Alf Gandson wrote:
> Thus the conflict between an interpretation as a
> single sound and a analysis into several sounds occurs
> at least with affricated stops, aspirated stops and
> diphtongs. But does it also occur with
> syllable-initial consonant clusters such as /tr/ or
> /kl/?

The Latin script as used for latin has at least such a letter: <X> (= [ks]).
The Greek alphabet as two more: <PS> and <Z> (= [ts]).

> The only phenomenon alike I know is the special
> diacritic mark for combinations with r which uses the
> devanagari script.

A Devanagari <subscript RA> is often analyzed as <virama> + <RA>. This
analysis certainly makes sense by the phonetic point of view, and is also
the basis for modern computer encodings for Indian languages.

I don't know whether it also makes sense from the point of view of the
history of that writing system: I am not sure whether this comma-like sign
is really derived from letter <RA>.

_ Marco