LINGUIST List 13.92

Wed Jan 16 2002

Sum: Phonotactics in Morphological Phenomenon

Editor for this issue: Marie Klopfenstein <marielinguistlist.org>


Directory

  1. John Alderete, phonotactics in morphological blocking and suppletion

Message 1: phonotactics in morphological blocking and suppletion

Date: Mon, 14 Jan 2002 14:38:12 -0500 (EST)
From: John Alderete <aldereteruccs.rutgers.edu>
Subject: phonotactics in morphological blocking and suppletion

Dear members of the list,

This message is an index and compilation of the messages I got in response
to my call (Dec. 5, 2001) asking for language examples in which it appears
that phonological constraints play a role in blocking morphological
processes and governing suppletion (the index combines messages from
my calls on both the Linguist List and the Optimal List). Thanks to all
those who participated (the names of the participants are listed below in
the index). You have helped to provide the larger linguistic community
with a wealth of interesting examples and a better bibliography than the
one I started with. Some contributors refer to their own work, available
in electronic format; please use the contact information in their message
to obtain this work. Finally, as this is a rather long document, without
hypertext links, take note that index message headers can be found
directly by using your favorite search tool: the index number #] should
help identify the index message headers in the actual correspondence below
since ']' is reserved from these headers.

Best wishes,

-John Alderete, Rutgers University

=========================================================

Index for correspondence:

1] John Alderete, Original appeal for references (updated) and examples 
2] Amanda Seidl, Example from Mende
3] Lee Bickmore, Example from Tahitian
4] Carson Schutze, English example: 'person from X /-er/'
5] Mike Maxwell, Examples from Tzeltal (Mayan), Shuar (Jivaroan, Ecuador),
Axininca Campa (Peru), and Cubeo (Tucanoan, Colombia)
6] Wayles Browne, Example of Slavic suffix -ba
7] Jose Elias Ulloa, Example from Shipibo and Capanahua (Panoan, Peru)
8] Gail Coelho, Example from Thompson River Salish
8.1] Suzanne Urbanczyk, Reference for Salish example
9] Brett Baker, Examples from Ngalakgan and other Australian languages
10] Donca Steriade, Examples from Latin and Greek declensions
11] Armin Mester, Examples from German (diminutive -chen / lein and A->N
suffix heit/keit) plus references
12] John Koontz, Example from Winnebago and discussion
13] Lisa Davidson, Example of Portuguese and Spanish diminutives
14] Yoonjung Kang, Example from Korean
15] Arto Antilla, References
16] Birgit Alber, Example from German
17] Todd Bailey, Examples from Spanish, Latin and Polish
18] Moira Yip, Examples from English and Cantonese
19] Jeremy Whistle, Discussion of English /-ful/
20] Andrew Spencer, Reference and discussion of general topic
21] Larry Trask, Examples from English (/en-) and Basque and discussion of
English /-ful/ and /-ize/
22] John Koontz, Examples from Siouan
23] Pius ten Hacken, Explanation of morphological dictionary in Word Manager
24] Anna Thornton, Examples from Italian (several examples)
25] Katherine Crosswhite, Example of verbal -s'a/-s' in Russian
26] Ingo Plag, Reference and discussion
27] Mark Aronoff, References
28] Eric Bakovic', Example of diminutives in Spanish
29] Andrew Carstairs-McCarthy, References

=========================================================
Individual messages (with repeated index messge headers):

1] John Alderete, Original appeal for references (updated) and examples 

Date: Wed, 5 Dec 2001 14:14:36 -0500 (EST)
From: John Alderete <aldereteruccs.rutgers.edu>
To: optimalucsd.edu
Subject: query: subcategorization in phonology and morphology


Dear friends and colleagues,

I'm currently working on a cross-linguistic study of subcategorization
effects in morphology and phonology, and I'd like to appeal to the members
of this listserv for examples and references. I will prepare a summary of
the responses of this query for future reference on the list.

The basic type of pattern I'm looking for is one in which a morpheme
'subcategorizes' for a phonological property of a neighboring morpheme,
i.e., it only occurs when a subcategorized phonological property is
present. For example, it is sometimes said that the adjective-forming
suffix /-ful/ in English subcategorizes for a stressed syllable at the end
of the stem to which it attaches, e.g., forgetful, cf. *forgettingful
(Siegel 1973; though this example is more complicated, see Brown 1958 and
Chapin 1970).

Subcategorization effects such as this can block the application of a
morphological process, as in the case with English /-ful/. Or it may be
the basis of the selection of a particular allomorph when two or more
morphemes compete for a particular collocation, e.g., noun-forming {-ei,
-erei} in German, where -erei is chosen if the stem ends in stress, and
-ei if the stem-final syllable is unstressed (Hall 1990, Hargus 1993); one
might say therefore that -erei subcategorizes for a stem that ends in a
stressed syllable. Examples such as these have contributed to the study of
interactionism in phonology and morphology (Siegel 1974, Hargus 1993,
Booij & Lieber 1993), the prosodic dependence/independence of morpheme
classes (Inkelas 1989), the nature of alignment constraints (McCarthy &
Prince 1993, Kager 1996), and the structure of OT grammars (Orgun &
Sprouse 1999).

If any examples like this come to mind (which are not already listed
below) in the languages you understand well, I'd greatly appreciate it if
you could take the time and respond with a brief description of the facts
and the relevant references. Many of the examples listed below are
prosodic in nature, but examples of subcategorizatin for segmental
properties are also very welcome. As far as whether an example
'qualifies', it is best to err on the side of inclusion. In fact, any
significant discussion of these phenomena (but see references below) or
relevant examples of phonologically governed blocking and allomorphy would
be most welcome. The list of examples and references below are my starting
place, but if you know of useful discussion that builds on this work, or
gives critical assessment, that would be helpful as well. Please address
all correspondence to this address (repeated at the end of the message).

Finally, please forward this message to colleagues who you feel would
be able to contribute but may not be on this listserv. 

Thank you for your time. -John Alderete


Examples:

Dutch -ig (N->A), only attaches to stems ending in stressed syllables (may
induce both blocking and stress shift, see Trommelen & Zonneveld 1989,
Kager 1996)

Dutch {-isch, -ief} (N->A), -ief is chosen when the final syllable is
stressed in the underived base and ends in -ie, -isch chosen elsewhere
(Booij & Lieber 1993)

English /-en/ (A->V), only attaches to heavy syllable stems; cannot attach
to stems ending in liquids, nasals, or vowels (Siegel 1974: 174-176)

English /-ful/ (N->A), only attaches to stems ending in stressed
syllables; cannot attach to stems ending in /v f/ (Siegel 1974: 164-174,
Brown 1958, Chapin 1970)

English /-al/ (V->N), only attaches to stems ending in c1V(c2)(c3), where
V is stressed, c2 (optional) is [+son] and c3 is either a coronal or a
labial (Ross 1972, Siegel 1974: 164-168)

English /-ize/ (N/A->V), only attaches to stems ending in an unstressed
syllable (Raffelsiefen 1996)

English {-eteria, -teria}, -eteria when stem-final syllable is stressed,
-teria elsewhere (Siegel 1974: 176-178)

German ge- (perfective participle), only attaches to stems that begin with
a stressed syllable (references welcome)

German {-ei, -erei} (V->N), -erei if the stem ends with a stressed final
syllable, -ei if stem ends with an unstressed final syllable (Hargus 1993, 
Hall 1990)

Lappish {-ide, -ida} (illative plural), -ide when stem has an even number
of syllables, -ida when stem has odd number of syllables (Hargus 1993,
Bergsland 1976)

Latin {-ia, -ie:s} (abstract nouns in 1st and 5th declension), -ie:s is
blocked if before a heavy syllable (i.e., if it leads to 'prosody
trapping', see Mester 1994)

Polish {-s, -ejs} (comparatives), -ejs is chosen if the underived stem
ends in an extrasyllabic consonant, -s otherwise


References: 

Anttila, Arto. 1995. Deriving variation from grammar: a study of Finnish
genitives. Ms., Standard University. ROA-63

Anshen, Frank, Mark Aronoff, Roy Byrd, Judith Klavans. 1986. The role of
etymology and word-length in English word formation. Ms., SUNY Stony
Brook/IBM Watson Research Center.

