Lecture 1: Morphologically conditioned phonology...Noun privilege: Smith’s (2011) survey finds overall that that nouns tend to exhibit more contrasts, while verbs are more prone

Morphophonology 07/07/2017

1

Lecture 1: Morphologically conditioned phonology (1) Morphologically conditioned phonology:

• the phenomenon in which a particular phonological pattern is imposed on a proper subset of morphological constructions (affix, reduplication, compounding) and thus is not fully general in the word-internal phonological patterning of the language.

• the inspiration for a number of influential theories of the phonology-morphology interface, including Lexical Morphology and Phonology, Stratal Optimality Theory, and Cophonology Theory.

(2) Today: survey

• the types of morphological information that can condition phonological patterns • the types of phonological patterns that can be conditioned by morphology.

1. ILLUSTRATIVE EXAMPLES

Mam Maya: a word may have at most one long vowel. Willard 2004, based on England 1983: ‘Dominant’ affixes cause long root vowels to shorten (3a); ‘Recessive’ suffixes preserve root vowel length (3b). Dominant vs. recessive status is lexical, not predictable. (3) a. Dominant suffix: shortens long root vowel facilitative resultant liich’- → lich’-ich’iin ‘break/breakable’ locative juus- → jus-b'een ‘burn/burned place’ directional jaaw- → jaw-nax ‘go up/up’ participial nooj- → noj-na ‘fill/full’ b. Recessive suffix: preserves root vowel length intransitive verbalizer muq- → muq-oo ‘bury (n.)/bury (v.)’ b’iitz- → b’iitz-oo ‘song/sing’ [b’liitza] instrumental luk- → luk-b’il ‘pull up/instrument for pulling up’ remainder waa- → waa-b’an ‘eat/remains of food’ Malayalam (Southern Dravidian): consonant gemination applies at the internal juncture of subcompounds (noun-noun compounds with head-modifier semantics) (4b). Gemination does not apply to cocompounds (noun-noun compounds with coordinate semantics) (4c) (Mohanan 1995:49): (4) a. meeša ‘table’ pet.t.i ‘box’ -kaḷ (plural suffix) b. [meeşa-ppet.t.i]S -kaḷ ‘boxes made out of tables’ c. [meeşa-pet.t.i]C -kaḷ ‘tables and boxes’


2

English: suffixes fall into two classes (e.g. Allen 1978, Siegel 1974, Chomsky & Halle 1968, Kiparsky 1982ab, Kiparsky 1985): those which shift stress (5a) and those which do not (5b): (5) Base (a) Stress-shifting suffix (b) Non-stress-shifting suffix párent parént-al párent-ing président prèsidént-ial présidenc-y áctive àctív-ity áctiv-ist démonstràte demonstrative démonstràtor

2. PHONOLOGICAL SENSITIVITY TO LEXICAL CLASS.

2.1 PART OF SPEECH (Tokyo) Japanese pitch-accent (McCawley 1968, Haraguchi 1977, Poser 1984, Pierrehumbert & Beckman 1988, Tsujimura 1996, Smith 1999): The distribution of accent is different in nouns and verbs. The location of accent in lexically accented non-derived nouns is unpredictable and must be learned individually for each such noun ((6a), from Poser 1984:46). The location of accent in an accented verb follows strict rules, falling on the first mora of the syllable containing the penultimate mora of the verb ((6b), from Poser 1984:52): (6) Accented Unaccented [Japanese] a. Nouns fu.ku.ro ‘bag’ hasira ‘pillar’ ta.ma.go ‘egg’ kusuri ‘medicine’ su.to.rai.ki ‘strike’ udoN ‘noodle dish’ b. Verbs ka.ke.ru ‘hang’ kakeru ‘be broken’ su.kuu ‘build a nest’ sukuu ‘rescue’ ue.ru ‘starve’ ueru ‘plant’ Polysyllabic words in Lenakel (Oceanic) have primary stress on the penultimate syllable; in verbs and adjectives, secondary stress falls on the first syllable and every other syllable thereafter, up to but not including the antepenultimate syllable (avoiding clash). In nouns, secondary stress is assigned to alternating syllables to the left of the primary penultimate stress. Data from Lynch 1978:18-20; see discussion in Smith 2011: (7) a. Verbs (four or more syllables) [Lenakel]

/r-ɨm-olkeikei/ [r̆!.̀mↄl.gɛ́y.gɛy] ‘he liked it’ /n-ɨm-r-olkeikei/ [n!.̀mɑ.r̆ↄl.gɛ́y.gɛy] ‘you (pl.) liked it’ /n-ɨm-m-r-olkeikei/ [n!.̀mɑ.mɑ̀.r̆ↄl.gɛ́y.gɛy] ‘you (pl.) were liking it’ /t-n-k-m-r-olkeikei/ [t!.̀nɑ.gɑ̀.mɑ.r̆ↄl.gɛ́y.gɛ́y] ‘you (pl.) will be liking it’ ~ [d!.̀nɑ.gɑ̀.mɑ.r̆ↄl.gɛ́y.gɛ́y]

b. Nouns (four or more syllables) /nɨmwakɨlakɨl/ [nɨ.mʷɒ̀.gə.lɑ́.gəl] ‘beach’ /tupwalukaluk/ [tu.bʷɒ̀.lu.gɑ́.lʊkʰ] ‘lungs’ ~ [du.bʷɒ̀.lu.gɑ́.lʊkʰ]

In Chuukese (Trukese; Micronesian), nouns must be minimally bimoraic, a condition which a monosyllabic noun can satisfy by possessing an initial (moraic) geminate (8a) or by undergoing


3

vowel lengthening (8b). (Coda consonants are not moraic in Chuukese.) By contrast, verbs are allowed to surface in monomoraic CVC form (8c). The data in (8a,b) exhibit vowel apocope (Smith 2011, citing Muller 1999:395 and Goodenough & Sugita 1980:xiv-xv):

(8) a. [kkej] ‘laugh’ (< /kkeji/) [Chuukese] [ʧar] ‘starfish’ (</ʧʧara/) b. [faːs] ‘nest’ (</fasa/) [fæːn] ‘building’ (</fæne/) c. [fan] ‘go aground’ [mær] ‘move, be shifted’

Noun privilege: Smith’s (2011) survey finds overall that that nouns tend to exhibit more contrasts, while verbs are more prone to neutralization. This finding is clearly consistent with the Japanese example in (6), though it is not as clearly applicable to Lenakel or Chuukese.

2.2 IDEOPHONES • a phonosemantic class of words whose meanings typically include color, smell, sound,

intensity, or (often vivid) descriptions of unusual appearance or activity. • can belong to various parts of speech, most often adjectives, adverbs or verbs. • of interest to the present discussion because in many languages they constitute a class of

words with distinctive phonology, often departing from prosodic or segmental norms. • for useful surveys, see Hinton et al. 1994 and Voeltz and Kilian-Hatz 2001.

