Communicative efficiency and institutional pressure on language

Abstract

How do languages become the way they are today? An influential view is that languages are shaped by pressures towards communicative efficiency: languages optimize getting the meanings across with only a few distinct words to memorize, and they minimize difficulty in production and comprehension. It has been proposed that the mechanism towards efficient communication is bottom-up: language users do not explicitly optimize for efficiency; instead, they favor a less costly form if it conveys a similar meaning, which becomes conventionalized through repeated individual choices.In this thesis, we investigate the robustness of this bottom-up mechanism in three studies, examining cases where top-down institutional pressures are imposed on language use.The first study is a baseline case, where we show how systems of spatial demonstratives (e.g., here'', from there'') vary across languages by pressure towards communicative efficiency. The second study investigates how regulations on personal name systems may have affected the information structure of names. Despite surface differences between East Asian and Western traditions, historical data reveal a shared information structure: the first element (e.g., Chen'', Edward'') is less informative than the second (e.g., Sihan'', Gibson''), enabling efficient individual distinction. However, sociopolitical processes transformed different components into hereditary patronyms, first in East Asia, second in the West, reshaping name informativeness. The third study concerns the effectiveness of government efforts to simplify legal language. Legal language is often difficult to understand because of long-distance syntactic dependencies, but it remains unclear why laws are written in such a convoluted way. We hypothesize that it is syntactically transparent to describe a complex legal concept in one complex sentence, which is harder to comprehend. Indeed, we found like modern American laws, legalese texts in ancient China also contain more convoluted structures than contemporary non-legalese ones, possibly because these legal texts often include the offense and the punishment in the same sentence. On the other hand, we found a much smaller difference in dependencies between the two genres in modern Chinese than previously observed in American English. This is possibly because China went through a drastic change from ancient China in legal tradition, and lawmakers adopted simpler alternative ways to convey the same meaning. Taken together, our results suggest the mechanism towards communicative efficiency can be interfered with by institutional constraints.

Publication
PhD thesis, August 2026

Related