2. 1 Contextual Vocabulary Acquisition: From Algorithm to Curriculum Michael W. Kibby, Ph.D. Department of Learning & Instruction and The Reading Center William J. Rapaport, Ph.D. Department of Computer Science & Engineering Department of Philosophy, and Center for Cognitive Science Karen M. Wieland Department of Learning & Instruction, The Reading Center, and The Nichols School NSF ROLE Grant REC-0106338

3. 2

4. 3

5. 4 Why Learning Word Meanings Is Important

6. 5 Why Learning Word Meanings Is Important Reason 1 National Assessment of Educational Progress- Reading (NAEP-Reading)

7. 6 Meaning Vocabulary Assessment on NAEP-R Meaning vocabulary is the application of one’s understanding of word meanings to passage comprehension.

8. 7 Vocabulary knowledge is considered to be one of the five essential components of reading as defined by the No Child Left Behind legislation.

9. 8 NAEP will not test definitions in isolation from surrounding text; i.e., students will not be assessed on their prior knowledge of definitions of words on a list.

10. 9

11. 10 Examples: Altruistic Magnanimously Dispersed Impetus Forage Soothing Lost in thought Huddled Abide Piqued Beholden Marathon journey Legacy Abated Social contract Grudge

12. 11 Three Reasons NAEP-R Does Not Test a Specific Word List 1.Knowledge of the explicit definition of a word is not what is required for reading comprehension.

13. 12 2.The meaning of a word is too often dependent upon the context. e.g. cast The fisherman cast his line. The members of the cast took a bow. They put a cast on my broken arm. The yard is littered with shells cast off by the cicadas.

14. 13 3.Writers often use words in a manner that goes beyond their concrete, familiar definition, but do so in ways that skilled readers can interpret effectively.

15. 14 Why Learning Word Meanings Is Important Reason 2 Learning new things and their words changes or increases our perception and organization of the world

16. 15 The Lego™ Notion of Learning New Things

17. 16 Why Learning Word Meanings Is Important Reason 3 Reading comprehension mandates knowing the meaning (i.e., concept, thing) associated with words in the text When students do not know meanings of words in a written text, comprehension often decreases.

18. 17

19. 18 Why Learning Word Meanings Is Important Reason 4 Learning new things and words facilitates students’ abilities to use words judiciously— which is much valued in our society

20. 19

21. 20

22. 21 Why Learning Word Meanings Is Important Reason 5 The Profound Effects of Limited Vocabulary

23. 22 Profound effects of limited vocabulary continued Limited vocabulary is associated with lower IQ scores. Limited vocabulary is associated with limited reading comprehension. –In grades 7+, vocabulary and reading comprehension correlate.75 to.85.

24. 23 Social Class and Meaning Vocabulary Hart, Betty, & Risley, Todd R. (1995). Meaningful differences in the everyday experience of young children. Baltimore, MD: Brookes.

25. 24 Studied 42 children’s vocabulary growth from their 9th month to their 36th month. Researchers –Visited each child’s home once a month. –Observed and tape recorded for one hour every word spoken to or by child. –Recorded 23-30 hours for every child.

26. 25 Actual and Estimated Number of Words Heard from 0 - 48 Months

27. 26 “The Invisible Curriculum”

28. 27 Cumulative Number of New Words (Hart & Risley, 1995)

29. 28 A Brief Background on the Counting of Words Carroll, Davies, Richman (1971), The American Heritage Word Frequency Book–called the WFI. A count of 5,088,721 different words (called tokens) in printed English for grades 3-9.

30. 29 Of 5,088,721 Words in WFI There were 86,741 different words. But the following 13 were counted as different words: addadditiveadditives addsadditionadditions addedaddendaddends addingadditionalADDITION as well as Add (capitalized).

31. 30 When Do Two Words Differ? Nagy & Anderson sampled WFI words. Put each word in 1 of 6 classes varying in semantic relation to other words. –Classes 0, 1, 2 closely related semantically. –Classes 3, 4, 5 progressively more distant. Estimated there are 139,020 different words in semantic categories 0, 1, & 2.

32. 31 But 45,453 of these are base words— knowing these 45,453 means a reader knows all 139,020. Adding 43,080 in classes 3, 4 & 5 brings total to 88,583 different word families in printed school texts for grades 3-9.

33. 32

34. 33 Learning New Words is Natural

35. 34

36. 35 Edna Heidbretter, The Attainment of Concepts. 1946 taught persons to associate nine pairs of visual shapes and pronounceable pseudo word told persons this was a memory task

37. 36 pran

38. 37 mulf

39. 38 relk...

40. 39 Test

41. 40 pran mulf relk I SETS Trials to Learn 27...

42. 41 pran mulf relk III SETS Trials to Learn 2732...

43. 42 pran mulf relk IIIIII SETS Trials to Learn 273211...

44. 43 pran mulf relk IIIIIIIV SETS Trials to Learn 2732114...

45. 44 pran mulf relk IIIIIIIVV SETS Trials to Learn 27321141.5...

46. 45 pran mulf relk IIIIIIIVV SETS Trials to Learn 27321141.5...

