1. Perfect auxiliary doubling

In syntactic doubling, a constituent (i.e. “a morphosyntactic feature, morpheme, word, or phrase”) is expressed two (or more) times (Barbiers Reference Barbiers, Barbiers, Koeneman, Lekakou and van der Ham2008:2, Koeneman et al. Reference Koeneman, Lekakou and Barbiers2011:35). The doubling of a constituent can exert a semantic effect, serve a discourse function, or introduce redundancy.

The doubling of auxiliaries ‘have’ and ‘be’ in perfect tense constructions is a European areal phenomenon. It is present in languages of different filiations that have been in contact for a long time. The doubling of perfect auxiliaries ‘have’ and ‘be’ (the latter typically with mutative verbs) produces what are known variously as double perfect constructions, auxiliary doubling (Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:47), double auxiliaries (Poletto Reference Poletto2009), perfect doubling (Koeneman et al. Reference Koeneman, Lekakou and Barbiers2011), supercompound forms (Thieroff Reference Thieroff and Dahl2000:284), passé surcomposé (Carruthers Reference Carruthers1994), and surcomposé past tense constructions (Saussure & Sthioul Reference Saussure, Sthioul and Binnick2012). Doubly composed periphrastic tense forms are usually treated as marginal features, which is understandable on grounds of low frequency.

Most of the languages that show supercomposed tense periphrasis are spoken in a contiguous area of Europe and belong to the Indo-European family. In the Romance branch the feature is distributed across an area that includes some regional varieties of French and Occitan (see especially Carruthers Reference Carruthers1994, Saussure & Sthioul Reference Saussure, Sthioul and Binnick2012), northern dialects of Italian, Rhaeto-Romance, Ladin, Friulian, some varieties of Catalan and Romanian, and with isolated attestations in Spanish (Thieroff Reference Thieroff and Dahl2000:286–287, Ammann Reference Ammann2007:196–199, Poletto Reference Poletto2009). Such forms also exist in continental West Germanic languages, the distribution of which I shall briefly survey in the following section. Formally similar constructions are attested in older stages of Slavic languages, in one modern Slavic language (Sorbian), and in Albanian (Litvinov & Radčenko Reference Litvinov and Radčenko1998:68–75, Thieroff Reference Thieroff and Dahl2000:286–87, Ammann Reference Ammann2007:199–200). Despite the formal similarities, the semantics of doubly composed periphrastic tense forms vary across languages.

Litvinov & Radčenko (Reference Litvinov and Radčenko1998:67–68) list Afrikaans among the languages that show doppelte Perfektbildungen, citing a Russian grammar to which I do not have access and (of all sources!) a mid twentieth-century primer in the Teach Yourself series. According to the latter: “The past perfect (pluperfect) ‘I had written’ is rendered by Ek het geskryf gehad, i.e., the auxiliary het [‘have’] followed by the past participle of the main verb, followed by the past participle of het (gehad)” (Burgers Reference Burgers1957:118). Supercomposed tense periphrasis is a marginal feature that has gone largely unnoticed in Afrikaans linguistics. But the fact that it exists at all in Afrikaans raises questions that are worth examining. Does this feature have a traceable history through the Cape Dutch philological record? On the assumption that the feature in question represents Netherlandic erfgoed (vis-à-vis autochthonous innovation), what is its likely metropolitan provenance? An answer to the second research question leads to a third: What implications would our finding have for hypotheses regarding the Netherlandic dialectal base of Afrikaans?

