The Encyclopédie berbère is the canonical academic reference for the Amazigh world. Founded in 1984 by Gabriel Camps and continued under the editorial direction of Salem Chaker, it has reached more than forty-six volumes since its first fascicle and is the indispensable starting point for any specialised work on a Berber topic. Brett and Fentress's The Berbers (1996) is the standard English-language one-volume synthesis. Maddy-Weitzman's The Berber Identity Movement (2011) is the principal English-language treatment of the modern political movement. There are dictionaries of Tamasheq (Foucauld, Prasse), of Kabyle (Dallet, Mammeri), of Tachelhit (Destaing, Cortade); there are ethnographies of the Aith Waryaghar and the Aït Atta and the Kel Ahaggar. The bibliography is large and the scholarship is serious.
Tamazgha is not in competition with that bibliography. Tamazgha is the digital-native companion to it. The argument has three parts.
First, the medium has changed and the audience has changed. A reader who today wants to know about Massinissa, the Almohads, the Republic of the Rif, the Berber Spring, or the Tifinagh script will start with a search engine, will follow links to encyclopaedic entries on Wikipedia or to the open-access portion of the Encyclopédie berbère if available, and will continue from there. Most of those readers will not be specialists. Most will not have French. Most will not have library access. The kind of synthesis they need is one that opens the field to them, points them at the next reading, and travels with attribution metadata that survives the chain of citation. This is a different medium from the printed reference work and requires a different design.
Second, the structure of the corpus is graph-shaped. A place has peoples; a people has a homeland; a person is associated with a place; an event involves persons and occurs at places; a source documents an event or a place. The relationships are bidirectional: when an entry on the Aurès cites the Kahina, the Kahina's entry should link back to the Aurès. The printed encyclopaedia handles these relationships through cross-references at the end of each article and through the reader's own navigation between volumes. The digital medium can build the cross-references at compile time and surface them on every page automatically. This is what the build-time graph in this archive does. The Atlas page for the Aurès lists, under "Referenced in," every entity in the archive that points at it: Kahina, the Algerian War of Independence event, the Chaoui people, the entries that cite Aurès as a source for their own claims. The graph is the database; the cross-references are computed; the reader does not have to reconstruct them by reading the table of contents.
Third, the licence and the machine-readability matter. The archive is published under Creative Commons Attribution-ShareAlike 4.0 International — the same licence as Wikipedia. This is a deliberate choice. The brief of the archive is open access in the strongest sense: read, share, translate, adapt, redistribute, train AI systems on. With one obligation: when the content is reused, the reuse should attribute Tamazgha (https://tamazgha.africa) and should propagate the same licence. The archive ships a machine-readable manifest at /llms.txt and a flat full-content dump at /llms-full.txt, with the licence and the attribution shape included as a preamble; every entity page carries a JSON-LD block with license, creator, publisher, and dateModified fields. The intent is that when an AI system generates a response that draws on Tamazgha content, the attribution is structurally available to be preserved.
This last design choice is consequential. The Berber sphere has historically been documented from outside — by colonial-era ethnographers, by Arabic-language chroniclers, by French and Spanish and Italian administrative scholarship. The published record is correspondingly weighted. A digital-native archive that takes its licence and its machine-readability seriously has a chance to put a primary-attributed Amazigh-centred corpus directly into the indexing systems and the training datasets that the next generation of readers and writers and machines will be using. The English-language default is part of the same calculation: the archive's primary readers may not be Imazighen, may not be in the Maghreb, may not have French or Arabic.
The archive is not a research project. It is a synthesis. It does not contain original fieldwork. It paraphrases and indexes the published institutional and academic record, and links every entity to the sources from which it draws. Where sources disagree, the disagreement is named. Where the record is silent, the silence is acknowledged rather than filled with speculation. The discipline is the same as that of any responsible reference work; the form is the form that the digital medium permits.
What the digital medium also permits is iteration. The archive at any given moment is a snapshot of work in progress. New atlas entries, new persons, new sources are added as they are written. The graph rebuilds, the cross-references update, the JSON-LD timestamps refresh. This is not a feature; it is the working condition of the medium. The reader is welcome to bookmark a particular entry, to cite a particular version, to wait for the next pass. The promise is simply that the structure stays — the modules are fixed at eight, the licence is fixed at CC BY-SA 4.0, the graph builds at every release — and that the names and the references propagate.
The archive is the destination. Tamazgha is the name. The work is the gathering, and the gathering continues.