47. 46 Definition of “CVA” “ C ontextual V ocabulary A cquisition” = def the acquisition of word meanings from text –“incidental” –“deliberate” by reasoning about –contextual clues –background knowledge (linguistic, factual, commonsense) Including hypotheses from prior encounters (if any) with the word without external sources of help –No dictionaries –No people

48. 47 CVA: From Algorithm to Curriculum 1.Computational theory of CVA –Based on: algorithms developed by Karen Ehrlich (1995) verbal protocols (case studies) –Implemented in a semantic-network-based knowledge-representation & reasoning system SNePS (Stuart C. Shapiro & colleagues) 2.Educational curriculum to teach CVA –Based on our algorithms & protocols –To improve vocabulary & reading comprehension –Joint work with Michael Kibby & Karen Wieland Center for Literacy & Reading Instruction

49. 48 Project Goals Develop & implement computational theory of CVA based on verbal protocols (case studies) Translate algorithms into a curriculum –To improve CVA and reading comprehension in science, technology, engineering, math (“STEM”) Use new case studies, based on the curriculum, to improve both algorithms & curriculum

50. 49 People Do “Incidental” CVA We know more words than explicitly taught –Average high-school grad knows ~45K words  learned ~2.5K words/year (over 18 yrs.) –But only taught ~400/school-year ~ 4800 in 12 years of school (~ 10% of total)  Most word meanings learned from context − including oral & perceptual contexts –“incidentally” (unconsciously) How?

51. 50 People Also Do “Deliberate” CVA You’re reading; You understand everything you read, until… You come across a new word Not in dictionary No one to ask So, you try to “figure out” its meaning from “context” How? –guess? derive? infer? deduce? educe? construct? predict? … –our answer: Compute it from inferential search of “context”, including background knowledge

52. 51 What does ‘brachet’ mean?

53. 52 (From Malory’s Morte D’Arthur [page # in brackets]) 1. There came a white hart running into the hall with a white brachet next to him, and thirty couples of black hounds came running after them. [66] 2.As the hart went by the sideboard, the white brachet bit him. [66] 3.The knight arose, took up the brachet and rode away with the brachet. [66] 4.A lady came in and cried aloud to King Arthur, “Sire, the brachet is mine”. [66] 10.There was the white brachet which bayed at him fast. [72] 18. The hart lay dead; a brachet was biting on his throat, and other hounds came behind. [86]

54. 53 Computational cognitive theory of how to learn word meanings From context –I.e., text + grammatical info + reader’s prior knowledge With no external sources (human, on-line) –Unavailable, incomplete, or misleading Domain-independent –But more prior domain-knowledge yields better definitions “definition” = hypothesis about word’s meaning –Revisable each time word is seen

55. 54 Cassie learns what “brachet” means: Background info about:harts, animals, King Arthur, etc. No info about:brachets Input:formal-language (SNePS) version of simplified English A hart runs into King Arthur’s hall. In the story, B12 is a hart. In the story, B13 is a hall. In the story, B13 is King Arthur’s. In the story, B12 runs into B13. A white brachet is next to the hart. In the story, B14 is a brachet. In the story, B14 has the property “white”. Therefore, brachets are physical objects. (deduced while reading; Cassie believes that only physical objects have color)

56. 55 --> (defineNoun "brachet") Definition of brachet: Class Inclusions: phys obj, Possible Properties: white, Possibly Similar Items: animal, mammal, deer, horse, pony, dog, I.e., a brachet is a physical object that can be white and that might be like an animal, mammal, deer, horse, pony, or dog

57. 56 A hart runs into King Arthur’s hall. A white brachet is next to the hart. The brachet bites the hart’s buttock. --> (defineNoun "brachet") Definition of brachet: Class Inclusions: animal, Possible Actions: bite buttock, Possible Properties: white, Possibly Similar Items: mammal, pony,

58. 57 A hart runs into King Arthur’s hall. A white brachet is next to the hart. The brachet bites the hart’s buttock. The knight picks up the brachet. The knight carries the brachet. --> (defineNoun "brachet") Definition of brachet: Class Inclusions: animal, Possible Actions: bite buttock, Possible Properties: small, white, Possibly Similar Items: mammal, pony,

59. 58 A hart runs into King Arthur’s hall. A white brachet is next to the hart. The brachet bites the hart’s buttock. The knight picks up the brachet. The knight carries the brachet. The lady says that she wants the brachet. --> (defineNoun "brachet") Definition of brachet: Class Inclusions: animal, Possible Actions: bite buttock, Possible Properties: valuable, small, white, Possibly Similar Items: mammal, pony,

60. 59 A hart runs into King Arthur’s hall. A white brachet is next to the hart. The brachet bites the hart’s buttock. The knight picks up the brachet. The knight carries the brachet. The lady says that she wants the brachet. The brachet bays at Sir Tor. [background knowledge: only hunting dogs bay] --> (defineNoun "brachet") Definition of brachet: Class Inclusions: hound, dog, Possible Actions: bite buttock, bay, hunt, Possible Properties: valuable, small, white, I.e. A brachet is a hound (a kind of dog) that can bite, bay, and hunt, and that may be valuable, small, and white.

