View
217
Download
0
Tags:
Embed Size (px)
Citation preview
Semantically-Vacuous Semantically-Vacuous Variable Affix Placement in Variable Affix Placement in Cajamarca Quechua: Some Cajamarca Quechua: Some
ImplicationsImplications
The 4th Newcastle Postgraduate Conference in Theoretical and Applied LinguisticsThe 4th Newcastle Postgraduate Conference in Theoretical and Applied Linguistics27 March 200927 March 2009
Neil Myler ([email protected])
Corpus Christi College
Cambridge
1. Cajamarca Quechua1. Cajamarca Quechua
2. Semantically-Vacuous Variable Affix-Placement 2. Semantically-Vacuous Variable Affix-Placement (SVAP)(SVAP)
3. Problems for Word-and-Paradigm Models3. Problems for Word-and-Paradigm Models
4. Problems for Stem-and-Paradigm Models4. Problems for Stem-and-Paradigm Models
5. Problems for ALIGN-based affix-ordering (Anderson 5. Problems for ALIGN-based affix-ordering (Anderson 2005)2005)
6. The Solution: Distributed Morphology 6. The Solution: Distributed Morphology
Talk Summary
Spoken in rural communities surrounding the northern Spoken in rural communities surrounding the northern Peruvian city of Cajamarca (PorcPeruvian city of Cajamarca (Porcón, Chetilla)ón, Chetilla)
Ethnologue reports that it has 30,000 speakersEthnologue reports that it has 30,000 speakers High bilingualism in SpanishHigh bilingualism in Spanish Extremely low prestige; in many cases the language is Extremely low prestige; in many cases the language is
no longer being passed on to childrenno longer being passed on to children Sources on the dialect:Sources on the dialect:
Coombs Lynch et al. (2003),Coombs Lynch et al. (2003),
Quesada (1976) (I have pdf)Quesada (1976) (I have pdf)
my own fieldwork (on which more later)my own fieldwork (on which more later)
Cajamarca Quechua
http://www.ethnologue.com/show_language.asp?code=qvc
Semantically-Vacuous Variable Affix-Placement
-llapa : The Personal Plural Morpheme-llapa : The Personal Plural Morpheme
-llapa pluralises person markers on both nouns and verbs. Note that it does -llapa pluralises person markers on both nouns and verbs. Note that it does not pluralise the noun itself.not pluralise the noun itself.
yanasa ‘friend’yanasa ‘friend’ yanasa-y ‘my friend’yanasa-y ‘my friend’ yanasa-yki ‘your (sg) friend’yanasa-yki ‘your (sg) friend’ yanasa-n ‘his/her/its friend’yanasa-n ‘his/her/its friend’ yanasa-nchiq ‘our (incl.) friend’yanasa-nchiq ‘our (incl.) friend’ yanasa-y-yanasa-y-llapa llapa ‘our (excl.) friend’‘our (excl.) friend’ yanasa-yki-yanasa-yki-llapallapa ‘your (pl.) friend’ ‘your (pl.) friend’ yanasa-n-yanasa-n-llapallapa ‘their friend’ ‘their friend’
--kuna: The Nominal Plural Morpheme kuna: The Nominal Plural Morpheme
yanasa ‘friend’yanasa ‘friend’
yanasakuna ‘friends’yanasakuna ‘friends’
-kuna uncontroversially follows the possessor person markers:-kuna uncontroversially follows the possessor person markers:
mishu-y ‘my cat’mishu-y ‘my cat’
mishu-y-kuna ‘my cats’mishu-y-kuna ‘my cats’
*mishu-kuna-y *mishu-kuna-y
(not in Cajamarca, but you do get this in some dialects) (not in Cajamarca, but you do get this in some dialects)
Semantically-Vacuous Variable Affix-Placement
So what if –kuna and –llapa appear in the same word?So what if –kuna and –llapa appear in the same word?
Depends who you ask…Depends who you ask…
Coombs Lynch et al. (2003): -llapa precedes -kunaCoombs Lynch et al. (2003): -llapa precedes -kuna
wasi-n-llapa-kunawasi-n-llapa-kuna
house-3house-3rdrd-pplu-plural-pplu-plural
‘ ‘their houses’their houses’
Quesada (1976:90): -llapa follows -kunaQuesada (1976:90): -llapa follows -kuna
wasi-n-kuna-llapawasi-n-kuna-llapa
house-3house-3rdrd-plural-pplu-plural-pplu
‘ ‘their houses’their houses’
Semantically-Vacuous Variable Affix-Placement
Actually, many speakers I’ve asked accept both orders!
