PRELIMINARIES

Before we embark on the argument itself, it needs to be set against a background of several viewpoints and theoretical positions. To start with, it is clear that literacy is not just part of “language” in general; it is a particular manifestation of language use, related to spoken language but different as a field of action (Hymes 1996:34ff.; see also Collins & Blot 2003). It requires the mobilization of highly specialized skills and techniques, and its products and patterns of reception are different from those of spoken language, both in type and in function. Whereas spoken language is a structured complex of meaningful sounds, the main feature of which is the fact that it disappears as soon as it has been produced, writing results in crafted artifacts that have – at least typically – the capacity to be lasting, to be archived. The orientations to meaning that people display in spoken versus written language are again different, as are the actions that they perform in order to retrieve meaning. We cannot attribute a different tone or intonation contour to spoken utterances we hear; but we can do that, and far more, while reading. Consequently, we can organize different complexes of reusing practices for written languages: quotation, copying, interpreting, translating, and so on (see Urban 1996 on the notion of “replication”). Writing reorganizes the cognitive, epistemic, affective, and identity frames of language.

A consequence of this is that written language should be addressed not as an object loosely attached to the study of speech but as an object in its own right, requiring the kinds of divisions into genre, style, register, and so forth usually reserved for spoken language. What is needed is an “ethnography of writing” (Basso 1974), in which writing is approached as a complex of language-organizing actions, stylistically and generically variable, and intrinsically connected to domains of use and social value attributions in people's lives. Little can be taken for granted here: Neither the conventional practices nor the attached functions or domains, and even less the specific place writing occupies in the repertoires of its users, are things that one can define a priori. They all need to be established empirically, through analysis.

One of the obstacles we are facing here is the strong historical “loadedness” of writing as both the “best” form of language – the most developed one, the most elaborate one, the “literary” one – which is manifest, for example, in the way unwritten languages used to be (and to some extent still are) catalogued as “primitive,” or at least “incomplete” (Fabian 1996, Rafael 1993, Errington 2001). Writing – alphabetical, normative, and generically and stylistically elaborate writing – is the prestige modus of language, and a language is not seen as “complete” unless it has acquired a standard orthography. Graphocentrism and textualism are, consequently, powerful language ideologies that organize a lot of what we pretend to detect in language and a lot of how we believe it functions (Collins 1996). We tend to project these images and functions onto writing at large, and thus take far too much for granted. Such rapid projections have given rise to widespread and widely supported assumptions about the empowering and liberating effects of literacy – assumptions underpinning, for example, UNESCO's view on education in development, many NGO activities involving alphabetization efforts, and many modern (or late modern) discourses on sustainable development in general (see Rogers 2001 for a critical appraisal). They also account for the fact that, wherever writing is introduced into a society's repertoire, it becomes a highly valued, prestige-bearing resource.¹

Within this tradition, the connection between speaking and writing was long addressed in a rather casual way. A language could be “improved” by introducing writing, and those who could write could also speak, and vice versa. Only recently have we witnessed the emergence of studies of literacy as a different complex of situated practices of language use, partly responding to what happens in spoken language but partly autonomous (Street 1995, Collins 1995, Collins & Blot 2003). The theses articulated in these studies revolve around the social embeddedness and situatedness of writing, and around the fact that literacy occurs in the shape of specific literacy practices performed in specific social contexts (Gee 1990, Besnier 1995, Street 1995, Barton 1994, Barton & Hamilton 1998, Graddol et al. 1991, Baynham 1995, Prinsloo & Breier 1996). Furthermore, apart from the genesis of this paradigm, the field of literacy has been divided into studies of writing (with a focus on orthography; e.g., Jaffe 2000), circulating (e.g., Barton & Hamilton 1998), multimodal text construction (e.g., Kress & Van Leeuwen 1996), and reading (e.g., Boyarin 1992), and studies have been devoted to the developmental dimensions of acquiring literacy-as-practices (Kress 1996, 2000) and to the politics of orthography and literacy (e.g., Schieffelin & Doucet 1992, Collins & Blot 2003).

The discussion in this article needs to be read against the background of this collection of works. I subscribe to broadly the same set of assumptions about literacy-as-contextualized-practice, although I would add more emphasis on the social evaluatedness of language practices in general, and on the specific “values” of writing in particular – writing as a mode of production of language. This will be thematized in the case analysis. I shall first introduce the particular documents on which it will be based.