Aronoff, Mark. 1976. Word formation in generative grammar. Cambridge: The
MIT Press.

Booij, Geert & Rochelle Lieber. 1993. On the simultaneity of morphological
and prosodic structure. In Hargus & Kaisse (eds.), pp. 23-44.

Brown, A.F. 1958. The derivation of English adjectives ending
-ful. Doctoral dissertation, University of Pennsylvania.

Carstairs-McCarthy, Andrew. 1988. Some implications of phonologically
conditioned suppletion. In Yearbook of Morphology.

Carstairs-McCarthy, Andrew. 1990. Phonologically conditioned suppletion. In 
Dressler et al (eds), Contemporary Morphology, Mouton de Gruyter.

Chapin, Paul. 1970. On affixation in English. In Manfred Bierwisch & Karl
E. Heidolph (eds.), pp. 51-63, Progress in linguistics. A collection of
papers. The Hague: Mouton.

Fabb, Nigel. 1988. English suffixation is constrained only by selectional
restrictions. Natural Language and Linguistic Theory 6: 527-539.

Hall, T. 1990. Syllable structure and syllable-related processes in
German. Doctoral dissertation, University of Washington.

Hargus, Sharon. 1993. Modeling the phonology-morphology interface. In
Hargus & Kaisse (eds.), pp. 45-74.

Hargus, Sharon & Ellen Kaisse. 1993. Phonetics and Phonology 4. Studies in
Lexical Phonology. San Diego: Academic Press. 

Inkelas, Sharon. 1989. Prosodic constituency in the lexicon. Doctoral
dissertation, Stanford University.

Ito, Junko & Jorge Hankamer. 1989. Notes on monosyllabism in Turkish. In
Junko Ito & Jeff Runner (eds.), Phonology at Santa Cruz. Santa
Cruz: Linguistics Research Center, pp. 61-69.

Kager, Rene. 1996. On affix allomorphy and syllable counting. In Ursula
Kleinhenz (ed.), pp. 155-171, Studia Grammatica 41, Interfaces in
Phonology. Berlin: Akademie Verlag. ROA-88.

Kager, Rene. 1996. Stem stress and peak correspondence in Dutch. Ms.,
Utrecht University.

McCarthy, John & Alan Prince. 1993. Generalized alignment. Yearbook of
Morphology 1993, pp. 79-153. ROA-7

Mester, Armin. 1994. The quantitative trochee in Latin. NLLT 12: 1-61.

Orgun, Cemil Orhan & Ronald Sprouse. 1999. From MParse to
Control: deriving ungrammaticality. Phonology 16: 191-224. ROA-224.

Prince, Alan & Paul Smolensky. 1993. Optimality Theory: constraint
interaction in generative grammar. RuCCS-TR-2, Rutgers Center for
Cognitive Science.

Raffelsiefen, Renate. 1996. Gaps in word formation. In Ursula Kleinhenz
(ed.), Interfaces in phonology. Berlin: Akademie Verlag, pp. 194-209.

Ross, John. 1972. A reanalysis of English word stress. In Michael Brame
(ed.), pp. 227-323, Contributions in generative phonology. Austin:
University of Texas Press.

Siegel, Dorothy. 1974. Topics in English morphology. Doctoral
dissertation, MIT.

Trommelen, Mieke, & Wim Zonneveld. 1989. Klemtoon en Metrische
Fonologie. Muiderberg: Coutinho.

=========================================================

2] Amanda Seidl, Mende example

Date: Wed, 5 Dec 2001 15:25:46 -0500 (EST)
From: Amanda Seidl <seidlvonneumann.cog.jhu.edu>
To: John Alderete <aldereteruccs.rutgers.edu>
Subject: Re: query: subcategorization in phonology and morphology


Hi John,
I'm wondering whether this qualifies:
In Mende (<SW Mande), it seems like a case-marker subcategorizes/triggers
the insertion of a phonologically mutated allomorph. If this sounds like
it qualifies I can send you my analysis.

The stuff you want is in a teeny chunk of my thesis. If you need more
details I have another paper. Let me know.

Best,
Amanda

### N.B.: contact author with info below for paper referred to above. ###

********************************************
Amanda Seidl
Department of Cognitive Science
Johns Hopkins University 
137D Krieger Hall
Baltimore, MD 21218

dept phone: (410) 516-4945

********************************************* 

=========================================================

3] Lee Bickmore, Tahitian example

Date: Wed, 5 Dec 2001 15:40:01 -0500
To: John Alderete <aldereteruccs.rutgers.edu>
From: Lee Bickmore <l.bickmorealbany.edu>
Subject: Re: query: subcategorization in phonology and morphology


Hi John. Long time no see. I hope all is going well. With regard to 
your request, one thing came to mind, though I don't think it's 
really what you're looking for--it certainly doesn't have to do with 
prosodic conditioning. In Tahitian, the causative prefix is fa'a- 
unless the root begins with a labial consonant, in which case it's 
ha'a-. The analysis of this, of course, is probably very 
straightforward: some constraint on multiple sequential labials with 
the default C being placeless. Of course, I'm not entirely sure that 
there's lots of independent evidence that h is the default. Seems 
like arguments could be made that perhaps ' (glottal stop), also 
placeless, is.

...

Best,

lee

******************************************************************
Lee S. Bickmore Snail Mail:
Associate Professor Dept. of Anthropology, SS 263
Dept of Anthropology University at Albany
Program in Linguistics & Cog. Sci. Albany, NY 12222
Associate Dean of Undergraduate Studies
Ph: (518) 442-4160
Fax: (518) 442-5710
Web pages: http://www.albany.edu/~lb527 (personal)
 http://www.albany.edu/~lb527/LOZ.html (Languages of Zambia)


=========================================================

4] Carson Schutze, English example: 'person from X /-er/'

Date: Wed, 05 Dec 2001 13:18:05 -0800
Subject: RE: query: subcategorization in phonology and morphology
From: Carson Schutze <cschutzeucla.edu>
To: <aldereteruccs.rutgers.edu>


Hi John,

There's an alternation in English for the N->A suffix meaning "person from",
as in

a New York-er
a Tex-an (or Texa-n?)
a Toronto-nian

There's a problem in the O'Grady textbook that mentions some of this data,
showing that vowel-final place names don't take -er (*Philadelphia-er), and
one datum suggesting that r-final ones don't either (*Denverer) [they don't
talk about allomorphs other than -er]. I actually don't know what the full
generalization is, i.e. what's wrong with *Torontan, *Torontoan. I keep
meaning to sit down with an atlas and check a more systematic collection of
cases. But I don't know if some of this is just conventionalized; maybe
actually it would be better to do a wug-test on oneself using made-up names.
Stress pattern probably has something to do with it I would think.

Please let me know if you find a source on this.

Best,
 Carson 

Carson T. Schutze Department of Linguistics, UCLA
Email: cschutzeucla.edu Box 951543, Los Angeles CA 90095-1543 U.S.A.

