Whoever Writes the Code Writes the Canon: Technology, Scripturalization, and the African Question of Authorship

Abstract. This essay argues that digital technology in Africa repeats the history of scripture. Each new medium of record, from the tablet and the printing press to the pulpit microphone and the social media timeline, has decided who writes, who reads, and whose reading counts as authoritative. Read through Comparativity Complex Theory, the platform, the ledger, the dataset and the language model appear as scripturalizing machines: they fix meaning, distribute authority, and extract value from those they inscribe. The essay tests the claim against Nigeria's 2021 suspension of Twitter, the #EndSARS protests of 2020, Kenyan mobile money and digital credit, outsourced content moderation in Nairobi, the popular claim that Ifá is a "binary system," digital catalogues of looted objects, and machine translation of African languages. It concludes that the prevailing policy vocabulary of "access" and "inclusion" mistakes the problem. The question for Africa is authorship: who writes the code, who audits the record, and who holds the keys.

Keywords: scripturalization, data colonialism, digital authorship, Nigeria, Heritage Security Intelligence, comparativity, African languages

The Tablet Before the Tablet

Long before the word "tablet" named a slab of glass and lithium, it named a slab of stone. The tradition that has Moses descending Sinai with the law written by the finger of God is, among other things, a story about a medium: the inscription is fixed, portable, and beyond the reach of the people who must obey it. They did not write it. They may not revise it. When the first tablets are broken, only the authorized mediator may go back up the mountain for a second copy. Whatever one believes about Sinai, the structure of that story has proved astonishingly durable. It reappears whenever a new technology of record arrives and a small class of people gains control of the stylus while everyone else is invited, graciously, to read.

This essay argues that the digital repeats the scriptural. The platforms, ledgers, datasets and models now spreading across Africa are not neutral pipes through which African voices at last flow freely. They are scripturalizing machines. Vincent Wimbush gave the name "scripturalization" to the process by which a society makes certain texts into scripture and, in the same stroke, makes certain people into those who read and those who are read.¹ This author has extended the term to the question of how meaning is fixed and policed across African religious and ethical life.² The extension proposed here is simpler and more provocative. Whoever writes the code writes the canon. The terms of service are a covenant, the content policy a catechism, the algorithm an exegete whose reasoning the congregation is not permitted to see.

The dominant view in development circles runs the other way. It holds that Africa's digital problem is one of access: too few smartphones, too little bandwidth, too many people "unbanked" and "offline." Fix the access, the argument goes, and the continent will leapfrog the slow industrial century that others endured. There is evidence for this view, and this essay will take it seriously when it comes to mobile money. But the access thesis quietly assumes that the thing being accessed is benign, that the only injustice is exclusion from it. The history of scripture suggests otherwise. The enslaved African in the Atlantic world was not only kept from the Bible; the Bible was also used to read him. Access to a canon one did not author has rarely been liberation. More often it has been a more intimate form of rule.

The question, then, is not access. It is authorship. Who writes the code, who keeps the ledger, who labels the data, who decides which words a machine will learn to hear, and who is permitted to audit any of it? These are old questions about scripture wearing new clothes. A decade ago, with Gbenga Dasylva and Saliu Funmi Imaledo, this author co-wrote a chapter whose title already bound "scriptures," "social media," and "social power" into a single phrase.³ What follows returns to that knot after ten years in which the timeline became, for millions of Nigerians and Kenyans, the first place they learned the news, the last place they said goodbye, and the place where the state learned to find them.

The method here is Comparativity Complex Theory, which distinguishes comparison, the act of setting two things side by side, from comparativity, the historical and power-laden conditions that make two things seem comparable at all.⁴ Technology is saturated with comparisons: Africa "behind" or "ahead," Ifá "like" binary code, Nairobi "the Silicon Savannah." The task is to dig beneath each comparison to what the theory calls thresholds of discomfort, where the flattering analogy gives way to the question of who profits from it.

