STEREOTYPES IN APPLESAMMY AND NAIDOO

The text of a popular radio comedy series aimed at listeners in Natal in the 1940s, involving two Indian characters, is of potential importance in adding SAIE data for a period in which information about colloquial speech norms is rare. Evolving from earlier sketches on radio involving Indian characters in the early 1940s, it stabilized as Applesammy and Naidoo (henceforth A&N). The leading roles were played by the scriptwriters, Ray Rich and Dusty Cracknell. About their performances, the creator of the sketches, P. B. Durnford (1946:3–4) wrote:

The text of the programs was reproduced in a booklet The Adventures of Applesammy and Naidoo, carrying the subtitle A book for young people, based on the stories behind the popular broadcasts of “Applesammy and Naidoo.” The booklet comprises 25 short chapters or episodes detailing the misadventures of its two heroes. Much of the presentation is dramatic, with their direct speech making up approximately 564 sentences (5681 words).⁴

It is apparent from excerpts like these that the comedy is based on (and/or builds up) a stereotype of Indians, more specifically of descendants of Indian indentured workers drawn into an urban environment. Typical misadventures involving fishing, bee-catching, buying cars, horse-racing, selling flowers, encounters with the bureaucracy, and so forth produce stock comedy with great opportunities for stereotyping. The comedy is not subtle; it demeans the character's actions, attributes lowly motives to their actions and—more insidiously—draws a firm line between the ignorant and less than law-abiding Indian characters and their rather better-informed white characters and young readers. For example, the very first chapter ends with a disastrous fishing outing for the characters, “they had lost their rods and their tackle, and were wet and miserable” (p. 8). The concluding paragraph speaks to the young reader:

The text certainly deserves the attention of the historical sociologist. Despite itself, it can be seen as serving a serious function—perhaps a reassurance to the dominant classes of the time of their own superiority in the face of the growing urbanization of (and therefore possibilities of greater competition from) people originally destined for the plantations. An essential part of this reassurance is in the very choice of medium. The two characters (and a nephew Chinsammy, who makes a brief appearance in chapter 22) speak in English jargon or “broken language”; the authorial voice is always in standard English, as is that of all the white characters. By this means a dichotomy is set up between “authentic” voice and that of the “other.”

In Mesthrie (2002) I argued that A&N is a caricature, rather than a representation that linguists might draw on for data on SAIE, from a period for which little other linguistic data exists. This caricature is one that was doing the rounds in the British Asian Empire since the 19th century. It is also parallel to the caricatures of African Americans in the U.S., in the radio series Amos and Andy (Anthony Kroch, personal communication, 2002). Linguistically, some of the features have their roots in Butler English, a variety spoken by house-servants in British India—the butler being a rather less dignified figure than that of upper-class Victorian England. Butler English was documented by Schuchardt (1891), who claimed that English employers also used the jargon in conversing with their employees. Hosali (2000) showed that this pidgin-like variety survives to the present day in parts of south India. In Mesthrie (1990) I showed that pre-basilectal SAIE shows many similarities with Butler English, but that it probably arose independently in Natal, where it is not a coherently defined entity in the way that the basilect is. However, because the existence of a similar variety to Butler English has been reported in at least one other territory (Burma) where British settlers employed immigrant Indians, it could well have been transported as a kind of foreigner talk by British settlers in Natal with prior experience of India. An illustration of such foreigner talk occurs in George Orwell's Burmese Days (1934):

I have quoted this passage at length because the employer uses one of the constructions studied in this article, as a kind of foreigner talk (“can't keeping ice cool” for “I can't keep the ice cool”). This example of foreigner talk involves pro-drop and the “bare -ing” form. That is, -ing is preceded by a modal, rather than by be; and the Verb + -ing is really a nonfinite form.

It is in the nature of stereotypes that they are based in a small measure on reality. Apte (1994:4349) cited Brown (1965) who spoke of a “kernel of truth” hypothesis in the literature on stereotypes. Some of the linguistic features of the characters in the text do cluster in Indian English dialect, and survive in the speech of older, rural, and less-educated people in the early 21st century. On the other hand, Apte (1994:4348) affirmed that “a stereotype is now considered to be an overgeneralization and an uninformed one.” These remarks apply rather well to the portrayal of SAIE by the A&N satirists.