2. Perfect auxiliary doubling in contemporary West Germanic languages

Supercomposed periphrastic forms are attested in literary German (1), though they are not frequent.Footnote ¹

Curme (Reference Curme1922:283–284) discerned a slight difference in meaning between regular and double past perfect (1c–d). The latter “expresses a state or condition of things instead of an action,” although “there is a tendency to use this past perfect tense form instead of the regular one, even where the reference is clearly to an act and not to a state or condition,” as in (1c). Litvinov & Radčenko (Reference Litvinov and Radčenko1998:112–144) identify contexts that license the supercomposed periphrastic forms in literary German, prominent among which are highlighted anteriority when the pluperfect functions as a narrative tense, combinations of anteriority and resultative states, and the marking of anteriority in the defective paradigm of the subjunctive. Curme’s first observation (above) anticipates the position of Poletto (Reference Poletto2009). Whereas supercomposed periphrastic forms express an anteriority value in Romance, in the Germanic area they realize a terminative aspect (i.e. the event is completed). At the same time, anteriority “generally implies terminative aspect” (Poletto Reference Poletto2009:32), and the two semantic values can be very close, if not at times indistinguishable.

According to Thieroff (Reference Thieroff and Dahl2000:286–287), supercomposed periphrastic forms “are only attested in languages where the present anterior is undergoing or has undergone a development toward a more general past category.” While this is not remarkable, given that in these languages preterital and imperfect tenses tend to be abolished, it is not possible to predict at which stage of present-anterior development supercomposed periphrastic forms emerge, nor whether they occur at all. Nonetheless, the loss of preterital tenses and the extension of the present perfect to the expression of (imperfective) past tense appears to be an areal phenomenon in modern Europe (Lindstedt Reference Lindstedt and Dahl2000:371).

As for nonliterary German, doubly composed periphrastic constructions of the kind under consideration “[sind] auch in der heutigen gesprochenen Sprache weit verbreitet” (Fischer Reference Fischer2018:217). In certain dialects the present perfect replaces the preterite, even for the perfect auxiliaries haben ‘have’ and sein ‘be’. In Upper German (i.e. German-speaking Switzerland, Bavaria, Upper Austria) the pluperfect is replaced by a special form made by the addition of the past participles gehabt or gewesen to the regular present perfect (cf. Abraham & Conradie Reference Abraham and Conradie2001:20–23, Poletto Reference Poletto2009:32). The resulting supercomposed periphrastic form that cooccurs with the loss of the preterite expresses the pluperfect – with E (the point of the event) < R (a point of reference) < S (the point of speech) in Reichenbach’s (Reference Reichenbach1947) time relation notation, albeit with specific modal nuances.Footnote ²

Figure 1.

Present anterior > more general past category and supercomposed pluperfects in European languages (Thieroff Reference Thieroff and Dahl2000:288).Footnote ³

Figure 2.

Netherlandic dialectsFootnote ⁴

Preterital loss and supercomposed periphrasis to indicate the occurrence of two events in the past in the time relation noted above are characteristic of two extraterritorial West Germanic languages, viz. Pennsylvania German, which is derived from German dialects spoken by immigrants from southwestern Germany and Switzerland to colonial Pennsylvania from the eighteenth century, and Yiddish, which shares many features with Bavarian German.

According to Litvinov & Radčenko (Reference Litvinov and Radčenko1998:61), “die einzige germanische Sprache, die dem Deutschen in dieser Hinsicht angeglichen werden kann, ist das späte Mittelniederländische (um das 15. Jahrhundert).” Stoett (Reference Stoett1977:181) gives a handful of examples, albeit free from context. They include:

According to Stoett (Reference Stoett1977), a construction like (3) presupposes a present-tense counterpart ic hebbe hem gevangen, the meaning of which would have been the equivalent of Ik houd hem gevangen ‘I hold him captured (i.e. captive)’, with the past participle gevangen ‘captured’ describing how a state of possession was achieved. Over time, the compositional meaning of hebben and gehad were no longer felt, whereby the meaning of the longer construction, ic hebbe hem gevangen gehad ‘I have possessed him captured’ became the same as that of the shorter construction, ic hebbe hem gevangen. The semantic bleaching of past participle gehad facilitated its spread to constructions with intransitive verbs, as in (4).