Hausa (Chadic): ca. 500 ideophones, which depart from the language’s phonological norms in two ways (Newman 1995, 2000:242 ff., 2001):

• Ideophones are pronounced with exaggerated intonation • Ideophones are usually consonant-final; the Hausa norm is to be vowel-final. • Ideophones can end in obstruent consonants, including plosives, impossible in the other

sectors of Hausa vocabulary (Newman 1995:776, Newman 2000:244,250): (9) fát fáríː fát ‘white IDEO = very white’ [Hausa] ʃár ̃ kóːrèː ʃár ̃ ‘green IDEO = very green’ ƙút àbóːkíː ƙút ‘friend IDEO = very close friend’ ták ɗájá ták ‘one IDEO = exactly one’ fár ̃át táː táːʃì fár ̃át ‘3SG.FEM get_up IDEO = she got up very fast’ túɓús yáː gàjí túɓús ‘3SG.MASC become_tired IDEO = he became very tired’ gàràrà súnàː jáːwòː gàràrà ‘3PL walk IDEO = they roamed aimlessly’


4

2.3 ETYMOLOGICAL CLASSES Japanese (Itô & Mester 1999:62): Sino-Japanese vocabulary distinguished by etymology and phonology No-DD bans voiced geminates; No-P bans singleton (onset) [p], and No-NT bans sequences consisting of a nasal consonant followed by a voiceless consonant:

(10) No-DD No-P No-NT Yamato Sino-Japanese violated Assimilated foreign violated violated Nonassimilated foreign violated violated violated

Itô and Mester (1999:70): the classification of lexical items into strata is not always technically etymologically accurate. For example, native (contracted) anta ‘you’ violates *NT.

2.4 ARBITRARY LEXICAL CLASSES: PATTERNED EXCEPTIONS Sacapultec (Mayan, Guatemala): some nouns undergo final-syllable vowel lengthening in combination with possessive prefixes (11a), while others do not (11b) (DuBois 1985):

(11) Plain Possessive a. ak ‘chicken’ w-:k ‘my chicken’ ʦ'eʔ ‘dog’ ni-ʦ'i:ʔ ‘my dog’ ab'ax ‘rock’ w-ub'a:x ‘my rock’ mulol ‘gourd’ ni-mulu:l ‘my gourd’ b. oʧ' ‘possum’ w-oʧ' ‘my possum’ am ‘spider’ w-m ‘my spider’ weʔ ‘head hair’ ni-weʔ ‘my head hair’ DuBois: (11a) vs. (11b) is lexically conditioned, modulo a weak semantic or pragmatic skewing; many of the (b) stems ‘do not often occur in possessed constructions’ (p. 396). Tagalog: paŋ- + C-initial stem creates environment for Nasal Substitution (NCi → Ni) (Zuraw 2000:19 ff.). NS is lexically conditioned: (12)

Undergoer of NS Non-undergoer of NS a. bugbóg pa-mugbóg bigáj pam-bigáj ‘wallo’ ‘wooden club used to pound clothes

during washing’ ‘gift’ ‘gifts to be distributed’

b. búlos pa-múlos buɁóɁ pam-buɁóɁ ‘harpoon’ ‘harpoon’ ‘whole’ ‘something used to produce a whole’

Zuraw: NS is statistically influenced by several factors (‘patterned exceptionality’):

• stems beginning with voiced consonants undergo Nasal Substitution in a much higher proportion than do stems beginning with voiceless consonants (Zuraw 2000:29).

• P(Labial-initial stems, NS) » P(Dental-initial stems, NS) » P(Velar-initial stems, NS)


5

Through overgeneralization and analogy, (arbitrarily) lexically conditioned patterns may become phonologically conditioned over time. See also recent work by e.g., Becker & Gouskova (2016), Moore-Cantwell & Staubs (2014), Shih & Inkelas (2015), and others.

3. THE ROOT-AFFIX DISTINCTION

(13) McCarthy & Prince 1995: roots are larger and more phonologically diverse than affixes • Sanskrit roots may contain consonant clusters but affixes never do • Arabic roots may contain pharyngeal consonants, but affixes cannot • English suffixes favor (unmarked) coronal consonants (e.g. Yip 1991). • roots are more resistant to undergoing alternations than affixes are.

3.1 ROOT FAITHFULNESS (14) Root-affix Faithfulness Metaconstraint (RAFM; McCarthy and Prince 1995): Root-FAITH >> Affix-FAITH (15) Ekuguusi vowel harmony (Guusi, Bantoid; Cammenga 2002): mid vowels in affixes

harmonize in [ATR] with mid stem vowels (o, e, ↄ, ɛ) a. o-mo-te ‘tree’ b. ↄ-rɛɛnt-ir-e ‘he has brought’ c. e-ñuↄm-ↄ ‘marriage’ d. tↄ-ɣɛɛnr-ɛ ‘let us go’

3.2 COUNTEREXAMPLES TO THE RAFM Casali 1997: survey of V-V hiatus in 87 languages revealed two strong preferences for deletion:

• V1 deletes… • affix vowels delete… • → stem-initial vowels should never delete to resolve VV hiatus across prefix-stem

boundary But: Karuk (Bright 1972); see discussion in Kenstowicz & Kisseberth 1979, Koutsoudas 1980)

(16) gloss No (or C-final) prefix V-final prefix gloss ‘cook’ (33) imniš ní-mniš ‘I cook’ (33) ‘to be cooking’ (62) imníˑštih ʔú-mniš ‘he cooks’ (33) ‘they’re cooking’ (62) kun-ímniˑštih ʔú-mniˑštih ‘he’s cooking’ (62) ‘head’ (50) axvâˑh mú-xvâˑh ‘his head’ (50) ‘money’ (44) išpuka mú-spuka ‘his money’ (44)


6

Cupeño (Alderete 2001a): When an accented root and accented affix co-occur in a word with culminative accent, root accent is expected to prevail (absent a directionality preference): (17) Accented root + accented affix(es): accent surfaces on root a. /pǝ́ + √míʔaw + lu/ pǝ-míʔaw-lu 3SG + COME + MOTION ‘He came’ b. /√ʔáyu + qá/ ʔáyu-qa WANT + PRES.SING ‘He wants’ Unaccented root + accented affix(es): accent surfaces on affix, not root c. /pǝ́ + √yax/ pǝ́-yax 3SG + SAY ‘He says’ d. /nǝʔǝn + √yax + qá/ nǝʔǝn ya-qáʔ 1SG + SAY + PRES.SING ‘I say’ But: in Yakima Sahaptin (Penutian; Hargus & Beavert 2006), accented affix wins over root (18c,d): (18) Accented root + unaccented affix(es): affix surfaces on root a. ʔiʔatɬ’áwiʃa /ʔi + ʔatɬ’áwi + ʃa/ 2SG.NOM + beg + IMPRF b. ‘he’s begging him’ wánpanim /wánp + ani + m/ sing medicine song + BENEFACTIVE + CISLOCATIVE ‘sing for me’ Accented root + accented prefix(es): accent surfaces on prefix c. páʔatɬ’awiʃa /pá + ʔatɬ’áwi + ʃa/ INVERSE + beg + IMPRF ‘he’s begging him’ Accented root + accented suffix(es): accent surfaces on suffix d. wanpáwaas /wánp + áwaas/ sing medicine song + INSTRUMENTAL ‘sing medicine song’ Summary

• There are clear cases in which roots are immune to alternations that affixes undergo; perhaps this is the majority pattern (but unclear)

• There are certainly clear examples that go in the opposite direction. • Is the correct dichotomy root morphemes vs. affix morpheme, or is it bases of affixation

vs. the affixes that attach to those bases? • Are the patterns obeying RAFM more general than the patterns violating it? Hard to

know given current data


7

4. BEYOND ROOTS: MORPHOLOGICAL CONSTRUCTION-SPECIFIC PHONOLOGY The bulk of morphologically conditioned phonology resides in the association of phonological patterns with the individual morphological constructions which derive and inflect words.

