Fernweh. Mamihlapinatapai. Sobremesa. Toska. These words exist in German, Yaghan, Spanish, and Russian respectively, and none of them translates cleanly into English. Fernweh is an ache for distant places — the longing to be somewhere other than where you are, closer to melancholy than to adventure. Sobremesa is the time after a meal when conversation continues, the food is finished, and no one wants to leave — a concept for which English has no word because English culture has not always considered it a distinct and valuable thing. Toska is, in Nabokov’s description, “a longing with nothing to long for.”
These examples are well-known, and there is a well-established genre of popular writing about untranslatable words that treats them as charming curiosities. This is not that genre. The argument here is structural, and it has consequences for how we understand the loss of linguistic diversity.
The phenomenon extends beyond individual words to entire modes of reasoning and entire intellectual traditions. The Kantian tradition in German philosophy developed in a linguistic context in which certain distinctions between types of knowledge — the Kantian distinction between Verstand and Vernunft, roughly understanding and reason — could be made with precision because German had separate words with separate connotations for concepts that English collapses into rough synonyms. Japanese has a rich vocabulary for relational concepts — amae describes a type of benign dependency and trust that has no precise English equivalent — that has generated a philosophical and psychological literature in Japanese that does not translate naturally into English-language frameworks.
New ideas are increasingly formed in English, by people trained in English, drawing on English-language intellectual traditions. The conceptual vocabulary available to them is the conceptual vocabulary of English.
This is not linguistic relativism in its strong form — the claim that language determines thought and that cross-linguistic translation is impossible. The weak claim, which the evidence supports, is that the linguistic categories available in a given language influence the conceptual distinctions that speakers find natural to make, and that when intellectual communities shift from one language to another, something is lost in the migration.
The shift in academic and intellectual discourse toward English as its sole serious medium has been rapid and comprehensive. Web of Science data shows that the proportion of scientific papers published in languages other than English fell from approximately 45% in 1990 to less than 10% by 2020. German, which was the dominant language of chemistry and physics through the early twentieth century — Einstein published in German, the quantum revolution was conducted in German seminars — is now peripheral in both fields. French, which carried an enormous intellectual tradition in philosophy, sociology, and literary theory, is rapidly retreating from the global academic conversation.
The consequence is not only that ideas are now expressed in English that were once expressed in German or French. It is that new ideas are increasingly formed in English, by people trained in English, drawing on English-language intellectual traditions. The conceptual vocabulary available to them is the conceptual vocabulary of English. The distinctions that feel natural, the problems that seem important, the framings that are ready-to-hand — all of these are shaped by the language in which the thinking is done.
What is lost in this homogenisation is difficult to measure precisely because its consequences are absences. We cannot easily inventory the intellectual moves that have not been made because the linguistic tradition that would have generated them is no longer producing scholarship at the necessary scale.
The technology most often cited as a solution to this problem is machine translation. AI translation systems have reached near-human performance on many language pairs and are continuing to improve. But translation, however good, is not formation. The question is not whether ideas expressed in German can be made accessible to English readers. They can, with effort and skill. The question is whether, when German intellectual communities stop producing new scholarship in German, the ideas that were being developed in German disappear along with the intellectual context that would have generated them.
The evidence from previous episodes of linguistic consolidation — the decline of Latin as a scholarly language, the marginalisation of Arabic after the medieval period — suggests that consolidation is generally accompanied by a narrowing of the intellectual tradition that survives the shift. The ideas that cross into the dominant language are those that fit the dominant conceptual vocabulary. Those that do not fit are not translated. They are lost.
