exploring karuk morphology in a parsed text corpus
TRANSCRIPT
Exploring Karuk morphologyin a parsed text corpus
Andrew Garrett, Erik Maier, Line Mikkelsen, Clare SandyUniversity of California, Berkeley
SSILA, Portland8 January 2015
Background Morphological parsing Example: Pluractional -va
Acknowledgements
ⶠKaruk elders, teachers, and activists: Tamara Alexander, LuluAlexander, Sonny Davis, Susan Gehr, Julian Lang, CrystalRichardson, Nancy Richardson, Bud Smith, Vina Smith, FlorrineSuper, Arch Super; and â Lucille Albers and â Charlie Thom, Sr.
ⶠTechnical collaboration and information: Ronald Sprouse
ⶠUC Berkeley students: Jeff Spingeld, Whitney White; Nico Baier,Shane Bilowitz, Kayla Carpenter, Anna Currey, Erin Donnelly,Kouros Falati, Matt Faytak, Nina Gliozzo, Morgan Jacobs, KarieMoorman, Olga Pipko, Melanie Redeye, Tammy Stark; and others
ⶠNSF (âKaruk [kyh] . . . syntax and text documentationâ, #1065620);Undergraduate Research Apprentice Program, UC Berkeley
ⶠThese slides: linguistics.berkeley.edu/âŒkaruk/resources/ssila-2015.pdf
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 2/34
Background Morphological parsing Example: Pluractional -va
Karuk text corpus
ⶠDatabase as of January 2015
ⶠ150 texts containing 23,511 words in 6,167 clausesⶠtexts recorded in work with numerous elders by A. L. Kroeber,
J. P. Harrington, Jaime de Angulo & Lucy Freeland, WilliamBright, Monica Macaulay, and the present authors
ⶠGenres
ⶠtraditional narratives, medicine textsⶠanecdotes, personal historyⶠprocedural texts, descriptionsⶠconversationⶠdocumentation sessions: responses to visual prompts (avoiding
English), responses to translation tasks (elicitation)
ⶠFor eventual incorporation
ⶠ>50 other legacy (transcribed) textsⶠ>160 hours of untranscribed Karuk recordings
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 3/34
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 4/34
Vina Smith in 2013
Photo: Florrine Super
Background Morphological parsing Example: Pluractional -va
Karuk lexicon database
ⶠHistory
ⶠ138-page lexicon in Brightâs The Karok language (1957)ⶠconverted to Shoebox by Susan Gehr & William BrightⶠWilliam Bright & Susan Gehr, Karuk dictionary (2005)ⶠconverted to XML format in 2000s; now a mySQL database
ⶠScope
ⶠ7,302 entriesⶠevery entry has a unique ID
ⶠLinked to text corpus
ⶠwords in texts are parsed morphologicallyⶠmorphological elements are tagged with lexicon ID numbers
ⶠAll lexicon and text resources are for (and are actively used by)academic and community researchers, learners, and teachers.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 5/34
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 6/34
Florrine Super in 2013
Background Morphological parsing Example: Pluractional -va
âKaruk Dictionary and Textsâ online: Home page
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 7/34
(http://linguistics.berkeley.edu/âŒkaruk)
Background Morphological parsing Example: Pluractional -va
âKaruk Dictionary and Textsâ: Dictionary search
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 8/34
Background Morphological parsing Example: Pluractional -va
âKaruk Dictionary and Textsâ: Search with examples
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 9/34
(continued)
Background Morphological parsing Example: Pluractional -va
âKaruk Dictionary and Textsâ: Text list
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 10/34
(continued)
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 11/34
Maggie Charley in 1948
Photo: Mary Jean Kennedy (Hearst Museum of Anthropology)
Background Morphological parsing Example: Pluractional -va
Text display: Paragraph mode
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 12/34
(continued)
Background Morphological parsing Example: Pluractional -va
XML text database fragment
ⶠFrom Maggie Charleyâs âIndian Foodâ (WB KL-68):
<txt lang=âkyhâ><w>
<m lemma=â4407â>pa</m> pa-âasiktavaan-sas<m lemma=â656â>âasiktavaan</m> âthe womenâ<m lemma=â5200â>sas</m>
</w><w>
<m lemma=â416â>apkaas</m> apkaas</w> âiris (leaves)â<w>
<m lemma=â3977â>kun</m> kun-âıshum-tih<m lemma=â3130â>âıshum</m> âthey scraped (it)â<m lemma=â6035â>tih</m>
</w>.</txt>
ⶠEach âlemmaâ tag links to a lexicon ID number.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 13/34
Background Morphological parsing Example: Pluractional -va
Text display: âWord componentsâ mode
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 14/34
(continued)
Background Morphological parsing Example: Pluractional -va
Non-hierarchical text parsing
ⶠText parsing with no internal hierarchical structure:
vıriso
vaaso
kıchjust
kun-kupı-tih-anik3S-do-DUR-ANC
pa=kun-par-ıshriih-va-naa-tih-anikNOMZ=3PL-ROOT-down-PLACT-PLUR-DUR-ANC
âThey were just doing this: twining.â (KS 41.003)
Even a word with seven internal morphemes and a procliticlacks internal hierarchical structure.
