THEORY: Tepa prosody [was: Estonian Quantity]

Hey.

What follows is best viewed in a monowidth font. It's also rather long.

At 3:50 AM +0000 11/09/01, And Rosta wrote:
>Doergkh:
>>              F                F
>>             / \              / \
>>            F   m            F   m
>>            |   |            |   |
>>            s   |            s   |
>>           /|\  |           /|\  | ...
>>          / m m |          / m m |
>>          |  \|/           | |  \|/  |
>>          s   a   [saa:]   l a   k   e   [lak:ke]
>>
>>  That is, a foot consisting of a monosyllabic bimoraic trochee with an
>>  adjoined mora. This conception of the foot and its role in stem
>>  gradation has (indirectly) inspired some changes in Tepa prosody,
>  > though I don't have overlong segments.
>
>What are the changes to Tepa prosody it has inspired?
I'm glad you asked. The changes in Tepa prosody are the result of the
introduction of exhaustive foot parsing and the restriction of some
phonological processes to foot-medial position.

The exhaustivity requirement of Tepa foot parsing means that all
syllables within a Tepa stem must belong to a prosodic foot; that is,
syllables may not adjoin directly to the prosodic word (this
restriction doesn't apply to clitics). There are three kinds of legal
foot shapes in Tepa:

1) the canonical moraic trochee

foot        F           F
             |          / \
syllable    s         s   s
            /|\       /|  /|
mora      / m m     / m / m
           | | |     | | | |
segment   c v x     c v c v

(x = vowel or consonant)

2) the trochee with resolution

     F
    / \
   s   s
  /|  /|\
/ m / m m
| | | | |
c v c v x

3) the augmented foot

       F
      / \
     F   \
    / \   \
   s   s   s
  /|  /|  /|
/ m / m / m
| | | | | |
c v c v c v

Insuring that a CV string is exhaustively parsed into licit feet
requires several kinds of adjustment: 1) suffixation of _-ka_ (this
is morphologically restricted), 2) final-vowel lengthening, 3)
light-syllable adjunction.

In Early Tepa, bound phase was marked by a final long vowel. In
Modern Tepa, unbound phase is marked by an initial heavy syllable
(moraic trochee). This heavy syllable is usually the result of
geminating the medial consonant, but may also be from a lengthened
vowel (voiceless fricatives and glides don't geminate). This prosodic
marker is accompanied by suffixation of _-ka_ in case parsing the
initial heavy syllable would leave behind a light syllable; _-ka_
suffixation thus fills out a prosodic foot:

tukana 'thrush:BOUND'
tukkana 'thrush:UNBOUND'

pite 'see:BOUND'
pitteka 'see:UNBOUND'

Secondly, the renewed emphasis on the prosodic foot has restricted
application of gradation, now properly seen as lenition. In  Early
Tepa, voiceless stops alternated with voiced fricatives
intervocalically; in the modern language, lenition only applies
within the foot; it may not occur across foot boundaries (somewhat
like American English flapping):

tapatapa [taBataBa]  'black widow:DIST'
pitepite [piD1piD1]  'see:DIST'
tipukankan [tiBukaNgan]  'sage hen:DIST'

In these forms, distributive number is marked by suffixal
reduplication of the final moraic trochee. The resulting form only
shows lenition between vowels when those vowels belong to the same
foot:

     F       F
    / \     / \
   s   s   s   s
  /|  /|  /|  /|
/ m / m / m / m
| | | | | | | |
t a p a t a p a  = [taBataBa]

     F     F     F
    / \    |     |
   s   s   s     s
  /|  /|  /|\   /|\
/ m / m / m m / m m
| | | | | | | | | |
t i p u k a n k a n  = [tiBukaNgan]

(In the form [tiBukaNgan] 'sage hen', the voicing of /k/ to [g] is
not the result of lenition, but is due instead to a separate process
of post-nasal voicing; post-nasal voicing is not restricted to
foot-medial position.)

When a form has an odd number of light syllables, one of two things
may happen: 1) the final vowel may lengthen, coercing a final moraic
trochee; or 2) the final light syllable may be adjoined to the
preceding foot to create a foot consisting of three light syllables.
Both options are illustrated with the form _hipite_ 'moon':

                                   F
                                  / \
     F     F                     F   \
    / \    |                    / \   \
   s   s   s                   s   s   s
  /|  /|  /|\                 /|  /|  /|
/ m / m / m m               / m / m / m
| | | | | |/                | | | | | |
h i p i t e   = [hiBit1:]   h i p i t e  = [hiBiD1]

In the first form, the final long vowel creates a monosyllabic moraic
trochee. There are two feet, thus lenition will not apply to /t/,
since it is not between vowels belonging to the same foot. In the
second form, the final light syllable is simply adjoined to an
existing foot to form a single augmented foot; since the /t/ is
between vowels belonging to the same foot, it is lenited.

As far as I can tell, both realizations are in free variation.

So while Estonian didn't provide a direct model, Prince's analysis of
overlength as a metrical phenomenon got me thinking about foot-based
morphophonology in Tepa.

>Where is up-to-date info on Tepa? It must be six or seven years since
>I last read the full description of Tepa.
Right now it's on my hard drive; Jeffrey Hennings has some stuff on
his Langmaker.com pages, but for some reason the inflectional
morphology of nouns and verbs never made it there. I'm still working
out some changes to the derivational morphology. Once they are done,
I'm planning on writing up the phonology and morphology and posting
it here. I'm still thinking about where to get web space; I don't get
any from BYU for personal pages. I'll probably have to go with a free
service.

>--And.
Dirk
--
Dirk Elzinga                              Dirk_Elzinga@byu.edu

"Speech is human, silence is divine, yet also brutish and dead;
therefore we must learn both arts."            - Thomas Carlyle

Replies