Scott (TEI-L) wrote:
> After having read through the pertinent sections of the P5 guidelines
> and the list archives, I'm having trouble deciding whether references to
> diphthongs (e.g., ae, au, ei, etc.) in a Latin grammar text would be
> properly encoded as characters (<c>) or morphemes (<m>).
> 
> They "smell" like character elements to me, but I am insufficiently
> schooled in linguistics to judge. The content model for <c> does not
> appear to limit element values to any particular length, but I'd rather
> not infer too much from that.


<c> every time for me.

diphthongs are not morphemes, they are characters. Indeed some of them 
are represented by a single Unicode character!