The baseline that I will use to illustrate the stereotyping effects in A&N is the English of the least competent speakers in my database of the 1980s. Such pre-basilectal speakers are difficult to follow, even for SAIE speakers who are thoroughly familiar with the rest of the dialect as a whole. Next I give an excerpt from a male pre-bas speaker:

The comparison between characters depicted in the 1940s and pre-bas speakers of the 1980s is not asynchronous, given that all six pre-bas speakers of my original database were elderly and loners (and therefore not likely to have undergone many changes in speech norms in their lifetime). They would have been around the age of 15 to 25 at the time of the A&N series—a perfect match with the characters satirized there. Linguistically, this group seems most appropriate for comparing SAIE usage with the language of the A&N scripts. Socially, however, there is a mismatch because the kind of adventurous persons who had regular (if unequal) contacts with whites would not have been pre-basilectal. Had Applesammy and Naidoo really been young, adventurous men living in the 1940s, they would probably be speakers of the basilect or lower mesolect and shift styles “upwards” in out-group communication with whites. In other words, the comparison I am making between pre-bas and A&N (rather than basilect or lower mesolect and A&N) is more than generous to the scriptwriter.⁵

In A&N, the two main characters have speaking parts in which -ing is applied as an ending to almost all verbs, irrespective of tense, mood, and aspect, as (8) shows.⁶

(8) But how me I can knowing when that bite he caming? (‘How could I have known when I got a bite’)

The range of functions of -ing in A&N is shown in Table 1.

The functions of -ing in A&N

All attestations of -ing in A&N are nonstandard. This includes the present progressives, which are standard in semantics, but nonstandard in form, because they lack a preceding be. Examples of each of these categories follow.

A&N thus gives the impression that its characters don't have a concept of tense, mood, and aspect. They use -ing simply as a marker of a verb, irrespective of whether it is progressive in semantics or not, and a fused object marker -it with got. Only three verbs are exceptions (see note 6 for these three and the fused form gottit). Table 2 (based on Mesthrie, 1992b) gives the use of -ing forms (with or without be) for the same functions among five pre-bas speakers by way of comparison.

The functions of -ing (with or without be) in pre-basilectal SAIE

Because -ing differs in its semantic function of progressive marker (with or without be) from its function as a bare suffix, it is necessary to separate these functions and examine them more closely. Table 3 details the use of (be) + -ing in these two functions in pre-bas and A&N.

The use of be + -ing and bare -ing in pre-bas and A&N

From Table 3 it is evident that A&N eschews regular forms for all categories of verb: almost all verbs have a Ø + -ing ending.¹⁰

Percentage of Ø + ing verb forms in relation to progressives in pre-basilect and Applesammy and Naidoo.

Percentage of Ø + ing verb forms in relation to nonprogressives in pre-basilect and Applesammy and Naidoo.

Figure 1A shows that the two databases do have something in common (the “kernel of truth” hypothesis): pre-bas does make high use of Ø + -ing forms in progressive contexts. At the same time, Figure 1A shows quantitative overgeneralization as the relatively high occurrences of Ø + -ing for progressives in pre-bas are made even higher in A&N. Figure 1B shows that pre-bas makes considerably less use of Ø + -ing in nonprogressive contexts. However, this is ignored in the A&N stereotype, which treats Ø + -ing in nonprogressive contexts the same as in progressive contexts.

Whereas Tables 1 and 2 give the raw figures for each environment (or function) for -ing, Table 4 gives the proportionate use of -ing to the verbs without -ing in each environment. Because we are in effect measuring the overgeneralization of -ing, this time it is necessary to factor out ordinary progressives requiring be + -ing in standard English.¹¹

A comparison of the proportionate use of nonprogressive Ø + -ing per verb function in A&N and pre-bas

Percentage of forms with nonprogressive -ing per tense category in Applesammy and Naidoo and the pre-basilect.

