As pointed out by Malkiel (1957: 14), many hypotheses have been proposed to account for the presence of the stop [d] instead of the expected lateral [l] of the etymological source laxare. The aim of this contribution is to review and discuss the various proposals put forth in the literature to explain the particular phonetic shape of this form. It will be argued that none of them offers a satisfactory account of the variation found in Romance. Our is that initial [d] in the forms of the verb ‘to let’ is the result of a fortition process: such a process aims at maximizing the syllable structure, especially at the beginning of the word (cf. Alber (2001)). On the other hand, it can be observed that the Romance varieties with initial d- in outcomes from laxare ‘to let’ don’t have d- as outcome of verbs like lavare (Castill. lavar, Port. lavar, Gasc. lavar, Sard. lavare, etc.). It will be argued, thus, that phonology is not the whole story: the fortition process witnessed by the verb forms dejar, dichà, deixar, dassare, dassari etc. must be put in relation with the particular semantics of the verb ‘to let’ (cf. Malkiel (1982)). As an auxiliary whose instruction is to signal causation or non-obstruction to the realization of a given state of affairs, its semantics – and its frequency – cannot be put on the same footing as that of verbs like “to sleep” or “to sing”. Moreover, the verb ‘to let’ is particularly frequent in Imperative / Prohibitive contexts. Given that Imperative contexts belong to those contexts in which verb forms bare special emphasis, it is suggested that strengthening of the initial consonant of laxare outcomes first appeared in the Imperative