Newsgroups: talk.origins,sci.skeptic,alt.religion.christian,alt.christnet,talk.religion.misc,alt.postmodern,sci.lang
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!newsflash.concordia.ca!news.nstn.ca!ott.istar!istar.net!van.istar!west.istar!n1van.istar!van-bc!nntp.portal.ca!news.bc.net!info.ucla.edu!newsfeed.internetmci.com!in3.uu.net!hearst.acc.Virginia.EDU!murdoch!usenet
From: dcs2e@darwin.clas.virginia.edu (David Swanson)
Subject: Re: Languages: Hard, Harder, Hardest
X-Nntp-Posting-Host: ara-mac-219.itc.virginia.edu
Message-ID: <Dv84K7.Gp9@murdoch.acc.Virginia.EDU>
X-Posted-From: InterNews 1.0.1@ara-mac-219.itc.virginia.edu
Sender: -Not-Authenticated-[9087]
Organization: University of Virginia
References: <Duys9H.1tB@murdoch.acc.Virginia.EDU> 
 <Dv46r3.G76@murdoch.acc.Virginia.EDU>  <4tdj8t$clf@csu-b.csuohio.edu>
Date: Sat, 27 Jul 1996 22:45:42 GMT
Xdisclaimer: No attempt was made to authenticate the sender's name.
Lines: 150
Xref: glinda.oz.cs.cmu.edu sci.skeptic:192777 sci.lang:58464

In article <4tdj8t$clf@csu-b.csuohio.edu>
b.scott@bscott.async.csuohio.edu (Brian M. Scott) writes:

> In article <Dv46r3.G76@murdoch.acc.Virginia.EDU>, 
> dcs2e@darwin.clas.virginia.edu (David Swanson) says:
> 
> >In article <4t5etk$jj5@news.ox.ac.uk>
> >patrick@gryphon.psych.ox.ac.uk (Patrick Juola) writes:
> 
> >> In article <Dv01wK.A8C@murdoch.acc.Virginia.EDU> dcs2e@darwin.clas.virginia.edu (David Swanson) writes:
> >> >In article <4t271h$cga@news.ox.ac.uk>
> >> >patrick@gryphon.psych.ox.ac.uk (Patrick Juola) writes:
> 
> >> >> Well, as a professional linguist I have to go with Mr. Silberstein here.
> >> >> To the best of my knowledge, there is no language into which people
> >> >> have tried and failed to translate the Bible, and it's commonly enough
> >> >> translated to serve as a very good base case.  (I can't think of a 
> >> >> language offhand that has a million words of text translated into it
> >> >> that do not include at least the Gospel of John.)
> 
> >> >For chrissake.  There are languages that have no written form.
> 
> >> That's certainly true.  These languages also tend not to have
> >> a million words of text translated into them.
> 
> >> I stand by my (original) statement.  If you can produce a counterexample,
> >> please do so.
> 
> >Your "statement" is cleverly ambiguous.  Do people try to translate
> >into languages quite at random and then either succeed or fail at it,
> >or do they first make sure it's feasible and then try, and of course
> >succeed, or is it something in between you have in mind?  That's
> >rhetorical, you understand.  I'm not particularly interested.
> 
> It seems very likely that the missionaries who produced so many 
> translations of the Bible simply worked with what was available.  
> I doubt that they chose their missions on the basis of feasibility 
> of translation.


Even with my lack of expertise or ability to produce the name of any
one of the langauges famous in similar discussions to these for having
only four or five numerals, I have encountered plenty of accounts of
absolutely hilarious Bible translations.  I guess you really are
imagining that absolutely anything claiming the name "translation" is
"satisfactory."  I have nothing to say to that.



> 
> [snip]
> 
> >> >I should have thought 30% was greatly.
> 
> >> So where do you define "greatly?"  
> 
> >Um, wherever I use it.
> 
> Doubtless accurate, but not very useful when the meaning is in 
> question.


It's not in question.  I meant more than about 20%, somebody else meant
more than 50%.  We've said so now, so where's the problem?



> 
> >If I strip out all the unused high-order
> >> bits 
> 
> >meaning ...?
> 
> He explained quite clearly: Maori has fewer than 64 (= 2^6) written 
> characters, so only 6 bits are needed to distinguish them.  English 
> needs fewer than 128 (= 2^7) characters,


I don't know how I got into this mess, but I suppose I'd better make
sure I know what's going on.  Is a character the following: a letter,
both lower and upper case, or a punctuation mark?


 so only 7 bits are needed.  
> Since the byte, the standard unit of computer storage, is 8 bits 
> long, normal storage routines waste 2 bits/character for Maori and 
> 1 bit/character for English.  The following calculation shows what 
> happens to the lengths of the Maori and English Bibles when these 
> unused bits are deleted.
> 
> >in Maori, I think that reduces the size by 25% (Maori having
> >> a fewer-than-64-character writing set), down to 3.48MB; a similar
> >> transformation on English (12.5% reduction) yields ~3.29MB, a 1%
> >> difference in Russian/Maori, and a just under 9% difference in 
> >> Russian/English.
> 
> >> Given that it's fairly easy to get a >%30 difference in translation
> >> sizes within the same language, depending upon the target readership,
> >> I'd be interested to know what you consider a valid measure of
> >> linguistic complexity.  
> 
> >What is the connection between what is "given" here and your question? 
> 
> The 'given' shows that length of translation can easily vary as much 
> within one language as it does between languages.  If it is a valid 
> measure of linguistic complexity, then the data already presented 
> suggest that most languages are of comparable complexity.  Since you 
> appear to disagree with that conclusion, one is entitled to wonder 
> what other measure of complexity you have in mind.


It strikes me as an extreme case: a thirty percent variation in the
same language.  It also would seem to follow from this that by taking a
translation in langauge A that varies by 30% from one in language B
(the original being in langage C) we might vary it within whatever the
acceptable bounds are for the data above, and come up with a variation
between A and B much greater than 30%.  If pages are taken as uniform,
let's say a book is 60 pages in A and (if the 30% figure is
conservatively applied to the smaller length) ca. 80 pages in B.  If we
can now produce another A version in, say, 45 pages, look what happens
to the A:B ratio.  How is this to be explained?  Only, I think, by
reverting to a non-extreme standard that I (too) had in mind.  That is,
a typical or average translation, in one language can be 30% different
from a typical or average translation in another.  That strikes me as
in line with what I said I expected.  I don't give a damn about the
"complexity" fetishing.






> 
> >I haven't the foggiest idea what in the world you could possibly mean
> >by "linguistic complexity," but I imagine that however one measures it,
> >that's what it is.
> 
> I'd imagine that anyone engaging in a discussion of linguistic complexity 
> would have at least a foggy idea of what it might be.
> 
> Brian M. Scott



The sort of fog one gets after a heavy rain with a lot of hot air.


David

"Heideggerian hope comes into question." J.D.