The overall difference between the two sets of bar graphs in Figure 2 is dramatic, revealing the exaggerative effect of the stereotype more fully than do Figures 1a and 1b:

Constructs from variation theory are useful in characterizing the nature of the stereotyping evident in A&N. It is clear from Tables 3 and 4 and Figures 1 and 2 that:

Furthermore the text as a whole promotes other effects, which are not presented in Figure 2:

The gross stereotyping text is thus an antithesis of natural language usage described in sociolinguistics (e.g., Bell, 1984; Chambers, 2003; Labov, 1972) in which variation is shown to be a normal form of linguistic behavior. My database shows that even pre-bas speakers of an L2 show variation in their grammar. Although they do not, by any means, have advanced abilities in the TL, the forms that do exist in their interlanguage are seldom invariant. Like the character, Ellis in Orwell's Burmese Days (1934) the authors of A&N attempt to suppress this variation and impose a unidimensional “otherness” to the speech of colonial subjects. It is interesting that the same grammatical form, -ing without be, should have been targeted. Orwell's character insists, “‘Please, master, can't keeping ice cool’—that's how you ought to talk.” Yet not a single occurrence of -ing in this context (after modals) occurs in my pre-bas data, despite 32 opportunities for such use.

Yaeger-Dror (1992) characterizes any instance of an individual using a linguistic variant more than those he/she is emulating, as a form of hypercorrection. Since “correction” towards a prestige norm is not always evident in studies of traditional (quantitative) hypercorrection of the sort described by Labov (1966), Yaeger-Dror proposed the more general term hyperaccommodation, based on the account of code convergence and divergence between interlocutors put forth by Giles, Bourhis, and Taylor (1977). Following from this account, Janda and Auger (1992) described various types of hypercorrection/hyperaccommodation, including humorous hypercorrection, which is of relevance here. They characterize it as a striking kind of downward hypercorrection, occurring when speakers in the emulating group have higher social status than members of the emulated group. Janda and Auger (1992:209) characterize this “mocking process” following from qualitative hypercorrection in the following way:

For examples from African American Vernacular English (AAVE), see Janda and Auger (1992:209). Baugh (1992) used the term hypocorrection in the case of adjustments “downwards” toward a less prestigious dialect variant by speakers of a more prestigious variety. The type of hypocorrection he described is not done for humorous reasons. His example is of African Americans brought up on standard American English attempting to deploy AAVE as a second dialect in certain contexts.

Although Janda and Auger's four-step account of humorous hypercorrection does apply broadly to the A&N text, the phenomenon underlying the parody involves so deliberate and invariant a process that neither the term hypercorrection nor hyperaccommodation convey the right nuance. The most appropriate term for the gross stereotyping in A&N seems to be “mock language”—that is, the use of a lower status language or dialect in an exaggerated or distorted way by speakers of a dominant variety, with the intention of stereotyping and parodying speakers of the dominated variety. Jane Hill (1993) used the term “mock Spanish” for the appropriation of presumed linguistic features of Spanish by English speakers in the southwestern United States in casual speech. Hill provided the example of Hasta la vista, baby, which parodies and devalues the formal Spanish phrase hasta la vista (literally, “until—the—seeing”), which signals a sincere hope for the pleasure of a future meeting. The juxtaposition with the slang term baby renders the original phrase colloquial and vulgar, as does the exaggerated intonation in which the phrase would be rendered. Similarly, Ronkin and Karn (1999) used the term “mock Ebonics” to describe the vast number of materials on the Internet that parody Ebonics or African American Vernacular English.

