Subject: sum : vowel length in orthographies

a few weeks ago i posted a query about the orthographic marking of vowel length , with special reference to korean . i received many very interesting replies , both on korean and on the general question , which i will summarise below - the whole topic raises some issues that people might like to comment further on . first though - thanks to everyone who took the trouble to send me information . it was my first linglist query and i ' m delighted with the way it works ! it would have taken me ages to assemble all these examples from library research . at first i was surprised to get a lot of replies suggesting as examples of languages that dont mark vowel length , english , german and so on . then i realised i should have been clearer in my original request . a . fox put the problem well with his question : ' your query about writing and vowel length is slightly odd , since it implies that in most languages with contrastive vowel-length the writing system represents it . but this depends on what you mean by ' representing ' vowel-length ; in the sense of having an explicit and consistent marker of length , such as a double letter , a diacritic or other length mark , this is not very common , at least in languages with which i am familiar . but length can be implicitly marked in a variety of ways , and therefore discoverable from the spelling for anyone ( such as the literate native speaker ) who knows the system . ' i 'd intended to include all of these things as ways of ' representing ' vowel length ; definitely not only cases where vowel length is represented as a feature in its own right by a macron , double symbol etc . so what i wanted was examples of languages with an extensive system of minimal pairs differentiated wholly or mainly by vowel length which have identical orthographic representations of the members of the pairs . this is the situation in korean , for quite a largish set of words of both one and two syllables . it 's definitely not the situation in english , german , canadian french , danish etc . i dont think its the situation in latin either , though i might be wrong , and would be glad of correction . here and in quite a few of the other languages that were suggested to me , vowel length is predictable from some other aspect of the word ( including its lexical identity ) so you dont actually get many or any minimal pairs differentiated only by length . right ? the same would appear to be the case for hausa ; and possibly for the other estonian vowel length contrast ? my estonian is a bit rusty , but would i be right in guessing that the full three way contrast is fairly limited ? however , several of the pacific languages that were mentioned to me do seem to fit the bill : tokelauan , samoan , maori , fijian and the ponapeic languages are all candidates for having what i would call a genuine length contrast ( although it seems that in maori at any rate the system of contrasts is quite limited ) , which is at least sometimes not shown in the orthography . so , it seems , does the australian language bardi , and its orthography doesnt show it . . the thing i found very interesting was the number of people who sent anecdotal evidence about how native speakers of these languages feel about marking vowel length . their writing systems are all of course relatively recent introductions based on the roman alphabet , and there are plenty of ways they could show vowel length . but they dont always use this potential it seems . apparently the speakers of ponapeic languages use length marking inconsistently , so do fijian speakers , maori speakers find it ' unnatural ' and dont like using it , tokelauan teachers have decided not to teach it , and in samoa use of a length marker is banned ! there seems to be something worth exploring here . the obvious hypothesis is that though the words are differentiated phonetically by length , this is not the distinction that is most salient to the native speakers . the next obvious hypothesis is that it is some kind of lexical distinction , similar to that of english ' good ' vs ' food ' , ' boot ' vs 's oot ' , ' look ' vs ' loop ' etc that is perfectly simple for native speakers and horrible for learners . ( which could also explain why gaelic has taken the opposite trend and started marking vowel length where it never used to , since it is used so much nowadays by less than fully native speakers . . . ) . thoughts on these hypotheses would be gratefully received . but back to korean . . . this still seems like a very unusual situation to me . it seems likely and indeed it was