5-1322msg1.txt

运用贝叶斯网编的一个识别垃圾邮件的程序

Subject: re : " snow " 1 / 2

about snow . . . first , i should apologize to tony for misunderstand his cocktail advice - - his derivation of the origin of my misunderstand be correct . second , i think that quotation from boa / whorf be very helpful and a good reminder to those of us in this debate . third , as to the question of whether or not sleet be relate to snow . i do n't agree . it be , i agree , relate ( and would be define in relation to ) water , but ( at least i and a few other student that i ask ) would n't define it as a form of snow ( ditto for freeze rain ) . fourth , i agree with tony that " count " have to be mediate by many consideration of grammatical structure in the compare language and spread of the form in the speech community . but , i do n't think that this wipe out the ( admittedly small , but original ) point . fifth , i ' m go to sidestep the issue over ' lexicalization ' v . ' complex construction ' because i do n't think that i share the same view as other on the importance / necessity of this distinction - - indeed , it be a bite ironic that another implication of sapir / whorf be that the view that our language be make up out of ' word ' and ' grammar ' ( construction ) be precisely the kind of objectification , which we would expect and which formal distributional analysis show to be a simplification ) . as jonathan state , " figure out just what count as a simple , lexicalise form be * very * hard in yup ' ik , give that it have a rich , higly productive derivational morphology " . i agree and the answer would eventually have to draw line along continua that i do n't think will be label ' lexical ' v . ' construct ' ( and sapir offer some nice theoretical machinery for these kind of comparative distinction too ) . i still see only four ' arbitrary and unmotivate ' form that deal specifically with 's now ' ( i . e . , snow , slush , blizzard , flurry ) . i ' ll leave it to other to decide whether or not various dialect of eskimo have more or less , but even if the point should fail here it still have life to it . so , i still find myself agree with the original insight . the point - - and not all that it have be use to argue - - have alway seem obvious to me . perhap if we narrow the scope of the relevant speech community and bring it closer to home , it be easier to see . would n't we all accept the idea that _ on average _ lawyer ( v . non-lawyer ) have more distinct form for legal concept than do other outside this community / culture ? ( and , of course , what we mean by distinct form imply all of the complex relative distinction hint at above ) . similarly to take an example i know more about , statistically speak arab have more arbitrary and unmotivate form for camel than english speaker ( even account for difference in the syntactico-semantic structure of the two language ) . why do this simple - - and to be honest relatively uninterest - - idea seem to bother people so much ? dougla j . glick department of anthropology vassar college doglick @ vassar . edu