DATA

The data I shall discuss here came to me by “structured accident”: a coincidence conditioned by my social position. In my capacity as professor of African linguistics and sociolinguistics, I am summoned rather frequently by official services – the police, the prosecutor's office, the immigration and asylum services – to provide linguistic expertise on African languages. Over the past few years, asylum applications have become a major domain in which African text material is being produced in the form of statements, testimonies and so forth. As such, in 2001 I received the set of seven documents shown in Figures 1, 2, 3, 4, 5, 6, and 7. The request added to the documents was to “translate these documents into Dutch” – a request in itself speaking to the pervasiveness of textualist ideologies. The request came from the prosecutor's office, and though I was not (and never am) informed about the specifics of the case (apart from the cursory mentioning of a male subject), I could see that this was a rather typical document from an asylum application. The applicant had claimed to be from a particular country, Burundi. Being from a country suffering from war or civil war is an advantage in the Belgian asylum procedure, which is highly selective and has the highest rejection rate in the European Union, approximately 96%. As often occurs, the applicant had no official documents substantiating his claim to Burundese citizenship, and the authorities almost automatically refused to accept this claim and demanded proof. The proof, in this case, was to produce “as much information about his country as he possibly could.” The product of this exercise would then be subjected to an examination of its “credibility” (a common practice in asylum applications; see Blommaert 2001a). This examination relies heavily on textualist criteria of coherence, transparency, and correctness – in short, those discourse features which we tend to associate with “truth.”

Document 1.

Document 2.

Document 3.

Document 4.

Document 5.

Document 6.

Document 7.

As for formal features, we here confront a heterogeneous collection of documents with different authors (or, at least, scriptores). The first five pages of the text are on official police case stationery, recognizable from the case register number and page number at the top right-hand corner. The two remaining pages are unnumbered and less structured, probably a set of “notes” separate from the more structured text. Furthermore, the first four pages are in one handwriting – Author A – and highly structured; the notes on document 6 are clearly written by someone else – Author B – and are messier. The map (document 5) and the notes scribbled above the address on document 7 are authored by yet another scriptor – Author C. Fragments of Author B's handwriting also appear in the first four documents (see, e.g., the bottom of document 1) as well as at the top of the map, document 5. We witness traces of different stages of collaborative text production, with Author B as the “desk editor” filling in and correcting here and there. We also witness issues of materiality here: Some of the documents are done on official police stationery, imposing a particular organizational grid on the documents. This is “special” paper, inviting “special” writing. Other documents are on different kinds of paper, and these differences are reflected in stylistic and generic differences, as we shall see. Notwithstanding these differences, the whole set of seven pages was sent to me as one document requiring one act of translation.

I will now embark on two series of reflections. First, how must we appraise the function of these documents? And second (deriving from the first), how can we comprehend such documents against the background of general, and different, economies of literacy on a worldwide scale?

DOCUMENTS MADE FOR READING?

In a society saturated by literacy, the typical set of activities connected to written langue is “reading”: a complex of physical and cognitive actions organized so as to extract “meaning” from a written text. The purpose of writing is to be read. The relation between writing and reading is assumed to be direct and unmediated: One “reads what is written.” If things are not written, they are not made for reading; few people would qualify the perceptual activity organized around a photograph, a building, or a painting as “reading.” We read written text, we look at photographs, buildings, and paintings.

This unmediated one-to-one view has been challenged by several scholars (see in particular the collection of essays in Boyarin 1992b). What is understood by “reading” can differ across communities and contexts, as well as across genres: We read an academic paper differently from a poem or a newspaper. Particular documents invite (and are designed for) particular kinds of reading – in fact, what we understand by “genre” in writing may be reflected in genres of reading as well. Looking at the documents under scrutiny here, we already sense that the question of whether these documents are designed for reading will not receive a simple answer.

Kress & Van Leeuwen 1996 discuss the complex, multimodal design of contemporary documents such as advertisements, textbooks, and video clips. New forms of literacy have emerged in which the visual and the textual combine in one sign. This forces text consumers to combine different activities – “reading” as well as “looking at” – and synthetic (the whole sign) as well as analytic (different constituent parts of the sign) decodings. Furthermore, such forms emphasize the primarily visual and material character of written text, and they advocate the visual as the point of entrance into any text: “Writing is only one way of visualizing meaning, a very exceptional one” (Kress & Van Leeuwen 1996:18). In fact, what we call alphabetical writing may be a residue of original, more complex multimodal ways of visualizing meaning, the result of a gradual restriction of the scope of visualizing meaning to writing. In the same move, writing became less and less an object of visual inspection – it became devisualized (and dematerialized) – and it became the object of a new, exclusive activity-type, reading. Kress 1996 expands the argument by looking at the development of writing skills in children, arguing that children move from highly multimodal representations of meanings (drawings with some written texts) to devisualized “text only” representations. Learning how to write is unlearning how to produce multimodal, visual meaning representations. This, it should be underscored, is an ideological process. Every written document is a visual document, and when we write we continuously deploy a wide range of meaningful visual tactics (differences in font and size, lines, arrows, indentation, etc.). Reading, similarly, involves the visual decoding of the document. Thus, visuality is not lost in practice, but it is lost in the ideological conception of the writing and reading process.