Office: Campbell Hall 2224B Deliveries/Courier: 3125 Campbell Hall
Campus Mail Code: 154302 Web: www.linguistics.ucla.edu/people/cschutze
Phone: (310)995-9887 Fax: (310)206-8595
 
=========================================================

5] Mike Maxwell, Examples from Tzeltal (Mayan), Shuar (Jivaroan, Ecuador),
Axininca Campa (Peru), and Cubeo (Tucanoan, Colombia)

From: "Mike Maxwell" <maxwellldc.upenn.edu>
To: <aldereteruccs.rutgers.edu>
Subject: subcat
Date: Wed, 5 Dec 2001 16:18:40 -0500


I'm a little confused about what you're looking for--

>...Or it may be the basis of the selection
>of a particular allomorph when two or more
>morphemes compete for a particular collocation...

Do you mean when two or more _allomorphs_ compete? But doesn't all
allomorphy work this way (assuming you can't just generate the surface
allomorphs from a common underlying form via phonological rules)?

At any rate, here are some examples from Tzeltal (Mayan):

1st person Ergative (on verbs) or Possessor (on nouns) is the prefix k-
before vowel-initial stems, and h- before C-initial stems

2nd Erg/Poss is aw- before V-initial stems, a- before C-initial stems

3rd Erg/Poss is y- before V-initial stems, s- before C-initial stems

So far as I know, there is no general phonological process in Tzeltal that
would generate these allomorphs from common underlying forms.

I believe many languages have similar affix allomorphs whose distribution is
governed by C- vs. V-initial (or -final) stems.

There is rampant allomorphy in Shuar (Jivaroan, Ecuador) suffixes
depending on whether the stem to which they attach is V- or C-final. As I
recall, the generalization seems to be that the language likes closed
syllables (except that the first syllable of the word has to be open). So
there are a lot of suffixes having the general form -CCV after a vowel,
and -CVC after a consonant. The morphology and phonology of Shuar is
described in a Ph.D. thesis by Glen Turner done around 1957 at (I believe)
the University of Indiana. (The title is s.t. like "Jivaro Grammar.") He
analyzes most of the allomorphy as metathesis, but you might argue
otherwise, particularly since the vowel in the two allomorphs is not always
the same. I'm sure his dissertation is hard to find now. Unfortunately my
copy is in storage pending completion of our move. A Spanish translation
(with minor editing, as I understand) was published a few years ago by SIL,
but I don't see it listed on their website. The Shuar Federation
(Federacion Shuar) in Ecuador published some work on their language as well.
But the last I saw (twenty years ago) it was not as linguistically analyzed
as Glen Turner's dissertation, so you'd be harder pressed to extract the
desired examples. (The language is fairly agglutinative.) And the last time
I tried, I could find no web presence for the Shuar Federation.

Axininca Campa (Peru) is another example of a language with rampant
allomorphy. David Payne's grammar ("The Phonology and Morphology of
Axininca Campa", published by SIL) accounts for most of it with phonological
rules, including epenthesis of both /t/ and /a/. How synchronically
productive those rules are, is another question. Axininca Campa has been
cited quite a bit in the phonology literature, particularly (I think) the OT
literature.

Finally, there is a causative suffix in Cubeo (Tucanoan, Colombia; I'm
taking the data from the "Cubeo Grammar" which I co-authored with Nancy
Morse, and which is published by SIL). There are four
allomorphs, -wA, -A, -O, and -OwA (the upper case vowels indicate that nasal
spreading affects these vowels). Which allomorph a given verb takes is
lexicalized (unpredictable), except that there are a couple generalizations,
namely: If the root is monosyllabic and ends in a back round oral vowel (u
or o), it takes -A. And most monosyllabic nasalized verbs take -O. But
those aren't very impressive generalizations...

 Mike Maxwell
 Linguistic Data Consortium
 maxwellldc.upenn.edu

=========================================================

6] Wayles Browne, Example of Slavic suffix -ba

Date: Wed, 5 Dec 2001 16:17:50 -0400
To: aldereteruccs.rutgers.edu
From: Wayles Browne <ewb2cornell.edu>
Subject: Slavic suffix -ba


On your Linguist query:
There is a suffix -ba in Slavic languages which forms nouns from
various stems, mostly verbs. In most of the languages (Czech is
an exception to some extent) it can't be added after a labial
consonant. Examples from Macedonian:
tvor-i 'to create' -> tvorba 'creation'
but not sram-i 'to shame' -> *sramba 'shame'
(there is another suffix, -ez^, which can be used instead: sramez^ 'shame').

This is not a constraint on the surface form, because one CAN
add -ba to a stem ending in a dental, even though the dental
(if it is n) will assimilate to the b:
e.g. Croatian
hin-iti 'to feign' -> hinba -> himba 'imposture' (old-fashioned).

This is not simply a matter of adjacent segments, either. The
version of -ba added to adjective stems is -oba. In Russian e.g.
xud-oj 'bad, thin' can form xudoba 'thinness', but rjab-oj
'speckled' cannot add -oba to form *rjaboba 'speckledness'.

This and several more examples (also one from English:
heavy -> heavily, but not friendly -> *friendlily)
are in my paper in Indiana Slavic Studies vol. 10 (1999);
shall I send you a copy?

### N.B. Please contact author with information below for above paper. ###

Yours,
- 

Wayles Browne, Assoc. Prof. of Linguistics
Department of Linguistics
Morrill Hall 220, Cornell University
Ithaca, New York 14853, U.S.A.

tel. 607-255-0712 (o), 607-273-3009 (h)
fax 607-255-2044 (write FOR W. BROWNE)
e-mail ewb2cornell.edu

=========================================================

7] Jose Elias Ulloa, Example from Shipibo and Capanahua (Panoan, Peru)

From: "Jose Elias Ulloa" <joseliaseden.rutgers.edu>
To: "John Alderete" <aldereteruccs.rutgers.edu>
References: <Pine.GSO.4.21.0112051413400.6235-100000ruccs.rutgers.edu>
Subject: Message from Beto
Date: Wed, 5 Dec 2001 16:46:32 -0500


Hi John,


... in Shipibo (a Panoan language in Peru), the ergative suffix
has two basic forms /-aN/ and /-niN/. /-aN/ is added to nouns with an even
number of syllables and /-nin/ to an odd number of syllables.

witax 'leg'
witax-an 'leg+ergative'
atapa 'hen'
atapa-nin 'hen+ergative'

(x stands for a coronal, continuant, voiced, retroflex consonant)

It's pretty similar to the illative plural suffix in Lappish (I think in
Estonian there's a similar case too. Sorry I don't remember the facts but I
think there's an article by Kager where he tries to account for it).

In Panoan languages, it's also common to find cases of vowel harmony
depending on the number of syllables. In Shipibo, again there's an
alternation between /-ribi/ and /-riba/. When the second syllable of the
ergative suffix is in an even syllable, the suffix is /-ribi/, when the
second syllable of the suffix is in an odd syllable the suffix will appear
as /-riba/.

In Capanahua, another Panoan language, something very similar occurs, but
the difference is that there's an alterning glotal stop. I'm sending you an
attachment with a paper written by Carolina Gonzales. This is only a draft,
she hasn't finished the paper yet. There you can find data about Capanahua
and reference to other Panoan languages that show alternations based on
number of syllables.

### N.B. contact author with information below for paper referred to above. ###
- ---------------------------------------
Jose A. Elias (Beto)
Department of Linguistics - Rutgers University
39963 RPO Way, New Brunswick, NJ 08901, USA.
Phone: (732) 565-9250

=========================================================

8] Gail Coelho, Example from Thompson River Salish

Date: Wed, 05 Dec 2001 16:27:53 -0600
From: Coelho <gailutxvms.cc.utexas.edu>
Subject: Re: query: subcategorization in phonology and morphology


Hi John,

One example that comes to mind are the inchoative markers in Thompson River 
Salish (and presumably several other Salishan languages). There are two 
inchoative allomorphs, an infix /-?-/ and a suffix /-*p/, where * is a 
schwa. The former attaches to strong roots and latter to weak roots. There 
are two views on what distinguishes strong roots from weak roots: (1) 
Strong roots have underlying accent, weak roots dont. (2) Strong roots 
have an underlying full vowel, weak roots have underlying schwa or no 
vowel, and you've to use a complicated set of environments to derive 
existing surface vowels from underlying schwa. I take the former view. 
Either way, the inchoative sub-categorizes for a phonological property, 
doesn't it? Incidentally, inchoative infixation turns strong roots into 
weak roots -- I think the reasons are phonological/phonetic, but I'm 
working on that.

