Re: Phonetic scripts and diphthongs ...

Quoting j_mach_wust <j_mach_wust@...>:

> I've always thought it's a problem of the notion of 'phoneme' that it
> only applies to sounds of the same complexity, since it assumes that
> there's no further division of the examined sounds. So _die_ and _tie_
> form a minimal pair because they share the same complexity: one sound
> only, but _die_ and _dry_ don't, because tradition says they don't
> have the same complexity (one sound: /d/ vs. two sounds /dr/). But why
> couldn't the pair _die_ vs. _tie_ be analysed as having different
> complexities: one sound /d/ vs. two sounds /dh/? It'd be all analogous
> to _die_ vs. _dry_.
While it, I guess, has a certain theoretical interest, it's not really relevant
to the orthographic questions I raised, is it?

> > Anyone's got any thoughts as to the possible causes of this apparent
> > asymmetry?
> >
> > Or is it just an accident of the development of the Latin script?
>
> Exactly. I believe it's a peculiarity of the English alphabet that
> certain affricates are represented by single letters.
By "the English alphabet", do you mean the Latin alphabet, or the English
orthography?

In the second case, you're patently wrong - affricates represented by single
letters are all over the place in plenty of languages written in the Latin
script; it's no more particular to English than is, say, the use of 'h' as a
modifier.

In the later, well, I'd like something a bit more concrete than a "belief". I
know for a fact that Cyrillic has single letters for affricates, but normally
uses digraphs for diphthongs. This, too, could be an "accident", or course.

> > What about other phonemic scripts? How do Indic scripts go about
> > indicating diphthongs, for instance? Affricates?
>
> The devanagari script, at least, has single representations for /ai/
> and /au/, and these come from old representations of long diphthongs
> /a:i/ and /a:u/, whereas the original short diphthongs /ai/ and /au/
> are nowadays the normal letters for /e:/ and /o:/.
Ah. Interesting.

> It has no letters
> for affricates, so they must be represented with ligatures and count
> as compound of two 'phonemes'.
If so, they aren't phonemic affricates.

                                                         Andreas