4.1 AN EARLY OBSERVATION: ‘DERIVED-ENVIRONMENT EFFECTS’, OR NONDERIVED ENVIRONMENT BLOCKING (NDEB)

(19) “Some phonological rules apply freely across morpheme boundaries, and morpheme internally where fed by some earlier phonological rule, but are blocked elsewhere, in what are referred to as ‘nonderived environments’. [Kiparsky 1993:277]

(20) Finnish Assibilation: t → s / _ i

a. /vete/ → veti → vesi ‘water (nom.sg.)’ b. /vete-nä/ → vetenä ‘water (ess.sg.)’ (*vesenä) c. /halut-i/ → halusi ‘want-3P.SG.PRET’ /halut-a/ → haluta ‘want-INF’ d. /tilat + i/ → tilasi ‘order-3P.SG.PRET’ (*silasi) /tilat-a/ → tilata ‘order-INF’ (*silata) e. /æiti/ → æiti ‘mother’ (*æisi)

(21) Hausa palatalization (Newman 2000) a. ‘steal’ [saːt-aː] [Hausa] ‘steal (before noun object) [saːʧ-i] ‘steal (before pronoun object) [saːʧ-eː] b. ‘street’ [tiːtiː] There are lots of palatalization examples like this! Polish…Korean…. (22) Recurrence of this general pattern → attempts to capture it in the form of a single

principle. But what principle?

• What kinds of phonological patterns are affected? • What environments count as ‘derived’? • How general is the condition cross-linguistically, or for that matter within any given

language? (23) Morphologically derived environment effect: “a process that takes place only when its

conditions are crucially met by virtue of material from two different morphemes.” (McCarthy 2003:21)


8

(24) Canonical case: trigger in stem, target in suffix (separated by morpheme boundary). But: • Canonical nonderived environment blocking is related to… • noncanonical cases of nonderived environment blocking, which are themselves part of… • a larger landscape of phonological patterns restricted to particular morphological

constructions (morphologically conditioned phonology)

4.2 NONCANONICAL NDEB EFFECTS Tohono O’odham (Fitzgerald 2001)(Yu 2000) (25) a. Nonderived words kí: ‘house’ pí:ba ‘pipe’ ʔásugal ‘sugar’ pákoʔòla ‘Pascola dancer’ b. Suffixation ʔásugàl-t ‘to make sugar’ (oddpar.;finalstress) hím-ad ‘will be walking’ (evenpar.;nofinalstress) číkpan-dàm ‘worker’ (oddpar.;finalstress) músigò-dag ‘to be good at being a musician’ (evenpar.;nofinalstress) pímiàndo-màd ‘adding pepper’ (oddpar.;finalstress) c. Reduplication tó-toñ ‘ants’ (evenpar.;nofinalstress) pí-pibà ‘pipes’ (oddpar.;finalstress) mú-msigò ‘musicians’ (cf. músigo ‘musician’) sí-sminǰùɹ ‘cemeteries’ (cf. síminǰuɹ ‘cemetery’) pá-pkoʔòla ‘Pascola dancers’ (evenpar.;nofinalstress) d. Suffixation and reduplication híː-him-àd ‘will be walking, pl.’ (oddpar.;finalstress) hí-hidòḍ-a ‘the cooking, pl.’ (evenpar.;nofinalstress) há-haiwàñ-ga-kàm ‘ones having cattle’ (oddpar.;finalstress) Differs from canonical cases:

• Final stress placement in odd-parity forms is not attributable to a nearby morpheme boundary (e.g. in prefixed forms like pí-pibà)


9

Turkish minimality (Itô & Hankamer 1989, Inkelas & Orgun 1995) (26) a. do ‘note do’ ham ‘unripe’

be ‘letter b’ yen ‘alight (!)’ ye ‘eat (!)’ ok ‘arrow’ de ‘say (!)’

b. *do-m ‘note do-1SG.POSS’ cf. araba-m ‘car-1SG.POSS’ *be-n ‘letter b-2SG.POSS’ cf. elma-n ‘apple-2SG.POSS’ c. *de-n ‘say-PASS = be said (!)’ cf. anla-n ‘understand-PASS’ *ye-n ‘eat-PASS = be eaten (!)’ cf. çine-n ‘chew-PASS’

Differs from canonical cases:

• There is not a clear phonological process involved • There is not a clear sense in which the environment for the imposition of the minimal size

constraint is crucially provided by two morphemes.

4.3 NDEB CASES ARE NOT ALWAYS AS NEAT AS THEY APPEAR In Finnish, Anttila (2009), citing Karlsson 1983, notes that Assibilation applies before some /i/-initial suffixes and not before others: (27) a. Suffixes triggering Assibilation [Finnish] Plural /-i/: /vuote-i-nA/ → vuosina ‘year-PL-ESS’ Past tense /-i/ /huuta-i-vAt-kO/ → huusivatko ‘shout-PAST-3P.PL-QUE’ Superlative /-impA/ /uute-impA- nA/ → uusimpana ‘new-SUP-ESS’ b. Suffixes not triggering Assibilation Instrument /-ime/ /lentä-ime-n/ → lentimen ‘fly-INST-GEN’ Conditional /-isi/ /tunte-isi/ → tuntisi ‘feel-COND’ c. Suffix that optionally triggers Assibilation N→Adj /-inen/ /vete+inen/ → vesinen ∼ vetinen ‘water-ADJ = watery’

(28) In Turkish, disyllabic minimality is not imposed by the aorist suffix -r: de-r ‘say-AOR’,

etc.

4.4 REFLECTION What do we make of all this noncanonicity? Suggestion: understand ‘nonderived environment blocking’ as a special case of a more general phenomenon, namely that morphological constructions differ in whether or not they are associated with particular phonological patterns of alternation.

