Attention Economies, Information Crowding, And Language Change

Language is a communication system that adapts to the cognitive capacities of its users and the social environment in which it used. In this chapter we outline a theory, inspired by the linguistic niche hypothesis, which proposes that language change is influenced by information crowding. We provide a formal description of this theory and how information markets, caused by a growth in the communication of ideas, should influence conceptual complexity in language over time. Using American English and data from multiple language corpora, including over 400 billion words, we test the proposed crowding hypothesis as well as alternative theories of language change, including learner-centered accounts and semantic bleaching. Our results show a consistent rise in concreteness in American English over the last 200 years. This rise is not strictly due to changes in syntax, but occurs within word classes (e.g. nouns), as well as within words of a given length. Moreover, the rise in concreteness is not correlated with surface changes in language that would support a learner-centered hypothesis: There is no concordant change in word length across corpora nor is there a preference for producing words with earlier age of acquisition. In addition, we also find no evidence that this change is a function of semantic bleaching: In a comparison of two different concreteness norms taken 45 years apart, we find no systematic change in word concreteness. Finally, we test the crowding hypothesis directly by comparing approximately 1.5 million tweets across the 50 US states and find a correlation between population density and tweet concreteness, which is not explained by state IQ. The results demonstrate both how Big Data can be used to discriminate among alternative hypothesis as well as how language may be changing over time in response to the rising tide of information.