61. 60 General Comments System’s behavior  human protocols System’s definition  OED’s definition: = A brachet is “a kind of hound which hunts by scent”

62. 61 Computational cognitive theory of how to learn word meanings from context (cont.) 3 kinds of vocabulary acquisition: –Construct new definition of unknown word What does ‘brachet’ mean? –Fully revise definition of misunderstood word Does “smiting” entail killing? –Expand definition of word used in new sense Can you “dress” (i.e., clothe) a spear? Initial hypothesis; Revision(s) upon further encounter(s); Converges to stable, dictionary-like definition; Subject to revision

63. 62 Motivations & Applications Part of cognitive-science projects –Narrative text understanding –Syntactic semantics (contra Searle’s Chinese-Room Argument) Computational applications: –Information extraction –Autonomous intelligent agents: There can be no complete lexicon Agent/info-extraction system shouldn’t have to stop to ask questions Other applications: –L1 & L2 acquisition research –Computational lexicography ** Education: improve reading comprehension **

64. 63 State of the Art Vocabulary Learning: –Some dubious contributions: Useless “algorithms” Contexts that include definition –Useful contribution: (good) reader’s word-model = updateable frame with slots & defaults Psychology: –Cues to look for (= slots for frame): Space, time, value, properties, functions, causes, classes, synonyms, antonyms –Can understand a word w/o having a definition Computational Linguistics: –Systems need scripts, human informants, ontologies Not needed in our system –CVA  Word-Sense Disambiguation Essay question vs. multiple-choice test

65. 64 State of the Art: Computational Linguistics Information extraction systems Autonomous intelligent agents There can be no complete lexicon Such systems/agents shouldn’t have to stop to ask questions

66. 65 State of the Art: Computational Linguistics Granger 1977: “Foul-Up” –Based on Schank’s theory of “scripts” (schema theory) –Our system not restricted to scripts Zernik 1987: self-extending phrasal lexicon –Uses human informant –Ours system is really “self-extending” Hastings 1994: “Camille” –Maps unknown word to known concept in ontology –Our system can learn new concepts Word-Sense Disambiguation: –Given ambiguous word & list of all meanings, determine the “correct” meaning Multiple-choice test –Our system: given new word, compute its meaning Essay question

67. 66 State of the Art: Vocabulary Learning (I) Elshout-Mohr/van Daalen-Kapteijns 1981,1987: –Application of Winston’s AI “arch” learning theory –(Good) reader’s model of new word = frame Attribute slots, default values Revision by updating slots & values –Poor readers update by replacing entire frames –But EM & vDK used: Made-up words Carefully constructed contexts –Presented in a specific order

68. 67 Elshout-Mohr & van Daalen-Kapteijns Experiments with neologisms in 5 artificial contexts When you are used to a view it is depressing when you live in a room with kolpers. –Superordinate information At home he had to work by artificial light because of those kolpers. During a heat wave, people want kolpers, so sun-blind sales increase. –Contexts showing 2 differences from the superordinate I was afraid the room might have kolpers, but plenty of sunlight came into it. This house has kolpers all summer until the leaves fall out. –Contexts showing 2 counterexamples due to the 2 differences

69. 68 State of the Art: Psychology Johnson-Laird 1987: –Word understanding  definition –Definitions aren’t stored –“During the Renaissance, Bernini cast a bronze of a mastiff eating truffles.”

70. 69 State of the Art: Psychology Sternberg et al. 1983,1987: –Cues to look for (= slots for frame): Spatiotemporal cues Value cues Properties Functions Cause/enablement information Class memberships Synonyms/antonyms –To acquire new words from context: Distinguish relevant/irrelevant information Selectively combine relevant information Compare this information with previous beliefs

71. 70 Sternberg The couple there on the blind date was not enjoying the festivities in the least. An acapnotic, he disliked her smoking; and when he removed his hat, she, who preferred “ageless” men, eyed his increasing phalacrosis and grimaced.

72. 71 State of the Art: Vocabulary Learning (II) Some dubious contributions: Mueser 1984: “Practicing Vocabulary in Context” –BUT: “context” = definition !! Clarke & Nation 1980: a “strategy” (algorithm?) 1.Look at word & context; determine POS 2.Look at grammatical context E.g., “who does what to whom”? 3.Look at wider context [E.g., search for Sternberg-like clues] 4.Guess the word; check your guess

73. 72 CVA: From Algorithm to Curriculum “guess the word” = “then a miracle occurs” Surely, we computer scientists can “be more explicit”!

74. 73 CVA: From algorithm to curriculum … Treat “guess” as a procedure call (“subroutine”) –Fill in the details with our algorithm –Convert the algorithm into a curriculum To enhance students’ abilities to use deliberate CVA strategies –To improve reading comprehension … and back again! Use knowledge gained from CVA case studies to improve the algorithm I.e., use Cassie to learn how to teach humans & use humans to learn how to teach Cassie

75. 74 Why not use a dictionary? Because: People are lazy (!) Dictionaries are not always available Dictionaries are always incomplete Dictionary definitions are not always useful –‘chaste’ = df clean, spotless /  “new dishes are chaste” –‘college’ = df a body of clergy living together and supported by a foundation Most words are learned via incidental CVA, not via dictionaries Most importantly: –Dictionary definitions are just more contexts!

76. 75 Why not use a dictionary? Merriam-Webster New Collegiate Dictionary: –chaste. 1.innocent of unlawful sexual intercourse –student: stay away from that one! 2.celibate –student: huh? 3.pure in thought and act: modest –student: I have to find a sentence for that? 4.a: severely simple in design or execution: austere –student: huh? “severely”? “austere”? b: clean, spotless –student: all right!: “The plates were still chaste after much use.” –Deese 1967 / Miller 1985

77. 76 Why not use a dictionary? Merriam-Webster (continued): –college. 1.a body of clergy living together and supported by a foundation 2.a building used for an educational or religious purpose 3.a: a self-governing constituent body of a university offering living quarters and instruction but not granting degrees… b: a preparatory or high school c: an independent institution of higher learning offering a course of general studies leading to a bachelor’s degree… –Problem: ordering is historical!