• Some constructions trigger patterns; others do not. • Monomorphemic stems, by virtue of not entering into other constructions, are naturally

exempt. • Let’s take a tour of morphologically conditioned phonological effects of this kind… • and then we’ll step back (next time) and try to make some sense of it all, in a theoretical

context


10

4.5 SEGMENT DELETION Turkish: vowel hiatus at stem-suffix boundaries is usually repaired by glide epenthesis, as illustrated below by /-Iver/. But vowel hiatus created by suffixation of the /-Ijor/ is resolved by vowel deletion: (29)

C-final root V-final root jap ‘do’ gel ‘come’ anla ‘understand’ søjle ‘say’ Facilitative /-Iver/: epenthesis jap-ɯver gel-iver anla-jɯver søyle-jiver Progressive /-Ijor/: deletion jap-ıjor gel-ijor anl-ɯjor søyl-yjor Nanti (Kampan; Michael 2008): VV hiatus at prefix-stem boundary is usually resolved by deletion of the prefix vowel (30a-b), but if the prefix is 1st person inclusive subject /a-/, the second vowel deletes (c-d):

(30) a. /no=am-e/ → name [Nanti] 1S=bring-IRREAL.I ‘I’m going to bring’ cf. /no=keNkitsa-k-i/ → nokeNkitsatake 1S=tell.story=PERF-REALIS.I ‘I told a story’ b. /pi=ogi-ratiNk-e=ro/ → pogaratiNkero 2S=CAUS-stand.up-IRREAL.I=3NMO [pogaɾatiŋkseɾo] (*piogiaratiNkero) ‘You will stand it up (e.g. a housepost) (polite imperative)’ cf. /pi=n-kem-e/ → pinkeme 2S=IRREAL-hear-IRREAL.I ‘you didn’t hear it c. /o=arateh-an-ak-i/ → aratehanake 3NMS=wade-ABL-PERF-REAL.I (*oratehanake) ‘She waded away’ d. /a=obiik-eNpa/ → abiikeNpa 1PL.INC.S=drink-IRREAL.A (*obiikeNpa) ‘let’s drink!’ e. /a=N-obiik-eNpa oburoki/ → abiikeNpa oburoki 1PL.INC.S=IRREAL-drink-IRREAL.A manioc.beer (*obiikeNpa oburoki) ‘Let’s drink manioc beer!’

4.6 GEMINATION Hausa: prefixing pluractional (31a) and intensive adjective (31b) CVC reduplication are associated with total assimilation (gemination); other prefixes are not (e.g., (31c)) (Newman 2000):

(31) a. ‘beat’ búgàː → búbbúgàː ‘be well repaired’ gʲàːrú → gʲàggʲàːrú ‘drink’ ʃáː → ʃáʃʃáː b. ‘brittle’ gáutsíː → gàggáutsáː ‘strong’ ƙárfíː → ƙàƙƙárfáː ‘salty, brackish’ zár ̃tsíː → zàzzár ̃tsáː c. ‘DIM-work’ ɗan-táɓà ‘PROHIBITIVE-2M.SG = don’t you!’ kár ̃-kà


11

4.7 VOWEL LENGTHENING Turkish: place name-forming -iye lengthens /a/ in stem-final open syllable; other suffixes do not

(32) Orthography UR Nominative Accusative (/-I/) as place name in /-Ije/ Murad (name) /murad/ [murat] [muradɯ] [muraːdije] refah ‘comfort’ /refah/ [refah] [refahɯ] [refaːhije] Ümran (name) /ymran/ [ymran] [ymranɯ] [ymraːnije]

4.8 TRUNCATION TO A PROSODIC CONSTITUENT Swedish nicknames: truncate to maximal syllable + suffix (Weeda 1992:121):

(33) a. alkoholist → alk-is ‘alcoholic’ laboratori:um → labb-is ‘lab’ b. mats → matt-e (proper name) fabian → fabb-e (proper name) Japanese illustrates another: truncate (or lengthen) to two moras, + /-ʧan/ (Poser 1984, 1990; Itô 1990) (34) Girls nickname formation: a. akira → aki-tyan (C)VCV megumi → megu-tyan b. syuusuke → syuu-tyan (C)VV taizoo → tai-tyan ti → tii-tyan c. kinsuke → kin-tyan (C)VC d. midori → mii-tyan, mit-tyan, mido-tyan (C)VV ~ (C)VC ~ (C)VCV kiyoko → kii-tyan, kit-tyan, kiyo-tyan Affix-triggered truncation is most often found in hypocoristics and vocatives, constructions where ambiguity is relatively unproblematic. But: Rarámuri (Uto-Aztecan) denominal suffix -tá truncates trisyllabic stems to two syllables (Caballero 2008:125-26, 310) (a-b). Not all suffixes do this (cf. c-e): (35) a. sipu-tá-a čukú (< sipúča ‘skirt’) skirt-VBLZ-PROG bend ‘(She is) putting on a skirt’ b. komá-ti-ma (< komáre ‘comadre’) comadre-VBLZ-FUT:SG’ c. tiyópi-či church-LOC d. banisú-ki-ni-ma pull-APPL-DESID-FUT:SG ‘will want to pull for’ e. wikará-n-čane sing-DESID-EV ‘it sounds like they want to sing’


12

4.9 ABLAUT AND MUTATION Vowel ablaut or consonant mutation: morphologically indexed featural alternations that are too complex or opaquely conditioned to be treated as simple phonology (e.g. German Buch ‘book’ ~ Büch-er ‘books’, Koch ‘cook’ ~ Köch-e ‘cooks’).

(36) Hua (Papuan; Haiman 1972, 1998): certain suffixes trigger the fronting of stem-final /o, u/ to /e, i/. Other phonologically similar suffixes do not. Data from Haiman 1972 (pp. 36-41)

Basic stem Suffixed stem gloss [Hua] ‘eat’ do- de-ra-’e ‘2 DL. have eaten’ de-na ‘when I eat (in the future)’ cf. do-ga ‘when (non-1st, non-sg) eat (future) cf. do-bai-na ‘when I eat (in the future)’ ‘do’ hu- hi-s-u (<hu-s-vu) ‘may do’ cf. hu-re-s-u (<hu-ro-s-vu) ‘may do (perfective)’ cf. hu-bai-s-u (< hu-bai-s-vu) ‘may do (progressive)’

4.10 DISSIMILATION AND ‘EXCHANGE’ RULES Morphologically conditioned ‘exchange rules’, ‘toggles’, dissimilation: one segment surfaces with a value opposite either to its own input value or to the the output value of another segment in the same word (Weigel 1993, Kurisu 2001, Baerman 2007, DeLacy 2012).

