ABSTRACT
It is crucial to realize that our claim that knowledge of language involves inferential competence does not entail that there are no inferential relations between sentences of a natural language that would be difficult to discern for the competent speaker of the language. Though the delimitation of the sentences of natural language that are mapped by the vertices of our logical maps is far from easy, what such a sentence is can be seen quite clearly at least in prototypical cases. These vertices are constituted by sentences or formulas of the regimented, formalized, or formal languages which we employ as the vehicles of logical analysis. The correspondence between these formulas and the sentences of the natural language that they represent might be one-to-one and hence quite transparent. The multiplicity of logical languages that appeared during the last century indicates that it would be unsubstantiated to assume that such logic would necessarily take the shape of our 'classical' logic.
