1. Introduction
“Post-creole continua” (Bickerton Reference Bickerton1975) or “Creole continua” form one frequent outcome of the colonial expansion of European languages. They are shaped by creolization, a process that combines simplification with “restructuring,” that is, the reanalysis of lexical items from the European superstrate, sometimes leading to the assignment of grammatical functions to lexical elements that previously had none (Bakker Reference Bakker, Kouwenberg and Victor Singler2007). Creole languages emerge against a societal backdrop of racialized hierarchies, with a European free class dominating a non-European indentured class. This linguistically translates into a polarization between the less restructured “acrolectal” varieties of the former and the more restructured “basilectal” varieties of the latter (Winford Reference Winford1997). Acrolects and basilects do not necessarily have rigid group associations as they may jointly form part of fluid stylistic repertoires. However, it is a historical feature of racialized (post)colonial societies that full access to the acrolect tends to be restricted to the European elite and socially mobile non-European classes.
Some creoles emerged without the population displacement involved by the transatlantic slave trade. “Fort creoles,” distinct from the “plantation creoles” that emerged in the Caribbean, developed nearby European colonial settlements on Africa’s or Asia’s coastlines and displayed strong indigenous influences, as bilingualism lasted longer around the forts (Bakker Reference Bakker, Kouwenberg and Victor Singler2007). In the same way as fort creoles, postcolonial varieties of former colonial languages – typically spoken as L2s – tend to display strong substratal influences. Unlike creoles in general, which historically developed in contexts of restricted access to education, postcolonial varieties of former colonial languages develop in a context of high societal pressure to acquire high-prestige varieties, made accessible via mass education. At the same time, postcolonial varieties of former colonial languages can be used to shape symbolic interactional spaces within which individuals have leeway for concealing their ethnolinguistic backgrounds via “substrate erasure,” a strategy that Mesthrie (Reference Mesthrie2017) described based on phonetic variation in Black South African English.
Afrikaans is conventionally considered a “semi-creole,” a cross between a plantation creole and a fort creole. The least restructured Cape Dutch variety was European Cape Dutch (ECD), while the slaves’ Dutch pidgin produced the most restructured variety, of which Southwestern Afrikaans is today’s offshoot. More aligned with ECD than the latter was the Dutch pidgin developed by the legally free Khoen, who would later nativize it into Orange River Afrikaans (ORA). ORA’s Dutch-like features, such as clause-final verb, may have been facilitated by analogies with Khoekhoe (Roberge Reference Roberge1994a, Den Besten Reference Besten2001). However, Dutch and Khoekhoe were not identical enough at a syntactic level to explain why ORA aligned with Dutch in specific areas. Upward social mobility in the Cape’s hinterland, Stell (Reference Stell2017) argued, is what drove the Khoekhoen to get as close as they did to the ECD norm.
ORA spread into late eighteenth-century precolonial Namibia, this article’s geographic focus, where it developed into a prestige language among local Khoekhoe speakers, nowadays known as the Namas and Damaras. As an L1, ORA was brought in by the Basters, while incoming White Boere and Cape Coloureds brought in varieties closer to the emerging Afrikaans standard variety and Southwestern Afrikaans, respectively. As a South African colony (1915–1990), Namibia widely used Afrikaans as an institutionalized lingua franca (Kleinz Reference Kleinz1981). Meanwhile, it was subjected to Apartheid, a regime that obstructed ethnic integration among “Whites,” “Coloureds,” and “Blacks” (Stell Reference Stell2021).Footnote 2 Using grammatical data, this article examines how this heritage of ethnoracial segregation is visible in grammatical variation in Namibian Afrikaans.
2. Methodology and data
This article aims to test the following three scenarios:
Scenario 1
Namibian Afrikaans shows the effects of L1 transfers at a grammatical level. As a result, Whites and Coloureds behave differently from Blacks, who are generally L2 Afrikaans speakers and possibly influenced by their native languages when speaking Afrikaans (if they know Afrikaans at all).
Scenario 2
In their variation, Namibian Afrikaans grammatical patterns mirror the heritage of Apartheid-era racial segregation. As a result, there is a question of three discrete varieties, namely, “White,” “Coloured,” and “Black” Namibian Afrikaans.
