phonological rules (summary)

Joan Bybee joan.bybee at gmail.com
Mon Nov 18 17:43:30 UTC 2013


Dear Fritz,

The reason you didn't get replies from people defending exemplar approaches
to phonology/phonetics is that they probably aren't on Funknet. Instead
they are the Laboratory Phonology people and the sociophoneticians. Most of
these researchers are finding exemplar theory indispensable.

The reason things work out so neatly for you, Labov and Kiparsky is that
you and they are ignoring a large and growing body of data that is in a
sense below the radar of phonemic theory. These are studies that use fine
phonetic detail and a large number of tokens from a large number of
speakers. This type of data goes way beyond your anecdote about how you say
*dog* and *frog*. Take a look at a special issue of the *Journal of
Phonetics*, vol. 34, 2006. And there is also Ogura 1995 (*Diachronica
12:31-53) *which presents data from the diphthongization of ME long /i/ and
/u/ (in the Great Vowel Shift) showing lexical diffusion of the
phonetically gradual change (passing through stages which were not already
represented by existing phonemes). Of course, Phillips 2006 (her book from
Palgrave: *Word Frequency and Lexical DIffusion*), who reviews many of
these studies.

Important evidence for exemplar storage comes from studies that show that
it is not just the frequency of the word but the frequency in which it
occurs in the contexts which condition the sound change/phonological
process that determine the phonetic shape of the word. See Bybee 2002
(in *Language
Variation and Change*) and studies by Esther Brown and colleagues  (Brown
and Raymond in *Diachronica 2012: 139-161 *and papers cited there). Neither
Kiparsky's theory nor Labov's taxomony accounts for any of these findings.

Joan




On Sun, Nov 17, 2013 at 2:48 PM, Frederick J Newmeyer
<fjn at u.washington.edu>wrote:

