Language and the Mind: Encounters in the Mind Fields John Goldsmith - PowerPoint PPT Presentation

Language and the Mind: Encounters in the Mind Fields John Goldsmith April 23, 2014 1

2

1. Strongest, best option: Discovery device Data Correct grammar of data 2. Next best option: Data Yes, or , No Verification device Grammar 3. Fallback position: Data G 1 is better; or , G 2 is better. Grammar 1 Evaluation metric Grammar 2 Chomsky’s vision of Generative Grammar (1955) 3

Generative position: a special case of Option 3 First, test grammars’ eligibility: Data Eligible? Yes, or , No Grammar1 Data Eligible? Yes, or , No Grammar2 If both grammars are eligible: Grammar 1 G 1 is better; or , G 2 is better. Evaluation metric Grammar 2 4

Three central questions: 1. Where do hypotheses come from? Answer: As far as Linguistic Theory goes, that’s none of your business. Ideas come from wherever they come from. As far as indi- vidual grammars go, hypotheses may come from anywhere, but mostly they come from looking at what linguists have said about other languages. 2. How do we determine the extent to which data support a hypothesis? Generative theory has no an- swer to this. 3. How do we determine the goodness of a theory, independent of data? Formal simplicity, but we have not yet found the right way to calculate this. 5

Machine learning: Back to Option 1 Discovery device; G Best grammar in G of data Data Generative grammar and Machine learning agree: • Growing the space of grammars when needed is a good thing. • Shrinking the space of grammars when we jettison unnec- essary possibilities is a good thing. Machine learning: • A linguistic theory requires a method to find the grammar (within the given hypothesis space) that best accounts for the data. 6

Two languages, two grammars, and a Universal Grammar The expected evolution of generative theory 7

A grammar is found that lies outside of Universal Grammar. The expected evolution of generative theory 8

A grammar is found that lies outside of Universal Grammar. Univeral Grammar is expanded, on empirical grounds. The expected evolution of generative theory 9

Revised Universal Grammar. The expected evolution of generative theory 10

Unused space in Universal Grammar is noticed. The expected evolution of generative theory 11

Universal Grammar is shrunk. The expected evolution of generative theory 12

Revised Universal Grammar. The expected evolution of generative theory 13

A grammar is found that lies outside of Universal Grammar. The expected evolution of generative theory 14

Univeral Grammar is expanded, on empirical grounds. The expected evolution of generative theory 15

Revised Universal Grammar. The expected evolution of generative theory 16

U 2 3 n data 1 Find the grammar within the Universe U of Universal Grammar which best models the data. Machine learning world 17

Example 1: Word learning Input: A million words without spaces, including: TheFultonCountyGrandJurysaidFridayaninvestigationo fAtlanta’srecentprimaryelectionproducednoevidenceth. . . Desired output: The Fulton County Grand Jury said Friday an investiga- tion of Atlanta’s recent primary election produced no evi- dence that any irregularities took place. Actual output: The F ult on County Gr and Ju ry said Fri day an investig ationof Atlan ta ’s recent primary election produc ed no evidence that any ir regular ities took place. 18

Iteration number 1 piece count th 127,717 he 119,592 86,893 in er 81,899 72,154 an re 67,753 61,275 on es 59,943 en 55,763 54,216 at ed 52,893 nt 52,761 st 52,307 nd 50,504 ti 50,253 48,233 to 19 47,391 or te 44,280

Iteration number 1 Iteration number 10 piece count piece count th 127,717 2,355 In he 119,592 vi 2,247 86,893 2,169 in some er 81,899 2,155 who 72,154 ical 2,130 an re 67,753 2,119 He ure 2,102 on es 59,943 ance 2,085 en 55,763 ty 2,061 54,216 edthe 2,061 at ed 52,893 sel 2,053 nt 52,761 2,053 its st 52,307 more 2,034 nd 50,504 2,023 form ti 50,253 fac 2,009 48,233 2,007 to act 20 47,391 cont 1,987 or te 44,280 ’t 1,970

Iteration number 10 Iteration number 1 piece count piece count 2,355 In th 127,717 vi 2,247 he 119,592 2,169 some 86,893 in 2,155 who er 81,899 ical 2,130 72,154 an 2,119 He re 67,753 ure 2,102 on ance 2,085 es 59,943 ty 2,061 en 55,763 edthe 2,061 54,216 at sel 2,053 ed 52,893 2,053 its nt 52,761 2,034 more st 52,307 form 2,023 nd 50,504 fac 2,009 ti 50,253 2,007 act 48,233 to 21 cont 1,987 47,391 or

