3.1 When the can indeed be used for possessive determiners in StdIntE

A fitting background to the discussion of the use of the definite article the for possessive determiners in the examples above is the discussion by Quirk et al. (Reference Quirk, Greenbaum, Leech and Svartvik1985: 270–272) of instances where, in StdIntE usage, the can indeed be used instead of possessive determiners. Below is a summary of the authors’ discussion:

Now, a close look at all the examples in (1) to (23), from our three KenE data sources will show that none of them refers to parts of the body and that none of them can be attributed to the ‘masculine style of speech’ either. Notice that in the examples in Quirk et al.'s (Reference Quirk, Greenbaum, Leech and Svartvik1985) discussion this latter possibility is illustrated by utterances directly addressed to someone, which is not the case with any of the 43 instances contained in examples (1) to (23): all of them refer to a third person, not to an addressee. For this same reason, none of them can be argued to be instances of ‘the habit of some men of referring to their wives, or children, by the’. So, all in all none of these 43 uses of the is expected to be idiomatic in StdIntE.

Biber et al. (Reference Biber, Johansson, Leech, Conrad and Finegan1999: 269) note that both the definite article and possessive determiners (together with demonstrative determiners) are ‘definite determiners’, in a paradigmatic relationship, and, hence, can be substituted one for another. However, the authors specify that ‘[t]he definite article is a more neutral determiner than the possessive forms, in that it marks an entity as known without specifying how it is related to other entities,’ while ‘[p]ossessive determiners are particularly associated with human beings, and they characteristically serve to identify entities by their relationship to human beings’ (p. 270). It becomes intriguing, therefore, why in KenE usage possessive forms are often replaced by the definite article in contexts where reference to human beings is involved. In fact, 84% (i.e. 36) of the 43 highlighted uses of the in (1) to (23) involve (direct) kinship nouns: father, mother/adoptive mother, son/sons, daughter, the stepbrother, the wife, and the husband (excluding the non-gender-specific nouns parents, child/children, orphans, and relatives).

3.2 Possible reasons why the is used for possessive determiners in KenE

The first reason to think of is that already alluded to by Sand (Reference Sand2004: 286), namely the influence of substrate languages, notably Swahili. However, Swahili influence, for example, cannot be an appealing explanation because Swahili does not have articles. And neither do most (or maybe all?) of the other indigenous languages of Kenya. Consider example (24), extracted from example (5) above.

1. (24) I looked at the father . . .

   Nilimuangalia baba yake . . .

In this Swahili translation, the equivalent of the father is baba yake , literally ‘father of his’. No other translation of it would be correct. In relation to this, in all the illustrative 43 phrases in examples (1) to (23), if the Swahili structure were to be transferred into English, we would expect to see an obligatory possessive determiner in all of them. Therefore, a different explanation, one not drawing upon substrate influence, should be sought for why KenE uses e.g. the father for his/her father.

One plausible explanation has been suggested to me by my colleague Zipporah Otiso (p.c., 21 Dec 2022). She argues that KenE speakers use terms such as the father and the mother instead of his/her father and his/her mother because using a possessive determiner would be semantically redundant, as it is clear that in a sentence like (25) the NP the father can only refer to John's father and not to the speaker'sFootnote ², and that in (26) the NP the mother can only refer to Mary's mother, and not to the speaker's.

1. (25) I met John and the father.

2. (26) I met Mary and the mother.

So, she continues, there is no need to redundantly signal the relationship between the two people referred to. Zipporah Otiso's suggestion implies that in KenE speakers assume that both the and possessive determiners (his/her/their/its) have the same grammatical meaning, namely ‘possession’, in a phrase like the mother, presumably as long as this is not postmodified (as in I met Jane and the mother of Andrew). Her suggestion is a reasonable proposition. What remains to explain is how this assumption (gradually) took root in KenE usage.

The third possible explanation (for the use of the mother for his/her/their mother) constitutes a plausible answer to this how: it is that such usage developed from classroom instruction. In this regard, Mwangi et al. (Reference Mwangi, Indangasi, Mukunga and Gecaga2011: 40–41), while advocating the use of ‘gender-neutral language when the gender is unknown or could be either masculine or feminine’ in their book meant for secondary schools in Kenya, give the following advice as one of the options: ‘Eliminate the pronoun’, as in ‘A careful writer proofreads the work before submitting it’. It is obvious that the article the in this sentence has replaced his, her, or their. So, it is not difficult to see how such advice appearing in a textbook could influence usage. However, given that Mwangi et al.'s (Reference Mwangi, Indangasi, Mukunga and Gecaga2011) advice is relatively recent, it might not be enough to explain the uses of the for possessive determiners in AGoW, a novel published in 1967, that is four decades earlier. Nevertheless, that advice is given support by anecdotes like the following: when I asked a former undergraduate student of mine (now in his forties) why he had just used the mother rather than his mother, he answered: ‘That's what I was taught by my primary school teacher, Mrs Maina’ (Tom Kituku, p.c. 26 Dec 2022). To paraphrase it even more appropriately, that is what he and other pupils picked up from their teachers, who themselves picked it up from their own teachers, without being able to determine when the usage started becoming Kenyan, though.

