Function (biology)

Problem of universals

Natural language

Linguistics

Semantic property

Philosophy

Biology

Computer science

Evolutionary biology

Natural language processing

Psychology

Mathematics

Linguistic universal

Psychotherapist

Epistemology

Domain (mathematical analysis)

Mathematical analysis

Artificial intelligence

Meaning (existential)

Property (philosophy)

Theoretical linguistics

Entropy

While the languages of the world vary greatly, they exhibit systematic patterns, as well. Semantic universals are restrictions on the variation in meaning exhibit cross-linguistically (e.g., that, in all languages, expressions of a certain type can only denote meanings with a certain special property). This paper pursues an efficient communication analysis to explain the presence of semantic universals in a domain of function words: quantifiers. Two experiments measure how well languages do in optimally trading off between competing pressures of simplicity and informativeness. First, we show that artificial languages which more closely resemble natural languages are more optimal. Then, we introduce information-theoretic measures of degrees of semantic universals and show that these are not correlated with optimality in a random sample of artificial languages. These results suggest both that efficient communication shapes semantic typology in both content and function word domains, as well as that semantic universals may not stand in need of independent explanation.

Quantifiers in Natural Language: Efficient Communication and Degrees of Semantic Universals