It is estimated that 'modern' human language has been with us for at least 100,000 years. Unfortunately, the historical evolution of languages is often poorly documented, and evidence does not go further back in time than the moment when writing was introduced. Moreover, many of the world's languages do not have a written record that allows us to reconstruct this temporal evolution.
The causes of linguistic changes have been studied since the birth of modern linguistic theory in the late 19th century. These changes are thought to come about due to two types of processes: vertical and horizontal. Changes that take place as language is passed on from parent to child are described as vertical (inter-generational), while changes that are produced by the influence of a neighboring language are described as horizontal (through contact between geographically close languages).
An interdisciplinary team of scientists, including Tobias Galla from IFISC (UIB-CSIC), has published an article in Science Advances entitled "Geospatial distributions reflect temperatures of linguistic features" in which they propose and analyze a model of linguistic evolution. Their model describes both the vertical and the horizontal processes, and from the model they define `Linguistic Temperature’ as the ratio of errors in transmission and the number of faithful copying events of features. Linguistic temperature is a measure of the propensity of a linguistic feature to undergo changes.
Galla and his co-authors simulated the evolution of different linguistic characteristics. Using these simulations, mathematical analysis and data collected in the World Atlas of Language Structures (WALS) the researchers then empirically estimated the linguistic temperature of different languages features such as interrogative particles, ordinal numbers, uvular consonants, and passive constructions. The analysis is based on regularities of the features’ distributions across languages on the globe, for example some features appear in clumped patterns in space, and others are more scattered.
This type of computational models are widely used in other fields such as the modeling of opinion dynamics in social sciences, or genetic evolution in biology; this work establishes a further step in the use of methods from statistical physics to problems in the computational social sciences.
reflect temperatures of linguistic features. Henri Kauhanen, Deepthi Gopal, Tobias Galla and
Ricardo Bermúdez-Otero. Science Advances 01 Jan 2021: Vol. 7, no. 1,
eabe6540. DOI: 10.1126/sciadv.abe6540