Another quirk: -llapa and the postpositionsAnother quirk: -llapa and the postpositions
Postpositions/case-markers practically always follow –llapaPostpositions/case-markers practically always follow –llapa
This is the only order reported in Quesada (1976) or Coombs-Lynch et al. This is the only order reported in Quesada (1976) or Coombs-Lynch et al.
(2003)(2003)
wasi-n-llapa-piwasi-n-llapa-pi
house-3house-3rdrd-pplu-in-pplu-in
‘‘in their house’in their house’
BUT check out this example from a folktale BUT check out this example from a folktale kimsa yanasakuna kimsa yanasakuna
‘‘the three friends’:the three friends’:
Semantically-Vacuous Variable Affix-Placement
manchay-ni-n-pi-llapa
fear-NI-3rd-in-pplu
‘in their fear’
Another quirk: -llapa and the postpositions continuedAnother quirk: -llapa and the postpositions continued
It turns out that both orders are possible. This is true for It turns out that both orders are possible. This is true for –pi –pi ‘in’ and a subset of ‘in’ and a subset of
the other postpositions (speakers vary on this point and I’m not yet sure what the other postpositions (speakers vary on this point and I’m not yet sure what
all the generalisations are) all the generalisations are)
Semantically-Vacuous Variable Affix-Placement
So, it seems –llapa can appear before or after –kuna. It can also appear before or after –pi, and some of the other postpositions.
BUT, when we have words which contain all three affixes, the orders that turn out to be allowed are more constrained than you might think…
Possible permutations with –llapa, -kuna and –pi combinedPossible permutations with –llapa, -kuna and –pi combined
a. wasi-n-llapa-kuna-pia. wasi-n-llapa-kuna-pi
house-3house-3rdrd-pplu-plural-in-pplu-plural-in
“ “In their houses”In their houses”
b. b. wasi-n-kuna-llapa-piwasi-n-kuna-llapa-pi
c.*wasi-n-pi-kuna-llapac.*wasi-n-pi-kuna-llapa
d.*wasi-n-pi-llapa-kunad.*wasi-n-pi-llapa-kuna
e.*wasi-n-llapa-pi-kunae.*wasi-n-llapa-pi-kuna
f.*wasi-n-kuna-pi-llapaf.*wasi-n-kuna-pi-llapa
Semantically-Vacuous Variable Affix-Placement
Blevins (2003, 2006)Blevins (2003, 2006)
““a rejection of the strict division between data and program, which is expressed a rejection of the strict division between data and program, which is expressed
by the familiar dichotomy between ‘entries and rules’ (Pinker 1999)” by the familiar dichotomy between ‘entries and rules’ (Pinker 1999)”
(Blevins 2003:743) (Blevins 2003:743)
Given WP’s eschewal of rule-based derivations, it relies purely on Given WP’s eschewal of rule-based derivations, it relies purely on
proportional analogy to account for productivityproportional analogy to account for productivity
But proportional analogy over surface patterns is not constrained enough to But proportional analogy over surface patterns is not constrained enough to
give us the grammaticality patterns we see in SVAP!give us the grammaticality patterns we see in SVAP!
Problems for Word-and-Paradigm Morphology
We observe that We observe that wasi-n-llapa-pi wasi-n-llapa-pi is a possible word, and that is a possible word, and that wasi-n-llapa-kunawasi-n-llapa-kuna
is too.is too.
So why not, by analogy, *So why not, by analogy, *wasi-n-llapa-pi-kuna wasi-n-llapa-pi-kuna when the three affixes arewhen the three affixes are
combined?combined?
Or given Or given wasi-n-pi-llapawasi-n-pi-llapa and and wasi-n-llapa-kunawasi-n-llapa-kuna, why not, in the absence of , why not, in the absence of negative evidence, *negative evidence, *wasi-n-pi-llapa-kunawasi-n-pi-llapa-kuna? ?
WP has no answer, and can gain no access to one, because the predictions ofWP has no answer, and can gain no access to one, because the predictions of
proportional analogy collapse at this point.proportional analogy collapse at this point.
Problems for Word-and-Paradigm Morphology
Anderson (1992); Stump (2001); Baerman et al. (2005)Anderson (1992); Stump (2001); Baerman et al. (2005)
Anderson (1992) and similar approaches reject the notion of ‘affix’; affixation Anderson (1992) and similar approaches reject the notion of ‘affix’; affixation
is just a special case of a morphological rule applying to a stem. is just a special case of a morphological rule applying to a stem.