Following Duinhoven (Reference Duinhoven1997:346–348), Barbiers et al. (Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:48–49) put forth a similar idea. Perfect auxiliary doubling in Dutch dialects (below) reconstitutes an older construction in which hebben ‘have’ is the main verb with the meanings ‘get, possess’ and a cooccurring past participle is an adverbial expression.

In (5) the construction [[hebben gehad ] + past participle] involves the perfect tense of a lexical verb focusing on the resulting state. As the construction in (5a) passed into obsolescence, so too did its perfect in (5b); further to the Middle and Early Modern Dutch evidence in Kern (Reference Kern1912:36–37).

Litvinov & Radčenko (Reference Litvinov and Radčenko1998:61) aver that “das Niederländische von heute kennt keine [doppelte Perfektbildungen], soviel wir wissen.” While that statement certainly holds true for Standard Dutch (Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:44), it turns out that the range of doubly composed periphrastic tense forms (figure 1) extends into Netherlandic-speaking territory, where, it should be noted, both the preterite and the pluperfect are available. The latter fact provides compelling argument against a significant correlation between Präteritumschwund and perfect auxiliary doubling (see Koeneman et al. Reference Koeneman, Lekakou and Barbiers2011:40–41 for discussion). According to Barbiers et al. (Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008a: map 40b, 2008b:48, 51), the distribution of supercomposed periphrastic forms is largely restricted to the southeastern part of Dutch-speaking Belgium, including Belgian Limburg, Vlaams-Brabant, and the eastern part of Antwerp.

In the Netherlands, the supercomposed periphrastic construction is only found sporadically (Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:51). If one compares map 40b in Barbiers et al. (Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008a) with map 1 in Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011:39), one finds a discrepancy in the distribution and number of cases of the perfect auxiliary doubling in North Brabant. Whereas the former indicates little to no perfect auxiliary doubling in this province, the latter shows that doubling does occur in seven communities in the central and eastern parts.

In these varieties the doubling of ‘have’ in sentences like (6) makes no independent semantic contribution; the construction “seems to be a genuine case of semantically empty doubling” (Barbiers Reference Barbiers, Barbiers, Koeneman, Lekakou and van der Ham2008:18; also Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:48). Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011) have tested the relevant dialects for anteriority (“a state resulting from an event that is located in the past relative to another past event,” p. 72), but these investigators have found no clear contrast between simple and supercomposed periphrastic tense constructions. Some Netherlandic dialects have the undative construction alongside perfect auxiliary doubling; other dialects have one but not the other; still other dialects have neither. Whatever their respective distributions, there is a significant meaning difference between undative and supercomposed perfect constructions, as in (8) beside (9), and it is clear that the two “do not immediately reduce to one” (Koeneman et al. Reference Koeneman, Lekakou and Barbiers2011:43, who give their examples in standard orthography).Footnote ⁵

Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011) take up the question of whether constructions involving perfect auxiliary doubling as found in certain southeastern Netherlandic dialects constitute genuine cases of doubling. Their answer to this question is no. Syntactic variation is attributable to variation in the lexicon (p. 37), which is of course in line with precepts of the minimalist program. What on the surface appears as doubling is actually a combination of functional and copular variants of hebben and zijn or, in other words, it is the perfect tense of a copular verb, the complement of which is an adjectival participle (p. 74). That being the case, there is no doubling in an analytical sense (p. 37), and the term “perfect doubling,” as applied to data of the kind in (6) and (8), is purely descriptive. Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011:72) believe that the semantic aspects of the phenomenon “do not, at this stage at least, provide a compelling argument in favour of or against a particular syntactic proposal.” With that conclusion in mind and given the specifically diachronic objective of this study, we shall not delve into the formal analysis of perfect auxiliary constructions in Netherlandic dialects.