There is a further violation of “naturalness” in A&N. In Mesthrie (2002) I show that A&N is an odd mixture of certain early-interlanguage features used frequently in the interests of stereotyping (e.g., -ing and the fixed, fused verb-form gottit, see note 6) with more advanced features (e.g., a range of prepositions and plural nouns) dictated by the need to be intelligible to its listeners. One should beware of circularity here, if early interlanguage features simply means “nonstandard forms more characteristic of A&N than pre-bas,” and advanced feature means “standard forms more characteristic of A&N than pre-bas.” There has to be an independent metric to decide which forms occur in early acquisition and which occur later. Such independent evidence can be found in studies of L1 and L2 acquisition. Ellis's summaries (1994:78–100) make it clear that (1) -ing is a form that occurs first (i.e., before other verb inflections) in child language and in L2 acquisition; (2) that among prepositions, in and on are acquired first; (3) that uncontracted copulas occur before contracted copulas (e.g., are before 're; is before 's); and (4) that can't occurs prior to, and is more frequent than, can. In all these respects, A&N is a poor reflection of a specific interlanguage level of its characters. There is a reason for this lack of naturalness, which I term “the interests of the text.” If A&N were really to reproduce the norms of pre-bas (or even the basilect), it would be unintelligible to a majority of its listeners or readers. As I indicated in Mesthrie (1992b:65), the pre-bas is difficult to follow, even for people familiar with SAIE as a whole. It is little surprise then that there are some features in which A&N performs in a more TL way than pre-bas. I illustrated this in relation to prepositions and noun plurals in Mesthrie (2002).

DIALECT REALISM IN THE LAHNEE'S PLEASURE

I now turn to The Lahnee's Pleasure (henceforth LP), a play written by Ronnie Govender, a member of the SAIE community. The play aims to faithfully represent the L2 working-class dialect of the mid-20th century. The broader question is whether a native speaker-writer, knowledgeable about the speech norms of a community, might nevertheless still be susceptible to the process of linguistic overgeneralization. In other words, is a degree of stereotyping inevitable as creative writers seek to impose some order on the multifarious utterances that speakers are capable of in everyday speech? The play LP is set on a sugar estate of the Natal North Coast in what appears to be the late 1960s.¹³

This excerpt draws largely on basilectal speech, emphasizing the norms of working-class people who evolved a variety of English radically different in many respects from the more standard varieties of colonial English spoken by the ruling classes and promulgated in classrooms. It is of some interest to ascertain exactly how authentic the representation of language in the play is. I will first comment on this issue broadly, before providing statistical comparisons with the categories analyzed in the A&N text. The playwright is sensitive to language variation of several sorts. Although the main character, Mothie, speaks mainly basilectal SAIE, his speech shows occasional style shifting, as one would expect. Thus, he switches to an acrolectal form (more or less standard) on p. 8:

(19) (Proceeds to table and says loudly) If I want to drink wine, it's my business.

This is an act of code divergence, in defiance of the newcomer in the bar, obviously an outsider (of Indian origin) looking “with kindly amusement” at Mothie's antics. Later (p. 11) Mothie repeats the same sentence aggressively to the barman:

(20) If I want to drink—my business.

The deletion of the dummy subject it and of the copula verb is in (20) is more in keeping with basilectal norms, showing a convergence back to the norms of the bar room. Another piece of variation involves Mothie's encouraging words to his young son, who is offered a soft drink by the barman:

(21) Go, uncle giving cold drink, go. (‘Go and receive the cold drink that the barman is offering you.’)

With the absence of the object pronoun you and the article a after the main verb giving, this counts as a piece of downshifting typical of adults speaking to children, signifying a kindly cajolement. In Mesthrie (2003) I showed how the basilect and pre-basilect is understood by working-class parents to be an appropriate variety for addressing children.

Another sociolinguistic point of relevance is that Mothie's language is basilectal, but does not include much slang.¹⁴

When the boss himself makes an appearance, his language involves a kind of talking down, in the direction of the mock language associated with the A&N stereotypes:

(22) Now you going to have one big, big wedding. Don't forget to bring me some curry and rice…. Hey Mothie have one nother one wine …. and give those chaps a drink too … (p. 41).

However, neither his language nor his attitude shows undiluted mockery. There is a gesture of goodwill in the offer of drink and the request for food from the wedding. Likewise, two of his utterances here are standard (Don't forget to bring me some curry and rice and give those chaps a drink too), rather than mock language. His attempt at joviality here is tinged with slight condescension, conveyed by a version of SAIE that unwittingly contains several mistakes. First, in the phrase big, big wedding, reduplication of the adjective big is used inappropriately with a singular noun. Compare this with Mothie's earlier utterance (p. 7):

(23) What big, big prayers we'll have.

