Summary

  • On 24 May 2004 in Dakar, participants asked what it would cost to translate AFRINIC documents and its website into languages other than English. Management said translation was under close consideration but that scarce bootstrap resources were being directed towards strengthening the organisation and obtaining formal recognition.
  • The surviving budget names two separately controlled budgets and many exact categories, but no translation, localisation, interpretation or language-services line. Its composite communications and training envelopes cannot be converted into evidence of translation funding.
  • External actors expressed willingness to support translation and said translation of the current meeting was already being provided. No amount, target language, medium, coverage, acceptance, contract or continuing obligation is recorded.
  • All members present adopted the whole 2004 budget by consensus. That was a bounded private corporate decision, not a separate translation appropriation, proof of delivery, regional legislation or authority over people who were not principals to the decision.

The revealing feature of AFRINIC’s founding translation debate is not a dramatic rejection. It is the distance between four ordinary institutional acts. A question was raised. Management explained a priority. Outside actors offered help and described a service at the meeting. The whole budget was adopted. Each act mattered, but none was interchangeable with the next. A question was not an amendment; close consideration was not an appropriation; willingness was not an accepted contract; and approval of the entire budget was not a receipt for a multilingual capability.

That separation is especially important because language access is easy to praise in the abstract and difficult to inspect in practice. Translating a document requires a source version, a target language, a quantity of text, a deadline, a qualified provider, terminology control, review and correction. Localising a website adds inventories, release coordination and continuing maintenance as the source pages change. Interpreting a meeting requires languages, hours, sessions, equipment or another delivery method, and some evidence of who could actually use the service.

The May 2004 record acknowledges the subject without supplying those capability measures.

The result is not evidence that no language assistance existed. The meeting report expressly says some translation of the current meeting was being provided. Nor is it evidence that nothing was ever spent. The result is narrower and more defensible: the surviving records do not expose a translation-specific amount, allocation, language list, output commitment or durable delivery mechanism. What can be audited is the priority statement and the named budget categories. What cannot be audited is the size or coverage of the language capability.

The question in the room

AFRINIC-1 was held in Dakar on 23 and 24 May 2004. Policy discussion occupied the morning of 24 May; the translation-cost question emerged in the administrative-document session of the afternoon annual general meeting. That immediate setting matters because the intervention was attached to a budget being presented for a young organisation, not to a later language strategy or a general cultural debate. Project manager Adiel Akplogan presented the 2004 budget. Participants voiced concerns about its administrative part and asked about the cost of translating AFRINIC documents and the website into languages other than English.

The recorded wording defines both the ambition and the evidential limit. The surfaces were documents and the website. English was the only language explicitly named, and it appeared as the reference point in “languages other than English.” No target language was listed. The identities, affiliations, voting status and particular language needs of the questioners were not recorded. The account calls what they did questions and concerns; it does not describe a motion, amendment, vendor request or agreed specification.

Separate questions in the same passage concerned whether the training provision was sufficient and whether the budget could cover travel, including Board travel. Those issues show that several demands competed for scarce bootstrap resources, but they should not be expanded into a history of the training programme or travel policy. Training is not a proxy for translation, and travel is not a language line. Their relevance here is limited to the budget pressure surrounding the translation question.

Akplogan’s response was a priority explanation. He said the budget rested mostly on expected donor funding, reflected expectations and covered the most important activities needed at that stage. Translation, he said, was under close consideration. Yet the route to making the bootstrap faster and easier was to allocate available resources to strengthening the organisation and obtaining formal recognition. He connected the budget’s design to the organisation’s main goals for 2004, compliance with ICP-2 criteria and recognition by ICANN as a regional registry.

That explanation tells us what management said it valued in sequence. It does not establish that recognition required translation to be deferred, that translation had been rejected permanently, or that a particular saving would be achieved. “Under close consideration” identifies attention, not an implemented workstream. “Available resources” identifies scarcity, not a priced trade-off. Formal recognition was management’s stated objective; the fact that an official report records the objective does not prove that recognition was necessary on the terms asserted, or that recognition could confer legitimacy or public authority.

