A humanities scholar applies Umberto Eco's framework to the MediaWiki extension at the center of the Wikimedia Foundation's multilingual future, and surfaces design assumptions about who the system is really for.
A new peer-reviewed paper reads the MediaWiki extension at the center of the Wikimedia Foundation's multilingual future as a contemporary attempt at what Umberto Eco called a "perfect language." The critique, surfaced last month in the community-run Wikipedia Signpost, comes from Michael Falk of the WikiHistories project, writing in AI & Society. It is a design argument about assumptions baked into a live system.
Eco's "perfect language" is a defined philosophical concept, not a compliment or a slur. In Eco's reading, perfect languages are systems that promise to recover the lost correspondence between words and the world: classifications, taxonomies, and structured representations that claim to reach the structure of reality itself. The cost of such systems is usually a presumption that the right code, given the right objects, will produce the right knowledge. Falk's paper, as summarized in the Signpost column, applies Critical Code Studies to Wikilambda, the MediaWiki extension underlying Wikifunctions and Abstract Wikipedia, and finds the project's stated universal-multilingual-knowledge goal hard to reconcile with the architectural choices being made on the ground.
The critique lands hardest on assumptions about who the system is implicitly designed for. Wikifunctions is, in the Wikimedia Foundation's framing, a collaboratively edited library of computer functions intended to support natural-language generation over a shared content model. The implicit user in that design is a contributor who can already think fluently in the shared model: someone who can read the function library, write new functions in a typed language, and trust that the renderer will produce accurate prose in any target language. That figure bears only a passing resemblance to the global volunteer editor who actually shows up to Wikipedia, and to the contributors whose local-language and local-frame knowledge the project is supposed to mobilize. Falk's argument, as carried in the Signpost summary, is that the gap between the implicit user and the actual user is structural. It runs through the function library, the content model, and the rendering pipeline. A universal-access goal does not survive that kind of mismatch.
The Signpost's column framing tends to crowd this out. The column runs under the title "WikiLambda the Ultimate," and the "one ring to rule them all" wordplay loads the story toward rhetorical dismissal before the design argument lands. The more useful question, the one Falk's paper is actually posing, is what a less "perfect-language"-shaped version of the project would require.
If the goal is multilingual access to a defined, well-structured subset of knowledge, the project has a fighting chance of saying what it is and shipping something a real user can use. If the goal is universal knowledge access via a single shared content model, the architecture is making a bet that the world's languages can be made to fit a single representational grammar. Eco spent a career explaining why that bet is hard to win. The paper's contribution is to point out that the bet is being made implicitly, not argued for explicitly. A serious response to the critique would name a metric tied to actual reader comprehension across languages, not to the throughput of the function library, and would either expand the population of contributors who can model knowledge in the shared representation or build tools that translate local-language contributions into that representation without losing them. Neither path is impossible, but neither is what the current design assumes.
None of this requires abandoning the project. It requires treating the perfect-language critique as a question about assumptions rather than a verdict. The Signpost column is right that Falk's paper is the kind of work the Wikimedia research community should engage with directly. It is also right that the column's "WikiLambda the Ultimate" framing risks turning a careful design argument into a punchline. The argument deserves better than the pun, and the project deserves better than a defense that ignores what the paper is actually saying.