78. 77 Why not use a dictionary? Merriam-Webster (continued): –infract:infringe –infringe:encroach –encroach: to enter by gradual steps or by stealth into the possessions or rights of another to advance beyond the usual or proper limits; trespass

79. 78 Why not use a dictionary? Collins COBUILD Dictionary –“Helping Learners with Real English” –chaste. 1.Someone who is chaste does not have sex with anyone, or only has sex with their husband or wife; an old-fashioned use, used showing approval. EG She was a holy woman, innocent and chaste. 2.Something that is chaste is very simple in style, without much decoration. EG …chaste houses built in 1732

80. 79 Why not use a dictionary? Collins COBUILD Dictionary –college. 1.A college is 1.1 an institution where students study for qualifications or do training courses after they have left school. … infract [not in dictionary] infringe. 1.If you infringe a law or an agreement, you break it. encroach. 1.To encroach on or upon something means to slowly take possession or control of it, so that someone else loses it bit by bit.

81. 80 Question (objection): Teaching computers  teaching humans! But : Our goal: –Not: teach people to “think like computers” –But: to explicate computable & teachable methods to hypothesize word meanings from context AI as computational psychology: –Devise computer programs that are essentially faithful simulations of human cognitive behavior –Can tell us something about human mind. We are teaching a machine, to see if what we learn in teaching it can help us teach students better.

82. 81 How Does Our System Work? Uses a semantic network computer system –semantic networks = “concept maps” –serves as a model of the reader –represents: reader’s prior knowledge the text being read –can reason about the text and the reader’s knowledge

83. 82 Fragment of reader’s prior knowledge: m3 = In “real life”, white is a color m6 = In “real life”, harts are deer m8 = In “real life”, deer are mammals m11 = In “real life”, halls are buildings m12 = In “real life”, b1 is named “King Arthur” m14 = In “real life”, b1 is a king(etc.)

84. 83 m16 = if v3 has property v2 & if v2 is a color & if v3  v1 thenv1 is a kind of physical object

85. 84 Reading the story: m17 = In the story, b2 is a hart m24 = In the story, the hart runs into b3 (b3 is King Arthur’s hall) – not shown (harts are deer) – not shown

86. 85 The entire network showing the reader’s mental context consisting of prior knowledge, the story, & inferences. The definition algorithm searches this network & abstracts parts of it to produce a (preliminary) definition of ‘brachet’.

87. 86 Implementation SNePS (Stuart C. Shapiro & SNeRG) : –Intensional, propositional semantic-network knowledge-representation & reasoning system –Node-based & path-based reasoning I.e., logical inference & generalized inheritance –SNeBR belief revision system Used for revision of definitions –SNaLPS natural-language input/output –“Cassie”: computational cognitive agent

88. 87 How It Works SNePS represents: – background knowledge + text information in a single, consolidated semantic network Algorithms deductively search network for slot-fillers for definition frame Search is guided by desired slots –E.g., prefers general info over particular info, but takes what it can get

89. 88 Noun Algorithm Find or infer: Basic-level class memberships (e.g., “dog”, rather than “animal”) –else most-specific-level class memberships –else names of individuals Properties of Ns (else, of individual Ns) Structure of Ns (else …) Functions of Ns (else …) Acts that Ns perform (else …) Agents that perform acts w.r.t. Ns & the acts they perform (else…) Ownership Synonyms Else do: “syntactic/algebraic manipulation” “Al broke a vase”  a vase is something Al broke –Or: a vase is a breakable physical object

90. 89 Verb Algorithm Find or infer: –Predicate structure: Categorize arguments/cases –Results of V’ing: Effects, state changes –Enabling conditions for V Future work: –Classification of verb-type –Synonyms [Also: preliminary work on adjective algorithm]

91. 90 Belief Revision Used to revise definitions of words with different sense from current meaning hypothesis SNeBR (ATMS; Martins & Shapiro 88) : –If inference leads to a contradiction, then: 1.SNeBR asks user to remove culprit(s) 2.& automatically removes consequences inferred from culprit SNePSwD (SNePS w/ Defaults; Martins & Cravo 91) –Currently used to automate step 1, above AutoBR (Johnson & Shapiro, in progress) & new default reasoner (Bhushan & Shapiro, in progress) –Will replace SNePSwD

92. 91 Revision & Expansion Removal & revision being automated via SNePSwD by ranking all propositions with kn_cat: most intrinsic info re: language; fundamental background info certain(“before” is transitive) story info in text (“King Lot rode to town”) life background info w/o variables or inference (“dogs are animals”) story-comp info inferred from text (King Lot is a king, rode on a horse) life-rule.1 everyday commonsense background info (BearsLiveYoung(x)  Mammal(x)) life-rule.2 specialized background info (x smites y  x kills y by hitting y) least certain questionable already-revised life-rule.2; not part of input

93. 92 Belief Revision: “smite” Misunderstood word; 2-stage “subtractive” revision Background knowledge includes: (*) smite(x,y,t)  hit(x,y,t) & dead(y,t) & cause(hit(x,y,t),dead(y,t)) P1: King Lot smote down King Arthur D1: If person x smites person y at time t, then x hits y at t, and y is dead at t Q1: What properties does King Arthur have? R1: King Arthur is dead. P2: King Arthur drew Excalibur. Q2: When did King Arthur do this? SNeBR is invoked: –KA’s drawing E is inconsistent with being dead –(*) replaced: smite(x,y,t)  hit(x,y,t) &  dead(y,t) & [dead(y,t)  cause(hit, dead)] D2: If person x smites person y at time t, then x hits y at t &  (y is dead at t) P3: [another passage in which ~(smiting  death)] D3: If person x smites person y at time t, then x hits y at t