Scenario 3
There is a grammatical continuum in Namibian Afrikaans as similar grammatical patterns are distributed across all ethnolinguistic groups. This suggests convergence dynamics, historical or contemporary, between ethnolinguistic groups.
This article uses an experimental dataset to test these scenarios. A sample of young Namibian informants from Windhoek, detailed in table 1, was invited to take part in an English-into-Afrikaans translation task. They all attended secondary school in Windhoek. Basters and Otjiherero speakers were not included as they could not be found in sufficient numbers at the data collection site. The sample mirrors the distinction between Namibia’s L1 and L2 Afrikaans speakers. The Whites and Coloureds represent the former while the Damaras and Ovambos represent the latter. The Damaras’ L1 is Khoekhoegowab, a Khoesan language, while the latter’s L1 is Oshiwambo, a Southwest Bantu language. The informants were all first-year students at the University of Namibia’s Windhoek campus and can be considered to have had equal exposure to English.
Table 1. Sample of informants for experimental dataset, stratified by ethnolinguistic background and gender (n=60)
Fifty in total, the English stimulus sentences read out to the informants were phrased in such a way as to prompt specific Afrikaans grammatical features likely to be contrastively realized by the various sampled ethnolinguistic groups. Out of these stimulus sentences, a subset was selected to show the effects of variation attributable to (i) different historical L1 Afrikaans dialect backgrounds in the case of the Whites and Coloureds and (ii) transfers into Afrikaans of Khoekhoegowab and Oshiwambo features for which Afrikaans varieties have no equivalents. Below is a list of features prompted by the stimulus sentences:
Past participle formation in verbal clusters
Central to Afrikaans past-tense formation is past participial prefix ge-, appended to the verbal stem, which may or may not be omitted in clause-final verbal clusters. In Donaldson’s (1993: 274–275) account, ge- can be appended to the first verbal element if it is not omitted. Other options are usually not mentioned in accounts of Standard Afrikaans. There is reported dialectal variation in past participle formation and Khoekhoe languages are reported to have played a role in this variation (see further section 3).
The progressive aspect
Afrikaans may use the simple verbal form (3) in section 3 below, the hendiadys with postural verb sit ‘to sit’, staan ‘to stand’, lê ‘to lie’, and loop ‘to go’ + INF (4), besig om te + INF (9), or aan’t/die + INF to express the progressive. However, simple verbal forms remain the default way of expressing it (Donaldson Reference Donaldson1993: 220–222; see further Cavirani-Pots Reference Cavirani-Pots2020 and Breed et al. Reference Breed, Brisard and Verhoeven2017). Khoekhoegowab and Oshiwambo grammatically mark the progressive, which may impact how the Damaras and Ovambos express the progressive in Afrikaans (see further section 4).
Complex initials
Afrikaans may form complex initials, that is, sequences of more than one verb in finite position. Donaldson (Reference Donaldson1993:364) mentions that order along with the more Dutch-like pattern of only one verb in finite position. The copula verbs that trigger complex initials are usually bly ‘to stay, keep on’, gaan ‘to go’, kom ‘to come’, laat ‘to let’, and loop ‘to go’. The origins of complex initials have been sought in Khoekhoe (see further section 5).
Subject–verb inversion
As does Standard Dutch, Standard Afrikaans applies inversion of the subject and finite verbs following clause-initial adverbials or subordinate clauses (Donaldson Reference Donaldson1993:302–305). Inversion occurs in Khoekhoegowab while it does not in Oshiwambo, which may impact how the Damaras and Ovambos treat this Afrikaans feature (see further section 6).
The marking of the pluperfect/remote past tense
Standard Afrikaans does not grammatically mark the remote past tense, while Khoekhoegowab and Oshiwambo do. This could warrant a prediction that Damara and Ovambo Afrikaans developed specific strategies to mark the remote past tense — strategies perhaps not found in the varieties spoken by Whites and Coloureds.
Sections 3 through 6 first provide a descriptive statistical account of how the variants of each selected Afrikaans feature appear to be distributed across ethnolinguistic groups and genders. In each case, they subsequently propose dialectal or contact-based interpretations of the occurring variant distributions while referring to available socio-historical literature. Finally, section 7 pieces the findings together while situating them in the broader picture of historical Khoekhoegowab influence on the formation of Afrikaans and ethnolinguistic polarization in Namibia.
