I'm not a trained linguist, but I do have an amature interest in the subject, and have studied Japanese from both a second language, and a linguistic perspectiv

I'm not a trained linguist, but I do have an amature interest in the subject, and have studied Japanese from both a second language, and a linguistic perspective. The basic building block of English phonetics is the syllable. In Japanese, it is the mora. If you are fammilar with the Japanese writing system, to a first approximation, each phonetic character maps to a single mora [0]. A mora comes in three forms: 1) (C)V - A vowel, optionally preceeded by a consonant (almost all phonetic characters, and compounds like しゃ) 2) The first part of a long consonant. (little っ). 3) Syllable-final nassal /n/ (ん). Although moras map closely to the writing system, their classification is based on an analysis of spoken japanese (including word games and common mistakes by native speakers). The relevent piece of this is that there is no way to express "des" in terms of moras/. Instead, we view is as "de-su" with a devoiced vowel. Devoicing is not common in all dialects of Japanese. In the dialects where it is common, a native speaker would recognize the voiced and de-voiced variations as being different ways of pronouncing the same thing. In dialects where devoicing occurs, there are well understood rules for predicting when it will occur. The particular rule prediciting desu and masu is that high vowels (い/i as in igloo, う/u as in goose) are devoiced when they are word ending and follow a voiceless consonant [1]. [0] Compounds like しゃ are still a single mora, but little っ is its own mora. [1] A voiceless consonant is one that is made without vibrating the vocal chords. This is the difference between, for example, ka/か and ga/が.