Hugo Cesar de Castro Carneiro, On 25/02/2012 22:53:
> Below is an idea I had:

Something's buggered the table. Maybe separate columns using spaces rather than tabs?

>     4. Four lexical sets were suppressed: BATH, LOT, PUT and THREW. They
>     were suppressed because they are pronounced as one lexical set in some
>     dialects and as another in others, e.g. BATH is suppressed due to PALM and
>     LAD, LOT due to CLOTH and THOUGHT, PUT due to STRUT and FOOT, and THREW due
>     to GOOSE and CUTE. The solution for these suppressed lexical sets would be
>     different orthografies depending on the dialect, like nowadays happens with
>     color/colour, gray/grey and jail/gaol;

That is unavoidable, yes.

>     6. I can't verify if NORTH would be a rhotic equivalent to CLOTH or to
>     THOUGHT;

Do you mean "LOT or THOUGHT"? It's not a matter of verification; it's a matter of your analysis and your orthographic scheme.

>     10. UE is also a possible orthography, that could be in GOOSE lexical
>     set, but only in the end of the word (TR*UE*);

I might have some of the sets popularized by Wikipedia wrong, but I'd have thought (10) applies to CUTE rather than GOOSE.

>     11. A friend of mine told me that alC when l became /w/, and then /aw/
>     became /ɔ/ could be associated to THOUGHT lexical set, so that ALL, WALK
>     and TALK would keep its spelling istead of becoming AUL, WAUK and TAUK.

_Pal_ /pal/, _talc_ /talk/.

> I still have some problems, the main ones are:
>     1. Demonstrate which vowels are reduced and which are not, without
>     changing their orthography or using diacritics;
>     2. Separate digraphs from diphthongs. Using diaresis, like in French?
>     "naïve"?
>     3. How to write words with stress in a syllable that is not the first
>     one?

Initial syllable is not the default position for stress.

> And how to write words with secondary stress?

If you are marking reduced vowels and primary stress, then you don't need to mark secondary stress.

Also, if you're marking reduced vowels, then location of primary stress becomes quite highly predictable, so you could mark it only when it is in nondefault place. (The degree of predictability depends on whether the prediction rules have access to morphological information rather than just phonological.)