In a description of the dialect of Kempenland (northeastern Belgium and parts of the southeastern Netherlands) De Bont (Reference Bont1962:430–434) characterizes perfect auxiliary doubling as a perfectum remotius (the term is from Bilderdijk Reference Bilderdijk1825:24) or “verwijderd voleindigde tijd” (distantly removed completed time), with copious examples, among which:

Swanenberg & van Spijk (Reference Swanenberg and van Spijk2024), who regard this perfectum remotius as a local North Brabantine shibboleth, explain to a lay readership how completion of the action or state in a sentence like (10) would have occurred at an indeterminate but remote point in time vis-à-vis simplex Hij heeft niet geschreuwd. Footnote ⁶ Additionally, as per Swanenberg & van Spijk, doubling of the perfect auxiliary may shift the focus to the result itself.

In (12a) the speaker has rented out an attic, the result of which is a space that is under contract and (one might infer) currently occupied. By deploying the perfectum remotius in (12b), the speaker implies that the space is no longer rented out and the lessee no longer has possession.

In an appendix to their study Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011:72–74) weigh in on “what looks like a meaning difference between doubling and non-doubling” (p. 68). In their assessment the doubling does not seem to make a clear temporal (or aspectual) semantic distinction but rather adds color or nuance, being “associated with a more vague, pragmatic contribution” (p. 73). A reading that their informants reported consistently is that the doubled construction involves what the investigators call a “reversal’ interpretation.” The implication of the sentence in (13) is that the bicycle is no longer missing (p. 74).

The data point in (13) is consistent with readings that Swanenberg & van Spijk (Reference Swanenberg and van Spijk2024) have reported. Koeneman et al. (Reference Koeneman, Lekakou and Barbiers2011:73) argue that Netherlandic varieties have only this “superperfect” use of perfect auxiliary doubling (vis-à-vis the anterior use), which (following Carruthers Reference Carruthers1994:172) marks a completed action or state that takes place in the distant past or at an indeterminate point in time, is unlikely to recur, and is in some contexts exceptional, with perhaps heightened involvement on the part of the speaker. Yet Swanenberg & van Spijk (Reference Swanenberg and van Spijk2024) provide one datum that does not seem consistent with the superperfect profile:

If there is a subject of high involvement in (14), it should be the experiencer in the embedded clause. However, ‘yesterday’ is time-specific and hardly remote, which suggests that the doubled perfect auxiliary is pleonastic in this instance.

3. Perfect auxiliary doubling in Afrikaans

Afrikaans has divested itself of the Dutch preterite (e.g. weak ik hoorde ‘I heard’, strong ik zong ‘I sang’) as an inflectional category, save for in the copula (was ‘was’), the verb ‘have’ (had ‘had’, now obsolescent), two “irregular” verbs (dink/dag ∼ dog ‘think/thought’, weet/wis ‘know/knew’), and the modals (sal/sou ‘shall, will’, wil/wou ‘want, would’, kan/kon ‘can, could’, mag/mog ‘may, might’, moet/moes ‘must, had to’). The modal pairs are arguably no longer present and preterite tenses of the same verbs, as, strictly, English shall/should, may/might, can/could, will/would aren’t either (cf. Donaldson Reference Donaldson1993:219). In contemporary colloquial Afrikaans even the copula favors an analytic construction (ek was gewees ‘I was’) over the historical preterite.

In Afrikaans the “past tense” is a periphrastic structure formally derivable from the Dutch perfect.Footnote ⁷ It is formed with a past participle of the verb plus het ‘have’ (the finite form of hê ‘have’), which has replaced forms of zijn ‘be’ as the tense auxiliary with mutative verbs. The (etymological) perfect has absorbed the functions of the preterite and, for that matter, the Dutch pluperfect in the active voice, with the obsolescence of preterital had ‘had’ in Afrikaans (Donaldson Reference Donaldson1993:240).

Because of the very common practice in Afrikaans of expressing past events in the present tense (cf. Donaldson Reference Donaldson1993:228–233), there are occasions where the etymological perfect has the function of a pluperfect, as in (16).