94. 93 Belief Revision: “dress” “additive” revision Bkgd info includes: (1)dresses(x,y)   z[clothing(z) & wears(y,z) (2)Spears don’t wear clothing (both kn_cat=life.rule.1 ) P1: King Arthur dressed himself. D1: A person can dress itself; result: it wears clothing. P2: King Claudius dressed his spear. [Cassie infers: King Claudius’s spear wears clothing.] Q2: What wears clothing? SNeBR is invoked: –KC’s spear wears clothing inconsistent with (2). –(1) replaced: dresses(x,y)   z[clothing(z) & wears(y,z)] v NEWDEF –Replace (1), not (2), because of verb in antecedent of (1) (Gentner) P3: [other passages in which dressing spears precedes fighting] D2: A person can dress a spear or a person; result: person wears clothing or person is enabled to fight

95. 94 Figure out meaning of word from what? context (i.e., the text)? –Werner & Kaplan 52, McKeown 85, Schatz & Baldwin 86 context and reader’s background knowledge? –Granger 77, Sternberg 83, Hastings 94 context including background knowledge? –Nation & Coady 88, Graesser & Bower 90 Note: –“context” = text  context is external to reader’s mind Could also be spoken/visual/situative (still external) –“background knowledge”: internal to reader’s mind What is (or should be) the “context” for CVA?

96. 95 Some Proposed Preliminary Definitions (to extract order out of confusion) Unknown word for a reader = def –Word or phrase that reader has never seen before –Or only has vague idea of its meaning Different levels of knowing meaning of word –Notation: “X”

97. 96 Proposed preliminary definitions Text = def –(written) passage –containing X –single phrase or sentence … several paragraphs

98. 97 Proposed preliminary definitions Co-text of X in some text = def –The entire text “minus” X; i.e., entire text surrounding X –E.g., if X = ‘brachet’, and text = “There came a white hart running into the hall with a white brachet next to him, and thirty couples of black hounds came running after them.” Then X’s co-text in this text = “There came a white hart running into the hall with a white ______ next to him, and thirty couples of black hounds came running after them.” –Cf. “cloze” tests in psychology But, in CVA, reader seeks meaning or definition –NOT a missing word or synonym: There’s no “correct” answer! –“Co-text” is what many mean by “context” BUT: they shouldn’t!

99. 98 Proposed preliminary definitions The reader’s prior knowledge = def –the knowledge that the reader has when s/he begins to read the text –and is able to recall as needed while reading “knight picks up & carries brachet”  ? small Warnings: –“knowledge”  truth so, “prior beliefs” is better –“prior” vs. “background” vs. “world”, etc. See next slide!

100. 99 Proposed preliminary definitions Possible synonyms for “prior knowledge”, each with different connotation: –Background knowledge: Can use for information that author assumes reader to have –World knowledge: General factual knowledge about things other than the text’s topic –Domain knowledge: Specialized, subject-specific knowledge about the text’s topic –Commonsense knowledge: Knowledge “everyone” has –E.g., CYC, “cultural literacy” (Hirsch) These overlap: –PK should include some CSK, might include some DK –BK might include much DK

101. 100 Steps towards a Proper Definition of “Context” Step 1: –The context of X for a reader = def 1.The co-text of X 2. “+” the reader’s prior knowledge Both are needed! –After reading: “the white brachet bit the hart in the buttock” most subjects infer that brachets are (probably) animals, from: That text, plus: Available PK premise: “If x bites y, then x is (probably) an animal. –Inference is not an enthymeme! (argument with missing premise)

102. 101 Proper definition of “context”: (inference not an enthymeme because): –When you read, you “internalize” the text You “bring it into” your mind Gärdenfors 1997, 1999; Jackendoff 2002 “Missing” premise might be in reader’s mind! –This “internalized text” is more important than the actual words on paper: Text:“I’m going to put the cat out” Misread as:“I’m going to put the car out” –leads to different understanding of “the text” –What matters is what the reader thinks the text is, Not what the text actually is Therefore …

103. 102 Proper definition of “context”: Step 2: –The context of X for a reader = def A single KB, consisting of: 1. The reader’s internalized co-text of X 2. “ + ” the reader’s prior knowledge

104. 103 Proper definition of “context”: But: What is “+” ? –Not: mere conjunction or union! –Active readers make inferences while reading. From text = “a white brachet” & prior commonsense knowledge = “only physical objects have color”, reader might infer that brachets are physical objects From “The knight took up the brachet and rode away with the brachet.” & prior commonsense knowledge about size, reader might infer that brachet is small enough to be carried –Whole > sum of parts: inference from [internalized text + PK]  new info not in text or in PK I.e., you can learn from reading!

