LINGUIST List 2.191

Saturday, 4 May 1991

Disc: Phonology

Editor for this issue: <>

Directory

1. , Re: Phonology
2. Mike Hammond, Rules
3. , Re: Phonology
4. bert peeters, Borrowing of phonological structures
5. John Coleman, Sanity
6. John Coleman, Formalism

Rick writes:

>I invite others who can verify Alexis' claim that Russians use bare [y] as an
>alternate letter name for "y" (traditionally called "yerih") to come forward.

Confirmed.

>All I am saying is that people have two distinct vowel symbols here--
>one for preceding soft consonants and one for hard consonants. Since it is
>unusual for alphabetic symbols to represent distinct allophones (i.e. they
>usually stand for phonemes or phonemic sequences), I am proposing that the
>existence of the symbols increased the salience of the allophones in the minds
>of speakers.

The Cyrillic alphabet was not designed to represent a language with the
phonemic structure of modern Russian. As Rick has noted elsewhere,
phonemically distinct "paired" palatalized and nonpalatalized consonants arose
as part of a rephonologization connected with the loss of the jers as
independent phonemes. Before this rephonologization, the <i> and <y> letters
represented /i/ and /y/ phonemes. I agree that orthography may the
perceptions of literate speakers, but orthography may not change to keep up
with phonological change.

Horace Lunt ("On the Relationship of Russian Spelling to Phonology," Studia
Linguistica Alexandro Vasilii Filio Issatschenko a Collegis Amicisque Oblata,
Lisse, 1978, 225-30) suggests that modern Russian orthography is actually quite
close to modern Russian phonological structure, that there is no pairing of
consonants for palatalization, and that Russian vowel letters pretty much
represent independent vowel phonemes. The system requires synchronic jers and
includes absolute neutralization. I mention this as an alternative analysis
that has its advantages and disadvantages.

--David

Harry Bochner maintains that the behavior of 'sanity'...

"...is only partly rule-governed. In particular, it is impossible to
predict the existence of the word 'sanity' from anything else in the
grammar; this is what it means to say that -ity is unproductive
(overall, that is; it is productive in two of its subenvironments)."

I think it would be a mistake to incorporate this notion of
productivity into our formal theory of morphology. Specifically, I
think that determining what the "possible words" of a language are is
a linguistic question and determining what the occurring words are
is largely a nonlinguistic question, hinging on history, technology,
etc.

mike hammond

John Coleman questions me on what I meant by the metaphorical expression
'intimacy between languages' in connection with borrowing. In general, it
means that there are a substantial number of bilinguals in the speech 
community. I don't think that pronunciations really get "borrowed" by 
monolingual speakers. Rather, the usual situation is for the borrower to
filter the foreign word through his native pronunciation. We sometimes refer
to this as the "nativization" of a word.

The issue between Harry Bochner and Mike Hammond on listing vs. rule-governed
behavior recalls Langacker's warnings about the "rule/list" fallacy that ties
linguists in knots from time to time (cf. p. 42 of Foundations_of_Cognitive_
Grammar). Langacker points out that there is no reason why speakers can't
memorize some plural forms as fixed units and still have a rule that derives
plurals. Why not have both lists and rules?

In the case of '-ity', it is crystal clear that English has a rule that tells
you how to construct words by sticking an '-ity' suffix on adjectives. We 
know what happens to the phonemes in the stem and the stress in the stem. We
know part of speech of the stem and the part of speech of the resulting word.
The rule can be used to make guesses about the meaning of an unfamiliar word
ending in '-ity', should you encounter one, and it allows you to make up a new
word ending in '-ity', even over the esthetic objections of other speakers.
But that does not mean that words like 'electricity' and 'sanity' have to be
stored in the brain in unassembled pieces. The '-ity' rule has a function--
to augment vocabulary. It is needed for unfamiliar words, not familiar ones.