Donaldson emends the communis opinio that “there is no difference between the perfect and the pluperfect in form, and therefore one usually talks of Afrikaans having no pluperfect” (Reference Donaldson1993:231). Variant constructions of the kind in (17) occur occasionally in both speech and writing when one wants to emphasize that a particular action has occurred prior to another: “The perfect of hê (in effect the equivalent of the English imperfect ‘had’) is employed as the auxiliary of the verb which is being put in the pluperfect” (Reference Donaldson1993:233).

A search of the Taalkommissiekorpus turned up twenty-two cases of perfect auxiliary doubling with the past participle of the lexical verb bearing the suffix ge-.Footnote ⁸ Of these, twenty are in the anterior role, for example:

Of the remaining pair of hits, one signals an event that took place at an unspecified moment in the remote past with high subject involvement; the other appears to be pure pleonasm. In view of the state of affairs in contemporary Netherlandic dialects in the metropole, the presence of perfect auxiliary doubling in the anterior guise in Afrikaans is striking. Yet one generally looks in vain for any discussion or even mention of this feature in the standard Afrikaans grammars, synchronic syntactic descriptions, normative handbooks, and diachronic studies.

4. The formation of a contact language at the Cape of Good Hope

In this section I briefly summarize a model of language formation at the Cape of Good Hope that I have laid out in detail elsewhere (e.g. Roberge Reference Roberge, Carstens and Bosman2024).

After the Dutch East India Company established a permanent outpost on Table Bay in 1652 and introduced an economy based on nonindigenous enslaved labor (from 1658), a mixed Dutch-English-Khoekhoe reconnaissance jargon of the precolonial period (from c. 1590) developed into a basic medium of interethnic communication (MIC), which, following Den Besten (Reference Besten and van der Wouden2012), we shall call the Cape Dutch Pidgin. The enterprise of language construction continued long after the crystallization of the basic MIC (±1713) in a complex multidimensional space. From the 1740s Cape labor-caste speakers – enslaved Asians and Africans and indentured Khoekhoe – developed a more elaborated form of the MIC, which would be the crucible for most of the creolelike features of Afrikaans.

Euro-Cape Dutch (settler Dutch) was a recognizable extraterritorial Netherlandic variety that was characteristic of the locally born patriciate of the Colony. It developed out of vernacular Early Modern Dutch imported to the Cape of Good Hope from 1652. Euro-Cape Dutch was shaped by koineization, internal change, and lexical and onomastic borrowing, principally from Khoekhoe, Malay, Creole Portuguese, French, and to a limited extent German. Euro-Cape Dutch moves away from metropolitan norms in morphology, most notably with the atrophy of grammatical gender in the noun and of personal agreement in verb inflection. Over the course of the eighteenth century and roughly the first third of the nineteenth century, Euro-Cape Dutch maintains a fairly stable opposition between finite and nonfinite forms of the verb and the use of zijn as a perfect auxiliary with mutative intransitives. The preterite remains available for expression of the narrative past but is under pressure from within and definitely “from below,” where it is a casualty of pidginization. In the former connection, Loubser (Reference Loubser1961) offers a robust empirical investigation of the development of the functions of the periphrastic verbal forms with hebben and (initially) zijn in Euro-Cape Dutch, at the expense of the preterite. Euro-Cape Dutch is also characterized by the absence of creolelike features such as we find in Afrikaans today.

Europeans who needed to make themselves understood by members of the labor caste as well as be understood among themselves could be expected to take part in the negotiation of the MIC as full participants. Because the communicative demands on the workforce were horizontal as well as vertical, non-Europeans made greater use of the Cape Dutch Pidgin than Europeans.