105. 104 Proper definition of “context”: But: Whole < sum of parts! –Reader can learn that some prior beliefs were mistaken Or: reader can decide that text is mistaken (less likely) Reading & CVA need belief revision! operation “ + ”: –input:PK & internalized co-text –output:“belief-revised integration” of input, via: Expansion: –addition of new beliefs from ICT into PK, plus new inferences Revision: –retraction of inconsistent prior beliefs together with inferences from them Consolidation: –eliminate further inconsistencies

106. 105 Prior KnowledgeText PK1 PK2 PK3 PK4

107. 106 Prior KnowledgeText PK1 PK2 PK3 PK4 T1

108. 107 Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization

109. 108 B-R Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization P5 inference

110. 109 B-R Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization P5 inference T2 I(T2) P6

111. 110 B-R Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization P5 inference T2 I(T2) P6 T3 I(T3)

112. 111 B-R Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization P5 inference T2 I(T2) P6 T3 I(T3)

113. 112 B-R Integrated KBText PK1 PK2 PK3 PK4 T1 I(T1) internalization P5 inference T2 I(T2) P6 T3 I(T3) P7 Note: All “contextual” reasoning is done in this “context”:

114. 113 Proper definition of “context”: One more detail: X needs to be internalized Context is a 3-place relation among: –Reader, word, and text Final(?) def.: –Let T be a text –Let R be a reader of T –Let X be a word in T (that is unknown to R) –Let T-X be X’s co-text in T. –Then: The context that R should use to hypothesize a meaning for R’s internalization of X as it occurs in T = def –The belief-revised integration of R’s prior knowledge with R’s internalization of T-X.

115. 114 This definition agrees with… Cognitive-science & reading-theoretic views of text understanding –Schank 1982, Rumelhart 1985, etc. & AI techniques for text understanding: –Reader’s mind modeled by KB of prior knowledge Expressed in AI language (for us: SNePS) –Computational cognitive agent reads the text, “integrating” text info into its KB, and making inferences & performing belief revision along the way –When asked to define a word, Agent deductively searches this single, integrated KB for information to fill slots of a definition frame –Agent’s “context” for CVA = this single, integrated KB

116. 115 Distinguishing Prior Knowledge from Integrated Co-Text So KB can be “disentangled” as needed for belief revision or to control inference: Each proposition in the single, integrated KB is marked with its “source”: –Originally from PK –Originally from text –Inferred Sources of premises

117. 116 Some Open Questions Roles of spoken/visual/situative contexts Relation of CVA “context” to formal theories of context (e.g., McCarthy, Guha…) Relation of I(T) to prior-KB; e.g.: –Is I(T i ) true in prior-KB? It is “accepted pro tem”. –Is I(T) a “subcontext” of pKB or B-R KB? How to “activate” relevant prior knowledge. Etc.

118. 117 Research Methodology AI team: –Develop, implement, & test better computational theories of CVA –Translate into English for use by reading team Reading team: –Convert algorithms to curriculum –Think-aloud protocols To gather new data for use by AI team As curricular technique (case studies)

119. 118 Problem in Converting Algorithm into Curriculum “A knight picks up a brachet and carries it away …” Cassie: –Has “perfect memory” –Is “perfect reasoner” –Automatically infers that brachet is small People don’t always realize this: –May need prompting: How big is the brachet? –May need relevant background knowledge –May need help in drawing inferences Teaching CVA =? teaching general reading comprehension –Vocabulary knowledge correlates with reading comprehension

120. 119 CVA & Science Education Original goal: CVA in & for science education –Use CVA to improve reading of STEM materials A side effect: CVA as science education –There are no ultimate authorities to consult No answers in the back of the book of life! As true for STEM as it is for reading about STEM –  Goal of education = To learn how to learn on one’s own Help develop confidence & desire to use that skill –CVA as scientific method in miniature furthers this goal: Find clues/evidence (gathering data) Integrate them with personal background knowledge Use together to develop new theory (e.g., new meaning) Test/revise new theory (on future encounters with word)

121. 120 CVA & Geography (  STEM) Use texts w/ unknown geographic terms –‘estuary’ –‘proximity’ (IGERT’s own Valerie Raybold Yakich ) L1 acquisition of spatial terms –Children’s concepts  adult concepts L2 acquisition of spatial terms –L2 spatial concepts  L1 spatial concepts –Especially spatial prepositions

122. 121 Conclusion Developing a computational theory of CVA, which can become … a useful educational technique for improving vocabulary and reading comprehension a model of the scientific method a useful tool for learning on one’s own.

123. 122 Participants Santosh Basapur (IE)Adam Lammert (Vasser/ugrad) Taha Suglatwala (CSE) Aishwarya Bhave (CSE) Amanda MacRae (LAI)Matthew Sweeney (ENG/CSE/ugrad) *Marc Broklawski (CSE) Brian Morgan (LAI)Matthew Watkins (CEN) Chien-chih Chi (PHI)*Scott Napieralski (CSE) *Karen Wieland (LAI) Justin Del Vecchio (CSE) Vikranth Rao (CSE/ugrad) Yulai Xie (CSE) Karen Ehrlich (Fredonia) Laurie Schimpf (LAI)Valerie Raybold Yakich (GEO) Jeffrey Galko (PHI)Ashwin Shah (CSE)& SNeRG members Christopher Garver (CSE) Stuart C. Shapiro (UB/CSE) (+ new students, Spring 2003 … Paul Gestwicki (CSE)Anuroopa Shenoy (CSE) … & Fall 2004 Kazuhiro Kawachi (LIN) Rajeev Sood (CSE)(* = supported on NSF grant)

124. 123 Web Page http://www.cse.buffalo.edu/~rapaport/cva.ht ml

125. 124 An Analysis of Think-Aloud Protocol of Good Readers Using CVA Strategies During Silent Reading

126. 125 Facilitating Vocabulary Growth Word fun. Roots, prefixes, affixes. Dictionary. Wide reading. Contextual vocabulary acquisition– CVA.

