Inscribing Difference: Code-Switching and the Metonymic Gap in Post-Colonial Literatures

In their seminal work The Empire Writes Back Ashcroft et al. (1989) identify code-switching between two or more codes in post-colonial literary texts as ‘the most common method of inscribing alterity’ (p.72). Ashcroft (2001) further develops the idea of installing cultural distinctiveness in the text and posits that, together with a wide range of other linguistic devices (e.g. neologisms, ethno-rhythmic prose), the use of code-switching – whether between the variants of the same language or between languages – has a metonymic function to inscribe cultural difference. In this chapter, I will examine the hybrid nature of post-colonial literary texts through the concepts of nativisation (Kachru, 1982a, 1982b, 1984, 1986, 1987, 1995) and indigenisation (Zabus, 1991, 2007). I will then focus on code-switching, adopting Myers-Scotton’s (1993) approach of matrix language vs. embedded language and considering that ‘EL [embedded language] material of any size, from a single morpheme or lexeme to several constituents, may be regarded as CS [code-switching] material’ (p.5). I will analyse examples of code-switching taken from modern Ghanaian English-language novels and short stories, and I will argue that a synecdochic relationship exists between the code-switched embedded language and the culture it originates from. I will contend that it is along the metonymic gap thus created by language variance that readers can be expected to be divided. I will briefly examine the types of authorial assistance that can be provided in order to make the text accessible to the reader, and I will illustrate, in Sperber and Wilson’s (1995) relevance theoretical framework, how different groups of readers cope with code-switched language left in the texts untranslated and/or unexplained. I will argue that by withdrawing assistance from the reader, the author makes it manifest that he concedes ‘the importance of meanibility’ (Ashcroft, 2001, p.76) and opts for the inscription of difference. I will conclude that the metonymic gap is not a simple bi-polar concept between coloniser and colonised culture but a multi-layered entity where the readers’ position in relation to the gap is indicative of their ability to interpret code-switched language unaided. Full appreciation of the writer’s meanings is shown by those readers who share both the writer’s cultural and linguistic experience. Other readers may be able to cross the metonymic gap to various degrees, but for them code-switched language will be the symbol of the writer’s difference of experience.

cultures, it stands as a hybrid product. Unlike other literary genres, it has no "predecessor" in the oral traditions of the peoples of Africa; it is a relatively recent and completely foreign import into the African literary scene. As Zabus (2007, p. 4) observes, the African novel is "looking 'inward' into African orature and literature and 'outward' into imported literary traditions. It is language, however, that has deflected the African novel [and by analogy, African europhone creative prose writing] from its inward course." When an African writer -or a writer in a similar sociolinguistic context in other parts of the world -decides to employ a historically alien language for literary creation, they are bound to be confronted by a conflict inherent in such a choice. The consequence for the written literary code is the African writers' attempt to mould the foreign language so that it can convey their cultural and linguistic experience, "thereby redefining and subverting its foreignness" (Zabus, 2007, p. 4), while operating "outside the boundaries of either [their] own society or that of [their] adopted language" (Nkosi quoted in La Pergola Arezzo, 1988, p. 40).
The process of moulding the European language, in our case English, to express African, Indian, Caribbean or other culturally and linguistically non-Anglo-Saxon vision is identified by Kachru (1982) as nativisation and by Zabus (1991Zabus ( , 2007 as indigenisation of the European language. It is "a process of linguistic and sociolinguistic change through which an external language becomes part of the culture of a community that uses it as an additional language, while it still retains many features of the language as it is used by [its] native speakers" (Owusu-Ansah, 1997, p. 24). In Ghana, for example, in the spoken language a continuum exists from the unmixed Ghanaian languages at one end to the unmixed English at the other end, both used in highly formal situations (e.g. the Asantehene's address to the state or board , 2007Ker, 1997;Booker, 1998) and African scholars and writers (e.g. Abiola Irele, 2009;Achebe, 1988) alike as a superordinate term to refer collectively to novels written by writers from Sub-Saharan Africa in a European language. The problematic nature of such generalisations is acknowledged here, but its examination falls outside the scope of the current discussion. meetings, respectively), with mixed varieties in between. In the written language, however, Ahulu's (1992) research suggests that indigenisation is a distinctive feature of the language of literature.