(37) Kↄnni Class 1 nouns pluralize by means of a tonally polar suffix (-a~ -e) (Cahill 2004):

gloss stem tone singular plural [Kↄnni] a. ‘fish’ H sí-ŋ sí-à ‘house’ H tígí-ŋ tíg-è ‘face mark’ H wí-ŋ wí-è b. ‘breast’ L bììsí-ŋ bììs-á ‘stone’ L tǎ-ŋ tàn-á (38) Dholuo: plural -e suffix triggers voicing dissimilation in the stem (Tucker 1994, DeLacy 2012):

gloss singular plural [Dholuo] a. ‘open space’ alap ӕlӕb-e ‘hill’ gɔt gɔd-ɛ ‘chest’ agɔkɔ agɔg-ɛ b. ‘book’ kitӕbu kitep-e ‘twig’ kɛdɛ kɛt-ɛ ‘year’ hɪga hik-e DeLacy 2012: it is hard to find convincing examples of morphophonological polarity. Most, including Dholuo, are ridden with lexical exceptions.


13

4.11 STRESS/PITCH-ACCENT (RE)ASSIGNMENT In Hausa, some morphological constructions replace stem tone with a tonal melody, composed of H and L tones, which associates to the syllables of the base in a predictable manner (Newman 1986, 2000):

(39) a. Suffixation with tone replacement (various plural classes) [Hausa] máːlàm → màːlàm-ái ‘teacher-PL’ -LH rìːgáː → ríːg-únàː ‘gown-PL’ -HL tàmbáyàː → támbáy-óːyíː ‘question-PL’ -H b. Suffixation without tone replacement (various) dáfàː → dáfàː-wá ‘cook-PPL’ -LH gàjéːréː → gàjéːr-ìyáː ‘short-FEM’ -LH hùːláː → hùːlâ-ř ‘hat-DEF’ -L In Japanese, ‘recessive’ morphological constructions preserve lexical stem accent; ‘dominant’ erase it (e.g., Poser 1984). If a recessive accented affix combines with an accented stem, ‘Leftmost Wins’ (b, c):

(40) Recessive affixes [Japanese] a. Unaccented (p. 49)

yóm- → yóN-da ‘read’ yob- → yoN-da ‘called’ b. Accented (p. 48) yóm- → yóN-dara ‘if he reads’ yob- → yoN-dára ‘if he calls’ c. Preaccenting (p. 54) áNdoo → áNdoo-si ‘Mr. Ando’ nisímura → nisímura-si ‘Mr. Nishimura’ matumoto → matumotó-si ‘Mr. Matsumoto’

(41) Dominant affixes a. Unaccented suffix (p. 72) kóobe → koobe-kko ‘an indigené of Kobe’ nágoya → nagoya-kko ‘an indigené of Nagoya’ nyuuyóoku → nyuuyooku-kko ‘an indigené of New York’ b. Accented suffix (p. 49) abura → abura-ppó-i ‘oil, fat/oily’ yásu → yasu-ppó-i ‘cheap/cheap, tawdry’ adá → ada-ppó-i ‘charming/coquettish’ c. Pre-accenting suffix (p. 55) nisímura → nisimurá-ke ‘the Nishimura family’ ono → onó-ke ‘the Ono family’ hára → hará-ke ‘the Hara family’ d. Post-accenting prefix (p. 57) futatu → map-pútatu ‘two/exactly half’ sáityuu → mas-sáityuu ‘amidst/in the very midst of’ syoozíki → mas-syóoziki ‘honesty/downright honest’


14

5. SUBSTANCE OF MORPHOLOGICALLY CONDITIONED PHONOLOGY What kind of phonology can be morphologically conditioned?

• Any kind of phonological pattern, other than the most low-level allophonic alternations • Morphological conditioning is norm for unnatural phonology (Spencer 1998). • Smith (2001, 2011): part-of-speech-sensitive phonology tends to be prosodic in nature.

(see similar observations re phonologically optimizing suppletive allomorphy (Paster 2009) and conditions on infix placement (Yu 2007))

6. CAN MORPHOLOGICALLY CONDITIONED PHONOLOGY REDUCE TO AUTOSEGMENTAL AFFIXATION?

Goal of many autosegmental analyses in the 1970’s and 1980’s: account for apparent morphological conditioning by complicating phonological representations

(42) Turkish: certain roots, suffixes are exceptions to palatal and rounding harmony

Clements & Sezer 1982: ‘we examine various types of exceptions to the principles of vowel and consonant harmony … and show that these follow from the existence of "opaque" vowels and consonants in phonological representations.’ [p. 221]

a. Norm: palatal and labial harmony in suffixes and epenthetic vowels i. al-dɯ ‘take-PAST’ gel-di ‘come-PAST’ ii. altɯ-da ‘six-LOC’ dørt-te ‘four-LOC’ iii. film-i ‘film-ACC’ filim ‘film’ iv. kojn-u ‘bosom-ACC’ kojun ‘bosom’

b. Exceptionally disharmonic suffixes i. al-ɯjor ‘take-PROG’ (progressive suffix with invariant [o]) gel-ijor ‘come-PROG’ ii. altɯ-gen ‘six-GON’ (polygon former) dørt-gen ‘four-GON’ c. Roots which trigger exceptional front harmony on suffixes infilaːk-i ‘explosion-ACC’ *infilaːk-ɯ harb-i ‘war-ACC’ *harb-ɯ saat-i ‘watch/hour-ACC’ *saat-ɯ d. Disharmonic epenthetic vowels (cf. aiv) vakit vakt-i ‘time(-3poss)’ kavim kavm-i ‘tribe(-3poss)’


15

(43) C&S’s solution: harmony accomplished by feature-filling association conventions, per Goldsmith 1976. Exceptional segments are prelinked to the features that they surface with; harmonic segments are lexically unspecified.

• Associate free (unspecified) P-features (“P” = [round], [back]) with free P-bearing segments in a 1-to-1, left-to-right fashion

• Associate remaining free P-bearing unit with leftmost available P-feature

→

ip-in, ip-ler, son-un, son-lar

a,b) Vowel disharmony in roots and suffixes: prelinking of [back], [round] values

(p. 228)

(p. 232)

c,d) Roots which trigger exceptional front harmony on epenthetic vowel, suffixes: prelink [-back] to appropriate root consonant

(p. 239)

(p. 245)

(p. 246)

• In the above example, floating and prespecified autosegmental features make it unnecessary to index the harmony rule to particular roots or suffixes.

• Can this approach work for all cases of morphologically conditioned phonology? o consonant gemination: affix a C? o vowel lengthening: affix (infix) a V? o ablaut, mutation: add a floating feature? o truncation: add a prosodic unit (bimoraic syllable)? o dissimilation, exchange: ? o stress deletion: ? o tone replacement: ? o segment deletion: ? (see Zimmerman 2013)

220 George N. Clements and Engin Sezer

(7) nom. singular: nom. plural: a. -R -R -R

I Ip In Ip IE r

-B -B

b. +R +R -R I

sEn In sEn IEr

+B +B

Here, in accordance with (4), nonhigh vowels are represented as opaque on the roundness tier. Otherwise, suffixes have no P-segments in their representation.s Each root, on the other hand, has a P-segment in its re-presentation for each autosegmental tier. Since root vowels have not been defined as opaque,6 there are no underlying associations between root vowels and root P-segments. Association Conventions (Sa) and (Sb) apply in succession to create the following output forms:

(8) a. -R -R -R . " I I I " , I I , I I

Ip In Ip I Er " I I I " I I , .-

I I " I I I"

-B -B

b. +R +R -R " I I , , , , , , " I

sEn In sEn I Er I ,- " I I' " I "

, I

" ,-

II' I I'

+B +B

Notice that no rule has applied in deriving the surface forms of (8) from the underlying forms of (7). The rules defining Turkish vowel harmony, as stated in (3) and (4), are structure-building rules rather than feature-changing rules.