3.3 The and possessive determiners used as variants across the three data sources

Whether any of the possible reasons suggested in the preceding section for the use of the for possessive determiners is solid enough or not may not be so important, after all. More important is the fact that, as the examples in (1) to (23) show, some (maybe all?) KenE speakers use both the definite article and possessive determiners in their speech. That is the case of the use of both the step brother and his father by the speaker in example (7) (from ICE-K), of the use of both the father and the mother vs. their African compatriots in example (13) (from GloWbE-KE), and of the use of both the mother and his father and his mother in example (16) (from AGoW). So, for these speakers, both the article the and a given possessive determiner co-exist in their idiolects and are used as variants. And it is the study of this user variability that is of greater interest.

But by way of background to an analysis of this variability, an overview of how the frequencies of occurrence of the two variants compare across our various data sources is needed. To this end, the pair of phrases his father vs the father and his mother vs the motherFootnote ³ will be used for illustration because they already appear in examples (1) to (23) more frequently than all the other pairs. For its part, the possessive determiner his will represent all the others (her, my, our, your, their, its) because corpus data reported by Biber et al. (Reference Biber, Johansson, Leech, Conrad and Finegan1999: 271) show that it is the most frequent of all. Table 1 gives the relative frequencies of the phrases concerned, normalized per million words (pmw).

Table 1. Occurrences of his father vs the father and of his mother vs the mother pmw in AGoW, ICE-K and GloWbE-KE

To put the relative frequencies in Table 1 into their right perspective, here are the sizes of the respective sources: a) the novel AGoW is 87,686 words-long (this is its main text, excluding all introductory material); b) the written component of ICE-K that will be considered in this study is 361,724 words-longFootnote ⁴; c) the spoken component of ICE-K is 289,625 words-long (Hudson–Ettle & Schmied, Reference Hudson–Ettle and Schmied.1999: 56); d) GloWbE-KE is 41,069,085 words-long (Davies & Fuchs, Reference Davies and Fuchs2015: 6).

Table 1 shows that in AGoW, the definite-article variant consistently occurred much less frequently than its possessive-determiner counterpart (that is 34 vs 125 and 57 vs 319), regardless of whether the noun used was father or mother. The picture is partially different in written ICE-K: while the definite-article variant the father was less frequently used than its possessive-determiner counterpart his father (that is 24.9 vs 55.3), the variant the mother was much more frequently used than its counterpart his mother (135.5 vs 44.2). (It is intriguing why the choice of either mother or father mattered significantly.) In spoken ICE-K, the picture becomes consistent again, but this time round the definite-article variant consistently occurred more frequently than its possessive-determiner counterpart (that is 48.3 vs 20.7 and 200.2 vs 31.1). Finally, in GloWbE-KE, the picture is also consistent, but in the opposite direction: in both cases the definite-article variant occurs less frequently than its possessive-determiner counterpart, although the differences between the respective figures (namely 23.4 vs 28.8 and 20.4 vs 22.2) are most likely not statistically significant.

One overall observation that emerges from Table 1 is that the frequencies for both variants are lower in GloWbE-KE than they are in all the other three data sources, except for one comparison slot (that in the case of his father when GloWbE-KE is compared to spoken ICE-K). Besides, not only are they lower, they are also, as already suggested, seemingly not statistically significant. From this double difference it could be claimed that despite GloWbE-KE being by far the largest of all the data sources considered, it is the least representative of KenE usage. And it seems to be so for the following reasons: first, in relation to both AGoW and written ICE-K, GloWbE-KE is composed of numerous, but only online texts, ‘[a]bout 60 percent of [which] come from informal blogs, whereas the other 40 percent come from a wide variety of (often) more formal genres and text types’ (Davies & Fuchs, Reference Davies and Fuchs2015: 4). Second, in relation to spoken ICE-K, GloWbE-KE is obviously different from it since it is all written while the other is all spoken. To this extent, it is puzzling that GloWbE-KE should contain much fewer instances of the definite-article structure (23.4 pmw for the father and 22.2 pmw for the mother) than spoken ICE-K (48.3 for the father and 200.2 pmw for the mother). One should expect any written corpus (even if it were composed of informal material) to contain fewer definite-article variants than any spoken corpus, since the father and the mother are assumed to be less international-English-like than his father and his mother in the logic of the present study. But while some of the frequencies of the two variants may seem to be puzzling, others do not: for example, the very high frequency (200.2 pmw) for the definite-article variant the mother (vs. only 31.1 pmw for his mother) is most likely attributable to the spoken nature of spoken ICE-K.