While "nativisation has been used to refer to the unconscious collective process of a speech community as well as to the conscious choice of an individual writer" (Egri Ku-Mesu, 2003, p. 83), Zabus (2007, pp. 3-4) defines indigenisation as the writer's attempt at textualizing linguistic differentiation and conveying African concepts, thought-patterns, and linguistic features through the ex-colonizer's language.
[…] For, when indigenized, it is no longer metropolitan French or English that appears on the page but another register reminiscent of the dominant European language, whether it be a Nigerian pidgin vaguely suggestive of a variety of English or a French that has 'something African about it'. Although Kachru and Zabus use different terminology, they essentially describe the same processes that work in the text and in the context: Kachru's (1986) nativisation Nativisation of context: text overloaded by cultural and historical presuppositions different from the traditionally expected cultural and historical milieu for English literature Nativisation of cohesion and cohesiveness: redefining the native English users' concepts of cohesion and cohesiveness, which affects lexicalisation (direct lexical transfer, hybridisation, loan translation), collocational extension and the use/ frequency of grammatical forms Zabus's (1991,2007) indigeni sation The use of Pidgin Code-switching Relexification: calquing (loan translation); "textual violence" (morpho-syntactic relexificatione.g. subjectless sentences, omission of function words, reduplication, tag questions, the use of the progressive aspect for mental processes, thematisation of complements and adjuncts; and lexico-semantic relexification -collocational innovation, repetition); ethno-text (e.g. proverbs, rules of address, riddles, praise names, dirges, prayers, greeting formulae, culturally-bound insults) My particular interest here is in code-switching, under which the use of pidgin and relexification will be subsumed. I will follow Myers-Scotton (1993, p. 3), who stipulates that "code-switching may take place on any level of linguistic differentiation," i.e. between languages, styles, dialects or registers, and I will examine three characteristic examples of code-switching between languages taken from modern Ghanaian anglophone fiction which illustrate how code-switching as an authorial device is used for a particular purpose and to achieve a particular effect. Although I will be dealing with written language, I will not be analysing morphosyntactic procedures. Nonetheless, Myers-Scotton's (1993) Matrix Language-Frame Model presents itself to be a suitable framework in that it considers code-switching to be "the selection by bilinguals or multilinguals of forms from an embedded variety (or varieties) in utterances of a matrix variety during the same conversation" (Myers-Scotton, 1993, p. 3). The matrix language is the main language in code-switching, and the embedded language is the other language that participates in code-switching, but with a lesser role. Myers-Scotton (1993) also considers embedded language material of any size as code-switching material.
Applying Myers-Scotton's Matrix Language-Frame Model loosely, I will consider English, the dominant language in which these hybrid literary works are written, as the matrix language, which is the unmarked, expected language choice for this type of text. Embedded code-switched language, on the other hand, is the authors' marked choice, whose aim, as will be demonstrated, is to install cultural distinctiveness.

Nativisation of rhetorical strategies: transfer to
English of strategies corresponding to patterns of interaction in the native culture of the author Linguistic realisation of thought patterns: nonlinear (spiral-like or circular) paragraph structure and non-sequential logic manifested in English Cushioning: the juxtaposition of an African-language word/ expression and its English equivalent to provide an immediate explanation or clarification Contextualisation by inference: the embedding of an Africanlanguage word/expression in a context that is suggestive of its meaning or in a dialogue to let the characters explain its meaning

Code-Switching between English and Nzema
So the people waited. That long wait that destroys a people's confidence in themselves. At last, one spoke. A deity spoke. Unheralded, he thundered through the village. He looked fierce, angry and weird. He tossed his head sideways, muttering incoherent things to the air. He threw his powerful bodua into the sky and caught it several times. Brown amulets and dark bracelets were fixed tightly around his powerful arms (Agovi, 1989, p. 6).
By virtue of their having been produced with an audience in mind, Ghanaian writers' English-language texts studied here can be considered, in Sperber and Wilson's relevance theoretical framework, as products of ostension, in which the writers "make manifest an intention to make something manifest" (Sperber and Wilson, 1995, p. 49). Being visible traces of the mother tongue or the most dominant language of the writer, African words in the English language text make it manifest that, indeed, such texts are, to use Zabus's metaphor, palimpsests and that "behind the scriptural authority of the European language, the earlier, imperfectly erased remnants of the African language can still be perceived" (Zabus, 2007, p. 3). They signal at least bi-, but in most cases multilingualism, and "a linguistic stratification, i.e. a multi-tiered system that differentially distributes the European language, the African language(s) and the languages in contact" (Zabus, 2007, p. 3).