Mothie's utterance in (23) shows the appropriate context for the reduplication, before plural nouns.¹⁵

I now examine the use made by the playwright of the be + ing construction in the utterances of Mothie. The broader questions, as already mentioned, are whether a degree of sociolinguistic veracity can be found in Mothie's use of a specific grammatical construction or whether even well-intentioned authors fall prey to stereotyping in this respect. Table 5 compares the statistics for A&N and pre-bas in Table 4 with those of LP and the basilect.

A comparison of the proportionate use of nonprogressive Ø + -ing per verb function in A&N, pre-bas, basilect, and LPIn this count I excluded present progressives with future meaning, as in I'm coming just now, which I place under present; futures with modal + be + V + -ing (I'll be standing next to you, man) are excluded.

For three categories (future, modal aux, and imperative) LP shows zero use of -ing, in keeping with the pre-bas and basilect, and unlike LP. In a fourth category (infinitives), LP reflects the zero usage of the basilect, rather than the sporadic use of pre-bas. The use of Ø + -ing in LP forms thus closely matches that of my basilectal sample, with one exception: The playwright eschews Ø + -ing for the simple present and past. The reason for this is that these are residual forms of the basilect (more characteristic of the pre-bas, which has overtones of being “broken”). The proportions for the basilect, as given in Table 5, are in fact 14/206 (6.8%) and 3/417 (0.7%). LP ignored this residue. Instead, the LP text favored a more sophisticated use of Ø + -ing for past habitual irrealis progressives (cf., everybody looking for us for “everybody would be looking for us” in the LP passage cited earlier).

Figure 3 presents the information given in Table 5 graphically. Figure 3 confirms that LP and A&N are mirror opposites for the nonprogressive Ø + -ing forms. A&N blows up the small number of such forms from pre-bas and basilect into a categorical presence. LP, on the other hand, ignores this small number of forms in favor of categorical absence. Hence, LP does not show up in the graph for Figure 3 at all. However, LP does not ignore dialect use of be + -ing. This is reflected in Table 6, which shows the proportion of progressive -ing forms without be to the number of -ing forms with be.

Percentage of forms with nonprogressive -ing per tense category in Applesammy and Naidoo, The Lahnee's Pleasure, the basilect, and pre-basilect.

Proportion of progressive Ø + -ing verb forms to total number of progressive verb forms with be + -ing or Ø + -ing in two texts and two lects

Table 6 shows how close Mothie's norms are to the basilect for progressives, rather than to pre-bas and A&N. This is a socially authentic portrayal of someone with minimal education, with community-based social networks, and little or no social contact with L1 English speakers. Tables 5 and 6 thus show that a native-speaker playwright with “realist” intentions can replicate a community's nonstandard norms with a high degree of accuracy. This includes representing the community's variable rules (in this case, for the use of the two types of -ing), as well as showing the effects of “audience design” (Bell, 1984), as Mothie style-shifts in accommodation to different speakers. The “interests of the text” are less evident in Mothie's speech, apart from his use of occasional past habitual used to rather than should. Here, it is possible that the playwright was constrained by the difficulties that should might pose to a wider audience. SAIE is the only dialect of English world-wide to use should (variably, but quite commonly) for the past habitual “used to.” There is another sense in which the interests of the text are invoked, that is, in the play as a whole. There is a necessary contrast between the standard speech of the Indian stranger returning from England, and the basilectal norms of the main character. The standard speech acts as something of a counterpoint to Mothie's language. Some standard speech is necessary to keep a sense of balance in order to make this a creative work with a reasonably broad appeal, and not simply a dialect tract. The standard speech also reminds the audience that not all Indians are basilectal. That is, LP avoids the social overgeneralization evident in A&N.