Another overall observation from Table 1 is that the frequencies for the possessive-determiner variants his father (125 pmw) and his mother (319 pmw) are much higher for AGoW than for any other data source. Such variation can easily be attributed to the written and edited nature of the novel. But it could also be explained from a different, diachronic perspective, in the following way: when the book was written in the mid-1960s, KenE was still in its first decade of its ‘nativization’ phase (to make reference to Schneider's (Reference Schneider2003, Reference Schneider2007) Dynamic Model for the development of New Englishes), according to e.g. Hoffmann (Reference Hoffmann, Schreier, Trudgill, Schneider and Williams2010: 290), who suggests that this stage started from the ‘late 1940s’. So, it could be argued that some of the structures that can today be called stable KenE features were just nascent then.Footnote ⁵ And much as authors such as Schneider (Reference Schneider2007: 196) and Hoffmann (Reference Hoffmann2021: 26) maintain that KenE is still in its nativisation phase, it is only reasonable to assume that each passing decade (three decades later, i.e. in the early-to-mid-1990s, in the case of ICE-K and, five decades later, i.e. in the early 2010s in the case of GloWbE-KE) entrenched the definite-article variant (the father/the mother) as a KenE feature.

3.4 Individual variation in the use of the for possessive determiners

The variation in the use of the for possessive determiners observed across ICE-K, GloWbE-KE and AGoW was attributed to ‘macro’ factors: medium (spoken vs written) and text type (informal blogs vs. formal texts). But, beyond this variation, it can assertively be re-stated from the frequencies reported in Table 1, with the lowest frequency being as high as 20.4 pmwFootnote ⁶, that both the possessive determiners and the are used as variants for the same grammatical meaning, ‘possession’. Arguably, though, more illuminating, in terms of really establishing the extent to which the use of the for his/her/their/its is entrenched in KenE usage, is to study the extent to which individual KenE speakers use both variants, and whether they use them in systematic or free variation, to borrow terminology from interlanguage variability studies (see e.g. Ellis, Reference Ellis1999, Reference Ellis2015). ‘Systematic’ variation would mean that the choice of the over his/her/their/its is dependent on linguistic, sociolinguistic, or psycholinguistic factors (Ellis, Reference Ellis1999: 464).

To illustrate this variability within an individual's production, four cases extracted from our data sources will be analysed. The first extract, in (27), is an excerpt from AGoW (p. 201), the only paragraph in the book where both his father and the father are used:

1. (27) But unexpectedly the day arrived. Muhoya, a young man newly circumcised, had come home and found his father at his favourite game. Suddenly the young man felt the moment had come. ‘If you value your life,’ he cried, ‘don't touch her again.’ At first the father was so surprised that his hand became numb in the air. Had he heard aright? He fell into a lion's rage. He lifted his hand to strike the boy, but Muhoya caught his father by the arm. The years of hatred and fear made him delirious with a fearful joy. Father and son were locked in a life-and-death struggle. The son did not see a father, but a perpetrator of unprovoked violence, a petty colonial tyrant who would extort money from even his closest relatives. And his father saw not a son, but a subject who had refused to be a subject. (p. 201)

In this extract it would be hard to attribute the single occurrence of the father to a predictable linguistic context. This cannot be its syntactic position (almost) at the beginning of a sentence, nor can it be its function as a subject, both of which it shares with his father in the last sentence. It is equally hard to see why where the father was used its counterpart his father should not have been used, instead. Furthermore, in the third occurrence of his father (in the last sentence), one can argue that the father should have been used instead of his father, to contrast with the son in the preceding sentence. All in all, it would be reasonable to conclude that the author used both his father and the father in free variation, though the former more frequently than the latter.

The second case, in (28), is an extract from written ICE-K, by author W2F017K. The passage comes from a published short story.

1. (28) Several times I found Gachara in a foul mood. <#/>He had told me his father had started the wrestling game after which he had pinned him down with the legs resting one on his head and the other on his shoulders. [ . . . ]

   <#/>It happened on a Saturday afternoon as we all sat having some tea. <#/>Gachara was sitting with the baby on the floor when the father casually moved and sat on the chair next to the two. <#/>they were so busy playing that Gachara did not notice his father approaching.