suggested by some of the korean respondents that the original hangul did have a length marker , but i have found no evidence for this in my research on hangul . ( that 's why i posted the orginal query , and i ve just had another look at ledyard 's thesis on the topic . ) maybe i just need to do more careful research : pointers on this also gratefully received . a couple of points of clarification : the distinction between the monosyllabic vowel length ' minimal pairs ' does seem to be dying out ( though it is the fact that it previously existed but was not marked in the writing system at that time that i find odd ; could it be that the lack of orthographic representation contributes to its demise ? ? ) . but there is a clear distinction in the sound of the two syllable pairs . to my ears these distinction seems to be much more one of pitch accent than of length . and : when i said that native speakers dont easily identify vowel length , i didnt mean that they say ' my language doesnt have a vowel length distinction ' - which would indeed be a statement to be cautious of ! i meant they dont know which word is supposed to have the long vowel and which the short . this is consistent with information sent by native speakers that it is something taught in schools as a prescriptive rule well for anyone who has read this far , let me reward you by sharing with you this response from fran karttunen ( i hope she doesnt mind ) . i dont know exactly how to interpret it with respect to my questions above , but i ' m sure there 's something in it for all of us ! ' both the maya and the nahua ( speakers of unrelated mesoamerican languages ) had achieved a partially-syllabic approach to writing before the arrival of europeans in the first quarter of the 16th century . from what we can tell of their writing , both made some use of rhebus principles in which a drawing of thing , the name of which was similar rather than identical , was used to suggest the intended word / syllable . for instance , a drawing of a a bare bottom ( tzin-tli " buttocks " ) was used for the honorific suffix - tzin . likewise a drawing of a banner , pan-tli was used for locative - pan . these two examples are suffixes , but there are similar cases for initial syllables of nahuatl noun stems , for instance . vowel length , at least for nahuatl , is not taken into account in these cases . the honorific suffix today has the reflex of a long vowel in most modern dialects of nahuatl , making it homophonous with the stem of " buttocks " , but attestation from the sixteenth century consistently show them to contrast in vowel length ( short for the honorific , long for " buttocks " ) . same for " banner " ( long vowel ) versus the locative ( short ) . ' thanks again to all respondents , who are listed below as acknowledgment . ( hope that 's ok ; preparing this summary has already taken so long that i ' m not keen now to go through and give proper ' who said what ' acknowledgments ! ) helen wechsler @ world . std . com ( allan c wechsler ) rcosper @ husky1 . stmarys . ca ( ronald cosper ) fran karttunen ( liar457 @ utxvms . cc . utexas . edu shelly @ uniwa . uwa . edu . au ( shelly harrison ) pulju @ ricevm1 . rice . edu gshin @ rs6 . chonnam . ac . kr ( gyonggu shin ) ian . green @ anu . edu . au ( ian green ) geoffn @ siu . edu ( geoffrey s . nathan ) mdr412 @ coombs . anu . edu . au ( malcolm ross ) jussi . karlgren @ sics . se kenneth de jong ( kdejong @ indiana . edu laurie . bauer @ vuw . ac . nz lance eccles ( lance . eccles @ mq . edu . au ) j . a . rea jarea @ ukcc . uky . edu jihualde @ ux1 . cso . uiuc . edu henry rogers ( rogers @ epas . utoronto . ca ) brian d joseph ( bjoseph @ magnus . acs . ohio-state . edu ) david fertig ( fertig @ acsu . buffalo . edu ) david gil ( ellgild % nusvm . bitnet @ cunyvm . cuny . edu stavros macrakis ( macrakis @ osf . org ) marc picard ( picard @ vax2 . concordia . ca ) charles scott ( cscott @ macc . wisc . edu ) lars mathiesen ( u of copenhagen cs dep ) ( thorinn @ diku . dk ) mark aronoff ( maronoff @ datalab2 . sbs . sunysb . edu ) caoimhin @ sabhal-mor - ostaig . ac . uk ( caoimhin p . odonnaile ) jaejung . song @ stonebow . otago . ac . nz ( jaejung song ) " a . t . c . fox " ( lnp6atcf @ lucs-mac . novell . leeds . ac . uk ) mjulien @ isl . uit . no ( marit julien ) blaine erickson ( erickson @ uhunix . uhcc . hawaii . edu ) _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ hfraser @ metz . une . edu . au ( 129 . 180 . 4 . 1 ) helen fraser ( dr ) dept of linguistics university of new england armidale nsw 2351 australia phone 067 73 2128 / 3189 fax 067 73 3735