The elaborated contact variety that developed at the Cape did not give rise to a medium of community of solidarity, which would be the next stage on the road to creolization. In other words the extended Cape Dutch Pidgin did not in itself become the first or primary language of a single, culturally cohesive community. Instead, the outcome was an unusual kind of mixing that arose at the interface between the extended Cape Dutch Pidgin and the extraterritorial Netherlandic variety that was Euro-Cape Dutch, specifically an intertwining of genetically different grammatical subsystems. This heterogeneous mixed variety, which I have called the Cape Dutch Vernacular, served as a medium of subaltern affiliation for most of its users of color. The Cape Colony’s underclass comprised many Europeans who, though never enslaved, were poor. For the white underclass, their versions of the Cape Dutch Vernacular came to serve as a kind of medium of local groundedness vis-à-vis the small but wealthy gentry of the southwestern Cape, Dutch East India Company officialdom, transient mariners and travelers, the merchant class in Cape Town, British colonial authorities (after 1808), and English settlers (after 1820).

What received the glossonym Afrikaans was a focused, later standardized variety that began to take shape only from the last quarter of the nineteenth century. Its (white) developers and partisans appropriated resources from the Cape Dutch Vernacular, with their own modifications. The initial model underwent a subsequent period of reshaping and (mainly lexical) borrowing from metropolitan Dutch and became the synecdochic variety.

According to the establishment view, the near total loss of the Dutch preterite as an inflectional category took place within the European Cape Dutch community during the eighteenth and nineteenth centuries. It is conventionally explained by speaker preferences for the historic present and the analytic perfect to relate past events.Footnote ⁹ Ponelis (Reference Ponelis1993: ch. 14), however, appears to assign greater weight to language contact. Preterite loss cannot be ascribed to gradual internal causes alone, such as competition between past tense forms: “The stable past tense variation that has been a feature of Dutch for centuries has not resulted in the elimination of the imperfect [i.e. preterite]” in the metropole (Ponelis Reference Ponelis1993:439). The loss of the preterite in Afrikaans, Ponelis continues, “is another aspect of verbal deflection resulting from verbal invariance under creolisation,” by which he understands restructuring due to “secondary proficiency” (imposition) on the part of the indigenous Khoekhoe, enslaved Africans and Asians, and other Europeans (Germans, French) at the Cape of Good Hope during the Dutch East India Company era (1652–1795).

Conradie (Reference Conradie1999; cf. also Abraham & Conradie Reference Abraham and Conradie2001: ch. 2) attributes preterite loss in Cape Dutch to mnemonic, systemic, and pragmatic factors, respectively, imperfect learning on the part of L2 speakers of Cape Dutch, simplification of the perfect (loss of the auxiliary zijn ‘be’, deflection), and the availability of alternatives to the preterite (historical present, the perfect itself, the temporal adverb and conjunction toe ‘then, when’ < Du. toen combined with the present tense). But Conradie acknowledges that the preterite must have disappeared much earlier in the vernaculars of L2-speakers or of speakers belonging to groups of former L2-speakers, whose access to metropolitan norms would have been limited.

No less clairvoyant a scholar than Den Besten seems unaware of the existence of perfect auxiliary doubling in Afrikaans (Reference Besten and van der Wouden2012:280, my emphasis):

Clearly, something of the kind did happen at the Cape of Good Hope, although along a trajectory that is different than the scenario that Den Besten imputes to southern German dialects and Yiddish. As should be clear from the preceding discussion, I fully agree that ascribing the loss of the Dutch preterite to a clash between tense systems in Euro-Cape Dutch and an extended Cape Dutch Pidgin “can hardly be called adventurous” (pace Conradie, Abraham & Conradie Reference Abraham and Conradie2001). At the same time, supercomposed periphrastic tense forms are in fact present in our Cape Dutch corpora of the eighteenth and early nineteenth centuries, even if only weakly so; they are the legacy of metropolitan Dutch of the early modern period; and they were available to express the anterior past as per Reichenbach’s formula (above).

