Data reduction typology and the bimodal distribution bias

Confronting low data reduction typologies, as established by using data from parallel texts, with the high data reduction typologies of WALS reveals a systematic bias of WALS typologies toward highly bimodal distribution. Properties with a distribution supporting a discrete feature analysis in many...

Full description

Saved in:
Bibliographic Details
Published in:Linguistic typology 2009-05, Vol.13 (1), p.77-94
Main Author: Walchli, Bernhard
Format: Article
Language:English
Subjects:
data reduction
doculect
imperative
linguistic atlas
methodology
number
parallel texts
word order
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Confronting low data reduction typologies, as established by using data from parallel texts, with the high data reduction typologies of WALS reveals a systematic bias of WALS typologies toward highly bimodal distribution. Properties with a distribution supporting a discrete feature analysis in many languages are likelier to be represented in WALS and to be represented accurately. This bias has important consequences when WALS typologies are interpreted theoretically or further processed statistically.
ISSN:1430-0532
1613-415X
DOI:10.1515/LITY.2009.004