Iteration number 1 Iteration number 10 Iteration number 399 piece count piece count piece count th 127,717 2,355 22 In divided he 119,592 vi 2,247 21 minimal 86,893 some 2,169 ender 21 in er 81,899 2,155 21 who Baltimore 72,154 ical 2,130 Memor 21 an re 67,753 2,119 21 He fever ure 2,102 WestBerlin 21 on es 59,943 ance 2,085 thickness 21 en 55,763 ty 2,061 21 contains 54,216 edthe 2,061 backin 21 at ed 52,893 sel 2,053 choiceof 21 nt 52,761 2,053 attentiontothe 21 its st 52,307 more 2,034 itthe 21 nd 50,504 2,023 21 form sophisticated ti 50,253 fac 2,009 21 sector 48,233 2,007 21 to act jungle 22 47,391 cont 1,987 Mid 21 or te 44,280 ’t 1,970 necessary. 21

Iteration number 399 Iteration number 1 Iteration number 10 piece count piece count piece count 22 divided th 127,717 2,355 In 21 minimal he 119,592 vi 2,247 ender 21 86,893 some 2,169 in 21 Baltimore er 81,899 2,155 who Memor 21 an 72,154 ical 2,130 21 fever re 67,753 2,119 He WestBerlin 21 ure 2,102 on thickness 21 es 59,943 ance 2,085 21 contains en 55,763 ty 2,061 backin 21 at 54,216 edthe 2,061 choiceof 21 ed 52,893 sel 2,053 attentiontothe nt 52,761 2,053 21 its itthe st 52,307 more 2,034 21 nd 50,504 2,023 form sophisticated 21 ti 50,253 fac 2,009 21 sector 23 48,233 2,007 to act 21 jungle 47,391 cont 1,987 or Mid 21

Example 2: Morphology learning NULL-s accomodation accomodations aunt aunt’s NULL-’s account accounted accounting accounts NULL-ed-ing-s afternoon afternoons afternoon’s NULL-s-’s e-ed-ing-es accuse accused accusing accuses ability abilities ies-y addition additional additions NULL-al-s NULL-ped-ping-s drop dropped dropping drops tried tries try trying ied-ies-y-ying guerrilla camera suburb electronic athletic poetic plastic characteristic hundred fluid field thousand ground method neighborhood standard toward afterward hazard cloud voice price device service 24

accomodation accomodations NULL-s according accordingly NULL-ly NULL-ed-ing-s account accounted accounting accounts afternoon afternoons afternoon’s NULL-s-’s accuse accused accusing accuses e-ed-ing-es ability abilities ies-y addition additional additions NULL-al-s NULL-ped-ping-s drop dropped dropping drops tried tries try trying ied-ies-y-ying proceed demand depend extend appeal reveal level dream remain train maintain question develop appear remember consider answer honor expect shift represent point print mount request consist exist review 25

Start econom-, techn- furi-, vigor- emotion- 67 31 45 81 -ic 38 80 -ate -ful 44 -ing -less 36 -ive -ous -al 4 vigor-ous-ly econom-ic-al (null) -ly 26 End

words jump jumped jumping jumps move moved moving moves stop stopped stopping stops try tried trying tries make made making makes buy bought buying buys We need a new device that will show us how words are used. . . a megascope . 27

28

29

30

Part 3: The Syntactic Megascope English Encarta 31

French Encarta feminine masculine singular nouns singular nouns xiisiecle cities simple past verbs infinitives years prenominal plural modifiers nouns 32

feminine masculine singular nouns singular nouns xiisiecle cities simple past verbs infinitives years prenominal plural modifiers nouns French English 33

A reminder about English parts of speech • Prepositions: to, from, up, down, in, out, of, off • Modal auxiliaries: Can I go outside? but not Speak you French? – I cannot speak Russian but not I speak not Rus- sian. – can, could, must, should, shall, will, would – Forms of be also invert, and there is a dummy do available as needed. 34

Dynamic view: English color codes Verbs: ‘bare’ verb ( jump ) red Verbs: past tense( jumped, bought ) blue Verbs: auxiliary ( should, can ) green Prepositions ( from, to, up, down aqua Adjectives purple Cities gray Nouns pink 35

Dynamic view: French color codes Infinitives red Prepositions light blue Past participles blue Adjectives purple Cities gray Masculine nouns pink Feminine nouns light green Inflected verbs light gray 36

founded formed discovered named established assumed carried considered initiated called influenced placed caused joined created adopted achieved supported forced brought provided ordered played 37

38

39