Print, Pulpit, Microphone: A Short Genealogy of Mediated Authority

Every medium of scripture has been a medium of power before it was a medium of meaning. Elizabeth Eisenstein's great study of the printing press showed how the fixity of print transformed European religion, science and scholarship, making possible the standardized text, the authoritative edition, and the Reformation's claim that every believer might read for himself.⁵ What Eisenstein's Europe exported, however, was not the democratizing press but the mission press, and it arrived in Africa with the catechism already set in type. The mission station printed grammars of African languages in orthographies chosen by missionaries, Bible translations whose key terms were settled by committees of foreigners and their converts, and primers that taught reading and obedience in the same lesson. V. Y. Mudimbe's account of the "invention" of Africa in European discourse is, among other things, an account of how the printed word made Africa legible to Europe on Europe's terms.⁶ Print did not simply record Africa. It authored an Africa that Africans were then required to read.

The pattern continued with broadcast. Brian Larkin's study of media in colonial and postcolonial Kano shows that radio networks, mobile cinema units and cinema halls in northern Nigeria were never simply channels for content; they were infrastructures through which the colonial state staged its own modernity and through which Hausa Muslim audiences negotiated what that modernity would mean.⁷ Larkin's larger argument, that infrastructures carry a politics and a poetics, that they promise as much as they deliver, is indispensable here.⁸ The loudspeaker on the colonial rediffusion pole in Kano and the fibre-optic cable landing on the Lagos shore belong to one genealogy. Both announce that someone far away has decided what may be heard.

Then came the pulpit microphone. Rosalind Hackett documented, as early as 1998, how charismatic and Pentecostal churches in Nigeria and Ghana seized electronic media, cassette, video, radio and television, as their preferred means of reaching urban Africa, and how that seizure both expressed and strained their theology.⁹ Marleen de Witte's study of a Ghanaian charismatic ministry's television programme shows how the medium itself manufactured charisma, fixing the fluid gift of the Spirit into a format and producing a new kind of believing subject in front of the screen.¹⁰ Asonzeh Ukah's work on the Redeemed Christian Church of God traces how a Nigerian church became a transnational enterprise whose media, publishing and branding were inseparable from its growth.¹¹ Under amplification, the pastor's voice becomes something close to scripture. The sermon is recorded, sold, clipped and circulated; the microphone that carries it also fixes it, and the fixed word acquires an authority that the living congregation cannot easily answer back to.

The social media timeline is the newest scroll. It is infinitely long, perpetually updated, and ordered not by chronology but by an algorithm whose logic belongs to a corporation in California. It seems to have abolished the distinction between author and reader: anyone may post. Yet this is precisely the illusion that the history of scripture should make one suspicious of. Everyone may write on the timeline, just as everyone might once whisper in the temple court. But the ordering of what is seen, what is amplified, what is buried and what is removed belongs to the keepers of the code. The congregation writes; the priesthood curates. That distinction is the whole of scripturalization.

Data Colonialism and Its African Accent

The most ambitious recent theory of this arrangement comes from Nick Couldry and Ulises A. Mejias, who argue that the capture of human life through data constitutes a new colonialism, not a metaphorical one.¹² Historic colonialism appropriated land, bodies and resources; data colonialism appropriates human life itself as it is lived, translating social relations into continuous streams of data that can be owned, processed and monetized.¹³ Michael Kwet, writing from South Africa, makes the imperial dimension explicit: American corporations exercise control at the level of the digital ecosystem's architecture, its software, hardware and network connectivity, so that the Global South rents the very infrastructure of its own public life.¹⁴ Abeba Birhane, writing specifically about Africa, calls the result algorithmic colonization: Western technology, built on Western values and assumptions, arriving on the continent with little scrutiny and often marketed as humanitarian rescue.¹⁵

To these one must add the American critics who showed that the code is not colour-blind. Safiya Umoja Noble demonstrated how commercial search engines reproduce racist and sexist representations, most notoriously of Black women and girls, while presenting their rankings as the neutral outcome of relevance.¹⁶ Ruha Benjamin named the phenomenon the "New Jim Code": designs that encode inequity while appearing objective, and sometimes precisely by claiming to correct for bias.¹⁷ Taken together, these works establish that the digital canon has authors, that those authors have interests, and that African users are, for the most part, read rather than reading.

Here, however, this essay parts company with a certain version of the data colonialism thesis. In its strongest form it risks reproducing the oldest error in the study of Africa: the assumption that Africans are only ever acted upon. Jean-François Bayart's concept of extraversion is the necessary corrective. Bayart argued that African elites have historically drawn their power from strategic relations with external actors, turning dependence into a resource and extracting rents at home from their brokerage with the outside world.¹⁸ Extraversion does not excuse empire. It shows that empire has always needed local partners, and that those partners have their own reasons.