ⶠThis is typically true of not only the visual display but theunderlying data structure, but it is not faithful to linguisticstructure . . .
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 15/34
Background Morphological parsing Example: Pluractional -va
Non-compositional meanings
ⶠComplex words transmit semantic idiosyncrasies (conventionalnoncompositional meanings) to their further derivatives.
ⶠThis most often occurs with less semantically predictable wordformationâe.g. compounds, denominal verbs, deverbal nouns,diminutives, but not e.g. causatives or productive directionals.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 16/34
Base Derivative word Further derivative
asiv âto sleepâ asım-chak âto close eyesâ asimchak-chak-veen-ach âwrenâ(âsleep-closing.upâ) (âclose.eyes-RED-AGT-DIMâ)
ih âto danceâ ıh-uk âto do a flower danceâ ıhuk-a âa flower danceâ(âdance-hitherâ) (âto.flower.dance-DEVRBâ)
ishvırip ishvınip-ich âplace nameâ peeshvınipich /pa=ishvınipich/âJeffrey pineâ (âJeffrey.pine-DIMâ) âmanâs nameâ (from ishvınipich)
(âDEF-ishvınipichâ)
vıitkira âridgeâ viitkır-ik âBald Hillsâ vitkirık-thuuf âRedwood Creekâ(âridge-LOCâ) (âBald.Hills-creekâ)
Background Morphological parsing Example: Pluractional -va
Non-compositional meanings: Compounds
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 17/34
Base Compound word Further derivative
aas âwaterâ as-iktav-aan âfemaleâ musmus-âasiktavaan âcowâiktav âcarryâ (âwater-carry-AGTâ) (âcattle-femaleâ)
apxaan âhatâ apxan-tınih-ich âwhite manâ apxantıich-puufich âsheepâ(âhat-wide-DIMâ), (âwhite.man-deerâ)reduced apxantıich apxantiich-tayiith âpotatoâ
(âwhite.man-brodiaeaâ)
asa ârockâ asa-xuus ârubberâ asaxus-kutrahara âraincoatâ(ârock-smoothâ) (ârubber-coatâ)
thuuf âcreekâ thuf-karoom thufkaroom-thuufâplace near Rock Creekâ âRock Creekâ(âcreek-upriver.aboveâ) (âthufkaroom-creekâ)
uuh âtobaccoâ uh-thaam âgardenâ uhthaam-hi âto plantâthaam âmeadowâ (âtobacco-meadowâ) (âgarden-DENOMâ)
uux âbitterâ ux-âaas âliquorâ uxâas-iyeeshrıhv-aan(âbitter-waterâ) âbartenderâ (âliquor-sell-AGTâ)
Background Morphological parsing Example: Pluractional -va
Non-compositional meanings: Denominatives, deverbatives
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 18/34
Base Derivative word Further derivative
pathak âkneeâ pathak-hi âto kneelâ pathakhı-ram âplace by Katiminâ(âknee-DENOMâ) (âto.kneel-placeâ)
pathakhı-ishrih âto kneel downâ(âto.kneel-downâ)
thukin âgall, bileâ thukin-hi âto tattooâ thukinh-a âa tattooâ(âtattoo-DENOMâ) (âto.tattoo-DEVERBâ)
ihruv âto useâ ihrooha /ihruv-ahi-a/ ihnooha-hiich âcommon-law wifeââwifeâ (âuse-ESS-DEVRBâ) (âwife-make.believeâ)
ithyuru âto haulâ ithyur-a âautomobileâ ithyura-âaah âheadlightâ(âhaul-DEVRBâ) (âautomobile-fireâ)
matnus âto burstâ matnus-a âcottonâ matnusa-vaasa âquiltâ(âburst-DEVRBâ) (âcotton-blanketâ)
Background Morphological parsing Example: Pluractional -va
Non-compositional meanings: Problems for searching
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 19/34
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 20/34
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 21/34
Background Morphological parsing Example: Pluractional -va
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 22/34
Background Morphological parsing Example: Pluractional -va
XML text database fragment
ⶠFrom Mamie Offieldâs âThe Greedy Fatherâ (WB KL-24):
<w><m lemma=â3745â>kari</m> kari âthenâ</w><w><m lemma=â6847â>xas</m> xas âthenâ</w><w><m lemma=â5506â>t</m> tuâasimchishrihvunaa
<m lemma=â6175â>u</m> âshe put (them) to sleepâ<m lemma=â661â>âasimachishrih</m><m lemma=â6648â>vunaa</m>
</w><w><m lemma=â4407â>p</m> p=aaxiich
<m lemma=â920â>aaxiich</m> âthe child(ren)â</w>.
