3-436msg2.txt

贝叶斯学习算法源码

Subject: re : 3 . 427 innateness

joe stemberger write : > i ' ve never understand why it make any difference at all to linguistic > theory whether highly language-specific information be innate or not . > [ . . . ] > innateness be usually use as a explanation for universal , or for > constraint on variation ( parameter ) . but it have alway seem to me that > what be important be that something be universal or that variation be > limit to a few option . what possible difference to linguistic theory > could it make whether the observe pattern be due to language-specific > innateness , or due to some more general feature of cognitive process , or > ( for that matter ) due to guidance from guardian angel or alien from > another dimension . the observe > pattern be real under any explanation of where they come from , and > language seem to abide by them . we can still rule out some potential > explanation because they may violate a universal , and still provide > explanation where two phenomenon be link because they be due to the > same parameter . > > so , why all this stuff about innateness ? i ' ve never understand why we care . > [ . . . ] > > doe innateness buy us anything for linguistic theory itself ? innateness be a * conclusion * from linguistics , not a premise . if one look on it as a premise , one indeed get into a logico / scientific muddle like the one you outline . but since it a conclusion , not a premise , linguistic theory buy us innateness , not the other way around . we care because its an interest conclusion , and because the more one learn about how language work in the child and adult , the more it look like the only plausible conclusion ( at least to me ) . it give neurophysiology / genetics some work to do , work which be begin to get do . in this respect it be superior to an appeal to guardian angel and alien , although in some other century , past or future , this judgment may be different . furthermore , language-specificity look more plausible than a " general feature of cognitive process " , for reason that be hash out during the flamefest on modularity early in the life of linguist . however , a negative can never be prove . thus , reduction to general cognitive principle of the ecp , the ocp , or categorial perception of point of articulation for stop consonant in neonate remains a possibility . and once again , we be deal with a conclusion , not a premise . - david pesetsky