Contrastive analysis is suggested as a tool in language teaching for such areas as: (1) deciding how much to focus on different aspects of the target language; (2) making generalizations about its structure; and (3) explaining texts or constructions which might otherwise be incomprehensible. The claim is made that such procedures need to be based on a motivated model of how different languages can be compared both in linguistic theory and in pedagogical application. Thus, both differences and commonalities between two languages need to be related to systematic properties of each, in terms of their overall structure and functioning. Two noncognate languages such as Hebrew and English can be compared along the following diameters: (1) as manifesting obligatory similarities in the form of linguistic univerals; (2) as having certain typological or other, nongenetically motivated similarities; (3) as differing in ways attributable to their distinct genetic origins; (4) as differing in other, systematic ways; and (5) as differing due to incidental, idiosyncratic properties typical of nonrule-bound features of any language. The relevance of these claims, illustrated by reference to properties of modern Hebrew and English, is then considered in relation to linguistic theory and language typology, on the one hand, and to the role of contrastive analysis in language pedagogy, on the other. (Author/AMH)