Gail

=========================================================

8.1] Suzanne Urbanczyk, Reference for Salish example

Date: Wed, 5 Dec 2001 15:18:11 -0800
To: aldereteruccs.rutgers.edu
From: Suzanne Urbanczyk <urbansuuvic.ca>
Subject: yep


Hi John,

That message from the OT list is right on the money (refers to [8] message
immediately above). I know that it is the same in St'at'imcets as well. 
The grammar by van Eijk (1997) The Lillooet Language: Phonology,
Morphology, Syntax. Vancouver, UBC Press, has the details in it.

Ciao for now,

Su

=========================================================

9] Brett Baker, Examples from Ngalakgan and other Australian languages

Date: Thu, 6 Dec 2001 09:59:37 +1100
To: aldereteruccs.rutgers.edu
From: "Brett J. Baker" <brettbsultry.arts.usyd.edu.au>
Subject: allomorphy


John,

My thesis has a chapter on morpheme selection in Ngalakgan according to
phonological properties of the stem. You get geminate-initial allomorphs if
the stem contains no geminates or clusters of obstruents,
singleton-initial allomorphs otherwise. The facts are published in an MIT
volume 'Working Papers in Endangered and less familiar languages' (if you
can believe it) volume 2, Rob Pensalfini and Norvin Richards eds. 2000. The
same pattern applies to two other (neighbouring) languages I know of:
Rembarrnga and Ngandi. The conditioning factors (prosodic structure) are
the same in Ngandi at least, I haven't checked out Rembarrnga. Warrumungu
has some similar things going on (Simpson in Goldsmith Handbook of
Phonological theory, and look at Evans' paper on phonology in Australian
languages for other stuff).

Other Australian languages have similar allomorphy patterns based on
stem-internal nasal-stop clusters (this pattern is much more common).
Caroline wrote a generals paper on this pattern. There's a published
account in Aboriginal Linguistics 1 (1988) by Patrick McConvell.

But I noticed you didn't mention the most common, and most famous,
Australian allomorphy pattern discovered by Ken Hale: the pattern whereby
ergative and locative suffixes take one of two allomorphs
-lV or -nggV depending on the number of syllables in the stem. This pattern
applies to lots of languages, especially in the centre, with some -lV or
-nggV depending on the number of syllables in the stem. This
pattern applies to lots of languages, especially in the centre, with some
variations on the prosodic conditioning. There's an account in Dixon's 1980
'Languages of Australia' (much as I hate to give him a plug - don't ask).
There's probably more recent stuff floating around.

Perhaps the most interesting patterns (apart from the geminate and nasal
cluster alternation ones) are the ones affecting apical-initial suffixes in
some Central and western languages. In this one (which is strictly
local), you get alternations in retroflexion, depending on whether the
preceding consonant in the stem is apical or not. In most languages, the
suffix is retroflex if the preceding consonant is apical (alveolar or
retroflex), and alveolar otherwise. E.g. (from Alan Dench: Martuthunira.
Pacific Linguistics, 1995).

kampa-layi
cook-FUT

wangka-layi
speak-FUT

ngaya-layi
cry-FUT

But:

ngarrarni-rlayi
get.stuck-FUT

thanta-rlayi
crawl-FUT

Kaytetye (H. Koch in Papers in Australian Linguistics No. 13. 1980. Eds
Bruce Rigsby and Peter Sutton) is similar, and Arrernte (J. Henderson, UWA
PhD thesis 1998).

But according to Rob Pensalfini (MIT PhD thesis 1997), Jingulu has almost
the *opposite* pattern: you get alveolars after apicals, retroflexes
otherwise. Mark Harvey and I are working on a paper discussing coronal
place, and we think we have an explanation for this. I'll send it to you
when we're done. Cheers,

Brett.

Brett Baker
Postdoctoral Fellow,
Dept of Linguistics F12
University of Sydney 2006
Ph: 02 9351 8763
Fax: 02 9351 7572
brettbsultry.arts.usyd.edu.au
http://www.sultry.arts.usyd.edu.au/brettb/index.htm

=========================================================

10] Donca Steriade, Examples from Latin and Greek declensions

Date: Wed, 5 Dec 2001 18:17:31 -0500
To: John Alderete <aldereteruccs.rutgers.edu>
From: Donca Steriade <steriadeMIT.EDU>
Subject: Re: query: subcategorization in phonology and morphology (fwd)


Dear John,

I have kept your morphological subcat handout on a visible end of my 
desk because it's intriguing. So far the cases that I know of involve 
exclusively Latin and Greek style declensional classes. The 
declension classes are all phonologically predictable: the a-final 
stems belong to one declension class, the o-final stems to another, 
the C-stems to the third etc. This relates to your story in that one 
can treat declensional classes by saying that certain sets of case 
endings (each set defining a particular declensional class) 
subcategorize for the final segment of the stem. E.g. the Latin the 
set defining the 1st declension {-s, -i, -m, -V (length), -i, -rum, 
-Vs...} subcategorizes for a-final stems. Not sure this is right but 
it may be worth playing with and it seems related.

...

Best regards,

Donca

=========================================================

11] Armin Mester, Examples from German (diminutive -chen / lein and A->N
suffix heit/keit) plus references

Date: Wed, 5 Dec 2001 17:05:49 -0800 (PST)
From: Armin Mester <mesterling.ucsc.edu>
To: John Alderete <aldereteruccs.rutgers.edu>
Subject: Re: query: subcategorization in phonology and morphology (fwd)


Hi John,
You already have collected a very impressive list! I'll keep thinking of 
other examples.
The best reference for German ge- is probably Kiparsky 1966 "Ueber den 
deutschen Akzent" (or words to that effect). To my knowledge he was the 
first to clearly state that the generalization was 
prosodic (after MAIN stress: trompeten - hat trom(2)pe(1)tet, 
*getrom(2)pe(1)tet 'play the trumpet'), not morphological.

A segmental case in German is the OCP-controlled behavior of diminutive 
-chen / lein: Among other things, -chen is avoided after ch-stems (Bach 
Baechlein, *Baechchen 'creek'), and lein (mostly) after l-stems (Saal 
Saelchen, *Saellein 'hall'). But the story is more complicated, with 
doublets such as Engelchen Engellein 'angel'. 

The A->N suffix heit/keit shows a similar segmental restrictions: 
Krank-heit *Krank-keit 'illness', Traeg-heit *Traeg-keit 'lethagy', 
Frech-heit *Frech-keit 'insolence (the latter case is interesting wrt. 
the question wheher the conditions are checked in the output, with 
allophones spelled out). The is also a variant -igheit which shows stress 
conditioning.

Fleischer's morphology book ("Wortbildung"), especially in the second 
edition with a second author, is full of such observations (it's written 
in the German tongue, however).

Greetings,
Armin

- ---------------------------------------------------------------------------
Armin Mester mesterling.ucsc.edu
Professor of Linguistics office: 408 459-3426
Linguistics Department msgs: 408 459-2555
Stevenson College dept: 408 459-2905
University of California, Santa Cruz, CA 95064 fax: 408 459-3334
- ---------------------------------------------------------------------------

=========================================================

12] John Koontz, Example from Winnebago and discussion

Date: Wed, 5 Dec 2001 16:35:47 -0700 (MST)
From: Koontz John E <John.KoontzColorado.EDU>
To: John Alderete <aldereteruccs.rutgers.edu>
Subject: Re: 12.3046, Qs: Ling Library Funds, Morpheme Subcategorization


I take it that a valid example would include cases where fairly trivial
allomorphs allow an ending to be added to set of stems determined by
non-phonological criteria, with only the allomorphs being phonologically
conditioned. For example:

> Lappish {-ide, -ida} (illative plural), -ide when stem has an even number
> of syllables, -ida when stem has odd number of syllables (Hargus 1993,
> Bergsland 1976)

Isn't that pretty much everything? I had thought at first you were
looking at a much more restrictive situation in which some bases logically
possible on non-phonolopgical grounds were excluded from taking
morphological category on phonological grounds.

