This paper cites the inadequacy of transformational generative grammar theories in their attempts to describe the meaning of a given sentence. The author sees the specification of meaning involving the recovery of the particular section or sections of the world model communicated or represented by the sentence. As a corollary, the author argues that sentences of English are essentially representations of "scenes" in the world model and not representations of objects. The author proposes a model for sentence analysis which seeks to recover meaning by rejecting the notion that sentences are "nothing but" objects and that their "disambiguations" are the representations of these objects. Recovery of scenes from sentential representations is considered equivalent to meaning-specification for language. (VM)