127. 126 Limited Data Describing CVA Processes Nation–guess. Ames. Deighton. Sternberg. Elshout-Mohr & van Daalen-Kapteijns. Harmon.

128. 127 Overview of Our Study We asked good readers to think- aloud when they encountered a word whose meaning they did not know as they silently read a set of 7-17 texts, each text containing at least one instance of the unknown word.

129. 128 Overview continued We analyzed the think-aloud protocol to gain understanding of the CVA processes. Besides understanding CVA processes, our goal was to build a more effective CVA curriculum.

130. 129 Beginning and Ending Research Questions What text cues are used for CVA? What are CVA reasoning processes? What sense or meaning of an unknown word is gained from CVA? How is information from prior encounters with a hard word used when the word is seen in new texts?

131. 130 Methodology: Hard Words Identified small set of “hard words.” –Some words from Ehrlich & Rapaport earlier work with Cassie. –Dale & O’Rourke and Carroll, Davis & Richman as a guide. –Some words came from scanning science and current event texts.

132. 131 Methodology: Text Sets For each word, identified a set of 7- 17 authentic texts, each with 1 or more instances of the word. Hard words were in boldface font. Sometimes, replaced real “hard word” with a neologism (a non- word): e.g., itresia for estuary

133. 132 Methodology: Participants High school students. Excellent or outstanding readers. Readers given pre-test.

134. 133

135. 134

136. 135 Methodology: Procedures Worked 1-1 with researcher. Read each passage, one at a time. Researcher provided meaning of other words in text reader did not know.

137. 136 Procedures continued When hard word encountered, reader thought aloud while trying to gain sense of word’s meaning. Audio tape recorded think-alouds. Transcribed taped think-alouds.

138. 137 Methodology: Analyses of Verbal Protocol Processing of texts and hard words. Use of external context cues in CVA. Reasoning processes in CVA. Sense of word meaning gained from CVA.

139. 138 Research Assumptions: Good Readers Will know and be able to apply multiple CVA processes. Are wide readers who have— incidentally or deliberately—learned many words from reading. Will have excellent comprehension of text.

140. 139 Research Assumptions: CVA Processes CVA processes are a set of sub- strategies activated by a disruption of text processing caused by encountering an unknown word.

141. 140 Research assumptions: CVA processes continued CVA shares many characteristics of reading comprehension (e.g., use of selective text cues, prior knowledge, reasoning), but triggers text processing strategies different than ordinary comprehension fix-up strategies.

142. 141 Research Assumptions: CVA for Word Learning and Reading Comprehension Most readers apply CVA processes to gain meaning from the text, therefore gain word meanings — incidentally. We ask readers to try to gain a sense of the unknown word’s meaning — deliberately.

143. 142 Research Assumption: Conditional Factors Needed for Applying CVA Disrupted comprehension is required, or reader may just skip word. Word awareness: i.e., reader must note that there is a hard word (Reed, 1957; Harmon, 1999).

144. 143 Overview of Findings: 1.Processing of texts and hard words in CVA. 2.Use of text cues in CVA.

145. 144 Overview of findings continued 3.Reasoning processes in CVA. a.Hypotheses or model building. b.Inferential / abstract reasoning from reading comprehension. c.Within-sentence language cues. d.Information processing / knowledge acquisition processes (Sternberg, 1987) e.Global strategies. f.Prior knowledge in CVA processes.

146. 145 Overview of findings continued 4.Sense of the word meaning from CVA.

147. 146 Findings–1:Processing Texts and Hard Words After encountering hard word, some readers seemed to continue to read the full passage, then returning to the word to work on its meaning. Some readers stopped reading upon the hard word (or read to the end of the sentence), and worked on the word meaning immediately.

148. 147 Findings–2: Use of External Text Cues First, we classified the textual cues in texts using Ames (1969), Deighton (1978), Sternberg & Powell (1983), Ehrlich (1995), and Sternberg (1987). Second, we analyzed the think-aloud protocol to see if these were the clues readers used.

149. 148 Use of external text cues continued When reading and encountering an unknown word, readers generally started CVA with reasoning processes. Sometimes they went back to text, sometimes they did not.

150. 149 Use of external text cues continued After forming a hypothesis, some readers reinspected text to find support for hypothesis. Some readers created a hypothesis using general passage and sentence meaning, but did not go back to text.

151. 150 Use of external text cues continued Others said there was nothing in text to help them gain a sense of word meaning. Other readers did not go back to text at all unless prompted.

152. 151 Use of external text cues continued Readers inferred a sense of the word on the basis of: –general passage meaning, –meaning of the specific sentence, –sentence language and syntax, –prior knowledge, and –prior passages.

153. 152 Use of external text cues continued When readers did refer to the passage for specific information: –Usually to confirm a hypothesis. –Did not generally select the sentences we had predicted they would use.

154. 153 These Findings Lead Us in Three Directions: An unpredicted, but–with hindsight– logical conclusion. Curricular implications. Abandoning coding “available” cues in the text to coding reasoning processes applied to word meaning.

155. 154 Why is it Readers Did Not First Reinspect Text for Cues? They did not know the word’s meaning, so those text cues had no particular salience for the reader. For the researcher’s, however, these cues were salient, because we already knew the word’s meaning. Readers varied in what they accepted as a sufficient hypothesis. We did not teach readers specific cues to look for–wanted to see what they did independently.