3. Past participle formation in verbal clusters
The translation task contained five stimulus sentences – presented below – with hendiadys constructions involving the verb to sit followed by the verb to watch in the past tense, as in (1). The point of inserting a past-tense hendiadys into the stimulus sentence was to create a trigger for a past-tense Afrikaans hendiadys structure, which in turn could reveal patterns of past participle formation in Afrikaans. There is a caveat: Beyond isolated accounts, such as Drew et al. (Reference Drew, Hakulinen, Heinemann, Niemi and Rossi2021), to sit in Inner-Circle English varieties may not possess the auxiliary-like status that Afrikaans sit ‘to sit’ has as a hendiadys trigger, which in turn may impact how the informants form their Afrikaans past participles if the informants are exposed to Inner-Circle English varieties.
Figure 1 sums up the distribution across ethnolinguistic groups and genders of three types of past participle formation. Type 1 is characterized by the absence of ge-affixation (1), Type 2 affixes ge- to the first verb in the verbal cluster (2), Type 3 affixes ge- to the last verb in the verbal cluster (3), and Type 4 affixes ge- to both verbs that constitute the verbal cluster (4).
Figure 1. Distribution in percentages (%) of past participle formation types (n=144). Combined individual means.
W: White; C: Coloured; D: Damara; O: Ovambo,
1: Type 1; 2: Type 2; 3: Type 3; 4: Type 4
The distribution of types shows ethnolinguistic polarization. Type 4 is dominant among the Whites and Ovambos (79.15% and 70.4%, respectively). In contrast, Type 2 is dominant among the Coloureds (45%) while Type 3 is dominant among the Damaras (80%). Type 1, preferred by Standard Afrikaans, is surprisingly near-absent, except marginally among the Damaras (1.6%).
From a historical perspective, the distribution of types across the informants is difficult to explain. Type 4 (with ge- affixed to the positional verb and the main verb) is described as only occurring where the positional verb and the main verb describe discrete events (Roberge Reference Roberge1994b:49; see further Donaldson Reference Donaldson1993:226). Alongside Type 1, which is the Standard Afrikaans form, one would expect Type 2 to be the dominant form among Coloureds and Whites, which is consistent with Type 2 being described as the “prototypical” form in ORA, Namibia’s initial Afrikaans dialect (Roberge Reference Roberge1994b:71). Importantly, Type 2 also represents the general “colloquial” form (Donaldson Reference Donaldson1993:226).
Given the nature of the translation task, Type 4 should not be interpreted as a new pattern in the process of dislodging others. Rather, it could be seen as an analysis of the English stimulus sentence as containing references to two discrete events, which Type 4 indeed does connote in Standard Afrikaans. The same interpretation might apply to Type 4 as used by the Ovambos. Based on the distribution of the other types among the Whites, Coloureds, and Ovambos, one could argue that Type 2 is in fact the dominant pattern among these ethnolinguistic groups, with some marginal tendency to use Type 3 among the Coloureds. Meanwhile, the fact that Type 1 does not occur at all suggests that it is being dislodged by Type 2 and Type 3.
The retention of past participle ge- in Cape Dutch has been explained as a calque of Khoekhoegowab preverbal particles ge or go, which was used as an unbound morpheme in the original varieties (Roberge Reference Roberge1994a:80–81). Type 2 seems to mirror Khoekhoegowab tense-marking patterns with serial verb constructions, which may have had an impact on ORA, and possibly on European Cape Dutch (Roberge Reference Roberge1994a). In Aikhenvald’s (Reference Aikhenvald, Aikhenvald and Dixon2006:17) examples, preverbal particles appear before the first verb of Khoekhoegowab verbal clusters, in line with the position of ge- in Type 2 (5).