Hence there is no formal difference between, say, marking an English plural Hence there is no formal difference between, say, marking an English plural
via vowel raising (e.g. footvia vowel raising (e.g. foot→feet) and marking one by suffixing ‘-s’ →feet) and marking one by suffixing ‘-s’
(e.g. hand→hands)(e.g. hand→hands)
Both are conceptualised as the result of a process applying to a stem.Both are conceptualised as the result of a process applying to a stem.
Problems for Stem-and-Paradigm Morphology
Anderson (1992); Stump (2001); Baerman et al. (2005)Anderson (1992); Stump (2001); Baerman et al. (2005)
Anderson (1992) relies on rule-ordering to capture affix-ordering.Anderson (1992) relies on rule-ordering to capture affix-ordering.
p.123: “the linear descriptive sequence of affixes is reflected in the rule system p.123: “the linear descriptive sequence of affixes is reflected in the rule system
as the relative ordering of the corresponding rules. That is, the fact that a word as the relative ordering of the corresponding rules. That is, the fact that a word
has the form has the form stem+affix1+affix2stem+affix1+affix2 reflects the fact that the rule attaching affix1 reflects the fact that the rule attaching affix1
applied ‘before’ the rule attaching affix2”applied ‘before’ the rule attaching affix2”
Problems for Stem-and-Paradigm Morphology
So, what do we say about SVAP?
Competing grammars?
Rule-inverting meta-rules?
Say the rules were ordered as follows:Say the rules were ordered as follows:
(1)(1) Add –llapaAdd –llapa
(2)(2) Add –kunaAdd –kuna
(3)(3) Add –piAdd –pi
Then, perhaps we could have special optional rule-inverting meta-rules.Then, perhaps we could have special optional rule-inverting meta-rules.
These would optionally invert rules 1 and 2, and rules 1 and 3 (in derivations These would optionally invert rules 1 and 2, and rules 1 and 3 (in derivations
where there was no nominal plural feature, in which rule 2 would not apply).where there was no nominal plural feature, in which rule 2 would not apply).
The absence of a meta-rule inverting 2 and 3 would correctly rule out all forms The absence of a meta-rule inverting 2 and 3 would correctly rule out all forms
in which –pi precedes –kuna: *wasi-n-pi-llapa-kuna,*wasi-n-pi-kuna-llapain which –pi precedes –kuna: *wasi-n-pi-llapa-kuna,*wasi-n-pi-kuna-llapa
*wasi-n-llapa-pi-kuna *wasi-n-llapa-pi-kuna
Problems for Stem-and-Paradigm Morphology
However, even this would not be constrained enough.However, even this would not be constrained enough.
If the meta-rule inverting rules (1) and (2) applied this would leave rules (1) and If the meta-rule inverting rules (1) and (2) applied this would leave rules (1) and
(3) ordered next to one another. Thus, the meta-rule inverting (1) and (3) could(3) ordered next to one another. Thus, the meta-rule inverting (1) and (3) could
also apply and we could get the following rule-order:also apply and we could get the following rule-order:
(2) Add -kuna(2) Add -kuna
(3) Add -pi(3) Add -pi
(1) Add-llapa(1) Add-llapa
But this would produce the word *wasi-n-kuna-pi-llapa, which is out.But this would produce the word *wasi-n-kuna-pi-llapa, which is out.
Problems for Stem-and-Paradigm Morphology
Some added observationsSome added observationsActually, Anderson’s theory allows for the existence of ‘rule blocks’- a rule block is a set of rulesActually, Anderson’s theory allows for the existence of ‘rule blocks’- a rule block is a set of rules
(minimally consisting of one rule only), only one of which can apply in the course of a derivation. If (minimally consisting of one rule only), only one of which can apply in the course of a derivation. If
one assumed that these rule-inverting meta rules were in the same block, then they would have to one assumed that these rule-inverting meta rules were in the same block, then they would have to
apply disjunctively. Thus rule 1 would never be able to feed rule 2, and this would get you theapply disjunctively. Thus rule 1 would never be able to feed rule 2, and this would get you the
grammaticality patterns.grammaticality patterns.