The insertion of an African language word or expression in the English text without any authorial assistance provided is significant because it becomes an ostensive stimulus that the writer has found most relevant in order to communicate the set of assumptions he intends to make manifest to the reader in a manner that is relevant enough to make it worthwhile for the reader to process this ostensive stimulus (cf. presumption of optimal relevance, Sperber and Wilson, 1995, p. 158). 2 The function of such an authorial device is the distancing of the mother tongue culture of the writer and 2 Presumption of optimal relevance (a) The set of assumptions I which the communicator intends to make manifest to the addressee is relevant enough to make it worth the addressee's while to process the ostensive stimulus. the other tongue culture represented by the language chosen for creation. In the space between, a gap is created, along which, as we shall see later, we can expect readers to be divided.

Code-Switching between English and Pidgin
'Massa, you tink you go like fried fillet of calf? Or a braised lamb liver? Yes, here a good one. An escalope of veal with onions and fried potatoes.' 'Zirigu, whom did you say you were going to cook for?' 'Yourself, Massa.' 'But that is not the food I eat.' 'But 'e be white man chop.' 'Zirigu, I no be white man. And that is the second time this morning I've told you that. And if you do it again, I'll pack up and leave.' (Aidoo, 1970, p. 16) In Ghana, Pidgin is predominantly an urban phenomenon, and it is not as widespread as in other English-speaking West African countries, for example, in Nigeria (Criper, 1971, p. 4;Huber, 1999, p. 156). It is usually associated with illiteracy and with the uneducated section of society, and as such, it is stigmatised (Huber, 1999, p. 140, 148, 154-63;Zabus, 2007, p. 56). Because of its inferior status, Pidgin in Ghana is rarely used in the printed or electronic media or in films. When it occurs, it is spoken by uneducated characters, or its association with a certain jocularity is exploited for popular humour. Some of the more political magazines use mock Pidgin in satire. In Ghanaian novels, Pidgin is statistically more infrequently and more episodically represented than, for example, in novels in Nigeria (Zabus, 2007. p. 56). 3 Prevalent sociolinguistic prejudice is complemented by technical difficulty. Being primarily a spoken medium, Pidgin has no written or orthographic tradition, so scripting it for use in literature may well prove demanding for the writer. Agheyisi (1984) demonstrates the various levels through which its written representation progresses to become what Zabus calls an "artefactual dialect" (2007, p. 194): Level 1: Andrew kari yu‫כ‬ tr‫כ‬bu k‫כ‬m‫כ‬t f‫כ‬ dE. yu n‫כ‬si se na w‫כ‬k i dè du. hEn! wetin yu dè t‫כ‬k? … if yu hiE di m‫כ‬ni we dEn tek bayam yu gÒ r‫כ‬n.
Level 2: Andrew, kari you trobu komot fo de. Yu no si se na wok i de du. Hen! Wetin yu de tok? … If yu hear di moni we den tek buyam yu go ron.
Level 3: Andrew, carry you trobu comot fo dere. You no see say na work i dey do? Hen! Wetin you dey talk? … If you hear de money wey dem tek buy'em, you go run.
Level 4: Andrew, carry you trouble come out for dere. You no see say na work he dey do? Hen! What ting you dey talk? … If you hear de money which dem take buy'em, you go run.
Level 5: Andrew, carry your mischief come out from there. You no see that is work he doing? Hen! What thing you talking? … If you hear the money that they paid, you go run.
Level 6: Andrew, cut out your mischief from there; can't you see that she is busy? Ha, what are you saying? … If you heard how much they paid for it, you would run. (Agheyeshi, 1984. p. 217) Even at the risk of stripping Pidgin of its essentially African elements and making it look like just a corrupted version of Standard English, for the sake of intelligibility the writer may opt for Level 3, which offers maximum accuracy of transcription with a moderate approximation to English. However, the transcription of Level 3 "may tax the patience even of a well-meaning local or international reader" (Mair cited in Zabus, 2007, p. 195), -with non-standard speech is Ken Saro-Wiwa's Sozaboy: A Novel in Rotten English. so most writers settle for the "interlanguage" of Level 4, whose intermediate approximation to English does not jeopardise ease of comprehension by a reader who is literate in Standard English. The writer's competence in Pidgin may also influence the way in which it is represented in a literary text.
