The three register tones are marked with the acute accent (high tone)
and the grave accent (low tone). The mid tone is unmarked. Nasals and
the approximant /j/ in coda position can bear a low
tone, which is why they can also have an accent mark. If the coda
consonant is unmarked, it means that it bears the same tone as the
syllable's nucleus. Syllabic nasals can bear all three tones and are
marked in the same way as the vowels. As the grammatical tones described
in the previous section and floating tones can cause shifts in the tones
of lexical roots and other processes, which cause a labor-intensive and
confusing transcription, they are not transcribed. Some examples in the
following sections, however, will transcribe grammatical high tones as H
when the tone is essential for a construction. As non-automatic
downsteps only occur due to deleted vowels or grammatical tones, they do
not have to be transcribed. The only exception are the paradigms for
interrogative and conditional subject pronouns. These are both
transcribed with the grammatical tones and a downwards arrow for a
non-automatic downstep to differentiate them from the indicative
paradigm. Automatic downsteps are not marked at all.