Stress shifting suffixes

[David Sanchez]

I think the more facts of the magnifical exposition of J. Campbell
can be satisfactorially explained by rules of epenthesis (Rule A),
and penultimate syllabe accent (Rule B).

Campbell, accept that in classical dialects the accent is always
on penultimate syllabe, this implies that Rule A precedes always
Rule B. But it is possible that in some contemporanial dialects of
classical dialect of Tenochtitlan the order of Rules was reversed ?
That is, if some modern dialects reflect that Rule B precedes Rule
A, is this a recent fact or may it to be an ancient fact?
________________________________________
Explaining Campell's data:

>                 Canoa           Tlaxcala
"Classical"
>
>                 tla'xcal        tla'xcal        tortilla        tlaxca'lli
>                 ca'xtol         ca'xtol         fifteen         caxto'lli
>                 ma'cuil         ma'cuil         five            macui'lli

0a) abstract phonological forms: /tlaxcal/, /caxtol/, /macuil/
0b) absolutive suffix for l-final forms: -Ø (none)
1) Rule A has the trivial efect (since the phonological forms already have
an acceptable surface form): <taxcal, caxtol, mauil>
2) Rule B produces trivially the correct accent: ['tlaS.kaL], ['kaS.toL] ...

>                 a'moxtli        a'moxtl         book            amo'xtli
>                 i'chpochtli     i'chpochtl      girl            ichpochtli
>                 tzo'htzomahtli  tzo'htzomahtl   clothing
tzohtzoma'htli
>                 ma'htlactli     ma'htlactl      ten
mahtla'ctli

Here there is a difference between San Miguel Canoa dialect and Tlaxcala
dialect, the second allowing more complex syllabic paterns at the end of
word.
But stress accent is equally well predicted by rules A and B, if we apply
first
rule B and then rule A.

As for the Michoacan dialects rules also predict correctly:
>
> a)    quichi'hua           he does it
> b)    quichi'c             he did it
>  c)   moca'hua             he remains
>  d)   moca'c               he remained

0) Abstract phonological forms: a) kichiwa, b) kichiak, c) mokawa & mokawak
1) Rule B: a) kichi'wa, b) kichi'ak, c) moka'wak
2) Rule A acts trivially, and has no effect.
3) Forms b) & d) undergoes an "apocope" and is reduced to b) kichi'k & d)
moka'k

The rest of "anomalous" examples (with respect "Classical nahuatl" accepted
forms),
are equally explained by the hierarchical application of rules in the order,
first B,
secondly A.

David Sánchez