5-1237msg1.txt

该程序实现的是一个垃圾邮件过滤系统

Subject: summary : basic order ( and remark on typology )

last week i ask for reference to discussion of a problem that come up in linguistic typology : when there be conflict or ambiguous criterion for decide whether a particular language be a particular ' type ' with respect to some feature ( word order , clause alignment , or whatever ) , how do one decide how to assign that language ? i would like to thank the follow for their helpful reply : george huttar , yehuda falk , dan everett , larry trask , jon aske , mike maxwell , mark newson , bill croft , georgium green , ingo plag , randy harri , and andrew carstair - mccarthy . i be quite surprise at the small amount of publish attention that there be to this problem . i be point to short discussion ( no more than a couple page ) in some of the major work devote to typology : the seminal greenberg paper , comrie 's ' language universal and linguistic typology ' , croft 's ' typology and universal ' , and hawkin ' ' word order universal ' . it be also suggest that i look at dori payne 's ' the pragmatic of word order ' and to papers on yagua by payne and dan everett and on tzotzil by judith aissen . what prompt my query be a read of johanna nichol ' linguistic diversity in space and time , which i find extremely impressive . but all through it i have an uneasy feel cause by her pigeon-holee language as ' svo ' , ' head-mark ' , ' active-stative ' , or whatever . since so many language be * not * transparently one particular ' type ' on the surface , i wonder what the basis for these type-characterization be . there be no general answer give to this question for an obvious reason : neither nichol or anyone else could have profound first-hand knowledge of more than a small handful of the 174 language in the datum base . i suspect that in most case nichol could not know what criterion be apply to type a language in the source she consult , because many source be insufficiently explicit on that point or take as self-evident some categorization that another would take as controversial or simply wrong . ( consider , for example , her type french as vso . ) there be , to be sure , case where nichol throw out some language from the sample of some particular feature because of its obvious ambiguous status with respect to that feature . but do so could have create more problem than it solve . as both aske and croft point out in their posting to me , if a language be ' inconsistent ' with respect to a particular feature , that too be typological datum ; datum moreover that could be highly relevant to conclusion about stability and diversity over time . in a sample of 174 language , misassignment of several language within a category with a 3 - way division could lead to rather different conclusion . likewise , so would postulate a different set of category or have category specifically for ' mix ' type . this be begin to sound like a critique of nichol , but i do n't mean it to be . rather , it be more a commentary on the shaky art of typological pigeon-holee that underlie not just conclusion about language prehistory , but also much functionalist theorize and - - increasingly - - generative theorize as well . there be also the question of sample * size * . typologist strive , quite reasonably , to correct for genetic and areal bias in their sample ( the most heroic effort along these line that i know of be dryer 's work ) . but how confident can we be of any attempt to eliminate bias from the sample , give nichol ' conclusion that influence can extend half-way around the globe ? and do n't that present a challenge to purport explanation of the relative frequency of some typological feature , which be common in the functionalist literature and increasingly so in the generative ? so much could be the result of historical accident on the one hand and contact and descent on the other , rather than the product of ' external ' functional force or the design of ug . the smaller the sample of language where mutual influence or common descent be not a possibility , the more likely that some implicational typological relation be artifactual . and the more reason we have to think that there be a lot of typologically possible but - - purely by chance - - nonexist language . fritz newmeyer fjn @ u . washington . edu ps : with respect to the last point , alan bell have show that if some feature appear in 1 % of the world 's language ( say , 40-50 language ) , it will show up only about 50 % of the time in a random sample of 75 language . you 'd need a sample of over 200 language before it could be count on to show up 90 % of the time . and we be assume here , utterly counterfactually , that there be no genetic relation or areal influence between language .