But if Type 2 is a calque of Khoekhoegowab, as proposed by Roberge (Reference Roberge1994b), why is Type 3 dominant among the Damaras, who are supposed to be more prone than Namibian Whites and Coloureds to effecting transfers from Khoekhoegowab into Afrikaans? I am proposing the following interpretation: Afrikaans positional verbs in the Damara variety undergo de-semanticization to the extent that they are no longer perceived as full-fledged verbal elements, that is, as elements apt to attract ge-. Some support for this interpretation could be sought in the behavior of the Khoekhoegowab imperfective auxiliary hã that follows the main verb. In the past tense, it may or may not combine with the auxiliary i (Olpp Reference Olpp1977:95–97). In any event, the tense particles ge (distinct from the subject particle ge) or go are omitted in sentences with hã, as in (6), where ge is a subject particle, not a tense particle. In other words, there may be a Khoekhoegowab trigger for exempting Afrikaans auxiliaries from ge-affixation, which may also affect sit if sit has become a full-fledged auxiliary in Damara Afrikaans.
Namibian Afrikaans past participle formation patterns in hendiadys contexts suggest that the standard form has been dislodged by the general “colloquial” (or ORA) pattern or, in the case of the Damaras, by an alternative pattern so far not described in the historical and variationist literature on Afrikaans. One might glimpse in this pattern a transfer from Khoekhoegowab. Meanwhile, the Oshiwambo speakers seem to follow the Whites and Coloureds in their use of the colloquial/ORA pattern.
4. The progressive aspect
The translation task contained English stimulus sentences with verbal elements that have an obvious progressive meaning. Four of these sentences are shown below. Out of these, ten were singled out for analysis. To not over-prompt Afrikaans hendiadys structures, the selected stimulus sentences do not contain English stative verbs, such as to stand, to sit, or to lie. (South African) English progressive structures with busy + ing were also omitted, such as to not over-prompt the analogous Afrikaans structures with besig + om te + INF (see Mesthrie Reference Mesthrie2002).
Figure 2 sums up the distribution of simple verbal forms, the hendiadys, and besig + om te + INF across ethnolinguistic groups and genders. Afrikaans simple verbal forms prove dominant in all ethnolinguistic groups. Examples (7) and (8) illustrate these simple verbal forms in the form of the uninflected verbal form kyk ‘to look’ (7) and a combination of the conjugated auxiliary het ‘have’ with the past participle form of kyk ‘to look’ with ge-affix (8).
Figure 2. Distribution of progressive markers. Combined individual means.
1: Simple forms; 2: Hendyadis; 3: Besig + om te + V.
W: White; C: Coloured; D: Damara; O: Ovambo (n=300)
Besig + om te + INF (9) most frequently occurs among the Whites and Coloureds (35.5% and 4.15%, respectively). The hendiadys structure, which always involves the Afrikaans verb sit ‘to sit’ (10), most frequently occurs with the Damaras (24.6%). To a lesser degree, it is also found among the Coloureds and Whites, especially among the former. No instances of besig + om te + INF or hendiadys are found among the Ovambos, who only use simple verbal forms.
In European Cape Dutch, progressive constructions with postural verbs and en ‘and’ + INF predate contact with English and may have partial Netherlandic dialectal origins (Roberge Reference Roberge1994b). Meanwhile, creolized non-European varieties of Cape Dutch developed (punctual or non-punctual aspect-marking) constructions with auxiliarized postural verbs directly followed by infinitives (and no connective en ‘and’), widely attested in ORA (Roberge Reference Roberge1994b). As mentioned earlier, the structure with besig + om te + INF has a cognate in South African English, namely, busy + V.ing.
The fact that the Whites use besig + om te + INF more frequently than the hendiadys could be interpreted as a strategy of convergence with English, given that White Afrikaans speakers have more historical exposure to it (Stell Reference Stell2025). The Coloureds assume an intermediate position between the Whites and the Damaras by exhibiting all three patterns. Besig + om te + INF possibly forms an incipient pattern among the Coloureds that is spreading under the influence of White patterns while the hendiadys might be receding as an effect of decreolization.
The fact that the Damaras more frequently mark the progressive than do any of the other ethnolinguistic groups could have to do with them using the hendiadys as a calque of Khoekhoegowab ge ra/ro/re + V, in which ra/ro/re act as progressive habitual preverbal particles (11).