There is still a potential problem for Anderson’s theory, though. Would such rule-inverting meta There is still a potential problem for Anderson’s theory, though. Would such rule-inverting meta rules rules
have any use in his theory other than to effect SVAP? Crucially, is there any evidence that they are have any use in his theory other than to effect SVAP? Crucially, is there any evidence that they are
used in process morphology? If not, then to invoke rule-inverting meta-rules would be to concede used in process morphology? If not, then to invoke rule-inverting meta-rules would be to concede that that
affixation is in some way formally distinct from process morphology. But this is precisely what affixation is in some way formally distinct from process morphology. But this is precisely what theories theories
like Anderson’s set out to deny.like Anderson’s set out to deny.
Anderson (2005), in the context of an OT account of clitic placement, outlinesAnderson (2005), in the context of an OT account of clitic placement, outlinesthe following approach to affix-ordering (p.the following approach to affix-ordering (p.135):135):
““Suppose a number of affixes are all constrained to be prefixes. That Suppose a number of affixes are all constrained to be prefixes. That means that for each of them, there is a constraint to the effect that it means that for each of them, there is a constraint to the effect that it should appear at the left edge of the word. Since these constraints are should appear at the left edge of the word. Since these constraints are strictly ranked, however, one of them will outrank the others […] In the strictly ranked, however, one of them will outrank the others […] In the optimal realization of a given form, the affix corresponding to the optimal realization of a given form, the affix corresponding to the constraint with the highest rank will actually be an initial prefix, the next constraint with the highest rank will actually be an initial prefix, the next highest ranked will occupy a position which violates minimally its highest ranked will occupy a position which violates minimally its requirement of being initial (i.e., as the second prefix), etc. Exactly requirement of being initial (i.e., as the second prefix), etc. Exactly similar considerations apply, similar considerations apply, mutatis mutandismutatis mutandis, for collections of , for collections of suffixes”suffixes”
Problems for an OT ALIGN-based Approach
ALIGNR(pi) ALIGNR(kuna) ALIGNR(llapa)
pi-kuna-llapa *!* **pi-llapa-kuna *!* *
kuna-llapa-pi *!**!* *llapa-kuna-pi ** **
llapa-pi-kuna *!*! ****kuna-pi-llapa *!*! ****
Problems for an OT ALIGN-based Approach
Of course, the derivation overleaf will turn out ok if we just tie ALIGNR(kuna) Of course, the derivation overleaf will turn out ok if we just tie ALIGNR(kuna)
and ALIGNR(llapa).and ALIGNR(llapa).
BUT, recall that, when –kuna is not present, -llapa can also invert with –piBUT, recall that, when –kuna is not present, -llapa can also invert with –pi
If we try to get this by also tying ALIGNR(llapa) with ALIGNR(pi), then If we try to get this by also tying ALIGNR(llapa) with ALIGNR(pi), then
effectively we end up with all three constraints tied together; we can’t avoid this effectively we end up with all three constraints tied together; we can’t avoid this
if we want to capture both generalisations.if we want to capture both generalisations.
But this would predict that all six logically possible orders should be But this would predict that all six logically possible orders should be
grammatical when the three affixes are combined. We’ve already seen thatgrammatical when the three affixes are combined. We’ve already seen that
this is wrong. this is wrong.
Problems for an OT ALIGN-based Approach
Yet more added observationsYet more added observationsThis is not to say that no OT approach to SVAP could be created. Since I gave this talk, Yuni Kim
has
drawn my attention to the work of Kevin Ryan, a PhD student from UCLA, who has a very impressive
bigram constraint scheme for SVAP (his data are mainly from Tagalog, along with some other
languages) which is implementable in Stochastic OT and a range of other constraint-based
frameworks. I haven’t investigated this in full yet, but on first look it seems as if Ryan’s scheme should
be able to get SVAP in Cajamarca.
The Solution: Distributed Morphology
Halle and Marantz (1993, 1994); Embick and Noyer (2001, 2006); Arregi and
Nevins (2008); Embick (2007)
• Narrow Syntax manipulates abstract lexical roots and bundles of formal
features encoded in terminal nodes.
• These nodes, which also receive the terms ‘head’ or ‘morpheme’, have no phonological content.
• The phonological matrices of surface morphemes, known as ‘Vocabulary
Items’, are inserted at Morphological Form before being submitted to
the Phonology.
• At MF, before, during and after VI-insertion, certain post-syntactic operations can
eliminate, split or permute morphemes.
