Thursday, June 02, 2005

Hugging the shoreline

Stephen R. Anderson points out why languages would prefer homonymy over filling out the space of accidental gaps in his 2004 book, Doctor Dolittle's Delusion. It's not the focus of his discussion, and I'm not sure it's necissarily what he is saying. But it sparked a thought in me.

The idea is in a short upfront section where he lists some of the characteristics of human language based on Hockett's design features, and he is describing Duality of Patterning.

His observation is:

This duality makes for a certain efficiancy in language, in that there are not very many different units that have to be kept apart in production and perception. If we needed to learn completely different signals for every word of our language , we would quickly reach the limits of our acoustic and auditory abilities. Compare alphabetic with ideographic writing systems, for example. A child learning to read and write a Chinese language, or the kanji characters necessary to read a Japanese newspaper, can testify to the immense burden of having to memorize all the individual signs separately, as opposed to a child learning the twenty-six signs of the roman alphabet.

So my thought was, it's easier to learn a language if it hugs the morphological shoreline, keeping the number of distinct word forms low. Because if the number of forms is large, you basically need to treat each one as a distinct form to be learned. So there is pressure from learnability to keep away from isolating languages.

I still think you could argue that there is corresponding pressure from perception to expand the space of word forms, kind of like what I was thinking in my previous post. So it's not a simple situation of one right way to design a language. Rather there are various reasons why you might want to design it one way or the other, and each choice involves some trade off.


cawshis said...

Does our potential to learn a language really become finite? I mean I've learned dozens of new words related to work and manufacturing, but I don't find myself straining to learn them. That's probably because I'm a genius though.

I can see the need to hug the coastline in the early development of language...but why do we still do so now? There is a vast land beyond the coast filled with beautiful sounds like blar (the feeling one gets when discussing linguistic ideas that are way above one's head) and dapperlin!

Dapperlin! I just love saying it! I wish it had a meaning!

liberal elite said...

I wouldn't say it's impossible to learn a completely new word, like blar. But it's harder than repurposing an old one, like mouse.

I'm assuming it's harder because so few of the new words seem to be acutally new forms. Except maybe borrowed words and some acronyms. I wonder how many of the new words you learned are new forms and not simply new meanings of old words or affixed versions of old words.

Site meter

Search This Blog