The special properties of opaque vowels become apparent if we consider the derivation of the genitive plural of these two forms. The underlying forms are as follows:


(7) nom. singular: nom. plural: a. -R -R -R

I Ip In Ip IE r

-B -B

b. +R +R -R I

sEn In sEn IEr

+B +B

Here, in accordance with (4), nonhigh vowels are represented as opaque on the roundness tier. Otherwise, suffixes have no P-segments in their representation.s Each root, on the other hand, has a P-segment in its re-presentation for each autosegmental tier. Since root vowels have not been defined as opaque,6 there are no underlying associations between root vowels and root P-segments. Association Conventions (Sa) and (Sb) apply in succession to create the following output forms:

(8) a. -R -R -R . " I I I " , I I , I I

Ip In Ip I Er " I I I " I I , .-

I I " I I I"

-B -B

b. +R +R -R " I I , , , , , , " I

sEn In sEn I Er I ,- " I I' " I "

, I

" ,-

II' I I'

+B +B

Notice that no rule has applied in deriving the surface forms of (8) from the underlying forms of (7). The rules defining Turkish vowel harmony, as stated in (3) and (4), are structure-building rules rather than feature-changing rules.

The special properties of opaque vowels become apparent if we consider the derivation of the genitive plural of these two forms. The underlying forms are as follows:


(20) opaque segments: [+syllabic, +root]

This ensures underlying representations like the following (cf. (12) ):

(21) -R +R +R -R +R -R /\/ III IstEkEz ErkInEs I V I 1\

-B +B 'lobster'

+B -B +B 'funny fish'

IzmErIt I I \

-B +B -B 'sea-bream'

This treatment is consistent with the analysis of opaque vowels given earlier. Recall that opaque vowels were characterized as "nonundergoers," "blockers," and "spreaders". Root vowels are nonundergoers; thus a root vowel does not harmonize with the preceding vowel. Root vowels are blockers; thus a suffix vowel cannot harmonize with a nonfinal root vowel, except coincidentally. Root vowels are spreaders; thus a final root vowel determines the harmonic category of the immediately following harmonic suffix vowel. The last two points are illustrated in (22) for the genitive singular of orkinos 'tunny fish':

(22) +R -R +R I I /\ ErkInEs In I I \/'

[

+B -B +B

It can be deduced from (3), (4), and (20) together with the universal Well-formedness Conditions that harmonic suffix vowels following a root will harmonize with the last root vowel, a claim which is correct for Tur-kish.

We now introduce the following constraint on vowel cooccurrence within single morphemes:

(23) The vowels Iii, 0, +1 do not occur disharmonically in VCo V se-quences, except that Ii, iii may occur in either order.

This statement admits roots of the type ICuCaC+I while excluding roots of the type ICuC+Cal and ICaC+Cu/, since only the latter involve harmony violations involving Iii, 0, +1 in disyllabic subsequences. This formulation seems intuitively correct, although we have not found examples of dishar-monic roots of the first type. This statement also provides for the subre-gularities involving Ii, iii (see (14) ). The first part of (23) may be restated formally as an if-then condition holding of single morphemes:


(32) -R I

gEl I

-B

+R I

lyEr 1m I

+B

Association Convention (5b) is applicable. Notice that the condition of precedence (6), incorporated into the statement of (5b), uniquely deter-mines the pattern of association shown below:

(33) -R +R I ''''''' / \

gEl lyEr 1m = \ '" \/ / /

,-

-B +B More generally, we find that (6) is always correct for Turkish: opaque segments, whether vowels or (as we shall see) consonants, always govern the harmonic category of a harmonic vowel to their right. We might call (6) the Principle of Inertia, according to which an articulatory state de-termined by a particular feature configuration is maintained until a new specification (or set of specifications) is encountered. This principle need not be stated as a condition on rule application. As noted earlier, this principle, together with the left-to-right mapping convention (Sa), ex-plains the common phonological bias toward spreading from the left. As a convention (rather than a language-particular rule condition), it expresses the "unmarked" case of spreading which can only be overruled by a language-particular statement taking precedence over it. l2

The constraint ruling out disharmonic sequences with /ll, 0, i:/ in roots holds of polysyllabic suffixes as well. A small number of such suffixes contain two opaque vowels. The suffix /-istan/, illustrated in (31c), is one, and the others are:

(34) -ane -vari -leyin -iyet

denominal, adjective-forming denominal, adverb-forming denominal, adverb-forming denominal, noun-forming

In these suffixes the disharmonic vowels are drawn exclusively from the set Ii, e, a/. More generally, opaque vowels in suffixes are always one of the following: Ii, e, a, 0, u/; the vowels Iii, 0, i:/ do not occur as opaque segments outside roots. We thus have the following condition holding of suffixes in underlying representation:

(35) /ll, 0, i:/ are prohibited in suffixes.

, I

:

1

I

I !' I i,

'I

II

1\

II [-

a

;-

II

" '" t

1-

1-

g :)

Disharmony in Turkish 239

(48) show that opaque consonants do not undergo the Consonant Har-mony rule (44), and are thus "nonundergoers". In this section we turn to evidence confirming our prediction that nonharmonic consonants are also blockers and spreaders.

Consider first the velars. We have seen examples of disharmonically palatal Ikl in initial and medial position in the word, but no examples of this con;onant word-finally. Indeed, this is a genuine gap in the surface distribu tion of this form in the dialects we have examined which remains to be accounted for.

There is a further idiosyncracy regarding roots with opaque palatal 115:1. We find a small number of Turkish words whose final syllables have back vowels and which govern front vowel harmony, and whose final consonant is [k] word-finally, or before a consonant-initial suffix, and

·before a vowel-initial suffix. A partial list includes the following:

(57) nom. sg. acc. sg. 'explosion' infilak infilaki 'perception' idr;k idrfkf ... 'alliance' ittifak ittifiiki 'participation' istirak istiraki .. 'fasting' imsak imsaki 'expropriation' istimlak istimlaki ,. .. .. 'real estate' emlak· emlaki

hel;k· ,. ,.

'exhaustion' helaki .. ... .. 'addiction' inhimak inhimaki ... 'consumption' istihlak· istihlaki ,. .. ..

Most of these words are ultimately of Arabic origin, and originally ended in nonemphatic velars. In most examples, the penultimate vowel is front and the final vowel is long back la/.

