Si hubiese un cambio en la morfología del imperfecto del subjuntivo, ¿cómo hubiera ocurrido?: a diachronic analysis of Spanish -ra/-se morphological variation

Abstract

Modern Spanish has two morphological paradigms at its disposal in the Imperfect Subjunctive, collectively represented by -ra and -se. The distribution of each morpheme has been much examined, both synchronically (Williams, 1982; DeMello; 1993; Kempas, 2011; Sorenson, 2016; Guzmán Naranjo, 2017) and diachronically (Martínez, 2001; Ramírez Luengo, 2001; Branza & van Heuven, 2005; Hanna, 2012), but questions still remain about how the use of these two forms changed over time. While synchronic studies conducted over the last 30 years have revealed that -ra is unequivocally the dominant form in all varieties of Spanish, diachronic analyses have shown that this was not always the case. Data collected from the 15th and subsequent centuries have shown a steady increase in the usage of -ra at the expense of -se over time. In attempts to trace its path, these studies have used corpus data to primarily report the relative frequency of each morpheme in the various syntactic contexts where they occur. Although a clear picture does not emerge upon comparing these findings, the general tendency for -ra to appear more frequently in conditional sentences (i.e. si tuviera…, podría… ‘if I had..., I could…´) and in other adverbial clauses has been documented (Martínez, 2001; Ramírez Luengo, 2001; Branza & van Heuven, 2005). In view of the need to go beyond relative frequency alone to explain -ra/-se variation, the present study employs variable-rule analysis in an attempt to more clearly hypothesize the nature of the generalization of -ra as it progressed as a grammaticalizing Imperfect Subjunctive morpheme by identifying the factors conditioning the selection of -ra over time. Data taken from a corpus of dramas and novels written in Spain in the 17th and 19th centuries were coded for syntactic context (conditional protasis, adverbial, adjectival, nominal clauses), verb frequency, verb polarity and use in dialogue or narration. Findings show that in the 17th century dialogue, conditional protases and other adverbial clauses, high-frequency verbs, and negative contexts favored the use of -ra. In the 19th century, adjectival clauses surpassed protases and adverbial clauses as the most favorable for -ra, and dialogue and high-frequency verbs remained favorable while polarity was no longer selected as significant. The key findings of this study suggest that the degree of syntactic embeddedness of clauses (Hopper & Closs Traugott, 2003) influences the ability of an innovative form to spread to new contexts and that routinization of high-frequency forms may help lead the way.