The documents produced by the Burundese man are overwhelmingly visual and distinctly multimodal. Textual features combine with drawings: the map (document 5), the national flag (document 3), and the banners of the political parties Frodebu and Uprona (document 4). But these are only the most striking visual items. In document 4 we also see how the part headed by mitaa ‘roads’ has drawings of two roads (or lanes of a highway) next to it, with the note bara bara kub[wa] ‘main road, highway’. Furthermore, the texts are replete with solidly visual structuring features: chapter headings completely or partially in capitals, with single or double horizontal lines marking them; double vertical lines separating columns (document 2); a play with different sizes of symbols, highlighting specific words or parts (e.g., INTAMBA in document 1, or OTRACO in document 3), and of course, the enormous numbers in the left margin, marking “chapters” or “sections” in the document. Note also the careful spatial alignment of parallel series, such as at the bottom of document 3:

In sum, we see how the author deploys several outspokenly visual-graphic techniques to provide clear, transparent structure and meaning in the documents.²

The documents are highly structured, and that gives us one clue as to what kind of text is being produced here. But in order to get the full picture, we need to turn to what the author is trying to express. In the Appendix, I provide a translation of the “textual” parts of the documents, which incorporates as many as possible of the textual/linguistic features of the Swahili originals.

The texts are written in a vernacular variety of Swahili, with traces of vernacular French and English in the parts written by Author A, and English in the parts written by Author B. Author A, for instance, uses loans from French reflecting local phonetics of spoken vernacular French. The same goes for a couple of English loans in Author A's text. Examples are:

Note that the French loans Museu and Independece are accompanied by standard Swahili glosses, makumbusho and uhuru respectively, indicating the fact that the author probably would use the French loans in everyday, mixed-language speech (see Blommaert 1999; similar “reverse glosses” appear in Fabian 1990). The French loans produced by Author A are almost all vernacular – that is, a graphic replica of spoken forms, marked, for example, by unrounding of vowels, as in avenue > avenii, campus > kampis, (French) bus > bisi (English bus would most likely result in basi).

Author B uses English terms such as died (document 1), green + white (document 2) and overthrown (document 6), and all of them are correctly spelled. Author C uses colloquial Swahili with embedded, unglossed loans such as cambi wamunisipal (le camp municipal), prison, and mosque, as well as the English loan petrol station. We are confronted with probably three different levels of literacy competence here – three different individuals with differing control over writing skills. Authors A and C use vernacular spoken language varieties as the basis for writing, which is a clear indicator of subelite literacy. This is most striking in the case of place names (avenii, shosii, cambi wamunisipal). The point of reference in writing here is not how they know these names are spelled, but how these names are pronounced.

This is important, for the collective effort of these three scriptores is one of remembering. It is a painstaking effort, the clearest sign of which is the fact that the whole set of documents – the four “structured,” numbered pages as well as the map and the notes – was submitted to the authorities as the answer to the initial assignment. Throughout the documents, we see a struggle with “getting things right.” Witness, for example, how Author B makes a list of street names on top of document 5, to be added to the map drawn by Author C. Similarly, document 6 is a list of the presidents of Burundi in chronological order, but with dates lacking for several of them. Incomplete lists are rather frequent: In document 2, Author A announces 16 districts but only provides a list of 13; the name of the beer factory in document 3 is not provided; in document 2, the section numbered 4, on money, is complemented by another one at the bottom of the page (marked by Author B with 4 – look); and Author B appears to have added elements of remembering in several places (see, e.g., document 2, the location of the Albatros Hotel). Author C has made notes on Hutu-Tutsi divisions on document 7, but they have not found their way into the text.

Everything taken together, the text is a sequence of 13 rather loosely ordered sections of “factual” remembering, strongly organized around naming practices and geographical location. Three people have collaborated in this reconstruction of a school-atlas kind of inventory of facts about Burundi and its capital, Bujumbura. The writing process itself structures remembering; numerous traces of this process of writing-as-remembering can be found in the form of notes, corrections, or additions in the text; remembering has to be done on the basis of highly fragmentary material, textually – visually – organized in a particular format (cf. Fabian 1996; Blommaert 2001b, 2003b, 2004a).

This brings us to another aspect. Apart from an effort at remembering, the texts are also an effort at generically regimented text production – a text that satisfies both the purpose of structured remembering and the generic requirements (real or perceived) of “official,” literate, “on record” discourse. The authors are not constructing just any text; they are trying to craft an ordered, Cartesian text full of tables, lists, and separate, neatly marked topical divisions in sections and chapters – an encyclopedic text. This encyclopedic text emerges in different formats complementing one another: “official” bits on official police stationery, and drafts, notes, and sketches supplementing the official parts. I am avoiding the verb “writing” here, for what they do is more than writing; it is “document design.” It is all there: textual, visual, and spatial architecture, generic differentiation. – Every available linguistic and communicative skill (including the skills of others) is mobilized in order to make sense.³

What we have to learn from this, I believe, is that the function of a set of documents such as these one is not primarily or exclusively reading. The documents are crafted in such a way that they have to be looked at, read, decoded, and reconstructed; the activities we need to deploy in order to make sense of them are as varied as the modes of activity that went into their production. Fabian is to be credited for his suggestion about orality being a function of literacy, and he has pushed us to recognize that the functions of literacy are potentially even more numerous. We need to keep this relativity of functions in mind (Hymes 1966; cf. Blommaert 2004b), especially when we are facing writing that originated in different economies of literacy, because such writing often confronts us with hetero-graphy from our point of view – that is, with the deployment of writing skills for functions we do not usually allocate to them. To this I now turn.

FREEDOM OR CONSTRAINT? THE NON-EXCHANGEABILITY OF LITERACY VALUES

As mentioned in the introduction, Fabian interprets the particular shapes of literacy in the Vocabulaire d'Elisabethville not as “deficient” writing, but as “liberating,” creative practice. By this, he means that “it is a literacy which works despite an amazingly high degree of indeterminacy and freedom (visible in erratic orthography, a great disdain for word and sentence boundaries, and many other instances of seemingly unmotivated variation)” (Fabian 2001:65). I certainly would not want to challenge the creative dimensions of writing practices such as the ones discussed here. What I want to qualify, though, is the general association of writing with opportunity and freedom. My argument will be that the question of whether or not writing offers opportunities for its practitioners is to be answered by ethnographic and sociolinguistic analysis.

Let us return to the issue of functions. I hope to have made sufficiently clear in the previous section that functions are a matter of uptake, of reception. I had to make this clear because it is amazing to see how often this fundamental aspect is overlooked or neglected, even in studies that claim to analyze interaction. Discourse has no intrinsic function; it is granted a function by others in a process that Bakhtin 1986 described as “dialogical.” Furthermore, there is no way in which we can detach “function” from “value”; discourse modes are made meaningful because of their insertion in stratified, indexical scales of social value-attribution (see Bourdieu 1982, 1991; Hymes 1996; see also the essays in Gal & Woolard 2001 and Kroskrity 2000). And like a second-hand car, a chunk of discourse is worth precisely as much as other people are willing to give for it.

In such scales of value-attribution, particular instances of discourse will receive function-value depending on the overall structure of the repertoires of language users. In contexts where few people have access to standard varieties of a particular language, the use of such standard varieties will typically be attributed much value. The reverse is also possible: Particular socially stigmatized slangs can be granted high value in small peer groups where the use of such slangs defines group membership and identity. Literacy – or at least particular forms of literacy – is often a crucial, highly valued resource offering elite status or membership to those who have access to it. In sum, particular formats of language use derive function-value from their place in hierarchies of linguistic resources, and from contrasts with other members of the hierarchies.

The problematic point, however, is, that such hierarchies are primarily local, even though they obviously connect with scales and hierarchies at many other levels. For instance, the high value attributed to slang in urban youth subcultures is valid only within such subcultures, but it derives its status from contrasts with the different norms held by the parents of members of such subcultures (which often stigmatize slang), and from society-wide ideologies of a standard (again, often stigmatizing slang). Yet such slangs often connect with transnational groups or networks in which ingredients of slang have high value and carry enormous prestige (e.g., Hiphop slang in East Africa; see Fenn & Perullo 2000).⁴

Examples of this are not hard to find. One obvious and well-documented domain of such transfer from one environment into another, with effects on function, is the courtroom. Colloquial speech and lapidary accounts of events couched in a world of experience that is local and articulates local relationships and conflicts get lifted out of their locality and moved to the level of law, society, the “case,” and so on – a contextual universe to which completely different rules of relevance and interpretation apply (e.g., Conley & O'Barr 1990, Jacquemet 1996). Consequently, what is a strong case in the neighborhood may not be a strong case in court. Similar phenomena are rife in bureaucratic environments, the education system, and many other places in society. One feature of contemporary societies is the enormous and intense traffic of discourse across contexts in so-called text trajectories, and each stage of such a trajectory recontextualizes, reinterprets, and re-creates the “original” discourse in other contexts and in spaces where different norms and conditions of uptake apply (Briggs 1997; Urban 1996; Blommaert 2001c, 2004b).

Let us now consider our present example. The text was produced in a transnational context: a Burundese subject producing text in Belgium addressed to Belgian legal officials. The Burundese subject clearly mobilized the resources he could mobilize: his own literacy skills, his own memory, and the skills and memories of two of his friends. But, as said above, this complex of resources betrays an economy of literacy that is different from the one that applies in societies with fully saturated literacy environments, in that both the writing itself and the remembering were “incomplete” when measured against the textualist norms used by the addressees. So what we have here is a clash of two different economies of literacy, one guiding the production of the documents, and the other guiding their uptake. The bridge between them is formed by the documents themselves: texts that are moving from one place into another, from Africa into Belgium, or to be more precise, from a subelite stratum of society in Burundi through a diasporic community of Africans in Europe (where the two other scriptores come in) to the core of the bureaucratic system in Belgium.

Let us assume for the moment that the form of writing displayed by the Burundese subject would qualify as acceptable writing skills in his place of origin and in his current place in his diasporic community; let us assume, in other words, that it is locally ortho-graphic in the etymological sense. This may be good, useful, and functionally adequate literacy in the subelite stratum of society in Burundi and in the African diaspora in Belgium. It may also be the “best possible” document in terms of the subject's available skills and competence (e.g., with respect to remembering place names, presidents, schools, etc., in Bujumbura). But in fact, it is not good, useful, and functionally adequate literacy in the Belgian bureaucratic world. When transferred from one place to another, the documents get repositioned in a different economy of literacy, and they lose function at a rather dramatic rate: They become hetero-graphic.⁵

The point of all of this is that, in the present globalized world, we encounter more and more instances of texts moving from the peripheries of the world system to its centers, and this move in space is also a move across different economies of literacy, involving differential allocation of function and value to texts as they travel across these economies, and a transition from ortho-graphy to hetero-graphy. Consequently, the relativity of functions needs to be placed against the wider frames of different economies of linguistic resources on a worldwide scale. The inferior value of texts from the peripheries – for instance, from Africa – is relatively predictable and systemic: Given Africa's peripheral position in the world system, resources that have exceptional value there do not necessarily have this value in Europe. The transfer of linguistic signs does not entail the transfer of their functions and values; the latter is determined by the general structure of the world system, by global patterns of inequality.

As linguists and anthropologists, we can reconstruct the value of such dislodged, displaced, hetero-graphic texts. In fact, perhaps we are the only ones capable of restoring and reconstructing non-local, ortho-graphic meaning in such texts, meaning understandable for us. The problem, however, is that we have to engage in expert practices in order to retrieve such meanings; they do not come automatically. The voice of the communicating subject has to be reconstructed and restored, for it is not in itself hearable.

Hence, the image of “freedom” attached to these literacy practices sounds literally out of place. The inconsistencies and different forms of coherence observed by Fabian in his Shaba texts may be a feature of freedom and may offer immense semiotic opportunities to their producers in Shaba. As soon as they start to travel across the world, however, all these features become objects not of difference but of inequality. The opportunities offered by particular, creative forms of literacy in Shaba or Burundi may turn into foci of discrimination, disenfranchisement, and injustice elsewhere. Opportunities, just like function and value, do not as a rule travel along with the texts; they are often left behind. In the global system, values of semiotic forms are not always exchangeable, and consequently, whereas writing may be a tremendously rich instrument for social mobility in the peripheries, it may be just a problem in the center – a problem of “fixing,” of tying subjects to their place of origin with its own economies of literacy. In sum, it may become a problem of denying mobility to communicative resources.