979: 15, Bickerton 1981: 51).
  2. The preference for syntactic rather than phonological marking of topic development may be a response to the contextual factors common to New Englishes, including BSAE: New Englishes emerge in situations of communicative stress resulting from intergroup communication, differing norms of usage, intergroup tensions etc. In such situations, Givón (1976) suggests, the ability to convey one's message may depend on redundancy, that is providing as many clues as possible to aid interpretation. He comments that what may seem from the perspective of first-language speakers to be 'wasteful over-use of discourse machinery, via "unnecessary" repetition may actually have a useful function: When the channel of communication is noisy, or when the communicative system is relatively frail, overuse may be just the right strategy to insure that the hearer knows what the speaker is talking about' (1976: 154). From this perspective pronoun copying serves to maximise the signalling provided and thus aids the interpretive task of the listener.

4.5. Summary

In summary, in this chapter I provide an overview of a number of features of ZE phonology and syntax, and in the process, I believe, go some way to demonstrating the systematicity (Platt et al. 1984) of BSAE/ZE. The features I consider are those that occur with some regularity in the spoken and written corpora I analysed. Many of these features clearly display grammatical regularity (rule-governedness), due either to the influence of the mother-tongue grammar, the workings of language universals or more general functional considerations. I argue that at least one feature of ZE, namely pronoun copying, can be shown to display considerable pragmatic regularity, and I argue further <114> that the function it serves and the frequency with which it occurs can be accounted for if the use of copy pronouns is understood as compensating for the absence of prosody in the signalling of topic development.

Having shown in this chapter that BSAE/ZE possesses the linguistic prerequisites to form the basis for a restandardised South African English, in chapter 5 I examine the social and political developments in South Africa that I believe will provide BSAE with the requisite status and influence to make some degree of restandardisation likely.


[Start] [Previous] [Next] [Contents] [References]

Created: April 1997. Rodrik Wade <wade@mtb.und.ac.za> URL: http://www.und.ac.za/und/ling/archive/wade_ch4.html