ⶠEach âlemmaâ tag links to a lexicon ID number.
ⶠThe underlined verbal stem is morphologically complex:asiv âsleepâ + causative -math + ishrih âdownâ.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 23/34
Background Morphological parsing Example: Pluractional -va
A morphologically complex stem in the lexicon
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 24/34
Background Morphological parsing Example: Pluractional -va
Morphological parsing: SummaryⶠLexicon database
ⶠEvery attested stem (base of inflection) has an entry.ⶠThe coded ingredients of any complex stem are its maximal
(attested) components. For example:
ⶠıhukvunaa âa flower danceâ = ıhukvunaa + DEVRB -aⶠıhukvunaa âto have a flower danceâ = ıhuk + PL -vunaaⶠıhuk âto do the flower danceâ = ıh âto danceâ + -uk âhitherâ
ⶠText databaseⶠThe coded ingredients of any complex form are its inflectional
affixes and its stem.ⶠThus tuâasimchishrihvunaa âshe put (them) to sleepâ =
PERF t- + 3SG u- + âasimachishrih âput to sleepâ + PL-vunaa (where âasimachishrih is a complex stem in thelexicon).
ⶠThereby search queries are possible on minimal morphs orcomplex forms.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 25/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va (Bright 1957, Conathan & Wood 2003)
ⶠRepresentative examples
ⶠxasthen
karithen
aseemfirrock:hot
t-u-turu-raamnih-vaPERF-3SG-handle-into-PLACT
âThen he put in hot stones.â (Julia Starritt, WB KL-81)ⶠchavura
finallyitaharaanten.times
taPERF
kun-p-arihıshriih-va3PL-ITER-sing-PLACT
âFinally they sang ten times.â (Nettie Reuben, WB KL-02)
ⶠHow is -va ordered with respect to other suffixes?
ⶠScope?ⶠOn the typological parameter space: Rice (2011)
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 26/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va position: Immediately after the root
ⶠRoot + PLACT (AM = Associated Motion)
ⶠxasthen
u-kfuy-v-uunish3SG-whistle-PLACT-at
âAnd she whistled (a signal) at him.â (Lottie Beck, WB KL-35)ⶠchımi
soonni-maah-v-ar-eesh1SG-see-PLACT-AM-FUT
âIâm going to go visiting.â (Bright & Gehr 2005 s.v.maahva)
ⶠInternally complex eventuality with sub-parts
ⶠikfuuy-va âwhistle a signalâ (complex whistling)ⶠmaah-va âvisitâ (complex seeing)
ⶠOther suffixes have scope over Root + PLACT
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 27/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va position: Surrounded by other suffixes
ⶠDirectionals + AM + PLACT + TAM
ⶠuknamxanahichplacename
chımisoon
nu-âıh-uk-an-v-eesh1PL-dance-hither-AM-PL.ACT-FUT
âWeâll go flower-dance at uknamxanahich.â(Nettie Reuben, WB KL-2a)
ⶠpa=kaanNOMZ=there
kun-âaxup-ruuprih-vu-ti3PL-put.deermeat-in.through-PLACT-DUR
pa=puufichthe=deer
t-ooPERF-3SG
mtaap-hadust-DENOM
âWhere they used to put in the dressed deer meat was dusty.â(Mamie Offield, WB KL-33)
ⶠMultiple event instances or participants in a temporal window
ⶠmultiple dancersⶠmultiple instances of putting deer meat somewhere
ⶠCompositional interpretation; no added conventional meaning
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 28/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va position: After TAM
ⶠTAM + PLACT
ⶠpeheeraha-âıpathe:tobacco-tree
pa=koothe=all
u-thvuy-tiih-va3SG-be.named-DUR-PLACT
pa=mu-svita-vathe=3SG.POSS-part-INDEF
âhow all the tobacco plantâs parts are (variously) namedâ(JPH TKIC-III.2)
ⶠkarumain.fact
nıka.little
apxan-yaamach-ascap-pretty-PL
taPERF
kun-p-ithxuna-tiih-va3PL-ITER-put.over.head-DUR-PLACT
In fact they wear pretty caps.â (Lottie Beck, WB KL-30)