• This means that DM is a piece-based, realisational and separationist
model of morphology
The Syntax of Cajamarca DPs
DP
Chay
DNumP
ishkay
-kuna
nP
atun
-llapa
AgrNumberP
AgrPersonP
-nChay ishkay atun wasi-n-llapa-kuna-pi
That two big house-3rd-pplu-plural-in
“In those two big houses of theirs”
n √wasi
PP
-pi
Morphological Lowering
-kuna
nP
atun
AgrNumberP
AgrPersonP
Chay ishkay atun wasi-n-llapa-kuna-pi
That two big house-3rd-pplu-plural-in
“In those two big houses of theirs”
n √wasi
√wasi -n
-n -llapa
-llapa
NumP
The morphology takes this output of syntax and converts it into a series of
concatenation statements, like so:
wasi*n*llapa*kuna*pi
This will be what we get if –kuna is present. If –kuna is absent but –llapa and
-pi are present, we will end up with:
wasi*n*llapa*pi
DM contains an operation, Local Dislocation, which exchanges the
linearisation statements between two adjacent elements, leaving them adjoined
to one another. I assume with Vaux (2008) that all PF operations can be
marked as being optional, and that SVAP is simply an instance of optional
Local Dislocation.
Preliminary Linearisation and Local Dislocation
(1)(1) llapa*kuna llapa*kuna → kuna+llapa (optional)→ kuna+llapa (optional)
(2)(2) llapa*pi → pi+llapa (optional)llapa*pi → pi+llapa (optional)
There is no rule inverting –kuna with –pi, so all the examples with –pi before There is no rule inverting –kuna with –pi, so all the examples with –pi before
-kuna are correctly ruled out. This includes *wasi-n-pi-llapa-kuna, since rule -kuna are correctly ruled out. This includes *wasi-n-pi-llapa-kuna, since rule
(1), like all instances of Local Dislocation, can only apply under strict (1), like all instances of Local Dislocation, can only apply under strict
adjacency. adjacency.
Local Dislocation
Since –llapa will be linearised next to –pi if –kuna is not present, we correctly predict both wasi-n-llapa-pi and wasi-n-pi-llapa to be grammatical
But how do we prevent *wasi-n-kuna-pi-llapa being derived from
wasi-n-kuna-llapa-pi? This is what would happen if rule (1) fed rule (2).
Underlying Representation wasi-n-llapa-kuna-piUnderlying Representation wasi-n-llapa-kuna-pi
Rule (1) wasi-n-kuna-llapa-piRule (1) wasi-n-kuna-llapa-pi
Rule (2) & Output *wasi-n-kuna-pi-llapaRule (2) & Output *wasi-n-kuna-pi-llapa
Why can’t rule (1) feed rule (2)?
In fact, this form is correctly ruled out, because rule (1) creates a form kuna+llapa from which –llapa cannot escape: it is bracketed with –kuna.
Hence rule (2)’s structural description is not met when rule (1) applies, and *wasi-n-kuna-pi-llapa is correctly ruled out. But rule (2) can apply if –kuna is never present in the first place, and so we correctly derive all the grammaticality patterns.
UR wasi*n*llapa*kuna*piUR wasi*n*llapa*kuna*pi Rule (1) wasi*n*kuna+llapa*piRule (1) wasi*n*kuna+llapa*pi Rule (2) -----cannot apply-----Rule (2) -----cannot apply----- Output wasi-n-kuna-llapa-piOutput wasi-n-kuna-llapa-pi
UR wasi*n*llapa*pi
Rule (2) wasi*n*pi+llapa
Output wasi-n-pi+llapa
SUMMARYSUMMARY
1a. wasi-n-llapa-pi (rule 2 optionally fails to apply)1a. wasi-n-llapa-pi (rule 2 optionally fails to apply)
b. wasi-n-pi-llapa (rule 2 optionally applies)b. wasi-n-pi-llapa (rule 2 optionally applies)
2 a. wasi-n-llapa-kuna-pi (rule 1 optionally fails to apply)2 a. wasi-n-llapa-kuna-pi (rule 1 optionally fails to apply)
b. wasi-n-kuna-llapa-pi (rule 1 optionally applies; rule 2 cannot apply)b. wasi-n-kuna-llapa-pi (rule 1 optionally applies; rule 2 cannot apply)
c.*wasi-n-pi-kuna-llapa (cannot be derived; no rule inverts –pi and –kuna)c.*wasi-n-pi-kuna-llapa (cannot be derived; no rule inverts –pi and –kuna)
d.*wasi-n-pi-llapa-kuna (cannot be derived; rule 1 is strictly local)d.*wasi-n-pi-llapa-kuna (cannot be derived; rule 1 is strictly local)
