Is “Canis Currit” Ungrammatical?

The main theme of Aristotle’s Sophistici Elenchi is false but apparently valid reasoning. Aristotle operates with a special class of fallacies depending on features of language. In that connection he touches on many questions of relevance to the grammarian. In their commentaries on the Elenchi, medieval scholars develop the hints given by Aristotle and introduce new problems of their own invention. Thus they discuss the generic use of singular nouns, the scope of conjunctions and modal expressions, suprasegmental phonemes, constructional ambiguities, the tense of participles, and parsing. Discussions of the last-named point revolve around the example canis currit. Canis is an equivocal term. Is it then one constituent of a sentence (one pars orationis) or is it more than one pars orationis? If it is several partes orationis, why is canis currunt ungrammatical? To save the grammatically of canis currit modistic logicians split the modi significandi into modi significandi ‘activi’ and modi significandi ‘passivi’, the latter attaching to the ‘things’ qua conceptualized in a certain way and signified as such by means of words, the former attaching to the words qua significative in some definite way. By letting the active modi significandi be the basis of construction they could save the grammaticality of canis currit. But the solution breaks with the fundamental idea of modistic grammar, the derivation of grammatical categories relevant to syntax from the conceptualized reality, and introduces an explanation of syntax founded on surface categorization of words. There was a reaction against this about 1300. Later Elenchi commentators seem to have lost interest in the subject.