cs182 - university of california, berkeleycs182/sp08/sections/week13.pdf · • last week –...
TRANSCRIPT
![Page 1: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/1.jpg)
CS182
Leon Barrett
with slides inspired by Eva Mok and Joe Makin
April 25, 2008
![Page 2: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/2.jpg)
Announcements
• a9 out, due Thursday
• BBS articles are assigned for the final paper!
![Page 3: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/3.jpg)
Schedule
• Last Week– Metaphor understanding using KARMA
– Grammars and parsing
• This Week– Construction Grammar, ECG
– Grammar Learning
• Next Week– ECG Learning
![Page 4: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/4.jpg)
Questions
• What are constructions?
• How does ECG use constructions?
• Why is learning language hard?
• What is the poverty of the stimulus?
• What is Gold's Theorem?
![Page 5: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/5.jpg)
Embodied constructions
construction HARRYform : /hEriy/meaning : Harry
construction CAFEform : /khaefej/meaning : Cafe
Harry
CAFEcafe
ECG NotationForm Meaning
Constructions have form and meaning poles that are subject to type constraints.
![Page 6: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/6.jpg)
Questions
• What are constructions?
• How does ECG use constructions?
• Why is learning language hard?
• What is the poverty of the stimulus?
• What is Gold's Theorem?
![Page 7: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/7.jpg)
A schema hierarchy of objects (Nomi)
schema Referentsubcase of Entityroles
categorydistributionboundednessnumbergenderaccessibilityresolvedref
schema Object subcase of Entity
schema Entity
schema Place
schema PhysicalObjectsubcase of Object, Place
schema Animate subcase of PhysicalObjectroles
animacyconstraints
animacy true←
schema Humansubcase of Animateroles
sexschema Nomi
subcase of Humansex female←
schema ManipulableObject subcase of PhysicalObject
schema Toysubcase of
ManipulableObjectschema Ball
subcase of Toy
slot filler
schema Cup subcase of ManipulableObject
![Page 8: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/8.jpg)
The schemas we just defined
Entity Place
Object
PhysicalObject
Referent
AnimateManipulableObject
HumanToy
Nomi Mother FatherBook
Cup
Ball Block
Food
Doll Bed
Cake Juice
![Page 9: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/9.jpg)
A schema hierarchy of actions (Nomi)
schema DirectedActionsubcase of Actionroles
patient : Entity
schema Actionroles
agent : Entity
schema Move subcase of Actionroles
mover : Entitydirection : Place
schema CauseMove subcase of DirectedAction, Moveroles
causer : Humanmover : PhysicalObjectmotion : Move
constraints motion.mover mover↔
motion.agent causer↔
motion.direction direction↔
agent causer↔
patient mover↔
identification constraint
type constraint
![Page 10: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/10.jpg)
The schemas we just definedAction
Directed Action
CauseMove
Move
JointMoveTransfer
![Page 11: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/11.jpg)
Constructions, finally (Nomi)construction RefExpr
form : SchematicFormmeaning : Referent
construction NomiCnlevel 0subcase of RefExprform : Word
self.f.orth "Nomi"←
meaningevokes Nomi as nself.m.category n↔
self.m.resolvedref n↔
local name
fancy way of sayingthat the category ofthe referent is Nomi
construction CupCnlevel 0subcase of RefExprform : Word
self.f.orth “cup"←
meaningevokes Cup as nself.m.category n↔
self.m.resolvedref n↔
![Page 12: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/12.jpg)
Constructions, finally (Nomi)
construction CauseMotionVerbsubcase of MotionVerbmeaning : CauseMove
construction MotionVerbmeaning : Move
construction GetCnlevel 0subcase of CauseMotionVerbform : Word
self.f.orth "get"←
lexical construction
![Page 13: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/13.jpg)
Constructions, finally (Nomi)
construction TransitiveCnlevel 2constructional
constituentsagt : RefExprv : CauseMotionVerbobj : RefExpr
formagt.f before v.fv.f before obj.f
meaningv.m.agent agt.m.resolvedref↔
v.m.patient obj.m.resolvedref↔
smaller constructions that it takes
ordering constraints on the constituents
![Page 14: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/14.jpg)
Questions
• What are constructions?
• How does ECG use constructions?
• Why is learning language hard?
• What is the poverty of the stimulus?
• What is Gold's Theorem?
![Page 15: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/15.jpg)
Difficulty of learning language
• What makes learning language difficult?– How many sentences do children hear?
– How often are those sentences even correct?
– Even when they're correct, how often are they complete?
– How often are children corrected when saying something wrong?
– How many possible languages are there?
![Page 16: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/16.jpg)
Larger context
• War!– Is language innate?
– Covered in book
![Page 17: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/17.jpg)
Questions
• What are constructions?
• How does ECG use constructions?
• Why is learning language hard?
• What is the poverty of the stimulus?
• What is Gold's Theorem?
![Page 18: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/18.jpg)
Poverty and Opulence
• Poverty of the stimulus– Coined to suggest how little information children have to
learn from
• Opulence of the substrate– Opulence = “richness”
– Coined in response to suggest how much background information children have
![Page 19: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/19.jpg)
Questions
• What are constructions?
• How does ECG use constructions?
• Why is learning language hard?
• What is the poverty of the stimulus?
• What is Gold's Theorem?
![Page 20: CS182 - University of California, Berkeleycs182/sp08/sections/week13.pdf · • Last Week – Metaphor understanding using KARMA – Grammars and parsing • This Week – Construction](https://reader033.vdocument.in/reader033/viewer/2022050610/5fb190be77e6934ee2343004/html5/thumbnails/20.jpg)
Gold's Theorem• Suppose that you have an infinite number of languages
– language = “set of legal sentences”
• Suppose that for every language Ln, there is a bigger language Ln+1
– makes every sentence, and then some
• There's some language Linfinity
– contains all the sentences in all other languages
• You can arrange data so that no one ever learns Linfinity
– http://www.lps.uci.edu/~johnsonk/Publications/Johnson.GoldsTheorem.pdf