An example of the more general situation would be the Winnebago
declarative, which is s^anaN after verbs ending in consonants and naN
after verbs ending in vowels. An example of the more restricted kind
doesn't immediately occur to me.

John Koontz

=========================================================

13] Lisa Davidson, Example of Portuguese and Spanish diminutives

Date: Wed, 05 Dec 2001 21:44:58 -0500
From: Lisa B Davidson <lbd1jhu.edu>
Subject: Re: query: subcategorization in phonology and morphology
In-reply-to: <Pine.GSO.4.21.0112051413400.6235-100000ruccs.rutgers.edu>
To: John Alderete <aldereteruccs.rutgers.edu>


Hi John--

If I understood your question correctly, then I think the Portuguese
diminutive is another example of this. The ending -inho/a is used when
the noun ends in an unstressed vowel (casa --> casinha) and the ending
-zinho/a is used when the noun ends in an accented vowel, a nasal vowel,
or a diphthong (pai --> paizinho, homem --> homenzinho, irma~ -->
irma~zinha [where the final m represents nasalized /e/ and the tilda
represents a nasalized /a/])

Spanish is also similar-- -ito follows stressless vowels (gato --> gatito,
planta --> plantita), and -cito follows stressed vowels and /n/ and /r/
(cafe' --> cafecito, Ramon --> Ramoncito) but my Spanish informant is not
so clear on this distinction (for example, /l/ takes -ito, and mama, at
least in Spain, is mamita or mamaita, not mamacita). If you're interested
in this one you should check on it (my Spanish informant, who is from
Spain but whose mother is Mexican, says it's different in different
countries.)

I'm sure someone has told you this by now, but if not, hope it helps!

Lisa

=========================================================

14] Yoonjung Kang, Example from Korean

From: "Yoonjung Kang" <yoonjungkworldnet.att.net>
To: <aldereteruccs.rutgers.edu>
References: <Pine.GSO.4.21.0112051413400.6235-100000ruccs.rutgers.edu>
Subject: Re: query: subcategorization in phonology and morphology
Date: Wed, 5 Dec 2001 23:04:13 -0500


Dear John,

...

I thought the following paper may be of interest to you.

Cho, Young-Mee Y. 1991. "A phonological constraint on the attachment of
particles in Korean" Harvard studies in Korean linguistics 4.

The basic fact is that certain particles/affixes can attach to a verbal stem
only when the stem is at least disyllabic.
If your library does not have the volume, I am sure that Professor Cho in
the East Asian department of Rutgers can give you a copy.

Best,
Yoonjung
yoonjungkatt.net

=========================================================

15] Arto Antilla, References
 
From: Arto Tapani Anttila <ellaatnus.edu.sg>
To: "'aldereteruccs.rutgers.edu'" <aldereteruccs.rutgers.edu>
Cc: Arto Tapani Anttila <ellaatnus.edu.sg>
Subject: subcategorization
Date: Thu, 6 Dec 2001 14:04:20 +0800 


Hi John,

Thanks for the biblio and the examples you sent out. Here are some
additional references that come to mind.

A general discussion of the typology of phonological/morphological
subcategorization can
be found in Kiparsky 1994. Finno-Ugric has quite a bit of prosodically
conditioned allomorphy. Estonian 
is discussed in Kager 1996 (the syllable counting paper you refer to) and I
believe also in Kager 1997 
in Rivista (see below). Kager's analyses are mostly based on Mati Hint's
very interesting original work, 
unfortunately mostly in Estonian. Dolbey 1996 is a cyclic OT analysis of
Saami allomorphy. I've heard 
Nganasan mentioned as the best example of syllable-counting allomorphy.
Helimski's paper (see below) 
is a good starting point, but I seem to remember the analysis is quite
abstract, so you won't see neat
lists of allomorphs. I also have some discussion of Finnish in a paper to
appear soon.

Best,

Arto

INCOLLECTION{kiparsky94a, 
AUTHOR = {Kiparsky, Paul}, 
TITLE = {Allomorphy or morphophonology}, 
BOOKTITLE = {Trubetzkoy's Orphan}, 
NOTE = {Proceedings of the Montr\'{e}al Roundtable on "Morphonology:
contemporary responses", 
 Current issues in Linguistic Theory 144}, 
PUBLISHER = {John Benjamins}, 
YEAR = 1994, 
PAGES = {12-31}, 
EDITOR = {Singh, Rajendra}, 
ADDRESS = {Amsterdam}}

ARTICLE{kager97, 
AUTHOR = {Kager, Ren\'{e}}, 
TITLE = {Generalized Alignment and morphological parsing}, 
JOURNAL = {Rivista di Linguistica}, 
VOLUME = {8}, NUMBER = {1}, 
YEAR = {1997}}

INCOLLECTION{dolbey96, AUTHOR = {Dolbey, Andrew}, TITLE = {Output
optimization and cyclic allomorph selection},
BOOKTITLE = {The Proceedings of the Fifteenth West Coast Conference on
Formal Linguistics (WCCFL 15)}, 
EDITOR = {Agbayani, Brian and Sze-Wing Tang}, PUBLISHER = {Stanford
Linguistics Association/CSLI}, 
PAGES = {97-112}, YEAR = {1996}} 

ARTICLE{anttila00a, 
AUTHOR = {Anttila, Arto}, 
TITLE = {Morphologically conditioned phonological alternations}, 
JOURNAL = {Natural Language and Linguistic Theory}, 
VOLUME = {}, 
YEAR = {in press}, 
PAGES = {}, 
NOTE = {Also {\tt http://roa.rutgers.edu/, ROA-425-10100}}}

INCOLLECTION{helimski98, 
AUTHOR = {Helimski, Eugen}, 
TITLE = {Nganasan}, 
BOOKTITLE = {The Uralic Languages}, 
SERIES = {Routledge Language Family Descriptions}, 
PUBLISHER = {Routledge}, 
ADDRESS = {London and New York}, 
EDITOR = {Abondolo, Daniel}, 
PAGES = {480-515}, 
YEAR = {1998}}

=========================================================

16] Birgit Alber, Example from German

Date: Thu, 6 Dec 2001 06:24:31 +0100
To: aldereteruccs.rutgers.edu
From: Birgit Alber <birgit.alberlett.unitn.it>
Subject: subcategorization etc.


Dear John,

very shortly because quite overworked (me):

Wiese (1996), The Phonology of German, CUP: here you find 
stress-dependent heit/keit allomorphy and possibly also discussion of 
ge-

Plag (1999), Morphological Productivity, Mouton de Gruyter: a study 
in derivational morphology on extensive corpus; here you find several 
cases of blocking etc. in English

Best

Birgit

=========================================================

17] Todd Bailey, Examples from Spanish, Latin and Polish

From: "Todd Bailey" <baileytm1Cardiff.ac.uk>
To: John Alderete <aldereteruccs.rutgers.edu>
Date: Thu, 6 Dec 2001 09:37:12 GMT0BST


I don't know if this falls in the realm of what you're looking for, 
but I'm aware of unstressed suffixes in Spanish, Latin and maybe 
Polish which are accompanied by shifts of stress in the stem to which 
they attach if the stem has antepenultimate stress (e.g. Spanish 
re'gimen (antepenultimate stress) -> regi'menes, and also the 
different pattern cha'racter -> characte'res). One might conceivably 
analyze these in terms of suffixes subcategorising for stress on one 
of the last two syllables of the stem, though this is not the usual 
analysis. I will send more details and references if this is of 
interest to you.