The use of pidgin involves the novelist's departure from the current oral usage and the creation of an artistic medium in which such an "artefactual dialect" (Zabus, 2007, p. 194), rather than reality in West Africa, is manifested. Pidginisation is meant to establish a character rooted in his or her supra-national or urban identity, as well as to represent attitudes towards pidgin speakers, solidarity and power relationships. The exchange quoted above between Zirigu, the middle-aged general keeper and cook of a government rest house and his only guest, a young medical doctor, spells out an intricate power and solidarity semantic.
Zirigu's pidgin signals both his lower education and his lower social status. For him the young doctor is not only the guest to whose service he is assigned, but the doctor's education reflected in his impeccable English and the status given to him by his highly valued profession make him a "big man" comparable to the former white colonisers. The doctor's attempt to neutralise the inequality between the cook and himself by changing to Zirigu's code fails perhaps because his standard English is a constant reminder of their different social standing and of the complexities of historical, political and cultural causes that lead to such a situation. Nevertheless, his changing to pidgin is significant because the "speaker can use codes for an identity shift: to obscure one identity and bring into the foreground another" (Kachru, 1984, p. 187). "Zirigu, I no be white man" not only separates the young doctor from the white race -with which he is identified by virtue of his education and status, but with which he disclaims all association because of the immediate connotations it has with colonisation and its consequences -but also renounces identification with the black masters who followed suit in exploiting their fellow black compatriots after independence. Using pidgin for this particular statement marks the young doctor's identification and solidarity with common people like Zirigu and emphasises his Africanness and his attachment to his cultural roots and heritage.

Code-Switching between English and an African Language in the English Medium
Next to the girl another, older seller wakes to her missed chance and begins to call out, 'Big man, I have fine bread. ' 'I have bought some already.' The voice of the suited man had something unexpected about it, like a fisherman's voice with the sand and the salt hoarsening it forcing itself into unaccustomed English rhythms. Why was this necessary? A very Ghanaian voice.
'My lord,' comes the woman again, 'my big lord, this bread is real bread.' Inside the big car the pointed female voice springs and coils around, complaining of fridges too full to contain anything more and of too much bread already bought. Outside, the seller sweetens her tones.
' Have you ever seen a big man without girls? Even the old ones,' the seller laughs, 'even the old men.' (Armah, 1988, pp. 36-37) Code-switching in post-colonial literary texts is perhaps most exciting when it happens between the coloniser European language, in this case English, and a colonised non-European language and is rendered in the medium of the European language. Embedded in the unmixed, hence unmarked, English of the author is the dialogue between a bread seller and a politician. It is enacted in the restricted code (Bernstein, 1964) of a bargaining relationship, in which the choice of linguistic code is determined by the social relationship of the participants. This relationship not only constrains the options available to them, but it also makes the status aspect of the relationship salient. The expressions the bread seller is using are, however, different from what would be expected in a bargaining routine taking place in a culturally and linguistically Anglo-Saxon context and, therefore, reflect the writer's marked choice to signal difference. Zabus calls the transference of such indigenous speech and thought patterns relexification and defines it using Loreto Todd's formulation: "'the relexification of one's mother tongue, using English vocabulary but indigenous structures and rhythms' -best describes the process at work when the African language is simulated in the Europhone text… [It is] the making of a new register of communication out of an alien lexicon" (Zabus, 2007, p. 112). Although it is related to notions of "transposition," "paraphrase," "translation," "transliteration," "transference" and "transmutation," it differs both from translation and auto-translation in that these take place between two texts -the original and the translated version -whereas relexification is characterised by the absence of an original. Relexification takes place between two languages within the same text. It is also defined in terms of power relationships: Although these two languages [the European language and the African language] are unrelated, they interact as dominant vs. dominated languages or elaborated vs. restricted codes, as they did and still do to some extent in West Africa where the European language is the official language and the medium of prestige and power. As it hosts such warring tendencies, relexification is a strategy in potentia which transcends the merely methodological. On the methodological level, it stems from a need to solve an immediate artistic problem: that of rendering African concepts, thoughtpatterns and linguistic features in the European language. On the strategic level, relexification seeks to subvert the linguistically codified, to decolonize the language of early, colonial literature and to affirm a revised, non-atavistic orality via the imposed medium (Zabus, 2007, pp. 118-119).