The Nigerian election of 2015 offers a stark illustration. In March 2018, Carole Cadwalladr reported in the Guardian that Cambridge Analytica and its parent, SCL Elections, had been engaged in support of President Goodluck Jonathan's re-election campaign, and that the firm's staff had allegedly been handed material hacked from the opposition candidate, Muhammadu Buhari; SCL confirmed it had been hired for the campaign but denied using hacked information.¹⁹ Whatever the full truth, the structure is instructive. The foreign data firm did not force itself on Nigeria. It was hired, reportedly with Nigerian money, by Nigerian political actors seeking advantage over other Nigerians. Data colonialism with an African accent is extraversion by other means. The scribes of the new canon are in Silicon Valley and London, but some of their patrons live in Abuja.

A further caution belongs here, and it arises from the method rather than the evidence. To call data extraction "colonialism" is itself an act of comparison, and Comparativity Complex Theory asks what conditions make that comparison available and whom it serves. For readers in London or New York, the analogy is illuminating: it lends moral gravity to the harvesting of clicks. For readers in Lagos or Nairobi, whose grandparents lived under the colonial ordinances, labour taxes and punitive expeditions of the first colonialism, the analogy can illuminate and trivialize at once. The data subject in Ohio and the data subject in Onitsha are not colonized in the same way, because the second arrives at the platform already carrying the weight of the first. This essay therefore uses the language of data colonialism with deliberate restraint. What makes the digital continuous with the colonial in Africa is not only that value is extracted, but that the extraction runs through the same channels of authority, the same division between those who write the rules and those who are written about, that scripture and empire had already cut.

This matters for the argument about authorship. If the problem were simply foreign domination, the solution would be simply national control. But national control of the canon is not the same as popular authorship of it, as Nigeria was about to demonstrate.

The Sovereign Who Could Not Delete a Tweet

In early June 2021, Twitter removed a post from the account of President Muhammadu Buhari. The post, which warned those the government blamed for unrest in the southeast in terms widely read as a threat of violence, was judged by the platform to breach its rules against abusive behaviour. On 4 June, the minister of information, Lai Mohammed, announced that the government was suspending Twitter's operations in Nigeria "indefinitely," citing the platform's use for activities said to threaten Nigeria's corporate existence.²⁰ The suspension lasted until January 2022, when the government lifted it after announcing that Twitter had agreed to conditions including the establishment of a local legal entity.²¹ In July 2022, the Community Court of Justice of the Economic Community of West African States ruled, in a case brought by the Socio-Economic Rights and Accountability Project and others, that the suspension had been unlawful and had violated freedom of expression and access to information.²²

Read as scripture, the episode is extraordinary. Here was a head of state, commander of an army, custodian of the nation's constitution, whose words were struck from the record by a private company headquartered in San Francisco. The sovereign could not keep his own text on the scroll. His response was not to contest the platform's exegesis on its own terms but to excommunicate the platform from the national territory. Two canons collided: the state's claim to author the national narrative, and the platform's claim to author the rules of speech within its own domain. The ECOWAS Court, in effect, reminded both that there was a third party to the covenant: the citizen, whose right to speak and to receive information neither the state nor the platform owned.

Consider, then, what had happened eight months earlier. In October 2020, young Nigerians took to the streets, and above all to the timeline, to demand the abolition of the police's Special Anti-Robbery Squad, a unit long accused of extortion, torture and extrajudicial killing. The hashtag #EndSARS organized marches, raised funds, documented abuses and carried the movement to the diaspora. On the night of 20 October, soldiers opened fire on protesters at the Lekki toll gate in Lagos. Amnesty International, which verified video footage and photographs circulating on social media, reported the following day that the army and police had killed at least twelve protesters at Lekki and Alausa.²³ The timeline had become an evidentiary archive. The footage that the state denied was preserved, time-stamped and geolocated by the very platforms the state would later ban.