Four entries in the decision ledger

The exchange becomes clearer when its components are kept in their exact procedural states.

Recorded item Exact status What the record supports What it does not support
Translation cost Question raised Participants asked about translating documents and the website into languages other than English. A formal motion, amendment, appropriation, procurement request or delivery order.
Management response Priority explanation Translation was under close consideration; available resources were prioritised for organisational strengthening and formal recognition. A cost estimate, permanent rejection, proof of zero spending or implemented language programme.
External support Willingness expressed; current meeting translation said to be provided A representative of the Agence Intergouvernementale de la Francophonie, together with the French minister of Foreign Affairs, offered support and described translation at the meeting. A named target language, accepted grant, contract, continuing service, document translation or website localisation.
Budget decision Whole budget adopted by consensus of all members present The complete 2004 budget received the recorded corporate approval. A translation-specific vote, acceptance of the outside offer, a language policy or proof of delivery.

These statuses are more than careful phrasing. They prevent an institution’s sequence of conversation from being compressed into an outcome it did not document. If a question becomes a proposal in the retelling, a proposal can soon sound like an approved policy. If willingness becomes a promise, and a promise becomes funding, an institution can acquire a capability on paper without anyone identifying its price or output. If whole-budget consensus is described as approval of every possible use within every category, the distinction between authority to spend and evidence of performance disappears.

The record instead leaves a clean chain of custody for claims. The question proves that language access was salient enough to be raised. The response proves that management considered it but placed two organisational priorities ahead of a named translation commitment. The offer proves that outside support was available in principle and that some service at that meeting was reported. The consensus proves that the members present accepted the overall budget. None of those facts supplies the missing links of quantified acceptance, procurement, production, maintenance and use.

Two budgets, not one language pool

The surviving presentation says the first year had two budgets. The management budget was under AFRINIC’s control and supported by donations and grants from the community. The operation budget was not under AFRINIC’s control; the deck attributed its support and management to South Africa’s Department of Communications, with CSIR as the hosting organisation. The control split matters. Even if two line items appear adjacent or bear similar labels, they cannot be silently added into one AFRINIC-controlled pot, much less a translation fund.

The management budget set out the following amounts:

Management budget item Amount
Total income USD 133,500
AFRINIC service, projected USD 12,250
Hosting countries USD 12,250
Local/international support, expected USD 109,000
Total operations USD 117,783
General office expenses USD 15,230
Telecommunications USD 4,000
Consulting USD 6,500
Human resource USD 22,353
Meeting/Training/Coms USD 12,000
Training USD 37,700
Travel USD 20,000
Taxes and other charges USD 2,696
Assets and initial setup cost USD 11,050
Total expenses USD 131,529
Projected gain USD 1,971

The operation budget set out a different structure under different control:

Operation budget item Amount or status
Total income USD 279,319
Hosting countries USD 279,319
Total operations USD 181,875
General office expenses USD 13,767
Telecommunications USD 20,000
Consulting USD 2,000
Human resource USD 108,108
Meeting/Training/Coms USD 10,000
Training No amount shown
Travel USD 28,000
Taxes and other charges USD 36,563
Assets and initial setup cost USD 60,880
Total expenses USD 279,319
Projected gain or loss No amount shown

Across the presentation’s eleven pages, no category is labelled translation, localisation, interpretation, language services, translated documents, translated website, translator, language staff or multilingual participation. The closest-looking labels are composite or adjacent envelopes. The management side has USD 12,000 for Meeting/Training/Coms, USD 37,700 for Training and USD 20,000 for Travel. The operation side has USD 10,000 for Meeting/Training/Coms, no amount shown for Training and USD 28,000 for Travel. None discloses a translation share.

