[...] substrings may be definable in terms of number of segments regardless of their consonantality and vowelhood; and substrings may be defined in terms of number of syllables, rather than number of segments. The number-of-segments hypothesis is easily refutable. If it were true, this would mean that in a language where the first CV of consonant-initial stems is reduplicated, in vowel-initial stems the first VC shoul be reduplicated. I have seen no language of this kind [...].
- Moravcsik (2004), a pag.319