> Thanks, Joan,
>
> Yes, you are right. Nobody suggested a publication that makes explicit
> comparisons. Such a publication appears not to exist.
>
> Interestingly, given that my posting was on Funknet, almost all of the
> replies provided some sort of defence of classical phonology. The big
> exception was that a few people told me to read Port and Leary's paper
> 'Against Formal Phonology'. I had read it, and I read it again. My problem
> with it is that what they consider to be the defining features of 'formal
> phonology' are so different from what I take them to be that I ended up
> being more puzzled by their critique than anything else.
>
> I am on admittedly shakey ground not being a phonologist and not
> commanding the literature as well as I wish that I did. But I find Labov's
> work on frequency effects and lexical diffusion as is has been extended by
> Kiparsky to be a convincing way of reconciling the empirical discoveries of
> usage-based phonology (yours and those of others) with the classical model
> of phonology. Kiparsky argues that 'being a redistribution of phonemes
> among lexical items, [lexical diffusion] cannot produce and new sounds or
> alter the system of phonological contrasts.' He goes on to show how lexical
> diffusion is a species of grammar simplification driven by the system of
> structure building rules in the lexical phonology. He gives examples
> including the shortening of long /u/ in English (in words like 'took' and
> 'good') and Labov' famous AE tensing. Point by point Kiparsky shows how
> lexical diffusion manifests the same properties as lexical analogy. In
> short, if he is right, we need only the mos!
>  t minimal extension of the Neogrammarian / classical phonology model.
>
> I can't resist giving an example from my own variety of English (impure
> Philadelphia) that seems to support Kiparsky's model. It pertains to the
> reflexes of Early Modern English short [o]. In this dialect, [o] became a
> low back rounded diphthong in frequent words and became [a] in
> less-frequent words:
>
> LOW BACK ROUNDED DIPHTHONG          [a]
> dog                                 frog, bog, log
> on, off                             honor, offal, don, doff, Goth
> loss, boss                          floss, dross
> strong, song, wrong                 gong, tong
> cost, frost                         Pentecost
>
> I'm not sure what the precise historical sequence of developments was
> here. But importantly, both sounds already existed in this dialect: Low
> back rounded diphthongs in 'saw', 'thought', 'bought', etc. and [a] in
> 'father', 'drama', etc. So it looks like what we have is an analogical
> extension of an existing phonological rule. Nothing new was created in the
> (presumably) word-by-word diffusion through the lexicon, nor do the
> frequency effects lead to calling into question classical phonology.
>
> Best,
>
> Fritz
>
> Frederick J. Newmeyer
> Professor Emeritus, University of Washington
> Adjunct Professor, U of British Columbia and Simon Fraser U
> [for my postal address, please contact me by e-mail]
>
> On Sun, 17 Nov 2013, Joan Bybee wrote:
>
> > Dear Fritz,
> >
> > Thanks for the summary. It seems you didn't come up with much in the way
> of
> > explicit comparison. Also you didn't mention the models that combine OT
> > with exemplar theory (by Joeren van der Weijer, James Meyers) or more
> > formal representations of exemplar theory (Paul Boersma). Rather than
> > compare, these theorists choose to incorporate features from both types
> of
> > models.
> >
> > As for the points made in your message, they unfortunately do not
> > constitute differences between exemplar theory and phonemic theory or
> > natural phonology (of Donegan and Stampe). I would like to comment on
> each
> > point.
> >
> > Point 1: Adopting exemplar theory does not mean there are no phonological
> > processes, as Bybee 2001 and Pierrehumbert 2001 note. They are in basic
> > agreement with Donegan and colleagues that phonological processes are the
> > neuromotor patterns (thus both cognitive and real time) that influence
> > actual pronunciations.
> >
> > Point 2: Exemplar theory is above all a theory of categorization and thus
> > what would have earlier been called phonemic categories are said to
> derive
> > from the organization of exemplars that are similar to one another.
> > Exemplar theory arose from studies of the categorization of
> non-linguistic
> > objects and when frequency in experience is taken into account, these
> > categories show prototype effects. Only if one is determined to believe
> > that linguistic categories are different from other cognitive categories
> > would one want to say that phonemic categorization is not based on
> > exemplars. The empirical evidence argues against this position: Joanne
> > MIller (*Cognition* 1994, 50:271-285) in various works has shown that
> > phonemic and even allophonic categories show prototype effects.
> >
> > In addition, exemplar theory allows for the gradual merger of phonemic
> > categories and their gradual split. Phonemic categories with their plus
> and
> > minus features are not compatible with gradual change.
> >
> > Point 3: It is not really an argument that Labov has overlooked the fact
> > that gradual phonetic change diffuses gradually through the lexicon.
> >
> > Point 4: Calling lexical diffusion analogy just kicks the can down the
> > road. What is analogy then? Why do morphophonemic analogies affect low
> > frequency items first while phonetic ones affect high frequency items?
> >
> > Point 5: This is by definition; it is not an empirical finding.
> >
> > I hope you find these points useful.
> >
> > best wishes, Joan
> >
> >
> >
> > On Wed, Nov 13, 2013 at 3:22 PM, Frederick J Newmeyer
> > <fjn at u.washington.edu>wrote:
> >
> >> Dear Funknetters,
> >>
> >> Last week I posted a request for literature references on the degree to
> >> which classical phonology (discrete rules, regular sound change) seems
> to
> >> be refuted by the 'usage-based' work of Bybee and others that stresses
> the
> >> importance of relative frequency of variants, 'exemplars' embodying
> >> multiple representations instead of discrete phonological units, and so
> on.
> >> I received several responses with references that seem to support more
> >> classical views. I summarise them below without personal comment:
> >>
> >> 1. Work in natural phonology (Donegan 2014; Donegan & Nathan 2014)
> defends
> >> the existence of 'phonological processes', which, like classical
> phonemic
> >> statements, are holistic characterisations of (non-morphological)
> >> phonological alternations, rather than the item-by-item representations
> >> provided by usage-based models.
> >>
> >> 2. Along these lines, work going back to Baudouin de Courtenay and Sapir
> >> and continuing to the present day supports the psychological reality of
> the
> >> phoneme (or its generative counterpart), not just that of individual
> >> phonetic elements.
> >>
> >> 3. Labov (1994) provides a typology of phonological change, in which the
> >> classic Neogrammarian view of exceptionless change plays a major role.
> >>
> >> 4. Kiparsky (1995) argues that lexical diffusion (which prima facie
> calls
> >> into question the classical model) is a form of analogical change and
> not
> >> 'sound change' per se.
> >>
> >> 5. (Degree of) gradience does not enter into statements of phonological
> >> processes or phonological change. For example, a phonological rule
> might be
> >> sensitive to whether a following segment is nasal, but not to its
> *degree*
> >> of nasalisation.
> >>
> >> Thank you all for your replies.
> >>
> >> Fritz
> >>
> >> REFERENCES
> >>
> >> Donegan, Patricia J. (2014), 'The emergence of phonological
> >> representation', in Brian MacWhinney and William O'Grady (eds.),
> Handbook
> >> of language emergence. Boston: Wiley.
> >>
> >> Donegan, Patricia J. and Geoffrey S. Nathan (2014), 'Natural phonology
> and
> >> sound change', in Patrick Honeybone and Joseph Salmons (eds.), Oxford
> >> handbook of historical phonology, Oxford University Press. Oxford.
> >>
> >> Kiparsky, Paul (1995), 'The phonological basis of sound change', in John
> >> A. Goldsmith (ed.), The handbook of phonological theory. Oxford:
> Blackwell,
> >> 640-70.
> >>
> >> Labov, William (1994), Principles of linguistic change. Volume 1:
> Internal
> >> factors, Language in Society, 20. Oxford: Blackwell.
> >>
> >>
> >>
> >>
> >>
> >> Frederick J. Newmeyer
> >> Professor Emeritus, University of Washington
> >> Adjunct Professor, U of British Columbia and Simon Fraser U
> >> [for my postal address, please contact me by e-mail]
> >>
> >>
> >>
> >
> >
> > --
> > Joan Bybee
> > HC 66 Box 118
> > Mountainair, NM 87036
> > 505-847-0137
> >
>
>
>



More information about the Funknet mailing list