The management budget’s USD 109,000 in expected local/international support is also important, but “expected” must carry its full weight. It cannot be reported as cash received. Likewise, the operation budget’s USD 279,319 income and USD 279,319 expenses cannot be folded into the management budget’s USD 133,500 income and USD 131,529 expenses and then described as AFRINIC’s discretionary capacity. The operation budget was separately controlled according to the deck. Control over classification and spending is part of capability; a number unavailable to the decision-maker is not its appropriation.

The budget presentation bears an event label of 24 May 2004, while slide footers show 31 May 2004. That mismatch does not justify a claim about when the deck was first published, exported or revised. It is sufficient to identify the instrument linked to the meeting and the amounts it displays. The near-contemporaneous public account was posted by Akplogan to the AfNOG list on 4 June 2004, eleven days after the annual general meeting. A later official nine-page report reproduces the same sequence, but its file metadata shows creation in 2012 and modification in 2019.

The later official copy corroborates the text; it does not become the contemporaneous file by institutional designation alone.

This source history illustrates a larger rule: official records prove the words, acts, amounts, dates, roles and institutional claims they contain within their evidential scope. They do not prove their own completeness, the legitimacy of the body that issued them, or unrecorded performance. A report is not a verbatim transcript. A presentation is not an expenditure ledger. A corporate vote is not a public-law appropriation. Treating each instrument according to what it can actually show is not scepticism for its own sake; it is the condition of honest institutional memory.

The capability that cannot be reconstructed

For documents, the evidence contains a demand expressed as a cost question but no named budget line, price, unit cost, funded volume, coverage or delivery evidence. No titles are identified. There is no page count, word count, source-version marker, target language, publication date, translator, reviewer or correction path. The supported finding is that the need was recognised in debate while quantified document-translation capability remains unproved.

For the website, the same question was raised and the same gaps remain. There is no inventory of pages, target language, release date, software or editorial owner, maintenance budget, update frequency, quality process or evidence that translated pages appeared. A website is not a one-time document. Its language capability decays whenever source text changes without corresponding updates. Even an initial translation would not establish continuing localisation unless version alignment and correction were owned and funded.

For the AFRINIC-1 meeting, the record goes further but still not far enough for measurement. External actors said translation of the current meeting was already being provided. That statement supports the conclusion that some meeting translation was reported.

It does not tell us whether the service was oral interpretation, written translation or both; which sessions or hours it covered; whether it was simultaneous or consecutive; whether remote participants could use it; who provided it; how many translators or interpreters were involved; what source material and terminology controls they used; whether accuracy was independently reviewed; how many participants needed or used it; what it cost; whether it was paid or in kind; or who bore that cost.

The organisation names in the report do not fill the language blank. A representative of the Agence Intergouvernementale de la Francophonie and the French minister of Foreign Affairs were associated with the willingness to support translation activities. Those identities do not establish the target language of the meeting service. The only language the exchange explicitly names is English, as the reference language from which “languages other than English” were contemplated. Geography, institutional title and later policy cannot be substituted for a missing language specification.

For future translation activity, willingness is the first rung of a possible arrangement, not the last. The record has no amount, in-kind valuation, term, condition, named output, acceptance, purchase order, contract, invoice or later performance. It does not say whether AFRINIC accepted continuing help, whether support covered documents or the website, or whether it ended with the meeting. An external offer can lower the organisation’s immediate cash requirement, but it can also place essential capacity outside the budget, outside ordinary ownership and outside the member’s ability to compare planned with actual delivery.

The composite management envelope of USD 12,000 and the separately controlled composite operation envelope of USD 10,000 do not close any of these gaps. “Coms” may invite speculation, but a category label is not an allocation record. The USD 37,700 management Training line is substantial, yet training has its own purpose and cannot be reassigned rhetorically to translation. Consulting and Human resource are equally unsuitable as inferred language lines. Absence of a named line does not prove zero spending. It proves that no translation-specific appropriation or auditable quantity is exposed in the surviving budget table.