ⶠMultiple (e.g. simultaneous) imperfective eventualities
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 29/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va position: A near-minimal pair
ⶠDUR (PLACT): multiple event instances in a temporal window
paythis
peethıvthaaneenthe:land
thaaneenaround
ni-p-thivruh-iroopith-vu-tih1SG-ITER-float-around-PLACT-DUR
âI float around and around this world.â(Chester Pepper, WB KL-52)
ⶠPLACT (DUR): multiple (simultaneous) imperfectiveeventualities
u-thivruh-tiih-va3SG-float-DUR-PLACT
â(Clouds) are floating.â (Bright & Gehr 2005 s.v. thivruh)
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 30/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va position
ⶠMultiple -va
taPERF
kun-âakun-v-an-va3PL-shoot-PLACTa-AM-PLACTb
âThey went hunting.â (Nettie Reuben, WB KL-72)
ⶠInterpretation
ⶠPLACTa: can add conventional (non-compositional) meaning
akun- âshootâ akunva âhuntâıimnih âloveâ iimnıhva âhave a love affairâiin âbe on fireâ ıinva âbe a forest fireâırih âdripâ ırihva âleakâmah âseeâ maahva âvisitâsir âdisappear, be lostâ sıinva âfail to recognizeâyuh âspitâ yuuhva âvomitâ
ⶠPLACTb: multiple event instances or participants
This is not canonical multiple exponence (Caballero & Harris 2012),because each instance of the affix is fully interpreted.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 31/34
Background Morphological parsing Example: Pluractional -va
Pluractional -va: Summary
ⶠSome locations for -va:
Root â va1 âDirectionals, AM â va2 âTAM â va3
ⶠMeaning effects of -va due to semantic composition:
ⶠ-va1 combines with Root: adds lexical aspect or some otherconventional meaning
ⶠ-va2 combines with Root + Directionals + AM: multipleinstances or participants of the associated event
ⶠ-va3 combines with Root + Directionals + AM + TAM:multiple events with shared temporal features
There is only one -va, with different kinds of interpretation acompositional effect of its placement.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 32/34
Background Morphological parsing Example: Pluractional -va
Morphological parsing and pluractional -vaⶠBrightâs (1957) templatic analysis of Karuk suffixes describes
about a dozen position classes. For example:
ⶠClass 3: directional suffixesⶠClass 4: AM -ar and several other suffixes
But his appendix lexicon is instead hierarchically structured,because complex word meanings are crucially compositional:
ⶠWe have shown that it is helpful to have direct access to thesecomplex units in a text corpus. A good analysis benefits frommore than just a list of all examples of a morpheme.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 33/34
Background Morphological parsing Example: Pluractional -va
Yootva!
References cited
ⶠBright, William. 1957. The Karok language. Berkeley.
ⶠBright, William, and Susan Gehr. 2005. Karuk dictionary. Happy Camp.
ⶠCaballero, Gabriela, and Alice C.Harris. 2012. A working typology ofmultiple exponence. Current issues in morphological theory: (Ir)regularity,analogy and frequency, ed. Ferenc Kiefer, Maria Ladanyi and PeterSiptar, pp. 163â188. Amsterdam.
ⶠConathan, Lisa, and Esther Wood. 2003. Repetitive reduplication in Yurokand Karuk: Semantic effects of contact. Papers of the Thirty-FourthAlgonquian Conference, ed. H. C.Wolfart, pp. 19â34. Winnipeg.
ⶠRice, Keren. 2011. Principles of affix ordering: An overview. WordStructure 4: 169â200.
SSILA âą 8 January 2015 âą Garrett, Maier, Mikkelsen, Sandy Exploring Karuk morphology in a parsed text corpus âą 34/34