Now it will easily be seen that if we permit opaque palatal velars to occur freely in underlying representations - and in particular, to occur finally in the underlying representations of the stems in (57) - the front vowel quality of the suffixes will be instantly accounted for. Thus, assuming the final consonant of idrak to be opaque, we have the following repre-sentations of the nominative and accusative singulars, respectively:

(58) -B +B -B I 1/

-B +B -B \ l /\

IdrEk IdrEk I

By Association Convention (5b), the P-segment associated with the final consonant of the accusative singular becomes associated with the suffix

ld .r, to Ie er :s. at 1t

ts

3. re a-

ly Le )-

IS s.

Disharmony in Turkish 245

(73) 'gunpowder' barut barutu baruttan 'strange' garip garibi garipten 'ambitious' haris harisi haristen 'outfit' k414k k4lt4 k4ltktan 'lifebuoy' simit simidi simitten

Permissable syllable-final clusters in Turkish are of the following types: 20

(74) a. sonorant + obstruent: tUrk 'Turk', gene 'young' b. voiceless fricative + oral stop: eift 'double', ask 'love' c. k + s: raks 'dance', boks 'boxing'

Under the deletion analysis there is no way to capture the relationship between the forms of (73) and the generalizations expressed in (74) short of postulating some sort of transderivational constraint on the deletion rule that would prohibit its operation just in case it would give rise to a cluster that, if syllable final, would be well-formed. Such a solution is clearly unacceptable.

In sum, traditional approaches place us in a dilemma. The epenthesis analysis forces us to introduce independent diacritic features to describe forms that are exceptional both in regard to epenthesis and to suffix harmony, and fails to explain why forms that are exceptional with regard to one are also exceptions with regard to the other. The deletion analysis is unable to relate a phonologically-motivated class of exceptions to de-letion to independently-needed constraints on syllable structure.

Under an autosegmental analysis, two solutions are potentially available in the face of data such as this. One solution is to postulate root-final float-ing P-segments in the case of the forms in (70); the other is to treat the second consonant of each root of (70) as opaque. Both solutions assume the epenthesis rule. Thus, the alternative analyses can be represented as follows:

(75) a. +B-B I

b +B -B 1/

vEkt vEkt

If all else were equal, the evaluation metric would select (75a) over (75b) as the simpler of the two representations. As it happens, however, all else is not equal. A consideration of the ablative forms shows us that only the second solution, positing an opaque consonant in C2 position, accounts for the surface form:


(76) +B-B I / '\

vEkt tEn (by (5b))

+B -B I A

vEkIt tEn (by (78))

+B-B /1\

vEklttEn (by (5b))

The first solution, positing a floating P-segment, fails:

(77) +B -B / \ ,

\

vEkt tEn (by (Sa))

+B -B I I

vEklt tEn (by (78))

+B -B (wrong I output)

vEklt ·tEn (by (5 b), (44)) .

The Epenthesis rule involved in the above derivations is the following:

(78) Vowel Epenthesis

(/J I I C_C' (C' = an extrasyllabic consonant)

(78) inserts an epenthetic high vowel, unassociated with any features on the autosegmental tiers involved in harmony, between two consonants if the second can form neither a syllable onset nor a syllable coda by the syllable structure rules of Turkish.21 The inserted vowel then undergoes the normal operation of the Association Conventions (5).22

Given this analysis, then, we see that a further class of monosyllabic noun roots of the structure CVCC must be recognized, where the second C is opaque. But it will be recalled that this is not in fact a new structural type of Turkish root, since such roots have already been motivated: see (63), and in particular (66), where we find a unique set of occurrences of palatal IrJ. Thus, roots of the type (7 5b) come cost-free in our analysis, and indeed fill in an otherwise unexplained gap in the distribution of opaque segments.

We have so far examined cases of disharmonic epenthetic "vowels inserted into root-final clusters. There is another source of epenthetic vowels in Turkish consisting of root-initial clusters, as the representative forms in (79) illustrate. It will be noted that each word (consisting of un-inflected roots) shows two or more variants. The first variant, containing the cluster, generally reflects a careful or learned pronunciation. The second and subsequent forms represent normal or colloquial pronuncia-tions. In the latter forms, which are the more usual, we observe a short epenthetic vowel between the two members of the cluster. Under contras-tive emphasis, these short vowels may receive the full value of normal (short) vowels.23


16

Next lecture: theories of morphologically conditioned phonology. • How many types of morphologically conditioned phonological patterns can there be in a

language? • How diverse can the patterns be? • If two affixes in the same word are associated with distinct phonological patterns, which

prevails? Some references

Alderete, John. 2001. Dominance effects as transderivational anti-faithfulness. Phonology 18(2). 201–253.

Allen, Margaret. 1978. Morphological investigations. University of Connecticut. Baerman, Matthew. 2007. Morphological reversals. Journal of Linguistics 43(1). 33–61. Becker, Michael & Maria Gouskova. 2016. Source-oriented generalizations as grammar

inference in Russian vowel deletion. Linguistic Inquiry. 391–425. Bright, William. 1972. The Karok language. (University of California Publications in Linguistics

13). Berkeley: University of California Press. Caballero, Gabriela. 2008. Choguita Rarámuri (Tarahumara) Phonology and Morphology.

University of California, Berkeley Ph.D. dissertation. Cahill, Michael. 2004. Tone polarity in Konni: an Optimality Theoretic account. Ohio State

University working papers in linguistics 51. 19–58. Cammenga, Jelle. 2002. Phonology and morphology of Ekeguusii. Köln: Rüdiger Köppe Verlag. Casali, Roderic. 1997. Vowel elision in hiatus contexts: which vowel goes? Language 73. 493–

533. Chomsky, Noam & Morris Halle. 1968. The sound pattern of English. New York: Harper and

Row. Clements, G. N & Engin Sezer. 1982. Vowel and consonant disharmony in Turkish. In Harry van

der Hulst & Norval Smith (eds.), The structure of phonological representations, part II, 213–255. Dordrecht: Foris.

DeLacy, Paul. 2012. Morpho-phonological polarity. In Jochen Trommer (ed.), The morphology and phonology of exponence, 121–159. (Oxford Studies in Theoretical Linguistics 41). Oxford: Oxford University Press.

DuBois, John W. 1985. Incipient semanticization of possessive ablaut in Mayan. International Journal of American Linguistics 51(4). 396–398.

Fitzgerald, Colleen M. 2001. The Morpheme-to-Stress principle in Tohono O’odham. Linguistics 39(5). 941–972.

Goldsmith, John. 1976. Autosegmental Phonology. Massachusetts Institute of Technology PhD dissertation.

Goodenough, Ward H. & Hiroshi Sugita. 1980. Trukese-English Dictionary. Philadelphia: American Philosophical Society.

Haiman, John. 1972. Ablaut in the Hua verb. Oceanic Linguistics 11(1). 32–46. Haiman, John. 1998. Repetition and identity. Lingua 100. 57–70. Haraguchi, Shosuke. 1977. The tone pattern of Japanese. Tokyo: Kaitakusha. Hargus, Sharon & Virginia Beavert. 2006. High-ranking Affix Faithfulness in Yakima Sahaptin.

