摘要：phonetic process of ‘height dissimilation’ is universally reconstructed as an intermediate stage between e.g. AGk þþþþà [ennéa] and MGK îþþþþ [eþá]. The same process is reconstructed for some Ancient Greek dialects (e.g. Boeot. îþþþà traditionally interpreted as representing [ennía]) and for other languages: e.g. Vulgar Latin. Allegedly, the existence of ‘height dissimilation’ is warranted by some modern dialects which seem to preserve the stage [enía]. In a previous paper I dealt with the data of Ancient Greek challenging this explanation. This paper deals with the evidence found in modern vernaculars. On closer inspection, the data at issue turn out to be illusory and call for an alternative reconstruction. ‘Height dissimilation’ is a mere artifact which fails to match any universal process type and lacks any real phonetic motivation either synchronically or diachronically. Synizesis (glide formation) gives a more satisfactory explanation: [eo], [ea] > [o], [a] > [jo], [ja] with non-syllabic [] turning into an optimal [j]-glide. In spite of appearances, the outcomes like [enía] are not a compelling argument for ‘height dissimilation’. There are good reasons to think that this type resulted secondarily from a process of dieresis with accent retraction: [enjá] > [enía].