The method of relexification, by virtue of conveying specific African thought and linguistic patterns, anchors a character firmly in a specific ethnicity. The use of English to convey this ethnicity may, however, conceal from the Western reader, 4 who is not acquainted with the cultural and social relationships depicted, that the bread seller in the above dialogue is not speaking English -as was the case with a group of Western, specifically British, literature students I analysed this dialogue with. This highlights the fact that the lack of cultural knowledge may prevent the Western reader from noticing the difference the African writer intends to install when relexifying African language expressions into English.

Authorial Intervention
The reader of African literature -and, by the same token, the reader of other post-colonial literatures -is involved in a double guessing game: the interpretation of literary texts within which there is the interpretation of culture-specific concepts expressed either in an African language or in English. In order to make his or her text linguistically accessible to the reader, the writer may choose to employ various methods of authorial assistance such as cushioning, contextualisation, ethnographic explanation and providing a glossary.
Cushioning is an authorial strategy whereby African-language words or phrases describing culturally bound concepts, objects and occurrences in a literary text are juxtaposed with their English equivalents to provide immediate explanation or clarification. It aims at "naming and identifying the -metonymic -gap between mother tongue and other tongue without necessarily bridging it" (Zabus, 2007, p. 7) as the short English tag may neither accurately convey the meaning of the African word nor fully encompass its cultural significance: I see my uncle among people who are dancing … Their movements are virile as they jump up and down, each man only covered by a small cloth as he would borrow from his wife or sister, tied around his waist over his shorts, leaving the torso bare. The drums and the adawuro (gong-gongs) are more intense in their rhythmic beating than I have ever heard them before … (Duodu, The Gab Boys, 1969, p. 51) I bought her a bottle as well as some of the shampoo soap which makes the hair longer. When she used the soap, her hair turned violet and everybody laughed at her and said her hair looked like mmefe -palm nut husks. (Duodu, The Gab Boys, 1969, p. 99) bilingual or polyglot, and at the time of the study, none of them had been to Sub-Saharan Africa.
Contextualisation is another authorial intervention whereby African-language words or phrases describing culturally bound concepts, objects and occurrences are embedded in an immediate context which provides clues as to their meaning, or sometimes in a dialogue, with the aim of letting the characters explain their meaning. The reader is involved in a guessing game and is expected to infer the meaning from the context. Similarly to cushioning, contextualisation identifies the gap between mother tongue and other tongue, but it does not necessarily bridge it: Once when the man was traveling to Cape Coast three different policemen had stopped the little bus and asked the driver for his quarter license. The driver had not bought it yet, and each policeman had said to him, in front of everybody, 'Even kola gives pleasure in the chewing.' In each case the driver had smiled and given the law twenty-five pesewas, and the law was satisfied. There was only one way. (Armah, The Beautyful Ones Are Not Yet Born, 1988, p. 95) The ethnographic explanation of African-language words referring to culture-specific objects and concepts in the Englishlanguage literary text is considered by Zabus (2007, pp. 8, 176) to be the ancestor of contextualisation. The reason behind such a rather intrusive authorial device is the need African writers feel to explain African culture to a Western readership and to correct previous misrepresentations of African culture: He, on the other hand, was sitting on his bampa. This was what some people slept on. It was part of the floor specially raised for sleeping. A reed quilt covered with plantain thrash and rags served as a mattress. (K. Aidoo, Of Men and Ghosts, 1991, p. 57) Providing a glossary either as footnotes or at the end of a book is the least frequently used assistance to readers. Very often it is not an authorial but an editorial intervention which the writer may have little control over. While cushioning, contextualisation and the ethnographic explanation are textual strategies that leave no choice for the readers to avoid them, a glossary is an extra-textual device giving the readers the freedom to use it when and only if they want to.
The research I conducted (Egri Ku-Mesu, 2003) on 20 volumes -novels and short stories -of Ghanaian fiction has shown that these authorial strategies and the usually editorial device of inserting a glossary all have some benefits for some of the readers in different reader groups -the Ghanaian, non-Ghanaian African and the Western. Eventually, however, none of them proves to be efficient because they all try to cater for the needs of a mixed, ill-defined readership.