The evidentiary role of that footage deserves more attention than it usually receives, because it shows that the timeline's authority depends on verification, not virality. A video is not evidence merely because it is shared; it becomes evidence when someone can establish when and where it was recorded, that it has not been altered, and that the chain from camera to archive is intact. Amnesty International's account rested on exactly this labour of authentication.²⁴ From the standpoint of security analysis, the Lekki footage is a case study in contested custody. The authorities disputed what had happened and who had fired; protesters held the only record; the platforms held the copies; and an international organization performed the verification that no Nigerian institution was then willing to perform. The record survived because its custody was distributed. It would have been far more fragile had it lived on a single server whose keeper had reason to lose it.

The state also turned to the older ledger. On 6 November 2020, TheCable reported that the Central Bank of Nigeria had obtained an order from the Federal High Court in Abuja freezing bank accounts linked to individuals associated with the protests.²⁵ Weeks earlier, when conventional payment channels had been closed to it, the Feminist Coalition, one of the movement's organizing groups, had begun accepting donations in bitcoin.²⁶ One need not romanticize cryptocurrency to see what had happened. When the state controlled the ledger, protesters migrated to a ledger whose code the state did not write. They did not escape authorship by others. They exchanged one set of authors for another.

This is why the Twitter suspension cannot be read as a simple morality tale in which the platform defends freedom and the state attacks it. The platform that preserved #EndSARS evidence also removed a president's post according to rules no Nigerian legislature had debated; the state that banned the platform also froze the accounts of its own citizens. Both were acting as keepers of a canon. Neither was accountable to the congregation. This author's position is that such episodes should be analysed through what might be called a comparativity of custodianship: the question is never simply which custodian is better, but by what historical conditions the citizen came to have only custodians to choose between.

Money That Speaks, Debt That Listens

No African technology has been more celebrated than M-Pesa, the mobile money service launched in Kenya by Safaricom. The celebration is not empty. In a widely cited study published in Science in 2016, Tavneet Suri and William Jack estimated that access to mobile money had increased per capita consumption and lifted some 194,000 Kenyan households, about 2 percent of the total, out of poverty, with effects most pronounced among households headed by women.²⁷ On the evidence, the access thesis has here a genuine success to point to, and it would be dishonest to deny it.

Yet the ledger that pays also reads. Kevin P. Donovan and Emma Park have shown how Kenya's digital credit economy grew on the foundations of mobile money, and how lenders converted the texture of everyday life, kinship, obligation, the social networks that a phone quietly records, into data for assessing creditworthiness.²⁸ Their account describes lenders capitalizing on intimacy itself: the relations through which Kenyans have long managed scarcity become raw material for scoring and for debt. What M-Pesa made possible as payment, the credit market made profitable as surveillance.

The scriptural parallel is precise. A ledger is a sacred text of a particular kind; it records obligations and determines who is in good standing. In the older Yorùbá moral economy, standing was established through reputation, through the witness of kin and community, through the long memory of exchanges honoured and dishonoured. The new ledger does not abolish that memory. It transcribes it into a format owned by a corporation and read by an algorithm whose criteria the borrower cannot inspect. The borrower becomes, in the most literal sense, a text read by others.

It is worth recalling that West African societies did not need Safaricom to invent trusted, distributed financial records. The rotating savings association, known among the Yorùbá as èsúsú or àjọ, has for generations pooled small contributions and released them in turn to each member, with the ledger held in the memory of the group and the collector's integrity guaranteed by the community that chose him or her. Such institutions have their failures and their frauds. But their rules are known to those who live by them, and their auditors are the members themselves. Set beside the rotating association, the digital lender is not simply more efficient. It is differently authored. The rules have moved from the circle of participants to the terms and conditions of a firm, and the audit has moved from the market square to a proprietary model. The question of whether digital finance "includes" the poor must be joined to the question of whether it has dispossessed them of an older authorship over their own credit.

The point is not that Kenyans should have refused mobile money. The point is that the celebrated success of access concealed the question of authorship until the debts came due. Who writes the scoring model? Who can audit it? Who can contest its judgment? A technology that gives the poor a wallet while taking from them the right to know how they are being read has not simply included them. It has scripturalized them.

The Hidden Scribes of Nairobi

Every canon needs copyists, and copyists have rarely been its authors. The medieval scriptorium was staffed by monks who reproduced texts they had no authority to amend; the colonial mission depended on catechists and interpreters whose labour made the Bible legible in African languages but whose names rarely appeared on its title page. The digital canon has its own scribes, and many of them work in Nairobi.