While grammatical progressive-marking in Damara Afrikaans is likely influenced by Khoekhoegowab patterns, Ovambo Afrikaans does not mirror any Oshiwambo progressive-marking pattern. However, the Ovambo informants mark the progressive habitual using a range of Afrikaans adverbials, such as net ‘just’, maar net ‘just’, and altyd ‘always’, which the other informants do not use (12). This suggests that the Oshiwambo speakers express by adverbial means grammatical categories that are morphosyntactically salient in Oshiwambo, such as – in this case – Oshiwambo habitual infix -ha- (13).
The polarization of progressive-marking patterns suggests a stronger English influence among the Whites and a transfer of Khoekhoegowab patterns among the Damaras. No categorically recognizable grammatical transfer occurs from Oshiwambo into the Ovambo Afrikaans varieties, which rather align with the Coloured varieties, marked by the default Afrikaans pattern of simple verbal forms for expressing the progressive. However, the Ovambo informants distinguish themselves by using adverbial means to mark the habitual progressive, as in (12).
5. Complex initials
The translation task contained six sentences with English verbal elements whose Afrikaans equivalents are copulas and main verbs combinable into complex initials. The elicited copulas are gaan ‘to go’, kom ‘to come’, and begin ‘to start’. The sentences that trigger gaan can sometimes also trigger the future auxiliary sal. Sentences with sal are left out of the count.
Figure 3 shows the ethnolinguistic distribution of these sentences that contain complex initials compared to the sentences where the copula and the main verbs are disjunct. The figure reveals sharp polarization between L1 and L2 speakers as sentences with disjunct copula and main verbs are dominant among the former while complex initials are dominant among the latter. Sentences (14) and (16) illustrate cases of the copula gaan forming a complex initial together with main verbs. Sentences (15) and (17) illustrate cases of these same copulas not being immediately followed by main verbs.
Figure 3. Distribution of complex initials. Combined individual means.
1: Complex initials; 2: Disjunct copula and main verb.
W: White; C: Coloured; D: Damara; O: Ovambo (n=152)
Complex initials occur least frequently among the Whites (5%) and most frequently among the Damaras (93.5%). Using complex initials in 21.6% of cases, the Coloureds assume an intermediate position between the Whites and Ovambos. Among the latter, complex initials occur in 75% of cases. Overall, kom occurs the least frequently as a component of complex initials and it appears as such only among the Damaras and Ovambos. Gaan and beginne as components of complex initials occur in all groups.
Den Besten (Reference Besten2001) suggests that there could be a parallel between Khoekhoegowab verbal compounding patterns and Afrikaans complex initials. However, he finds only fused V+AUX patterns in Khoekhoegowab (e.g. V+tsâ ‘to try’, V+!gû ‘to go’, quoted from Olpp Reference Olpp1977:123–124), which, in their syntactic order, do not mirror the AUX+V pattern of Afrikaans complex initials. Khoekhoegowab influence in verb compounding might be more directly visible in early ORA, Den Besten (Reference Besten2001:8) argues, in the form of some instances of V+AUX. However, Den Besten is reluctant to see in these early ORA forms a direct predecessor to contemporary Afrikaans verb compounding patterns.
Oshiwambo has a directive particle ka ‘to go, going to’ that expresses futurity or intention, followed by the stem of the main verb (18).
The verb ujaa/ya ‘to come’ is not recorded as associating with other verbs as a copula. In short, the only Afrikaans copula that has a categorial Oshiwambo equivalent, that is, an Oshiwambo equivalent that is directly followed by a verbal stem rather than a full infinitive with Class 15 concord oku ‘to’, is gaan ‘to go’. This might explain why 5 out of 10 Ovambo informants use gaan as a component of complex initials versus only one who uses kom and beginne as components of complex initials. This contrasts with the Damaras, who all use both gaan and beginne as components of complex initials (while only one uses kom as a component of a complex initial).
It is not certain whether complex initials may be historically considered a Khoe transfer. In any event, they occur most frequently with the Damaras, even if they appear more compatible with Oshiwambo syntax (at least gaan+INF) than with Khoekhoegowab syntax. Meanwhile, the pattern with disjunct copulas and main verbs occurs more frequently among L1 speakers, especially among Whites, which suggests that this pattern is considered more standard.