The Metonymic Gap and its Consequences
In their seminal work, The Empire Writes Back, Ashcroft et al. (1989, p. 72) identify code-switching between two or more codes in post-colonial literary texts as "the most common method of inscribing alterity." Ashcroft (2001) further develops the idea of installing cultural distinctiveness in the text and posits that, together with a wide range of other linguistic devices (e.g. neologisms, ethno-rhythmic prose), the use of code-switching -whether between the variants of the same language or between languages -has a metonymic function to inscribe cultural difference. According to him, language variance in the English text creates a metonymic gap, which is that cultural gap formed when appropriations of a colonial language insert unglossed words, phrases or passages from a first language, or concepts, allusions or references which may be unknown to the reader. Such words become synecdochic of the writer's culture -the part that stands for the whole -rather than representations of the world, as the colonial language might. Thus the inserted language 'stands for' the colonized culture in a metonymic way, and its very resistance to interpretation constructs a 'gap' between the writer's culture and the colonial culture. The local writer is thus able to represent his or her world to the colonizer (and others) in the metropolitan language, and at the same time, to signal and emphasize a difference from it (Ashcroft, 2001, p. 75).
Indeed, when I asked the group of Western readers in my study to explain the meaning of African-language words for which no authorial assistance is provided in the text (Egri Ku-Mesu, 2003), the data gathered provided evidence which supports Ashcroft's claims. These Western readers' inability to understand untranslated/unglossed words, or their uncertainty about being able to infer the meaning of these words from the context confirms the cultural distance between them and the writer: they live in a different world with different experience, traditions, habit, understanding and expectation. Applying Sperber and Wilson's (1995) relevance theoretical framework, the set of facts the writer has in his cognitive environment is not manifest to these Western readers -they do not share the writer's cognitive environment either mutually or partially. The explanations they provide, none of which is consistent with the writer's intended meaning, confirm that in the inferential process Western readers rely, as all readers would do, on their cognitive environment -a cognitive environment from which the writer's physical, social, cultural and linguistic reality is absent. As the set of assumptions manifest to them is different from that of the writer's, they do not have encyclopaedic schemata, for example, about African deities, Ghanaian means of transport or the people associated with them. They may have some marginal, but by no means specific, knowledge of the "kind of thing," which is evidenced by explanations that come closest to the intended meaning (e.g. "a kind of taxi driver," "taxi drivers" for aplankey -bus conductor). The fact that drawing on assumptions in their cognitive environment cannot lead them to fully correct explanations confirms their cultural difference and distance from the writer. This distance and difference locate them on the other side of the metonymic gap, as shown in Figure 1 below.
While the concept of the metonymic gap is plausible, the suggestion that the inserted unexplained African language words resist interpretation seems to have less validity. On the one hand, While the concept of the metonymic gap is plausible, the suggestion that the inserted unexplained African language words resist interpretation seems to have less validity. On the one hand, such a point of view takes account only of a readership which excludes those who share the author's cultural and/or

Ghanaian writer's culture
Western/ metropolitan/ colonial (readers') culture such a point of view takes account only of a readership which excludes those who share the author's cultural and/or linguistic background, thereby tacitly agreeing that modern Englishlanguage African literature is, indeed, written for and read by a non-African audience. On the other hand, although the Western reader may find such words unintelligible and incomprehensible as defined by Enkvist (1991, pp. 7-8), 5 the very fact that she recognises them as synecdochic of the writer's culture and as signals of an intricate linguistic reality proves that she is able to build a text world around them in which they make some kind of sense, albeit not necessarily literal sense. Further investigation of a Ghanaian and non-Ghanaian African readership reveals a complexity that Ashcroft's original concept does not account for. In my study (Egri Ku-Mesu, 2003), Ghana is identified both as a multitude of ethnic groups and also as a supra-ethnic national entity, and the Ghanaian readers in my survey come from seven ethnic groups and are bi-, trilingual or polyglots. Relative to each individual writer's background, there are readers who share the writer's physical, cultural and linguistic environment -they share a cognitive environment, which implies that they are capable of making the same assumptions. Readers who do not belong to the writer's ethnic group do not share a mutual cognitive environment with the writer. However, because of similarities of their physical environment, overlaps in culture and their linguistic knowledge, the cognitive environments of these readers can still intersect with that of the writer, i.e. they will share sets of facts that are manifest to them all. 5 According to Enkvist (1991, pp. 7-8), "a piece of text is intelligible to those who can recognize in it phonological, lexical and syntactic structures. Intelligibility thus presupposes pattern recognition, the correct perception of structures. A text is comprehensible to those who can assign to it a definite meaning, a semantic structure. And a text is interpretable to those who can build around that text a scenario, a text world, a set of states of affairs, in which that text makes sense.