Language access as an economic mechanism

Why does the missing capability matter? A participant who cannot work comfortably in the meeting’s language faces an information-access cost. That cost might be paid in money for independent translation, in time spent deciphering documents, in dependence on an intermediary, in slower preparation, or in silence when an intervention becomes too difficult to make precisely. A website in an inaccessible language can make basic procedural knowledge costly before a meeting even begins. Translation and interpretation can reduce those frictions.

The May 2004 record contains no participant-level dataset, so it cannot tell us who stayed away, who misunderstood a paper, who paid for help or whose intervention changed. The economic mechanism should therefore not be inflated into a historical outcome claim. It is enough to say that language can alter the cost of informed participation and that an institution wishing to claim practical access should be able to show the capability it purchased or received.

A dedicated line creates visibility. It makes the amount and priority legible, though a line alone does not prove delivery or quality. A composite line creates flexibility. Managers can adjust among meetings, training and communications, but members cannot identify what quantity of translation was funded from the category. External support can reduce near-term expenditure. It can also make continuity dependent on the provider, obscure the in-kind value and weaken direct control over language coverage, timing and standards. None of these arrangements is inherently improper. Each produces a different audit surface.

The management budget’s donor dependence sharpens the issue. USD 109,000 of USD 133,500 in management income was expected local/international support. Future translation assistance was also offered externally. A young organisation using expected support may reasonably sequence its expenditure around what it believes will secure operational standing. But expected funds are not received funds, and willingness is not an obligation. When core access depends on either, the member cannot know the durable capability without a record of acceptance, conditions, ownership and actual output.

Language access is also recurring rather than static. A translated page can diverge from its source. A changed policy document can remain current in one language and stale in another. Terminology can drift between a meeting, a website and later written records. The organisation therefore needs more than an initial purchase if it wants to claim continuing capability. It needs source-version matching, publication dates, correction procedures, authoritative-status labels and a responsible owner. The 2004 documents do not show those controls; that is a capacity gap, not evidence that a specific translation failed.

The strongest case for the founding choice

The most persuasive defence of management’s sequence starts with the organisation’s age and constraints. A start-up registry needed a corporate base, people, infrastructure and the formal recognition it believed necessary to begin independent service. Translation can be expensive, recurring and version-sensitive. Spending scarce donor-backed resources on a wide language programme before the institutional core was stable might have purchased outputs that quickly became obsolete. Keeping money in composite categories could preserve flexibility while priorities were changing.

The record supports several parts of that defence. It describes a bootstrap stage. It says the budget was mostly based on expected donor funding. It identifies two budgets with different control. Management openly stated that strengthening the organisation and obtaining recognition came first. Outside participants reported that translation was already being provided at the meeting and expressed willingness to support future activities. The management budget also contained USD 37,700 for Training, showing that a significant category adjacent to broader capacity building was not ignored.

This is a serious case, not a straw man. The absence of a dedicated line cannot prove bad faith, hostility to multilingual access or total non-delivery. A small organisation may properly prefer a dated sequence to simultaneous promises it cannot execute. It may accept in-kind support for a meeting rather than create permanent staffing. It may decide that a flexible communications envelope is more useful than a narrow sub-line while the pattern of demand remains unknown.

But the defence has a limit: flexibility cannot be counted as measured capability. Nothing in the record gives translation a price, language matrix, output commitment, owner or delivery receipt. The training line is not translation. The composite communications labels cannot be allocated after the fact without a ledger. The reported service at one meeting does not establish document translation, website localisation or continuing support. Recognition-first may be a rational sequence, but without a dated trigger for the next language step it remains an indefinite priority statement rather than a controlled deferral.

The appropriate conclusion is therefore not that management made an obviously wrong choice. It is that the choice left an accountability gap between acknowledged need and auditable capability. A governance analysis should be able to accept the force of the bootstrap defence while still refusing to award performance for unpriced flexibility. Those propositions are compatible.

