A paradox?

A major morphological innovation in the Latin-Romance transition was the appearance of alternations in the root of the verb. This study examines the complex morphological evolution of the alternants which arose from proto-Romance palatalization and became characteristic of the present subjunctive together with the first person singular (and in some places also the third person plural) present indicative. I offer a brief historical-comparative sketch of these phenomena, and focus on an initially paradoxical situation. The alternant is predominantly aligned with the category ‘present subjunctive’. But what Romance speakers almost never do is to favour and protect that alignment by keeping the alternant in all and only the present subjunctive cells of the paradigm. In cases of analogical change, on the one hand the 1sg (and 3pl) present indicative almost always also participate in modifications affecting the present subjunctive, and on the other the alignment is frequently ‘broken’ by the elimination of the alternant from the 1pl and 2pl present subjunctive. I shall show that we are dealing rather with ‘autonomously morphological’ structures, which can actually override ‘common sense’ expectations that speakers should favour transparent relations between morphological form and grammatical meaning. What counts for speakers, I shall claim, is above all simply the predictability of alternation patterns within inflectional paradigms, whose basis may often be recurrent, but autonomously morphological, patterns.