Proceedings of the 25th West Coast Conference on Formal Linguistics, 177–185. Cascadilla Press.


17

Inkelas, Sharon & Cemil Orhan Orgun. 1995. Level ordering and economy in the lexical phonology of Turkish. Language 71(4). 763–793.

Itô, Junko. 1990. Prosodic minimality in Japanese. In Michael Ziolkowski, Manuela Noske & Karen Deaton (eds.), Papers from the twenty-sixth regional meeting of the Chicago Linguistics Society. Volume 2: the parasession on the syllable in phonetics and phonology., 213–239. Chicago: Chicago Linguistics Society.

Itô, Junko & Jorge Hankamer. 1989. Notes on monosyllabism in Turkish. In Junko Itô & Jeff Runner (eds.), Phonology at Santa Cruz 1, 61—69. Santa Cruz: University of California, Santa Cruz Syntax Research Center.

Itô, Junko & Armin Mester. 1999. The structure of the phonological lexicon. In N Tsujimura (ed.), The handbook of Japanese linguistics, 62–100. Malden, MA: Blackwell.

Kenstowicz, Michael & Charles Kisseberth. 1979. Generative phonology: description and theory. San Diego: Academic Press.

Kiparsky, Paul. 1982a. Lexical morphology and phonology. In I. -S. Yang (ed.), Linguistics in the morning calm, 3–91. Seoul: Hanshin.

Kiparsky, Paul. 1982b. Word-formation and the lexicon. In Frances Ingemann (ed.), 1982 Mid-America linguistics conference papers, 3–32. Lawrence, Kan: University of Kansas.

Kiparsky, Paul. 1985. Some consequences of lexical phonology. Phonology Yearbook 2. 85–138. Kiparsky, Paul. 1993. Blocking in non-derived environments. In Sharon Hargus & Ellen Kaisse

(eds.), Phonetics and Phonology 4: Studies in Lexical Phonology, 277–313. San Diego: Academic Press.

Koutsoudas, Andreas. 1980. The question of rule ordering: some common fallacies. Journal of Linguistics 16(1). 19–35.

Lynch, John. 1978. A grammar of Lenakel. (Pacific Linguistics B55). Canberra: Australian National University.

McCarthy, John. 2003. Comparative markedness. Theoretical Linguistics 29. 1–51. McCarthy, John & Alan Prince. 1995. Faithfulness and reduplicative identity. In Jill Beckman,

Laura Dickey & Suzanne Urbanczyk (eds.), University of Massachusetts Occasional Papers in Linguistics 18: Papers in Optimality Theory, 249–384. Amherst, MA: GLSA.

McCawley, J. D. 1968. The phonological component of a grammar of Japanese. The Hague: Mouton.

Michael, Lev. 2008. Nanti evidential practice: Language, knowledge, and social action in an Amazonian society. University of Texas, Austin.

Moore-Cantwell, Claire & Robert Staubs. 2014. Modeling morphological subgeneralizations. In John Kingston, Claire Moore-Cantwell, Joe Pater & Robert Staubs (eds.), Proceedings of the 2013 Annual Meeting on Phonology. http://journals.linguisticsociety.org/proceedings/index.php/amphonology/article/view/42.

Muller, Jennifer. 1999. A unified mora account of Chuukese. In Sonya Bird, Andrew Carnie, Jason Haugen & Peter Norquest (eds.), Proceedings of the 18th West Coast Conference on Formal Linguistics, 393–405. Malden, MA: Cascadilla Press.

Newman, Paul. 1986. Tone and affixation in Hausa. Studies in African Linguistics 17(4). 249–267.

Newman, Paul. 1995. Hausa tonology: complexities in an “easy” tone language. In John Goldsmith (ed.), Handbook of Phonological Theory, 762–781. Cambridge, MA: Basil Blackwell.


18

Newman, Paul. 2000. The Hausa language: an encyclopedic reference grammar. New Haven: Yale University Press.

Paster, Mary. 2009. Explaining phonological conditions on affixation: Evidence from suppletive allomorphy and affix ordering. Word Structure 2(1). 18–47.

Pierrehumbert, Janet & Mary Beckman. 1988. Japanese tone structure. Cambridge, MA: MIT Press.

Poser, William J. 1984. The phonetics and phonology of tone and intonation in Japanese. MIT PhD dissertation.

Poser, William J. 1990. Evidence for foot structure in Japanese. Language 66. 78–105. Shih, Stephanie & Sharon Inkelas. 2015. Morphologically conditioned tonotactics in multilevel

Maximum Entropy grammar. University of British Columbia. Siegel, Dorothy. 1974. Topics in English morphology. Massachusetts Institute of Technology. Smith, Jennifer. 1999. Noun faithfulness and accent in Fukuoka Japanese. In Sonya Bird,

Andrew Carnie, Jason Haugen & Peter Norquest (eds.), Proceedings of WCCFL XVIII, 519–531. Somerville, MA: Cascadilla Press.

Smith, Jennifer. 2011. Category-specific effects. In Marc van Oostendorp, Colin Ewen, Beth Hume & Keren Rice (eds.), Companion to Phonology, 2439–2463. Malden, MA: Wiley-Blackwell.

Spencer, Andrew. 1998. Morphophonological operations. In Andrew Spencer & Arnold Zwicky (eds.), Handbook of Morphology, 123–143. Oxford: Blackwell.

Tsujimura, Natsuko. 1996. An Introduction to Japanese Linguistics. 1st ed. Blackwell. Tucker, A. N. 1994. A grammar of Kenya Luo (Dholuo). Köln: Rüdiger Köpper Verlag. Weeda, Donald Stanton. 1992. Word truncation in prosodic morphology. University of Texas,

Austin. Weigel, William. 1993. Morphosyntactic toggles. Papers from the 29th Regional Meeting of the

Chicago Linguistics Society, 467–478. Chicago: Chicago Linguistics Society. Willard, Rainbow. 2004. Dominance effects in a dialect of Mam Maya. UC Berkeley, ms. Yip, Moira. 1991. Coronals, coronal clusters and the coda condition. In C. Paradis & J. -F.

Prunet (eds.), Phonetics and Phonology 2: the special status of coronals. Academic Press. Yu, Alan C. -L. 2000. Stress Assignment in Tohono O’odham. Phonology 17(1). 117–135. Yu, Alan C. -L. 2007. A natural history of infixation. Oxford: Oxford University Press. Zimmerman, Eva. 2013. Non-concatenative allomorphy is generalized prosodic affixation: the

case of Upriver Halkomelem. Lingua 134. 1–26. Zuraw, Kie Ross. 2000. Exceptions and regularities in phonology. University of California, Los

Angeles.

Documents

Lecture 1: Morphologically conditioned phonology...Noun privilege: Smith’s (2011) survey finds overall that that nouns tend to exhibit more contrasts, while verbs are more prone