156. 155 That is, when a reader knows a word’s meaning, that word’s connection to all the cues in the text is obvious. But when one does not know the meaning of the word, one does not readily discern the cues that provide insight to the word’s meaning.

157. 156 Our Conclusion Using Context for CVA is Easy When You Already Know Meaning of Word!

158. 157 Curricular Implications Teachers should model CVA with words they do not know in texts they have not previously seen. Students should practice CVA with words they do not know in texts they have not previously seen.

159. 158 Findings-3: Reasoning Processes in CVA Hypotheses or model building Inferring / abstract reasoning from reading comprehension Language cues Global strategies Background knowledge Conclusion

160. 159 Findings: Hypothesis Building All readers hypothesized a meaning of the hard word.

161. 160 On Further Encounters With Word, Readers Confirmed hypothesis if congruent with text, usually stating rationale; Revised hypothesis if not congruent, usually stating rationale; or If hypothesis not congruent with text, but not enough information in text to revise it, readers questioned hypothesis.

162. 161 Like Elshout-Mohr & van Daalen-Kapteijns’ Good Readers Our readers generally modified hypotheses in keeping with text. Our readers seemed to be aware that they did not really “know” the word’s meaning, that what they knew was a hypothesis.

163. 162 Unlike Deegan (1995 ) Our readers rarely altered the meaning of text to stay in keeping with prior hypothesis.

164. 163 Within Sentence Language Cues Familiar expressions Figurative language Connected series

165. 164 Inferring / Abstract Reasoning from Reading Comprehension Encoding selected information. Combining selected text information. Comparing selected information from text to background knowledge.

166. 165 Global Comprehension Strategies Visualizing. Summarizing. Clarifying. Self questioning. Insight. Confirming / confidence.

167. 166 Background Knowledge Essential Idiosyncratic Pervasive

168. 167 Findings–4: Sense of the Word Meaning from CVA Right and wrong are not useful descriptors of appropriateness of word meanings from CVA processes. Rational and defensible are better descriptors of appropriateness than is right or wrong. Gradual and cumulative over texts.

169. 168 Sense of the word continued 0. No meaning provided. 1.Don’t Know (dk) 2.Incorrect–No logical justification for sense of word. 3.Incorrect–Reasonable justification for sense of word proffered. 4.Incorrect–Based on language patterns, not general text meaning.

170. 169 Sense of the word continued 5.Vague or partial word meaning sense. 6.Approximate word meaning sense. 7.Nearly correct word meaning sense. 8.Correct sense of word meaning.

171. 170 How to Improve CVA in Classrooms Guess? Magical Mathematical Formula? CVA Strategies!

172. 171 How to improve CVA continued Teacher modeling of CVA strategies in texts with words whose meanings are not known. Scaffolding groups as they together think-aloud when applying CVA strategies in texts with words whose meanings are not known.

173. 172 How to improve CVA continued Guiding small groups in think-alouds of CVA strategies in texts with words whose meanings are not known. Student’s independent application of CVA strategies in texts with words whose meanings are not known.

174. 173 Protocol Study Limitations Reading done in a research environment. We required reader to think about the unknown word and its meaning. –Readers ordinarily might choose to skip or ignore word. –Readers ordinarily might not note word.

175. 174 Protocol limitations continued Our readers encountered word in multiple, consecutive texts. –Therefore, readers had immediate memory of the previous encounter. –This is atypical, as readers ordinarily might not encounter new word a second time for a long period.

176. 175 Protocol limitations continued We sometimes ask leading questions: e.g., –To activate background knowledge. –To direct reader back to text. –To elicit reasoning processes. –To ask reader to think again.

177. 176 Protocol limitations continued At times, we used non-words in place of real words: e.g., schmalion for tatterdemalion vedosarn for taciturn itresia for estuary.

178. 177 Protocol Study Strengths Used authentic texts. Words were, generally, difficult and not previously known by readers. Used one hard word in repeated texts, not multiple hard words each in a different text. Gained knowledge and insight on teaching CVA.

179. 178 Developing Meaning Vocabulary in Classroom There is no one way to teach meaning vocabulary, there is no best method. Helping children build their meaning vocabularies is a philosophy that shapes classroom culture and interaction, not a specific lesson plan.

180. 179 Preliminary to Recommendations CVA. Dictionaries. Roots and affixes. Direct teaching of vocabulary.

181. 180 Preliminary to recommendations

182. 181 Rationale for Our Recommendations On average, students learn 3000 words a year (Nagy & Herman, 1987) At most, students can be directly taught 300-400 words a year. Ergo, students learn 2,700 words per year outside of direct instruction.

183. 182 Rationale for recommendations continued That is, 10% of the words students learn per year on average may be directly taught. And, no less than 90% (and probably more) are learned by means other than direct teaching.

184. 183 Rationale for recommendations, continued This means that the extra effort in direct teaching of words will result in only an additional 50± words. Very little gain for so great an amount of instructional time.

185. 184 Recommendations Therefore, to promote vocabulary growth, maximize methods of learning word meanings other than dictionaries, morphology, direct instruction, and–even–deliberate contextual vocabulary acquisition.

186. 185 Overview of Recommendations Word consciousness. Word curiosity. Word fun. Teacher modeling and guided student practice in CVA. Wide reading.