6. Two ethnolinguistic markers: missing inversion and remote past-tense marking
The translation task contained seven stimulus sentences with adverbials or subordinate clauses in initial position. These sentences could thus be used to test whether the Standard Afrikaans subject–verb inversion after initial adverbials or subordinate clauses is equally effected across ethnolinguistic groups. As it turns out, instances of missing inversion were only found among the Ovambos, as in (20). The expected Standard Afrikaans inversion is missing in 44.7 percent of cases among them. All but three Ovambo informants produced Afrikaans sentences without inversion.
Oshiwambo does not require subject–verb inversion when adverbials and other elements appear in initial position (21). In contrast, Khoekhoegowab requires subject–verb inversion where some elements are topicalized, namely, nouns and adjectives (22). Even though topicalized adverbials do not trigger subject–verb inversion in Khoekhoegowab, the fact that Khoekhoegowab has subject–verb inversion suggests that Khoekhoegowab speakers are in a privileged position to acquire Afrikaans subject–verb inversion. In turn, this could explain why there is no instance of missing subject–verb inversion in the Afrikaans sentences produced by the Damara informants.
An unexpected development during the experiment was the occurrence of a nonstandard grammatical pattern in the Damara Afrikaans sentences, which was possibly triggered by the distinction between recent and remote past tense that one finds in both Khoekhoegowab and Oshiwambo. Khoekhoegowab distinguishes between a remote and a recent past tense via the structures ge ge + V and ge go + V, as in (23) and (24), respectively. According to Zimmermann & Hasheela (Reference Zimmermann and Hasheela1998:50–51, 118–122), Oshiwambo (at least in its Oshikwanyama variety) forms a recent past tense by adding the suffix -a to the subject concord while keeping the verb in its present-tense form (25) and a remote past tense by adding the suffix -a to the subject concord and -ile or its variants -ele, -ine, or -ene to the verb (26).
While there is no indication that the Ovambo informants’ Afrikaans sentences mirror the Oshiwambo distinction between recent and remote past tenses, the way that the Damara informants use the past tense in their Afrikaans sentences seems to mirror the Khoekhoegowab distinction between recent and remote past tenses. The past-tense Damara Afrikaans sentences occasionally display the past participle form gehad ‘had’ of auxiliary hê ‘to have’ following the main verb’s past participle form (27). Overall, gehad following main verbs’ past participles occur in 13.2 percent of past-tense sentences. It is used by 6 out of 10 Damara women and by 4 out of 10 Damara men.
None of the English stimulus sentences contained English pluperfect forms. Therefore, gehad may not be ascribed to attempts at rendering an English pluperfect meaning. Indirect support for interpreting gehad as a remote past-tense marker can be found in historical indications that ORA – possibly as an effect of Khoekhoegowab influence – retained means to express distinct past tenses, unlike Standard Afrikaans (if one overlooks the “occasional” pluperfect pattern that Donaldson (Reference Donaldson1993:233) reports on). This could be the reason why Roberge (Reference Roberge1994a:88) observed that ORA retained the old preterite form had of auxiliary hê, while the other Afrikaans varieties had dropped it to only retain the present-tense form.
This section demonstrated that L2 Namibian Afrikaans varieties are likely to show the effects of transfer from native languages. This is most clearly the case when it comes to subject–verb inversions, predictably missing in Ovambo Afrikaans. However, the concentration of the (probable) remote past-tense marker gehad in the Damara Afrikaans sentences and the absence of any equivalent in the Ovambo Afrikaans ones despite Oshiwambo having remote past-tense marking calls for a socio-historical contextualization of variation in the Namibian Afrikaans continuum.
7. Discussion
The study found some suggestions of intergroup differences attributable to contact. One can glimpse possible Khoekhoegowab grammatical transfers among the Damaras. This applies perhaps most visibly to their use of overt progressive structures with sit en + V and past participles with gehad, while the other distinctive features used by the Damaras can at least partially be related to Khoekhoegowab features. L1-transfers among the Ovambos are most visible in their tendency to omit subject–verb inversions. These observations seem to argue in favor of Scenario 1 (i.e. intergroup variation based on the L1/L2 distinction). However, Scenario 3 (i.e. “there is a grammatical continuum”) finds confirmation in the fact that some of the features that are dominant among the Damaras may occasionally also be found, albeit marginally, with the Coloureds (and sometimes with the Whites). This creates the impression of a continuum with a White pole and a Damara pole, between which the Coloureds occupy an intermediate position. In contrast, and contra Scenario 3, the Ovambos do not form part of that continuum by not displaying at all some of the features that are dominant among the Damaras, such as Type 3 past participle formation, overt progressive structures with sit en + V, and (remote?) past-tense marking with gehad, as well as by distinctively displaying missing subject–verb inversion.
