Re: Compiling a list of Semitic transliteration characters

[Bill Poser]

There is another reason for romanizing, namely where it is desired to
represent constituents below the level of analysis of the native writing
system, e.g. individual segments when the native system is suprasegmental.
For example, many Japanese verbs have stems that end in consonants that are
not permitted in syllable-final position on the surface and therefore have
no representation in the native writing systems. Restricting oneself to the
native writing systems of Japanese, it is not possible to say that the stem
of the verb "to bite" is /kam/, that the stem of "to take" is /tor/, and so
forth. In discussions of the phonology and morphology of Japanese written
in Japanese romanization is therefore used for such examples.