SummaryIn the Latin language, there were extensive reductions of noninitial short vowels in the period immediately preceding the literary period (starting with Plautus around 200 B. C.). One of these reductions is discussed here from several different points of view. First, a philological analysis of the available data shows that the traditional picture of the development is incorrect in important respects. Secondly, an attempt is made to assess the phonetical plausibility of a revised picture of the sound change. Thirdly, it is demonstrated that the data of early literary Latin should be seen as reflections of a process of lexical diffusion that was just about finished. Lastly, it is pointed out that the rules describing the situation in literary Latin can hardly in any sense be taken to reflect any part of the speaker's behavior, but may well be conceived to mirror a strategy of perception.