[…] Intelligibility is thus the syntactic component of interpretability (which includes phonology, lexis, and syntax); comprehensibility is its semantic component including syntax plus semantics; and interpretability its pragmatic totality involving pragmatics as well as semantics and syntax." Indeed, my research findings confirm that there exists a multi-tiered distribution of Ghanaian readers. At the supra-ethnic national level, they all share cultural and linguistic experience with the writer(s) and each other. This is the level which embraces them all and situates them on one side of the metonymic gap -the writer's. At the level of ethnic groups, subdivisions occur according to cultural and linguistic proximity to the writer. There are readers who are both culturally and linguistically distantthey share experience with the writer at the supra-ethnic national level. Readers who know the writer's language but are not fully conversant, or not familiar at all, with his culture are able to share in the writer's experience to the extent their linguistic and encyclopaedic knowledge allows. This highlights the overriding importance of cultural knowledge in communication between different ethnic groups even within the same national boundaries. Full appreciation of the writer's meanings is shown only by those readers who share both the writer's cultural and linguistic experience. Figure 2 below illustrates this complex situation.
By virtue of sharing a common African heritage (see Note 1), it is reasonable to expect that the non-Ghanaian African reader would exhibit some affinity, if not with the Ghanaian writer's linguistic experience, at least with his cultural experience. Like the Ghanaian reader group, the non-Ghanaian African readers are all polyglot, although none of them speaks either Akan, Ga Readers who share the writer's language but not culture or Pidgin, or any other indigenous Ghanaian language. Although the smallest reader group in my study, its members represent East, South, Central and West Africa -they come from six ethnic groups which are distant from the writers'. While it is obvious that these readers cannot share a fully mutual cognitive environment with the Ghanaian writer(s), the similarities of their physical environment and overlaps between their cultures make it possible for them to have some sets of facts that are manifest to them all, with regard to which they are able to make the same assumptions as the writer(s). My findings provide evidence that, for the non-Ghanaian African readers, the lack of linguistic knowledge may well be a barrier to activating and accessing cultural knowledge. At the same time, similarly to those Ghanaian readers who share neither the writer's cultural nor her linguistic background, when working out the meaning of the untranslated/unglossed words, the non-Ghanaian African readers rely not so much on the context as on their existing cultural knowledge. Information in the texts triggers off these readers' schemata in their encyclopaedic memory about such concepts, for example, as deities, vehicles used for transporting people and their crew, which get filled in with their assumptions based on their knowledge of their own culture and on their knowledge of other African cultures. The fact that non-Ghanaian African readers are able to provide either good approximations of the intended meaning or they indeed provide the intended meaning of the untranslated/unglossed words is evidence that their cognitive environment contains assumptions about the writer's physical and social reality, i.e. they share cultural knowledge with the writer and they are capable of making the same assumptions as the writer. That this cultural knowledge is accessed by the members of the non-Ghanaian African reader group who represent a wide ethnic and regional variety testifies to such cultural knowledge being supra-ethnic, supranational, pan-African in nature. As Figure 3 shows, this cultural knowledge places the non-Ghanaian African reader on the same side of the metonymic gap as the writer.
Such a picture clearly shows that the metonymic gap is not a simple bi-polar concept between coloniser and colonised culture but a multi-layered entity where the readers' position in relation to the gap is indicative of their ability to interpret the untranslated/ unexplained words.
Leaving these words without any authorial assistance to the reader, the writer makes it manifest that she concedes "the importance of meanability, the importance of a situation in which meaning can occur" (Ashcroft, 2001, p. 76; italics as in original) and opts for the inscription of difference, distance and the absence of writer and metropolitan reader from each other as a result of them being located in different cultures. The readers' position in relation to the gap thus created is indicative of their ability to interpret these untranslated/unexplained words. For the majority of the Western readers in my study who may be able to cross the metonymic gap partially, these words do remain symbols of the writer's difference of experience, while the Ghanaian and other African readers can access their culture-bound meaning fully or to varying degrees depending on the similarity of their cultural experience to the writer's. Whether the reader understands these words or not is not crucial for the understanding of the story -if it were, there would probably be an authorial strategy present in the text to assist the reader. Such words become synecdoches for the writer's culture. Their function is to signal cultural difference and to contribute to the creation of the Africanness of the text. Readers who share writer's language and culture Readers who share writer's language but not culture Supranational, pan-African cultural heritage Non-Ghanaian African readers