5-1333msg1.txt

贝叶斯学习算法源码

Subject: re : sum : native speaker intuition

marilyn silva recently summarise response deal with " native speaker intuition " , especially with respect to the sentence ' john ask mary to leave ' . she note that several respondent point out that the interpretation of this sentence be determine , at least in part , pragmatically , and the view that the interpretation be syntactically fix be wrong . this point be demonstrate by the ) pair of example : ) ) [ 1 ] the teacher ask the child to leave the room . ) [ 2 ] the child ask the teacher to leave the room . i be a native speaker of english ( new zealand english ) , and i be not a syntactician , so i do n't think my syntax be be mould by my , or anyone else 's , rule or theory , but , in both [ 1 ] and [ 2 ] , i can only interpret the person be ask ( the child in [ 1 ] , the teacher in [ 2 ] ) as be the person require to leave the room . no amount of ) intuition fatigue ( which could be sum ) up as " say a weird string 50 time and it sound good ; say a ) good string 50 time and it sound weird " ) have be able to change my interpretation of this kind of sentence . perhap for some people the interpretation of this sentence * is * syntactically determine . the summary conclude with : ) in his post to me , carson contend that judgment datum " can ) provide real , useful datum , if we would just be more systematical ) and careful about how we collect them . of course speaker will ) differ on certain point . . . . the interest question be whether ) the range of variation we find be systematically constrain in ) way that interest theory of grammar can explain . " i agree that ( native ) speaker will differ on certain point - - my syntax probably differ from that of most people , and i think this be probably typical ( in my case the follow ( a ) sentence be ungrammatical and would normally be the ( b ) sentence : 1 . ( a ) i promise mary to leave . ( b ) i promise mary i would leave . 2 . ( a ) candidate be recommend to obtain the booklet . ( b ) it be recommend that candidate obtain the booklet . 3 . ( a ) john require to pass the examination . ( b ) john need to / must pass the examination . john require a pass in the examination . s . o . require john to pass the examination . 4 . ( a ) i see the barn red . ( b ) ? ? ? it be important that syntactition realise that one there be probably thousand of syntactically distinct english out there , and i be not talk about socially or geographically definable varity . there may not be many new zealand english speaker who share the above judgment , but * my * english be no less base on a natural language grammar than anyone else 's . because the interpretation of a certain construction may be influence by pragmatic factor be some , or even in the majority , of speaker , do not disqualify a purely syntactic account if there are speakers whose language data are accounted for by it . where there be variation among langauge user , it be very important not to conflate these difference and try to base a grammar ( syntax / phonology . . . ) on an impure sample . untimately , a grammar be an account of an individual ' s language . student 's who disagree with datum be rarely be contentious - - i contend that they have a slightly different grammatical structure which be , of course , no less valuable in the search for constraint on syntactic structure than the teacher 's or a " receive " english grammar . i heartily endorse carson 's call for more systematic and careful collection of syntactic datum . think about it - - be your english really syntactically the same as your colleague 's , or your father 's ? if not , care must be take when conflate datum from these source , and even more care must be take when conflate datum from even more diverse source . wayne p . lawrence