|
沈老反复让我给出提示或者拿出“哪怕一小段有力的话语”,我是想让各位自己判断的,总体文字也不算太长,这样可以完整了解作者的观点和论证,无奈沈老盛情难却,只好转录一段“有力的”在此:“The problem with this approach is that such a search process must assume the existence of huǒchē as a compiled lexical entry,i.e.,as a word:why would a word identification algorithm bother to identify a morpheme as a preliminary step in identifying the word? Identification via the word gives you the whole item "for free",simply because it is that gestalt item, and not its subcomponents,that is stored and accessed. Since the identification of morphemes is, as we saw above, a time consuming procedure that benefits little from contextual syntactic, semantic and pragmatic information, it would be quite inefficient for morpheme identification to occur as a way of accessing a precompiled word.(P291中段) ”。 |