This article builds upon previous studies that contrast the Lexical Aspect Hypothesis (LAH) and the Discourse Hypothesis (DH) in order to establish principles of tenseaspect morphology acquisition. Whereas the LAH predicts that (beginning) language learners use verbal morphology to mark situation types (cf. Vendler 1967), the DH maintains that learners use verbal morphology to mark discursive features. I argue that both hypotheses lack sufficient theoretical/empirical evidence to rule out the possibility that verbal morphology is also sensitive to temporal deixis, especially given the widely-attested usage of the present tense in L2 past-time narratives. Therefore, the current study seeks to replicate Salaberry (2011) without controlling for tense. Twenty-eight learners of French as a second language judged the acceptability of various verbal inflections in a cloze narrative task and then engaged in a retrospective protocol. The results were submitted to a repeated measures ANOVA, which revealed a significant interaction among morphology, lexical aspect and discursive environment that are not envisioned by either the LAH or the DH. Moreover, the systematic (and not random or ‘default’) use of the present tense lends support to a fractal model of French morphology acquisition in which learners ‘soft assemble’ their (inter)language with the resources at hand.