e.*wasi-n-llapa-pi-kuna (cannot be derived; no rule inverts –pi and –kuna)e.*wasi-n-llapa-pi-kuna (cannot be derived; no rule inverts –pi and –kuna)
f.*wasi-n-kuna-pi-llapa (cannot be derived, rule 1 cannot feed rule 2)f.*wasi-n-kuna-pi-llapa (cannot be derived, rule 1 cannot feed rule 2)
Anderson, Stephen R. (1992) Anderson, Stephen R. (1992) A-Morphous Morphology A-Morphous Morphology (Cambridge: CUP)(Cambridge: CUP) Anderson, Stephen R. (2005) Anderson, Stephen R. (2005) Aspects of the Theory of Clitics Aspects of the Theory of Clitics (Oxford: OUP)(Oxford: OUP) Arregi, Karlos and Nevins, Andrew (2008) Arregi, Karlos and Nevins, Andrew (2008) A Principled Order to Postsyntactic Operations A Principled Order to Postsyntactic Operations (LingBuzz/000646: retrieved (LingBuzz/000646: retrieved
02/11/08)02/11/08) Baerman, Matthew, Dunstan Brown and Greville Corbett (2005) Baerman, Matthew, Dunstan Brown and Greville Corbett (2005) The Syntax Morphology Interface: A Study of Syncretism The Syntax Morphology Interface: A Study of Syncretism
(Cambridge: CUP)(Cambridge: CUP) Blevins, James P. (2003) Blevins, James P. (2003) Stems and ParadigmsStems and Paradigms (Language 79.4:737-767) (Language 79.4:737-767) Blevins, James P. (2006) Blevins, James P. (2006) Word-based Morphology Word-based Morphology (Journal of Linguistics 42:531-573)(Journal of Linguistics 42:531-573) Coombs Lynch, David, Heidi Carlson de Coombs and Blanca Ortiz Chamán (1997) Coombs Lynch, David, Heidi Carlson de Coombs and Blanca Ortiz Chamán (1997) Rimashun Kichwapi. Una introducción Rimashun Kichwapi. Una introducción
al quechua cajamarquino al quechua cajamarquino (Lima: Atares artes y letras)(Lima: Atares artes y letras) Embick, David (2007) Embick, David (2007) LLinearization and Local Dislocation: Derivational Mechanics and Interactionsinearization and Local Dislocation: Derivational Mechanics and Interactions
(Linguistic Analysis 33 p.303-336)(Linguistic Analysis 33 p.303-336) Embick, David and Rolf Noyer (2001) Embick, David and Rolf Noyer (2001) Movement operations after syntax Movement operations after syntax (Linguistic Inquiry 32: pp. 555-595)(Linguistic Inquiry 32: pp. 555-595) Embick, David and Rolf Noyer (2006) Embick, David and Rolf Noyer (2006) Distributed Morphology and the Syntax-Morphology Interface Distributed Morphology and the Syntax-Morphology Interface (Oxford Handbook of (Oxford Handbook of
Linguistic Interfaces, p.289-324)Linguistic Interfaces, p.289-324) Halle, Morris and Marantz, Alec (1993) Halle, Morris and Marantz, Alec (1993) Distributed MorphologyDistributed Morphology, in , in The View From Building The View From Building 20 20 (Cambridge MA:MIT Press (Cambridge MA:MIT Press
pp.111-176)pp.111-176) Halle, Morris and Marantz, Alec (1994) Halle, Morris and Marantz, Alec (1994) Some key features of Distributed Morphology Some key features of Distributed Morphology (MIT Working Papers in Linguistics (MIT Working Papers in Linguistics
21:275-288)21:275-288) Quesada, Félix (1976) Quesada, Félix (1976) Gramática Quechua-Cañaris Gramática Quechua-Cañaris (Lima: Ministerio de Educación)(Lima: Ministerio de Educación) Stump, Gregory T. (2001) Stump, Gregory T. (2001) Inflectional Morphology: A Theory of Paradigm Structure (Cambridge: CUP)Inflectional Morphology: A Theory of Paradigm Structure (Cambridge: CUP) Vaux, Bert (2008) Vaux, Bert (2008) Why the Phonological Component must be Serial and Rule-BasedWhy the Phonological Component must be Serial and Rule-Based , in Vaux, Bert and Andrew Nevins (eds.) , in Vaux, Bert and Andrew Nevins (eds.)
Rules, Constraints and Phonological Phenomena Rules, Constraints and Phonological Phenomena (Oxford:OUP pp.20-60)(Oxford:OUP pp.20-60)
List of Reference Materials