Best,

todd


### second, follow up message with references: 

From: "Todd Bailey" <baileytm1Cardiff.ac.uk>
To: John Alderete <aldereteruccs.rutgers.edu>
Date: Fri, 7 Dec 2001 11:10:38 GMT0BST


I don't know if I have any non-canonical references, but I'll give 
you what I've got, picking them out of Ch. 3 of my thesis on primary 
stress (in case it's of interest you can find that chapter by chapter 
at http://www-cogsci.psych.ox.ac.uk/~todd/PAPERS/thesis.html, among 
other places). Actually, that's probably as good a place as any to 
find a summary of what I know about the "stress-shift" phenomena in 
these languages, so I won't repeat the facts here.

For Spanish nouns, three patterns of shift are illustrated in Figure 
2d, p. 69 of my thesis (also see Figure 10, p. 82).

Some references on Spanish stress shift in nouns:
Harris, J. W. (1983). Syllable structure and stress in Spanish. A 
nonlinear analysis. Cambridge, MA: MIT Press.

Roca, I. (1988). Theoretical implications of Spanish word stress. 
Linguistic Inquiry, 19, 393-423.

Roca, I. (1990). Diachrony and synchrony in word stress. Journal of 
Linguistics, 26, 133-164.

Den Os, E., & Kager, R. (1986). Extrametricality and stress in 
Spanish and Italian. Lingua, 69, 23-48.

Otero, C. P. (1986). A unified metrical account of Spanish stress. In 
M. Brame & H. Contreras & F. J. Newmeyer (Eds.), A Festschrift for 
Sol Saporta (pp. 299-332). Seattle, Washington: Noit Amrofer.

Stress also moves around within Spanish verbs depending on the 
following inflectional endings (Figure 14, p. 98). The main 
reference for this is Harris (1987).

Harris, J. W. (1987). The accentual patterns of verb paradigms in 
Spanish. Natural Language and Linguistic Theory, 5, 61-90.

Polish nouns with (exceptional/marked) antepenultimate stress also 
undergo a stress shift when followed by inflectional suffixes, as 
illustrated in Figure 22, p. 112. 

Other refs on Polish stress shift:
Halle, M., & Vergnaud, J.-R. (1987). An Essay on Stress. Cambridge, 
MA: MIT Press.

Franks, S. (1991). Diacritic extrametricality vs. diacritic accent: A 
reply to Hammond. Phonology, 8, 145-161.

Idsardi, W. J. (1992). The computation of prosody. Ph.D. thesis, MIT. 

Enclitic stress in Latin also exhibits interesting interactions 
between stress and morphology. Basically, they force stress onto the 
final syllable of the stem to which they attach (i.e. they 
"subcategorize" for stems with final stress). See Figure 32a, p. 
126.

References on Latin enclitic stress include:
Halle, M., & Kenstowicz, M. (1991). The free element condition and 
cyclic versus noncyclic stress. Linguistic Inquiry, 22, 457-501.

Mester, R. A. (1994). The quantitative trochee in Latin. Natural 
Language and Linguistic Theory, 12, 1-61.

Steriade, D. (1988). Greek accent: A case for preserving structure. 
LI, 19(2), 271-314.

I hope some of this is helpful.

Best,

todd
Todd M. Bailey, Ph.D.
Lecturer of Psychology, Cardiff University
Email: BaileyTM1cardiff.ac.uk
Smail: School of Psychology, Cardiff University
 PO Box 901, Cardiff CF10 3YG, United Kingdom
Ph: +44 29 2087 5375
Fax: +44 29 2087 4858

=========================================================

18] Moira Yip, Examples from English and Cantonese

Date: Thu, 06 Dec 2001 14:19:53 +0000
To: John Alderete <aldereteruccs.rutgers.edu>
From: Moira Yip <moiralinguistics.ucl.ac.uk>
Subject: Re: query: subcategorization in phonology and morphology


The most obvious cases that come to mind, and erring on the side of
over-inclusion, are these:

English: -ion/-tion/-ition, and same for ive/tive./itive: The -tion
allomoproph only attaches to non-coronals: decieve/decep-tion, and the
others are chosen for coronals: add-ition, or divis-ion.

Englsih: -er, only on monosyllabic adjectives, or disyllables ending in -y.

Cantonese: Vocative prefix a-, used only on monosyllables to give a
bisyllabic output form. Bi-syllables do not take it (although I found a
native speaker last week who disagrees, and says he can add it to anything).

Moira

- -----------------------

Dr. Moira Yip
Dept of Phonetics and Linguistics
University College London
Gower Street
London WC1E 6BT
ENGLAND

Tel: +44-20-7679-3158
Fax: +44-20-7383-4108

=========================================================

19] Jeremy Whistle, Discussion of English /-ful/

Date: Thu, 06 Dec 2001 14:19:56 +0000
Subject: query: subcategorization in phonology and morphology
To: aldereteruccs.rutgers.edu


Hi,

I don't want to appear nit-picking but is the following statement true?

"The basic type of pattern I'm looking for is one in which a 
morpheme 'subcategorizes' for a phonological property of a 
neighboring morpheme, i.e., it only occurs when a subcategorized 
phonological property is present. For example, it is sometimes 
said that the adjective-forming suffix /-ful/ in English subcategorizes 
for a stressed syllable at the end of the stem to which it attaches, 
e.g., forgetful, cf. *forgettingful (Siegel 1973; though this example is 
more complicated, see Brown 1958 and Chapin 1970)."

I agree that /-ful/ is more often than not attached to a stressed 
syllable but how do you account for:

beautiful
colo(u)rful
meaningful
plentiful
powerful
wonderful?

Are the restrictions grammatical/semantic (ie only nouns) rather 
than phonological? Or have I misunderstood something?

Best wishes,

### N.b.: see Siegel 1974 for discussion of these 'exceptions' ###

***********************************************************
Jeremy Whistle, School of Business and Management
University College Northampton, Moulton Park, Northampton NN2 7AL, GB
tel: +44 (0)1604 735500 ext: 2182
e-mail: jeremy.whistlenorthampton.ac.uk

=========================================================

20] Andrew Spencer, Reference and discussion of general topic

From: "Spencer, Andrew J" <spenaessex.ac.uk>
To: "'John Alderete'" <aldereteruccs.rutgers.edu>
Subject: RE: query: subcategorization in phonology and morphology
Date: Thu, 6 Dec 2001 15:03:54 -0000 


Of course, whether you get 'subcategorization' depends on your theory of
representations ('underlying forms' or whatever). One could in principle say
that the -z allomorph of the English plural selects/is selected by
sibilant-final stems. Or you could handle the distribution in different
ways.

Carstairs-McCarthy has written about these things in various places. An
overview can be found in his chapter in Spencer and Zwicky 1998 Handbook of
Morphology, 144-148. Maybe you should contact him directly for further
examples.

Best wishes,

Andrew Spencer

=========================================================

21] Larry Trask, Examples from English (/en-) and Basque and discussion of
English /-ful/ and /-ize/

Date: Thu, 06 Dec 2001 15:07:15 +0000
From: Larry Trask <larrytcogs.susx.ac.uk>
To: aldereteruccs.rutgers.edu
Subject: Re: Q: Morpheme Subcategorization


I can suggest one more English case. I think that the English verb-forming 
prefix (or prefixes) <en-> can only be added to a base with initial stress:

 enable
 empower
 enslave
 ennoble
 encourage
 enrich
 enmesh
 entomb
 ensnare
 enfeeble

Can't think of any counterexamples, but I haven't checked a dictionary.