Access does not manufacture authority

There is another temptation in language debates: to treat multilingual provision as proof that an organisation represents a region. That conversion must be rejected. Language access is a practical good because it can reduce information friction and widen the range of people able to understand or contribute. It does not authorise a private corporation to speak for absent operators, countries, users or language communities. Participation supplies evidence, expertise and sometimes consent from an identified principal. It is not a shortcut to a continental mandate.

AFRINIC was and remains a private registry bookkeeper and coordinator. Its proper role is narrow: maintaining records and coordinating services needed for unique, accurate and continuous number-resource administration through private arrangements. It may budget for, procure and deliver language services that improve access to those functions. It is not a state, legislature, regulator, court, police force, prosecutor, punitive body, confiscator, territorial authority, continental authority or owner of Africa’s number resources.

A translated room, donor contribution, corporate consensus or recognition by another private coordinator cannot enlarge those powers.

The 24 May consensus was a corporate act under the rules applicable to the members present. The report does not provide their number, names, individual positions or a roll call. It does not prove that they represented every member, every operator, every Internet user, every country or any language community. Describing the act accurately does not diminish it. The members present adopted the whole budget. That is enough to establish the bounded decision and not enough to transform it into regional legislation.

The same discipline applies to other institutions cited for analytical support. NRS belongs to research, advocacy, convening and the representation of members that explicitly authorise it; it does not operate a registry, RPKI, WHOIS or RDAP, appeals, settlement, elections, custody or continuity, and it supplies no witness to this 2004 exchange. BTW.Media provides a bounded mechanism for understanding how language, travel, time, money and procedural fluency shape practical participation, but it does not prove the historical facts of the budget session.

LARUS provides an operator-facing explanation of the narrow registry and recordkeeping role, not testimony about the debate or expenditure. The 2004 facts stand on the meeting accounts and budget presentation; doctrine and analysis keep the claims within their proper authority.

This division of roles protects both access and accountability. If translation is expected to validate authority, evidence about actual delivery becomes politically awkward: admitting gaps would seem to weaken legitimacy. If translation is treated instead as a service capability within a private coordinator’s limited role, its quality and coverage can be inspected without constitutional theatre. Members can ask what was funded, what was received and what changed, while leaving sovereignty where it belongs—with none of the registry’s corporate rituals capable of creating it.

The finding the record can bear

The historical finding is precise. On 24 May 2004, participants in the Dakar annual general meeting asked about the cost of translating AFRINIC documents and its website into languages other than English. English was the only explicitly named language; no target-language list survives in the record. Akplogan said translation was under close consideration but explained that available bootstrap resources were being directed towards strengthening the organisation and obtaining formal recognition.

He described a budget mostly dependent on expected donor funding and tied its priorities to the organisation’s main goals, ICP-2 criteria and recognition by ICANN.

A representative of the Agence Intergouvernementale de la Francophonie, together with the French minister of Foreign Affairs, expressed willingness to support translation activities and said translation of the current meeting was already being provided. The amount, target language, medium, sessions, staffing, quality, payer and continuing coverage were not recorded. All members present then adopted the whole budget by consensus. No separate translation motion, amendment, vote, appropriation or acceptance of the external offer is recorded.

The budget deck contains exact amounts and two control domains but no translation-labelled line. The amounts cannot be repurposed by inference. The record neither proves zero spending nor allows a claim that AFRINIC funded, contracted, delivered or rejected a particular document or website programme. It supports an acknowledged need, a stated priority away from a named line, an unquantified offer, some reported meeting service, and a gap between formal openness and measurable language capability.

That narrow conclusion deliberately does not reconstruct or judge the entire 2004 budget, which is a distinct subject. It does not analyse the five fee categories, their prices, objections, dissent, abstention or later revisions. It does not retell April document publication or the visibility of language versions. Nor does it import later language policies, later operational developments, disputes, elections, litigation, receivership, scarcity, transfer rules or enforcement into the motives of a May 2004 budget conversation.

The translation-cost question deserves its own accounting precisely because it disappears when made to carry those other histories.