December 4, 2009 § Leave a comment
It appears it isn’t just me. Lots of linguists also seem to like maps. So combine maps and languages and you’ve already got me going. Make it interactive and I’m in heaven. Bernard Comrie and the team at World Atlas of Language Structures have done exactly that. An online version of their book, the site allows you to manipulate language features and map them – a typologist’s dream!
Using data from upto 2650 languages, the maps can plot which of these have specific linguistics features from all levels of enquiry – phonology, syntax and even a little on the lexicon.
Here for example is a map about hand~arm distinctions. As encoded in the previous sentence English lexicalises this distinction. Other languages, as you can see do not.
The following map demonstrates how you can manipulate and combine features.
Some of the languages I am interested in don’t have the segment /p/. When looking at the missing consonant charts I noticed that /g/ is also mentioned as a possble gap. I am also intrigued by the velar nasal, i.e., the last sound in sing as it tends to have odd distributions. For no good reason, many languages do not allow it in syllable initial position. So I decided to combine the two features to see if there were languages that had no /p/ but the velar nasal (including #_ … that’s short hand for first position), no /g/ and no velar nasal – these two sounds are made in the same place, no /g/ but the velar nasal etc etc etc …
It is interesting that they langauges that languages in the WALS sample that are without either velar nasal and hte velar stop seem to cluster in two places. West Coast US, and South America. The Bolivian languages, geographically somewhat adjacent, and sharing these features, we should presume to be related to each other,. In fact the ethnogluge calls them Tacanan family, listing only a handful of them. Likewise the red dot near the end of Argentina represents Teheulche – a language with a single sister.
Looking now at the yellow dots in South East Asia, they look close enough together to be a possible bunch of relatives. In fact they are not. We have Thai, a member of the Tai-Kadai family, Bawm and Kayah Li which are distant cousins in quite different parts of the Tibeto-Burman family, and Khasi from the Austro-Asiatic family.
Languages, that have neither /p/ nor /g/ and no #_ sing are rare. Una on Mainland New Guinea, and Ket an extremely endangered language of the Russian Federation seem to be the only two in the sample.
Go take a look at WALS and play around with the features. I tell you, it’s addictive!