This study attempts to characterize vocoids, i.e. vowels and semivowels, as a unified class of segments. In order to do so, it investigates the main phenomena concerning the quantitative distribution of these sounds, namely syllabic alternation, length alternations, deletion and insertion. Such phenomena are best analyzed by making reference to prosodic structure, and syllable structure in particular. Therefore, both frameworks adopted in this thesis take into consideration this type of representation. The main approach, which I refer to generally as Derivational Theory (DT), is based on the notion that surface phonetic forms are derived from underlying forms through a series of structural changes taking place at different levels of representation. This model is contrasted with the recently introduced (Prince and Smolensky 1993) Optimality Theory (OT), an output-oriented paradigm based on the parallel evaluation of candidate forms by means of universal but violable constraints. This thesis shows that OT offers some valuable insights into the phenomena under analysis, although there are areas in which it requires integration with derivational tools. This study also makes specific reference to two languages: Ammani Arabic and Standard Italian. These diverge in their treatment of vocoids, but clear general trends may be detected which have also been found in other languages.