TRANSCENDING DIALECT REALISM—THE WEDDING

I now turn briefly to the final text, The Wedding (henceforth TW), a novel by Imraan Coovadia, also a member of the SAIE speech community. The novel's plot is simple: It deals with the chance meeting of a city man and a village woman in India, their hurried wedding (against the woman's wishes) and their eventual journey to a new life in South Africa. More anti-romance than romance, this tale is told in an effusive style, described on the front cover blurb by 2003 Nobel laureate, J. M. Coetzee as “prose of dazzling comic wizardry.” Much of the novel is in the form of dialogue and the free indirect style of the characters. An authorial voice using standard English is rather less evident. There is one way that this text differs from A&N and LP: The English is not meant to represent English as such (though the author is silent about this). Rather, it is intended as a representation of the L1 of the characters (either Marathi, Gujarati, or the Meman dialect of Sindhi). The following excerpt deals with the attempts by a poor and miserly father-in-law to get his new son-in-law to accept two train tickets to Hyderabad as a wedding gift. This show of generosity is mitigated by the fact that he did not have to pay for the tickets, and that Hyderabad—an unlikely honeymoon spot—is in the wrong direction from the newlyweds destination, Bombay. In the following extract, Yusuf is the father-in-law and Ismet is the son-in-law besotted with his new village-girl wife, Khateja, who does not reciprocate such feelings in the slightest.

Dialect uses of -ing in TW are woven into a wider fabric of poetic language. The effect of such transformation of dialect into “poetic diction” is not to stereotype (as in A&N) or to be a faithful representation of a particular dialect (as in LP), but to transcend the particular. In this regard, the most noteworthy feature is the absence of bare -ing forms. Whenever -ing occurs in TW, it is accompanied by some form of be. Examples of -ing in the TW text are therefore entirely of an aspectual nature, capturing the well-known Indian English predilection for extension of progressive -ing to habitual and stative contexts (Trudgill & Hannah, 1985:110). Some examples of be + -ing in stative contexts are given next (all citations are taken from TW, chapt. 3):

(24) And he is still wishing to marry me.

(25) We are wanting you to look your best.

(26) You are looking so handsome.

It is also noticeable that Coovadia exploited the ambiguity in some be + -ing forms, where the construction can be taken as nonstandard or just about acceptable as colloquial standard. However, because from the viewpoint of formal standard English these do not take be + -ing in the sense intended, they are counted as habituals (27–28) or stative (29), rather than as progressives in Table 7.

(27) She is always talking about it.

(28) That's what they are saying.

(29) I'm looking forward to meeting it.

As Table 7 shows, the use of be + -ing is outnumbered by the use of simple past and present forms. TW has no instances of be + -ing for future, simple past, perfective, or habitual past. It does, however, occasionally use be + -ing after imperatives (30–31) and modals (32), with special effect:

(30) Better be kissing that chicken goodbye.

(31) Do not be forgetting us, eh.

(32) Can't be squandering the minutes…

The effect of sentences (24) to (32) is to stylize the Indianness of the thought processes, going back to the Marathi/Gujarati/Sindhi of the speakers, rather than to stigmatize incompetent English via bare -ing forms. This effect is strengthened by having the vast majority of these forms used variably with the standard form predominating, as shown in Table 7.

Use of nonstandard be + -ing versus standard forms of verbs in The Wedding (chap. 3)

Standard present progressives (n = 5) and past progressives (n = 2) with be + -ing in TW are excluded from Table 7, but are included in Table 8, which gives a comparison of the proportion of be + -ing progressives to all progressives in the varieties studied in this article.

Use of be + -ing progressives in three texts and two lects

Table 8 shows a continuum in the use of be + -ing in all five varieties studied. A&N, which draws on and exaggerates pre-bas norms, steadfastly refuses to bestow any tokens of the standard variant on the speech of its main characters. At the other end, TW joyously acknowledges the use of be + -ing with all progressives (Table 8) and with some habituals/statives, as well (Table 7), in keeping with the norms of Indian English and SAIE. Intermediate figures occur in pre-bas, basilect, and LP, reflecting an interplay between bare -ing forms and be +-ing forms for the progressive. This situation is reflected in Figures 4A and 4B, which draw on Tables 5 and 6. These figures summarize the use of Ø + -ing in progressive use and nonprogressive use, respectively, in all the varieties studied in this article.

Percentage of past and present progressive Ø + ing verb forms to total number of progressives in three texts and two lects.

Percentage of past and present nonprogressive Ø + ing verb forms to total number of verb forms in three texts and two lects.