Perhaps I might mention the Basque noun-forming suffix <-tze>. This forms 
nouns from nouns and verbs, but it is subject to the restriction that it 
cannot be added to a stem ending in a sibilant (fricative or affricate). 
Stems in sibilants must take instead the quite different suffix <-te>. The 
two suffixes have wholly different origins. When added to nouns, they also 
have different functions. But, when added to verb-stems, they have the 
same function: deriving verbal nouns and gerunds. Examples of verbal 
nouns:

 <etor-> 'come', <etortze> 'coming'
 <sar-> 'enter', <sar-> 'entering'
 <hel-> 'arrive', <heltze> 'arriving'
 <kanta-> 'sing', <kantatze> 'singing'
 <ken-> 'remove', <kentze> 'removing'

But:

 <ikus-> 'see', <ikuste> 'seeing'
 <idatz-> 'write', <idazte> 'writing'
 <egos-> 'cook', <egoste> 'cooking'
 <ikuzi> 'wash', <ikuzte> 'washing'
 <ebatsi> 'steal', <ebaste> 'stealing'

Basque <z> spells a voiceless lamino-alveolar fricative similar to English
/s/, while <s> spells a voiceless apico-alveolar fricative similar to 
Castilian Spanish /s/ in northern Spain; <tz> and <ts> are the 
corresponding affricates.

By the way, I query the statement that the English adjective-forming suffix 
<-ful> must follow a stressed syllable. There seem to be lots of 
exceptions:

 wonderful
 powerful
 pitiful
 merciful
 beautiful
 plentiful
 meaningful

But perhaps I've missed something.

Anyway, I think *<forgettingful> is excluded because no suitable lexical 
item *<forgetting> exists to serve as a base anyway. Compare 'meaningful', 
above.

> English /-ize/ (N/A->V), only attaches to stems ending in an unstressed
> syllable (Raffelsiefen 1996)

I broadly agree with this. But there's a tiny problem with 'realize'. 
Though 'real' is historically two syllables, and still so for some 
speakers, it is now commonly one syllable in both British and American 
English. But then I suppose most speakers no longer associate 'realize' 
with 'real' anyway.

Moreover, some speakers fail to observe this restriction in their coinages. 
I have often encountered 'basquize' ('make Basque'), which I consider 
horrible, and Laurie Bauer cites 'Vietnamize', which is also bad for me, 
unless I pronounce the second syllable with schwa, which I don't do in 
'Vietnam'.


Larry Trask
COGS
University of Sussex
Brighton BN1 9QH
UK

larrytcogs.susx.ac.uk

Tel: (01273)-678693 (from UK); +44-1273-678693 (from abroad)
Fax: (01273)-671320 (from UK); +44-1273-671320 (from abroad)

=========================================================

22] John Koontz, Examples from Siouan

Date: Thu, 6 Dec 2001 08:47:49 -0700 (MST)
From: Koontz John E <John.KoontzColorado.EDU>
To: John Alderete <aldereteruccs.rutgers.edu>
cc: Koontz John E <John.KoontzColorado.EDU>
Subject: Re: 12.3046, Qs: Ling Library Funds, Morpheme Subcategorization

...

Perhaps you didn't have the example of suppletion between more/most and
-er/-est in English based on length of the adjectival stem.

There are certainly other examples of this sort of thing in Siouan, e.g.,
the allomorphs of the causative in Mandan and also in Hidatsa. I still
haven't thought of any cases where phonological form prevents application
of a category. The genious of the family extends more to pleonastic
application of categories or analogical recision of the canonical or
morphosyntactic form of stems.

Best of luck with your project.

JEK

=========================================================

23] Pius ten Hacken, Explanation of morphological dictionary in Word Manager

Date: Thu, 06 Dec 2001 17:47:26 +0100
From: Pius ten Hacken <pius.tenhackenunibas.ch>
Subject: Re 12.3046: morphophonological subcategorization
X-Sender: hackenigor.urz.unibas.ch


Dear colleague,

Regarding your question in the Linguist-List, it might be interesting 
to know the resources offered by Word Manager (WM). WM is a system 
for morphological dictionaries. A WM-dictionary is an object-oriented 
database with information about inflection and word formation both in 
terms of classes of items and in terms of instructions. From the 
database, dedicated tools can be derived, giving information about 
for instance how many and which word were formed according to a 
particular morphological pattern.

For German, a large WM database has been developed, and a few 
dedicated tools are available at http://www.canoo.net. For English 
and Italian, the database is under development.

As your query refers to phonological information not directly encoded 
in WM, WM resources can be used either to check a hypothesis or to 
look for generalizations by identifying the full class of lexemes 
produced with a particular word formation suffix and checking the 
phonological properties manually. I think, however, that also in such 
a case, the availability of the classes will be helpful.

More information about Word Manager can be found at 
http://www.unibas.ch/LIlab/projects/wordmanager/wordmanager.html

Best wishes,
Pius ten Hacken

=========================================================

24] Anna Thornton, Examples from Italian (several examples)

Date: Fri, 07 Dec 2001 12:15:48 +0100
To: John Alderete <aldereteruccs.rutgers.edu>
From: "Anna M. Thornton" <thorntoncc.univaq.it>
Subject: Re: query: subcategorization in phonology and morphology


John,

I have a couple of examples from Italian.

1. prefix s- with negative semantics attaches to adjectives

e.g. fortunato 'lucky' sfortunato 'unlucky'

BUT: a.: s- doesn't attach to vowel-initial adjectives
 b.: it doesn't attach to adjs whose inital segment cannot be preceded by
s- by normal phonotactic rules of the language (that is, there is no
repairment, like epenthesis or the like --- it just doesn't take those bases).

Examples:
 a. utile 'useful' *sutile 'not useful' (we say inutile)
 b. giusto 'right' *sgiusto 'not right' (we say ingiusto)

Reference: S. Scalise, Generative morphology, Foris, Dordrecht, 1984, pp.
47-48.

2. In my own doctoral dissertation on action nouns in Italian
(unfortunately unpublished and written in Italian), I have found that there
is a general constraint in Italian suffixation, by which suffixes
containing an affricate do not attach to bases whose root ends in the same
affricate:

so you don't have nouns in -dZaddZo (orthographic -ggiaggio), which could
be formed by attaching suffix -aggio to verbs in -giare, -gere, -gire
(which are plenty)

and you dont' find nouns in -ttattsjone, ttsittsjone (orthographic
-zzazione, -zzizione) formed with the suffix -ttsjone (orth. -zione, cfr.
-ation) from verbs in -ttsare (of which there are a few, over 20).
Note that you do have plenty of words in -ddzattsjone (orthographic
-zzazione, like above) from verbs ending in -iddzare (ortho. -izzare, cfr.
-ize) -- so the constraint only applies to identical affricates, a
difference in voicing is enough to allow the attachment.

In the same work, I noted that suffix -mento (another action noun suffix)
does not attach to monosyllabic bases, such as those of the verbs dare
'give', stare 'stay', dire 'say', fare 'do,make', while the almost
synonymic suffix -zione is found with such base:

fare fazione *famento
stare stazione *stamento
dire dizione *dimento
dare dazione *damento

All the -zione derivatives are in fact words inherited from Latin, and have
undergone some degree of lexicalization, so that fazione 'party' and
stazione 'station' (but also 'stazione eretta' = standing on two legs, said
of humans vs. apes) are no longer semantically connected to their base
verb, but dizione "the way one speaks (which can be taught, for instance to
actors)" and dazione "act of giving, only used in legal terminology" are
still connected to the base verb.

The reference to this is
Anna M. Thornton, Sui nomina actionis in italiano, Ph. D. Diss. University
of Pisa, 1988, pp. 356-358 (I have given you more details than are in the
actual text, as I had to gloss and discuss the semantics of dazione, etc.,
so if you need to quote these data you can add personal communication from
me as a source).