In February 2022, Billy Perrigo reported in TIME on an office in Nairobi where nearly two hundred young people from across Africa, employed by the outsourcing firm Sama, reviewed content for Facebook, watching videos of murder, rape, suicide and child abuse so that users elsewhere would not have to; some, he reported, took home as little as about $1.50 an hour.²⁹ In January 2023, Perrigo reported that OpenAI had used Kenyan workers contracted through Sama, earning between roughly $1.32 and $2 an hour, to label graphic text so that ChatGPT could be trained to avoid producing it.³⁰ Mary L. Gray and Siddharth Suri have called this invisible labour "ghost work": the human judgment on which supposedly automated systems depend, hidden so that the machine can appear autonomous.³¹

The Nairobi moderators are the scribes who purify the canon. They decide, case by case, what counts as violence, what counts as hate, what may stay on the scroll. Yet they do not write the rules they enforce; those arrive from policy teams elsewhere, in languages and with categories that may bear little relation to the cultural contexts in which the content was produced. The pattern is the catechist's: indispensable, underpaid, invisible, and denied authorship of the text whose integrity depends upon them.

There is a cruelty in this arrangement that the analogy with the scriptorium does not fully capture. The medieval copyist handled texts he revered; the Nairobi moderator handles the worst that human beings record of one another, hour after hour, so that the canon may appear clean to its readers. The purity of the feed is purchased with the psychological wounds of workers whose names the feed will never display. One is reminded that every scripture has had its underside: the heresies burned, the texts suppressed, the people who were made to carry away what the sanctuary could not hold. That the carriers of this burden are once again young Africans, paid a fraction of what the same judgment would command in California, is not an accident of the labour market. It is the latest expression of a long comparativity in which African bodies absorb the costs of a global order whose benefits accrue elsewhere.

There is, however, a development that distinguishes this story from the colonial one. In 2022, the former moderator Daniel Motaung sued Meta and Sama in Kenya, alleging exploitation and union-busting. In February 2023, a judge of Kenya's Employment and Labour Relations Court rejected Meta's argument that it could not be sued in Kenya and held that the company was a proper party to the case.³² The significance of that ruling for the argument here is considerable. A Kenyan court asserted that the author of the canon could be held answerable where its scribes labour. Authorship, in other words, carries liability, and liability can be claimed from below. That is a small crack in the tablet, but it is the right crack.

Is Ifá Binary? The Temptation of the Flattering Comparison

No discussion of African technology can avoid a claim that now circulates widely in popular writing and on social media: that Ifá, the Yorùbá system of divination and knowledge associated with Ọ̀rúnmìlà, is a binary system and therefore an ancestor of the digital computer. The claim has a scholarly root. Ron Eglash's African Fractals argued that African sand divination, particularly as practised among the Bamana, used binary operations, and proposed that geomancy travelled through the Islamic world into medieval Europe, where it formed part of the intellectual background to Leibniz's binary arithmetic.³³ The structure of Ifá lends itself to such readings. UNESCO, which inscribed the Ifá divination system on its Representative List of the Intangible Cultural Heritage of Humanity in 2008, describes a corpus of 256 odù, each with its specific divination signature determined by the babaláwo with sacred palm nuts or the divining chain.³⁴ Two states on each of eight positions do produce 256 combinations. The arithmetic is not in dispute.

What is in dispute is what the comparison accomplishes. Here the distinction between comparison and comparativity does its work. As a comparison, "Ifá is binary" is a formal observation about combinatorics. As an event of comparativity, it is something else: a bid for prestige in which an African knowledge system is validated by its resemblance to the most powerful technology of the present. The claim flatters, and flattery is a form of extraversion. It borrows the authority of Silicon Valley to authenticate Ilé-Ifẹ̀. The unspoken premise is that computing is the measure, and Ifá matters because it anticipated the measure.

That premise should cause discomfort. The odù is not a bit string. Its power lies in the ẹsẹ, the verses attached to each sign, which UNESCO notes are unknown in total number and continually growing, and which the babaláwo must know, chant and interpret for a particular client in a particular predicament.³⁵ Wande Abimbola's exposition of the Ifá literary corpus makes clear that this is an immense oral literature, carrying history, ethics, medicine and cosmology, with the diviner as its living interpreter.³⁶ The binary mark on the ọpọ́n Ifá opens the text; it is not the text. To reduce Ifá to its binary signature is to keep the index and discard the library. It is, in a precise sense, to do to Ifá what the digital does to everything: extract the machine-readable layer and leave behind the meaning that required a human reader, a lineage, and a relationship of trust.

This is not to reject Eglash, whose work did much to unsettle the assumption that Africa lacked mathematical sophistication. It is to insist that the question "Did Africa invent binary?" is less important than the question "Who today is authorized to interpret?" Ifá's genius was never a data format. It was an institution for governing interpretation: a corpus that grows, a priesthood that trains for years, and a practice in which the reading is always addressed to a person. If Africa has something to teach the age of algorithms, it is that a canon without an accountable interpreter is not wisdom but mere output.

Scanned, Not Returned

The same logic applies to the digitization of African objects held in foreign museums. In 2016, the artists Nora Al-Badri and Jan Nikolai Nelles announced that they had covertly scanned the bust of Nefertiti in Berlin's Neues Museum and released the three-dimensional data online under an open licence, an intervention they framed as a challenge to the museum's claim of possession.³⁷ Since November 2022, the Digital Benin platform, hosted by the Museum am Rothenbaum in Hamburg, has catalogued more than 5,000 objects looted from the Kingdom of Benin, drawing on data from 131 institutions in twenty countries and presenting Edo names and knowledge alongside the records.³⁸

Both initiatives are valuable. Neither settles the question of authorship. A scan is a new text, and texts have owners. Who holds the file of a looted object: the museum that scanned it, the artist who leaked it, or the community from whom it was taken? Kimberly Christen, writing about Indigenous knowledge in Australia, has warned that the open-access creed that "information wants to be free" can reproduce colonial collecting, because it assumes that all knowledge should circulate to all audiences, whatever the protocols of the communities that produced it.³⁹ Some knowledge is restricted for reasons of kinship, initiation or sacredness. The digital copy can become a second looting, conducted with better lighting.

From the perspective of security, which this author approaches as a heritage-intelligence analyst, the digital object raises questions that are also ethical: who can verify that a record is authentic, who can trace its provenance, and who can see who has accessed and altered it? The restitution debate is taken up elsewhere; the narrower point here is that digitization shifts the site of heritage from the vitrine to the database, and the database, too, has keepers.

Diacritics and the Machine

Language is where the question of authorship becomes most intimate. Pratik Joshi and colleagues, surveying the state of natural language processing in 2020, classified the world's languages according to the data and tools available for them and found an overwhelming concentration of resources in a small number of languages, with most of the world's languages, including the great majority of African languages, left far behind.⁴⁰ A language model that has read little Yorùbá, Hausa or Kiswahili will not simply speak them poorly. It will speak them through English, importing categories and assumptions from the language on which it was mostly trained.

The Yorùbá case is instructive at the level of the smallest mark. Yorùbá is tonal, and its standard orthography uses diacritics to distinguish words that would otherwise look identical; much digital Yorùbá text circulates without them, because keyboards, platforms and habits of typing were designed elsewhere. Iroro Orife, treating the restoration of Yorùbá diacritics as a machine-translation problem, noted that these marks are essential for disambiguation and pronunciation and for speech and language technologies, and released his models and data openly.⁴¹ A people whose tones are stripped by the keyboard is a people whose language is being rewritten by the tools they use to write it. The difference between Ọ̀rúnmìlà and "Orunmila" is not cosmetic. It is the difference between a name spoken and a name transliterated for someone else's convenience.

Here, too, there is a counter-example that points toward authorship rather than access. The Masakhane community, a grassroots network of African researchers, published a study in 2020 describing a participatory approach to machine translation for African languages, involving dozens of contributors and producing datasets and benchmarks for more than thirty languages.⁴² The significance exceeds the technical. It lies in the fact that African speakers, including many without formal training in the field, wrote the data, set the benchmarks and named themselves as authors. That is the difference between being included in someone else's model and co-authoring one's own.

Authorship in language technology is more than a matter of data, however. It requires computing power, storage, licensing arrangements and governance over how the resulting datasets may be used. A community that contributes its proverbs, its oríkì and its everyday speech to an open dataset may find that the largest beneficiaries are firms with the capital to train models at scale, who can then sell the resulting capabilities back to the speakers who supplied the words. The lesson of Christen's critique applies here as well: openness is not the same as justice, and a corpus given freely can be enclosed by whoever holds the machinery to exploit it. Authorship must therefore extend to the terms under which African language data is shared, and to institutions on the continent that can hold, govern and use it.

Security as Ethics: Verified Identity and Auditable Records

The question of authorship leads, finally, to the question of security, which is too often treated as a merely technical matter. In December 2020, the Nigerian Communications Commission directed telecommunications operators to link subscribers' SIM cards to their National Identification Numbers, suspending new SIM sales and initially giving citizens two weeks to comply.⁴³ Keith Breckenridge's history of biometric government in South Africa shows that the dream of a total, centralized, fingerprint-based register of persons has a long colonial pedigree, and that the biometric systems now spreading across the continent inherit that genealogy.⁴⁴ Identity is scripture of a different sort: the register determines who exists before the state, and the person who is missing from it may find themselves unable to bank, travel, telephone or vote.

Nigeria's Data Protection Act of 2023, signed in June of that year, established the Nigeria Data Protection Commission and placed the processing of personal data on a statutory footing.⁴⁵ It is a real step. But law is a text; enforcement is a practice; and the question that matters is whether citizens can see, contest and correct what is written about them.

The framework this author has developed as Heritage Security Intelligence begins from the proposition that security is not the opposite of openness but the condition of trustworthy memory. A verified identity, an auditable record, a documented chain of custody: these are technical controls, yes, but they are first of all ethical commitments. They answer the questions any community should ask of a canon. Who wrote this? Who changed it, and when? Who may read it, and who decided? An unauditable system is, in scriptural terms, a revelation without witnesses. Heritage Security Intelligence insists that the record of a people, whether an object's provenance, a citizen's identity or a community's language data, must be secured in a way that the community itself can audit. Security without accountability is surveillance; openness without security is extraction. The ethical middle is authored custody.

Authorship, Not Access

The argument can now be gathered. From Sinai to Silicon Valley, the arrival of a new medium of record has repeatedly divided people into authors and readers, custodians and congregations. The printing press did so; the mission press exported the division; radio and the pulpit microphone amplified it; the timeline disguised it by letting everyone post while reserving to a few the power to order, rank and remove. The Nigerian state and Twitter fought over who owned the scroll while citizens needed a regional court to remind both that the covenant had a third party. Kenyan mobile money gave wallets and took readings. Nairobi's moderators cleaned the canon without writing it. The binary reading of Ifá flattered Africa by the measure of the machine while discarding the interpreter. Scans multiplied objects without returning them. Language models learned African languages without their tones.

The policy vocabulary of access and inclusion misreads each of these cases. Access to a canon one did not author is admission to a congregation, not a seat at the council. This is the threshold of discomfort that the digital question poses for Africa: that the celebrated arrival of connectivity may deepen, rather than dissolve, the old scriptural division, now with the added intimacy of data. Heritage morphs or it dies, and technology is the medium through which heritage is now morphing. Whether it morphs into something Africans author or something they merely consume is not a technical question. It is a question of power.

Two objections deserve an answer. The first is that authorship is a utopian demand in a world where the capital, the chips and the data centres belong to a handful of corporations and states, none of them African. The objection is correct about the distribution of power and wrong about its permanence. The printing press was once a European monopoly too, and African presses, newspapers and publishing houses, from Lagos to Nairobi, became instruments of nationalist authorship within a few generations. The second objection is that the call for African authorship risks a new nativism, a demand that only Africans may write about Africa. That is not the argument. Authorship here means accountable custody, not ethnic ownership: the right of those who are written about to see, contest, amend and govern the record. It is a demand that Africans be authors among others, not that others be forbidden to write.

What would authorship look like? It would look like Masakhane: speakers building the datasets of their own languages. It would look like a Kenyan court holding a platform answerable where its scribes labour. It would look like data protection that citizens can invoke, identity systems they can audit, and digital heritage whose keys are held by the communities it describes. It would look, in short, like the institution Ifá has always been: a growing corpus with accountable interpreters, addressed to persons rather than processed about them. Whoever writes the code writes the canon. Africa's task is to write.