 -Rick Wojcik (rwojcikatc.boeing.com)

To my idea, Andre Martinet is still an authority to be read and reread about
the borrowing of phonological structures. He has written about it in several
places, only one of which I have a precise reference for here and now. The
publication I have in mind is Martinet's 1975 Evolution des langues et
reconstruction (Paris, PUF). I hope in a not too distant future to publish a
paper which will have details about Martinet's stand in this matter.
Bert Peeters <peeterstasman.cc.utas.edu.au>

I agree with Harry Bochner <bochnerdas.harvard.edu>
that we need a theory that allows us to
express generalizations about lexically listed facts.

Mike Hammond statement that 
> Rules are rules. What else is there? 
is too glib. There are different kinds of rules and rule-systems, of 
different generative power. "Rules" in phonology no longer means
simply "productive generalisation" or "licensor of a step in
a derivation". It means something more like the Hallean kind of
A --> B / C ___ E, even in theories like Autosegmental Phonology.
 D

There are types of rules that are much less powerful than this,
such as structure-building and redundancy rules, which, being
monotonic, need not be extrinsically ordered. To take the "sane/sanity"
example, a declarative, lexicalist analysis could proceed as follows:

The stem (san-) is a lexical entry, as is the bound suffix (-ity). 
san- surfaces with two forms with vowels of different length and 
vowel quality. But the change in vowel quality is parasitic on the
length alternation (vowel shift). So if we factor out the vowel
shift relation, which is general, and express it as a structure-building
redundancy rule, it might look like this:

[+long] -> [+shift]

(read -> as `implies', not `rewrites as').

The stem (san-) must be underspecified for length, in a structure-building
account, because feature-changing rules aren't allowed. Let's represent
it as (sAn-), say. Now, in its unsuffixed form, its long. Let's state this
as a default:

DEFAULT: V -> [+long]

Defaults can't *change* features so the analysis is still monotonic.

Now for -ity. There are two ways of handling this, I think, one specific
to the representation of -ity, one more general. The general one first.
Let's establish a constraint on ternary feet which says that a syllable
before two unstressed syllables is short (or that the head of a ternary
foot is short). This is a declarative/metrical implementation of
trisyllabic shortening (except that it only ADDS the feature [-long],
of course; it doesn't change anything). I propose this analysis first
as I believe it's fairly general, even though there may be specific 
exceptions. If it turns out to be specific to just a few suffixes like
-ity, then the [-long] feature which it introduces could be built into 
the lexical representation of -ity, e.g.:

[-long]([+cons]) ] ity

This can be unified with the stem (sAn-) to form sanity, with short
and consequently unshifted stem vowel.

My point in raising this is to show that even in a lexicalist approach
of the sort that Harry Bochner may be interested in, there seems to be a need 
for "rules" of a sort if generalisations such as vowel shift and trisyllabic
shortening are to be expressed, even if they are not the powerful kinds of
rewriting rules in common use in most phonological analyses.

--- John Coleman

sakacogsci.Berkeley.EDU (Paul Saka) points out that
> If algorithm is formal in Manaster-Ramer's
> "form" sense (in the sense that its operations are defined over shapes),
> then the algorithm is also formal in the "rigorous" sense: there is much
> less room for disagreement about the SHAPE of an object than there is
> about the INTERPRETATION of an object. 

I'm not sure this is correct; in the absence of agreement about the 
interpretations of the forms which a formal theory proposes, there is
plenty of room for disagreement about what the shape of objects in the
theory should be. Consider, for example, SPE phonology. By contemporary
standards this was a fairly precisely defined (i.e. `rigourous-formal')
model, and in fact even the intended *interpretation* of the minimal
objects in the theory was fairly tightly defined, yet there was (and 
still is) considerable debate about what the phonological representation
of lexical entries should be. This is because as well as the formalism,
every linguist has unformalised metatheoretic views regarding what they
expect or would like the objects of their study to be like. In addition,
phonological analyses are precisely that - ANALYSES - i.e. the 
result of things you do with some observations, thoughts, hunches etc.
which is (perhaps) not particularly affected by the formalism.

--- John Coleman

[End Linguist List, Vol. 2, No. 191]