more thoughts on MLU

[Alcock, Katherine]

There are almost certainly translations of the Christian Bible into more languages than any other document, and one of these is Modern Arabic.  Languages that have an orthography often have it because it has been used to translate the Bible or portions of it.  When this is done, it is always into the version of the language that is current at the time, although this can date extremely rapidly in many cases, so there should not be a problem with recent translations, about the version of the language used being up-to-date.  However, the style in different translations in English varies considerably and therefore I would imagine also the MLU.
 
For example the first half verse of the gospel of John in (first) the King James version and (second) the Contemporary English version:
 
In the beginning was the Word
In the beginning was the one who is called the Word. 

As translators who have written the (only, or most recent) translation of a bible into a particular language have had to make a decision about what style to write in, some will have chosen a more literal or literary style (as in the KJV) and some a more casual, spoken style or a more explanatory style (as in the CEV).  So in some languages only one or the other will exist.
 
Many languages in addition do not have orthographies - one of the languages that I have worked on, and tried to get measures of MLU (Kigiriama), does not have an orthography of its own, and literacy specialists are trying to construct one more suited to its phonology.
 
However, most of the languages that we work on do have orthographies and extensive portions of ancient and modern literature.  In looking at verb frames in Kiswahili, which does have some ancient and modern literature, we used three corpora: an oral history that had been written down (current, conversational language), newspaper articles (current, more literary language) and the published version of 1001 Nights, that we chose because we would also be able to find an English version (older, more literary language).  Although newspaper articles are not going to have exactly the same concepts in them, given the variety of ways of expressing the same concept that exist in different English translations of the bible, these might be a helpful resource for comparison, too.
 
Katie Alcock

	-----Original Message----- 
	From: info-childes at mail.talkbank.org on behalf of Ginny Mueller Gathercole 
	Sent: Fri 2/20/2004 11:46 AM 
	To: Brian MacWhinney; info-childes at mail.talkbank.org 
	Cc: 
	Subject: Re: more thoughts on MLU
	
	

	>How about this method:  You take the Bible in its various translations,
	>morphemicize it, devise a counting system, compute Bible MLU and then
	>compute corrected child MLU as the ratio of raw child MLU over Bible MLU.
	
	Hi, Brian,
	
	The Bible may not work for all languages.  An important case would be
	Arabic.  First of all, the religious text is the Koran, which is
	written in classical Arabic, not the vernacular.  So even getting MLU
	counts from the Koran may not reflect the MLU counts in vernacular
	speech of adults (whether adult-adult or adult-child).  It is
	possible that there are translations of the Bible into Arabic, but
	(apart from any concerns for cultural sensitivities in using the
	Bible instead of the Koran) it is likely that such a translation
	would still be written in classical Arabic, not the vernacular.
	
	Ginny
	
	--
	
	Virginia C. Mueller Gathercole, Ph.D.                       
	Reader
	
	Ysgol Seicoleg                                  School of Psychology
	Prifysgol Cymru, Bangor                         University of Wales, Bangor
	Adeilad Brigantia                               The Brigantia Building
	Ffordd Penrallt                         Penrallt Road
	Bangor LL57 2AS                                 Bangor LL57 2AS
	Cymru                                           Wales
	                         |      /\
	                         |     /  \/\
	Tel:  44 (0)1248 382624 |  /\/    \ \
	Fax:  44 (0)1248 382599 | /  ======\=\
	                         |  B A N G O R