In this extract, as in that in (27), it would also be difficult to argue that the father, rather than his father, was used because of the linguistic context around it. Of relevance is the fact that the father, like his father used before it, is a subject in a dependent clause.

The third case, from GloWbE-KE, appears in (29).

1. (29) As for those from the immigrant Asian and European communities, it was felt that the longer they had been settled there, the greater would be their sense of attachment to the land and its people and so those of them who were, at the least, second generation born, whether through the father or the mother, were put on the same footing as their African compatriots. (awaazmagazine.com)

What is notable in (29) is the use of two successive definite-article structures, the father and the mother, and then a possessive-determiner one, their African compatriots. One wonders why the author, after writing through the father or the mother, did not continue to use the article the, and, thus, write through the African compatriots. Here there seems to be a case of an identifiable linguistic context, because, in (29), unlike in (27) and (28), the noun father contrasts with compatriot, which is not a kinship term. Based on this case, and on the fact that in 84% of its 43 occurrences highlighted in bold type in examples (1) to (23), the appears before the kinship terms father, mother, son, daughter, brother, sister, wife, and husband, it can be concluded that the use of the definite article for possessive determiners is predictable before kinship terms. In this case we can say that the variability in the use of the for possessive determiners is systematic.

The fourth illustrative case is an entire 1743-words-long ‘monologue’ by speaker S2B021K in spoken ICE-K. It is one of the ‘broadcast talk’ texts (see Hudson–Ettle & Schmied, Reference Hudson–Ettle and Schmied.1999: 18). (Note in passing that example (4) was extracted from the same text.) Summary information on the frequencies of occurrence of different kinship terms in this monologue will be enough to make the point: the mother: 20; the father: 7; his father: 0; her father: 0; his mother: 0; her mother: 0; this mother: 5; her husband: 1; the husband: 3. Clearly, for the speaker in this case the figures are unambiguous: 0 frequencies for the possessive-determiner variants his mother, his father, her father and her mother, against 20 for the mother and 7 for the father. In fact, one would have been tempted to conclude that this speaker (who is a female) does not have the possessive-determiner option in her idiolect at all, if it were not for the sole occurrence of her husband. Still, her case is strong evidence of how entrenched the use of the for possessive determiners is for some KenE speakers: for her, hardly can we talk of the two structures being variants of each other. But if we were to do so, we would conclude that the use of her husband was random, and, then, that the author used the two variants in free variation, although with the definite-article being by far the more frequent.

It transpires from the four illustrative cases that individual variation in KenE usage of the for possessive determiners is both systematic and free: systematic in relation to the semantic field of the noun (whether a kinship or non-kinship term) to take either the or a possessive, and free in relation to using either the or a possessive before kinship terms.Footnote ⁷ However, discussion of the fourth illustrative case shows that this free variation masks an uneven distribution of the two variants, with the definite article being the more likely of the two. This reinforces the suggestion made earlier that there seems to be strong collocation between the definite article and kinship terms in KenE usage. Here is further illustration:

1. (30)

   1. a) Jane came with the father

   2. b) Jane came in the car

While the father in (a) automatically means ‘Jane's father’, the car does not (automatically) mean ‘Jane's car’. In fact, it is highly unlikely that KenE speakers would simply say Jane came in the car without specifying whose car it was.

This apparent collocation between just an article and a noun is rather unusual because English collocations typically involve lexical words: verbs, nouns, adjectives, adverbs (see O'Dell & McCarthy, Reference O'Dell and McCarthy2008: 10). Still, it seems that from one generation of KenE speakers to another, phrases like the father and the mother have been transmitted as ‘formulaic sequences’ (Keck & Kim, Reference Keck and Kim2014), of the kind of ‘constructions’ described in Hoffmann (Reference Hoffmann2021). And in the way those phrases have been stored in their users’ brains, the typically means his/her/their/its, as further illustrated by (31):

1. (31)

   1. a) Andrew came with the mother.

   2. b) ?I came with the mother.

In (a) the mother automatically means ‘Andrew's mother’ while it does not in (b), where my mother, not the mother, is actually expected in KenE. This association of third person with the can be considered another aspect of systematic variability. In this regard, it is instructive to note that none of the uses of the father/the mother identified in examples (1) to (23) has a first-person or a second-person reference. The phrases my mother/my father and your mother/your father appear on their own in the respective sources. (For example, in AGoW there are 21 occurrences of my mother, 5 of my father, 5 of your mother, and 3 of your father.)