The internal organization of the continuum of grammatical variation confers some validity on Scenario 2, which held that the Apartheid-era ethnoracial hierarchy predicts variation. Apart from past participle formation, the Whites always use the patterns that are the most compatible with Standard Afrikaans. This is a long-term repercussion of the strong ideological association between the Afrikaners and the more Eurocentric, Dutch-aligned varieties of Afrikaans, on which Standard Afrikaans, the socioeconomically highest-ranking Afrikaans variety, is historically based (Ponelis Reference Ponelis1993). The fact that the Coloureds occupy an intermediate position between Whites and Blacks, and that the Damaras occupy an intermediate position between the Coloureds and the other Blacks, speaks to the social dynamics described in Stell’s (Reference Stell2020) ethnographic account of Apartheid-era Windhoek: The Coloureds acted as social intermediaries between Europeans and non-Europeans, while the Damaras, notoriously proficient in Afrikaans, acted in turn as intermediaries between the White and Coloured Afrikaans-speaking elite and the rest of the African population. Varying degrees of social proximity between these population groups may be what lies behind the gradual distribution of features noted above, or in other words, the mutual convergence predicted by Scenario 3. The peripheral position occupied by the Ovambos in this continuum may be ascribed to their historical underexposure to Afrikaans, which is itself caused by their mass urbanization beginning only after English became institutionally dominant.
This study’s findings might also tell us about the formation of Afrikaans. The high frequencies of certain features in Damara Afrikaans offer support for seeing a partial Khoekhoe origin in them. The widespread use of the progressive-marking hendiadys in Damara Afrikaans partly mirrors the Khoekhoegowab use of progressive-marking particles. This, in turn, could indicate that Khoekhoe aspect marking lies behind the spread of the hendiadys in the historical Cape Dutch basilects. Similarly, high frequencies of complex initials in Damara Afrikaans strongly suggest that Khoekhoe verb serialization has something to do with the emergence of this phenomenon in Afrikaans. Finally, the generalized occurrence of subject–verb inversion pattern in Damara Afrikaans, which one could call an adaptation of an already extant Khoekhoe pattern, could explain why Afrikaans, be it in its acrolectal or basilectal form, retained this Dutch-like feature while the other Dutch-based creoles dropped it. More problematic is the Damara Afrikaans past participle formation pattern in hendiadys contexts, which – unlike the other Damara Afrikaans features – does not quite look like the “classic ORA” pattern. On this point, I argue, Damara Afrikaans produced an innovation that is bringing about divergence from earlier ORA stages, which one still finds represented among the Coloureds and Whites.
8. Conclusion
There is a grammatical continuum between the Afrikaans varieties of ethnolinguistic groups that have historical exposure to Afrikaans. This continuum includes the Damaras: Their grammatical features are generally compatible with ORA ones, which one finds among the Whites and Coloureds, and much less among the Ovambos. Historically, there is grammatical convergence around ORA within the Namibian Afrikaans continuum. Seeds of fragmentation might be found in the (contact-induced or other) innovations shown by Damara Afrikaans and Ovambo Afrikaans. Whether it is the case that convergence and fragmentation dynamics co-occur at the grammatical level, or that fragmentation is turning into the chief dynamic, cannot be categorically answered based on this study’s dataset. However, Stell’s (Reference Stell2021) findings on sociophonetic variation in Namibian Afrikaans suggest that interethnic divergence may be at work as a result of post-independence dynamics of decreasing exposure to Standard Afrikaans and use of Afrikaans as an interethnic lingua franca between Blacks, with potentially little input from L1 varieties, as suggested by the form of Afrikaans-based Namibian youth languages discussed in Stell (Reference Stell2020).
Open access