3. Easy example: the hypochoristic suffix /i/ (orthographically -i/ -y/
-ie) only attaches to disyllabic trochaic bases. The bases can be
disyllabic trochees in their own right (a), or have become such by a
previous operation of prosodic morphology (b):

Ambra --> Ambry
Roberta --> Robe --> Roby

Plenty of data and discussion in my paper:
Thornton, Anna M., "On some phenomena of prosodic morphology in Italian:
accorciamenti, hypocoristics and prosodic delimitation", Probus, 8,1, 1996,
pp. 81-112. 


I am currently teaching a class in basic Itaian morphology, so I might
think of more examples in the next couple of weeks.
I hope you'll find my examples helpful. I look forward to receiving your
summary on the topic.

Anna Thornton
Associate professor of Linguistics
University of L'Aquila, Italy

=========================================================

25] Katherine Crosswhite, Example of verbal -s'a/-s' in Russian

Date: Thu, 06 Dec 2001 15:13:51 -0500
From: Katherine Crosswhite <crosswhiling.rochester.edu>
Reply-To: crosswhiling.rochester.edu


Dear John,

...

Also, here's an example for your subcategorization query: There is a
verbal morpheme in Russian that has two allomorphs, depending on the
shape of the stem. This is usually called a reflexive marker, although
some have pointed out the it is more accurate to call it an
intransitivity marker since verbs using it can be passive, middle,
reciprocal, or just intransitive. The ending is -s'a if the verb
otherwise ends in a consonant, but is -s' if it otherwise ends in a vowel:

ja ulybajus' I am smiling
on ulybajets'a he is smiling

ona ulybnulas' she gave a smile
on ulybnuls'a he gave a smile

A complicating factor is the behavior of this ending when attached to a
verb that has been made into a participle (=morphologically an
adjective). The ending always shows up as -s'a in participles, even is
the form otherwise ends in a vowel:

ulybajushchijs'a mal'chik the smiling boy
ulybajushchajas'a devochka the smiling girl

Hope that's the type of thing you were looking for!

-K.

=========================================================

26] Ingo Plag, Reference and discussion

From: "Ingo Plag" <plaganglistik.uni-siegen.de>
To: John Alderete <aldereteruccs.rutgers.edu>
Date: Sun, 9 Dec 2001 12:49:57 +0100
Subject: phonoligical restrictions


Dear John, 

you find a very detailed OT-analysis of the intricate prosodic properties 
and restrictions of English verbal suffixes (IZE, ATE, IFY) in my 1999 
book (chapters 6 and 7). Furthermore, you will find more examples of 
phonological restrictions in English suffixation in chapter 4 of that book 
or in my 1996 article in LINGUISTICS.

Here is the full reference of the book:

Plag, Ingo. 1999. Morphological Productivity. Structural Constraints in 
English Derivation. Berlin/New York: Mouton de Gruyter.

One the basis of my analyses I argue that such phonological 
restrictions should not be viewed as akind of subcategorization but 
rather as restrictions on possible output forms (i.e. possible derivatives, 
i.e. the morphological category) and on the relation of these derived 
output forms to their bases (=OO-fauthfulness). In this way certain 
things can be explained that remain unexplained in an input-centered 
subcategorization theory of phonological restrictions. More detailed 
discussion of this can be found in my book. 

Comments and discussion welcome!
Looking forward to your summary.

Best regards,
Ingo Plag

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Prof. Dr. Ingo Plag
English Linguistics
Fachbereich 3
Universitaet-Gesamthochschule Siegen
Adolf-Reichwein-Str. 2
D-57068 Siegen

http://www.uni-siegen.de/~engspra/
tel. 0271-740-2560
tel. 0271-740-2349 (secretary)
fax 0271-740-3246
e-mail: plaganglistik.uni-siegen.de
tel.: 06422-2817 (home)

office: room AR-K 103
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

=========================================================

27] Mark Aronoff, References

From: maronoffnotes.cc.sunysb.edu
Subject: Re: query: subcategorization in phonology and morphology (fwd)
To: John Alderete <aldereteruccs.rutgers.edu>
Date: Mon, 10 Dec 2001 08:54:23 -0500


John,
If you don't know Carstairs-McCarthy's piece on phonologically conditioned
suppletion (Yearbook of Morphology some 10 years ago), it is the best piece
I know on the subject. I am sure that Geert Booij's very recent book Dutch
Morphology should have examples.
Mark

=========================================================

28] Eric Bakovic', Example of diminutives in Spanish

Date: Mon, 10 Dec 2001 11:48:47 -0500 (EST)
From: Rutgers Optimality Archive <roaruccs.rutgers.edu>
Reply-To: bakovicling.ucsd.edu
To: John Alderete <aldereteruccs.rutgers.edu>
Subject: Re: query: subcategorization in phonology and morphology


Hey Aldo, it's Eric. (I'm writing from ROA because I don't subscribe to
Optimal personally, but you can reply to my account.) The diminutive
suffix in Spanish displays allomorphy between -ito/a and -(e)cito/a ('c' =
[s] in some dialects, theta in others) depending on a complex set of
factors, both phonological (prosodic and otherwise) and morphological. I
think Megan Crowhurst wrote a paper about the prosodic conditioning
(appeared in Phonology?) in the early 90's, and of course Harris wrote
some sort of reply to it (maybe also in Phonology; don't remember),
probably berating her for not considering the other conditions. But I
think that one of the original sources is the following excellent paper:

Jaeggli, Osvaldo. 1980. "Spanish Diminutives." In F. H. Nuessel,
 Jr. (ed.), Contemporary Studies in Romance Languages, 142-158. IULC.

Just to give you a taste, my own name ['erik] takes -ito [eri'kito], but
Karen's name ['karen], which is prosodically identical, takes -cita
[karen'sita]. Likewise, my brother's name ['boris] takes -ito
[bori'sito], while my mom's name ['nansi] takes -cita [nansi'sita]. 
(Ignore the fact that these particular facts happen to fall along gender
lines; this is irrelevant.) It seems to be the final segment that is
determining the allomorphy in these cases; consonants like [k] that don't
typically make good codas tend to take -ito/a, while good codas like [s]
or [n] take -cito/a. There's a morphologically-defined split among
vowel-final words; if the vowel is a gender-marking -o/a, then the word
takes -ito/a (e.g., ['marta] > [mar'tita], *[marta'sita]), but as you saw
with my mom's name above, other vowel-final words take -cito/a.

I forget the other details about the allomorphy, but it's all certainly
interesting and well worth looking at (if you haven't already done so).

Good luck! Let me see this work when you have it ready for viewing.

- Eric

=========================================================

29] Andrew Carstairs-McCarthy, References

Date: Sun, 16 Dec 2001 21:04:59 -0500 (EST)
From: Andrew Carstairs-McCarthy <a.c-mccling.canterbury.ac.nz>
Subject: Re: query: subcategorization in phonology and morphology


Dear John

I discussed a few such instances in my article 'Some implications of 
phonologically conditioned suppletion', in Yearbook of Morphology 
1988. See also my article 'Phonologically conditioned suppletion' in 
Dressler et al (eds) _Contemporary Morphology_ (Mouton de Gruyter, 
1990), and my article in the Spencer-Zwicky _Handbook of Morphology_.

Best
Andrew
- 
Andrew Carstairs-McCarthy
Professor and Head of Department
Department of Linguistics, University of Canterbury, Private Bag 
4800, Christchurch, New Zealand
phone (work) +64-3-364 2211; (home) +64-3-355 5108
fax +64-3-364 2969
e-mail a.c-mccling.canterbury.ac.nz
http://www.ling.canterbury.ac.nz/adc-m.html

=========================================================
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue