Re: Math/Phonological formulae

Hey. Just popping in for a moment, and seeing that phonology
formalisms are being discussed ...

On 2/6/07, Roger Mills <rfmilly@...> wrote:
> Eric Christopherson wrote:
>
> > I thought I understood the notation used by sound change rules, but
> > there are a few areas I don't understand in this discussion.
>
> I'll try to answer these questions, and I assume David will too....
> >
> > On Feb 6, 2007, at 2:15 PM, David J. Peterson wrote:
> >
> > > Roger wrote:
> > > <<
> > > Plan X is to continue writing things all on one line (OK for feature
> > > matrices e.g. [+XX +YY -ZZ], clear enough). Curly braces could if
> > > necessary
> > > be done like this:
> > >
> > > X --> Y / {1.abc, 2.def 3.mno} or this possibility--
> >
> > What do braces mean? What is 1.abc, etc.?
>
> In my ex. I simply used "abc, def" as hypothetical distinct environments.
>
> Curly braces are used to indicate a variety of usually unrelated
> environments. Suppose vd.stops > fricatives in 1.intervocalic and 2.after
> nasals. There's no (simple) way to combime those into a single env., so--
>
> bdg > BDG {1.V_V 2.[nasal]_}
>
> That's one of the reasons why McCawley considered the use of { } spurious--
> you are probably combining two separate rules into one, making a
> generalization that probably isn't correct.
In practice they're used to abbreviate rule disjunction; either the
first rule applies, or the second, but not both. I think that McCawley
overstates his case, since I don't think that the rule notation found
in SPE was intended to be interpretable in first-order logic, which is
what he tries to do.

> > > For my own personal purposes, I've always found one line to
> > > be sufficient--especially if you condense stuff.  As long as you
> > > know what your features are, and can define them in terms of
> > > widely accepted features, you can do whatever you want.  So,
> > > for example, say you had a rule like this:
> > >
> > > C[+cons, -cont, +voice, +lateral] > [+cont, -voice] / #_
> > >
> > > This'd be a rule that changes an /l/ to a [K] at the beginning of
> > > a word.
> >
> > Isn't /l/ also a continuant ([+cont])?
> Yes. David get a D on that assignment :-)
Not so fast. In Spanish voiced stops do not spirantize following [l],
but they do following [r] and vowels. You can use [±cont] to
distinguish the environment for spirantization; that is,

C[-son, -cont, +voice] -> [+cont] / [+cont] __

Or, voiced stops are realized as voiced continuants following a
continuant (vowel or /r/). [l] ends up being [-cont] by this rule.

> > What's C&H?
> Chomsky and Halley (Sound Pattern of English, where they invent several rule
> formatting conventions that are controversial, to put it mildly)
No, C&H is pure cane sugar, from Hawaii, growing in the sun. When you
cook, when you bake, for goodness sake use C&H pure cane sugar, that's
the one.

But seriously ...

I've always referred to Chomsky and Halle as SPE.

[snip]

> > d. What do the parentheses mean?
> Presence of that segment is optional. For ex. if vowels reduce to [@] in 1.a
> final open syllable or 2. a final closed syllable, that can be abbreviated
> to:
>
> V --> [@] / __(C)#
>
> or say, V(:) would refer to long or short V
Yes. more precisely, the parentheses are another abbreviatory device.
The first rule includes all elements in parentheses, the second
excludes them. Brame 1974 has the following rule for stress in
Palestinian Arabic (the same rule works for Maltese):

V --> [1 stress] / __ C_0((VC)VC_0^1)]

_0 is a subscript zero
^1 is a superscript 1
] is the word boundary
C_0 means zero or more consonants
C_0^1 means zero or one consonant

Brame says that this rule abbreviates the following three rules, which
apply in this order:

i.  V --> [1 stress] / __ C_0VCVC_0^1]
ii.  V --> [1 stress] / __ C_0VC_0^1]
iii.  V --> [1 stress] / __ C_0]

These rules work, but they sure are a clumsy way to talk about stress!

[snip]

> I don't know how much in fashion old-style Generative Phonology is, anymore;
Not much. It's pretty much been replaced by Optimality Theory. For
those who just can't abandon derivations, there's still Autosegmental
Phonology, which makes different representational assumptions but
still allows for all of the fun of feeding and bleeding ordering
relations, the cycle, and Lexical Phonology.

> but C&H "SPE" has a long section on rule-writing conventions, plus a very
> handy guide is Robert T. Harris "Introduction to Phonological Theory" (pb.,
> Prentice Hall 1968), which serves me well as a guide.
You mean *Harms*, don't you?

Dirk

Reply