5. Perfect auxiliary doubling at the Cape of Good Hope

Perfect auxiliary doubling resides in archival materials representing two genres, the earliest attestation of which is from 1712 (example (20), below). One corpus consists of official colonial government trial records, some of which were excerpted by van Oordt (Reference Oordt1959–1962) and others collected and published as whole documents by Worden & Groenewald (Reference Worden and Groenewald2005). The language of these records reflects the formal Dutch of the era, albeit interspersed with localisms that are of linguistic interest. A second corpus, the Kaapse taalargief (van Oordt Reference Oordt1947–1956), is a compilation of reports from 1712–1831 that are written by unlearned (sometimes only marginally literate) field cornets, who in many cases appear to have struggled simply to compose something intelligible for their magistrates. Individually, these letters can vary significantly in terms of their structural distance from metropolitan norms. Collectively, they represent a uniform text type, the language of which is Euro-Cape Dutch.

The source material in van Oordt (Reference Oordt1959–1962) contains six tokens with the auxiliary hebben that the archivist has detached from context and from which it is not possible to determine whether the doubled auxiliary bestows an anterior or remote reading or is itself a pleonasm. Our source material contains two isolated sentences in which the doubled auxiliary hebben is arguably a pleonasm; see (19–20).

In nine other occurrences the doubled auxiliary hebben either clearly bestows an anterior reading or may be reasonably assumed to do so.

Turning now to the doubling of zijn in the perfect, let us consider the three cases in which the lexical verb is mutative or zijn itself. A one-off escape in (23) precedes alleged enticement and flight in the sequence of events. Example (24) is cited out of context in van Oordt (Reference Oordt1959–1962), which renders moot a determination of the doubling’s semantic contribution. In (25), however, the act of abscondment occurs prior to the time in focus, which is the alleged misconduct on the part of two bandits.

In the perfect tense of passive constructions the auxiliary worden ‘become’ is replaced by the perfect auxiliary zijn ‘be’, as in present tense Het huis wordt vandaag verkocht ‘The house is (being) sold today’ beside perfect Het huis is vandaag verkocht ‘The house has been/was sold today’. As a variant of the perfective passive, some Dutch dialects have (26), in which addition of the past participle of zijn yields a double participle.Footnote ¹⁰

This use of geweest, which is ordinarily considered superfluous in Standard Dutch, is commonly heard in the southern Netherlandic dialects of Belgium, is scattered across North Brabant and Dutch Limburg, and is sporadically attested in Utrecht, North Holland, Groningen, Drenthe, Overijssel, and Gelderland (Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008a: map 41a). In Standard Dutch (26) is also possible, but it can only mean that the house had been sold but is no longer so; that is, it could be back on the market (Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:44). In our Cape Dutch source material there are eleven cases in which copular zijn combines with the past participle of the lexical verb and, tautologically, its own past participle geweest. Of these, four can be interpreted as a past participle in a passive construction with an implied agent, as in (27):

The remaining seven constructions refer to states of mind (‘frightened’, ‘intoxicated’, ‘aware’) or physical quality (‘armed’ in four occurrences) resulting from an event that is located in the past relative to another past event, as in (28).

Such cases should be analyzed as the perfect tense of an adjectival predicative construction denoting a property of the subject, not an event (cf. Barbiers et al. Reference Barbiers, van der Auwera, Bennis, Boef, De Vogelaer and van der Ham2008b:44), on the order of (29), and are not, sensu stricto, cases of auxiliary doubling.

Perfect auxiliary doubling is absent from Euro-Cape Dutch writings of a private nature. A later diary from 1836–1838, namely, that of Louis Trigardt (1783–1838), a prominent settler figure along the Eastern Frontier of the Colony and Great Trek leader, shows only an isolated instance (twice repeated) of a supercomposed periphrastic construction with zijn plus past participle of the lexical verb as adjectival and geweest.

To date I have extracted a mere thirty-two examples of perfect auxiliary doubling in source material spanning a period of approximately a century and a quarter (1712–1838). Seven of these data points are culled from the same document (CJ 373, 1760) memorializing, inter alia, the interrogation of an enslaved male.