Like other linguistically diverse institutions, the European Union promotes conflicting values äs it chooses official languages. Increased activism and the admission of new states threaten a language-policy crisis. One approach to a solution is to analyze more carefully than before the universe of possible language regimes for institutions like the EU and the justifications for deeming any alternative "optimal." We can define a language regime äs a set of official languages and a sei ofrules permitting complete mutual comprehension in a "deliberation" among representatives of language groups. If only the groups' languages and one synthetic language are relevant, ifthe representatives are monolingual, if translators are bilingual, and if other simplifications are assumed, there are almost five times äs many potentially optimal language regimes äs language groups (e.g. with nine groups, 40). We can partition the language regimes into ten classes, distinguished by their official languages and whether they translate directly, via intermediate group languages, or via the synthetic language.
The duration and urgency ofthe deliberation, the capacity of translation facilities, the importance of language equality, the distribution of group sizes, and the relative cost of learning the synthetic language are some conditions that determine which language regime is optimal. The prevailing conditions in the EU create a clear choice between twofamilies of language regimes. One family satisfies the professed norm of equal language treatment by making either none or all of the groups' languages official. The other family, by making only the largest languages official, systematizes the common EUpractice ofsacrificing language equality for cost reduction.
A victoryfor one normative position wouldfavor a corresponding language regime, but a continued daily normative struggle would tend to produce complex Variation in language regimes across EU agencies.