Language change as a (random?) walk in entropy space

Duration: 29 mins 19 secs

Description:	Talk by Dr Christian Bentz, Assistant Professor, University of Tübingen

Created:

2019-11-25 14:44

Collection:

Language Sciences Annual Symposium 2019 - Perspectives on Language Change
Cambridge Language Sciences

Publisher:

University of Cambridge

Dr Christian Bentz

Language:

eng (English)

Keywords:

language variation and change; language distance; information theory; computational linguistics;

Credits:

Producer:	Mike Dowds (Cambridge Assessment)
Producer:	Dan Vickers (Cambridge Assessment)


Abstract:	This talk explores language change from an information-theoretic perspective. The word entropy of more than 1000 written languages from over 100 families is mapped along two dimensions to measure lexical diversity and runs of the same word combinations. Both turn out to be highly constrained in the sense that languages of the world fall into relatively narrow ranges in both dimensions. However, zooming into these ranges, there is also considerable variation between languages. As languages change over time, they move around in the entropy space. A fundamental question is whether this walk in entropy space is purely random, directed, or a combination of both. In a first step to answer these questions, I explore how ancient languages within the same language family compare to their modern counterparts.

Abstract:

This talk explores language change from an information-theoretic perspective. The word entropy of more than 1000 written languages from over 100 families is mapped along two dimensions to measure lexical diversity and runs of the same word combinations. Both turn out to be highly constrained in the sense that languages of the world fall into relatively narrow ranges in both dimensions. However, zooming into these ranges, there is also considerable variation between languages. As languages change over time, they move around in the entropy space. A fundamental question is whether this walk in entropy space is purely random, directed, or a combination of both. In a first step to answer these questions, I explore how ancient languages within the same language family compare to their modern counterparts.

Format	Quality	Bitrate	Size
MPEG-4 Video	640x360	1.24 Mbits/sec	274.01 MB	View	Download
WebM	640x360	498.36 kbits/sec	107.07 MB	View	Download
iPod Video	480x360	491.19 kbits/sec	105.47 MB	View	Download
MP3	44100 Hz	249.78 kbits/sec	53.70 MB	Listen	Download
Auto *	(Allows browser to choose a format it supports)