| { |
| "paper_id": "J77-4002", |
| "header": { |
| "generated_with": "S2ORC 1.0.0", |
| "date_generated": "2023-01-19T02:50:02.690881Z" |
| }, |
| "title": "", |
| "authors": [ |
| { |
| "first": "Robert", |
| "middle": [], |
| "last": "Young", |
| "suffix": "", |
| "affiliation": { |
| "laboratory": "", |
| "institution": "University of Texas", |
| "location": {} |
| }, |
| "email": "" |
| } |
| ], |
| "year": "", |
| "venue": null, |
| "identifiers": {}, |
| "abstract": "", |
| "pdf_parse": { |
| "paper_id": "J77-4002", |
| "_pdf_hash": "", |
| "abstract": [], |
| "body_text": [ |
| { |
| "text": "1, T h e C o n t e n t of-C o n n e c t e d D i s c o u r s e , . , , . , , . , , . , , . .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "I n t r o d u c t i~n~. .~~-~~~~--~~~~,", |
| "sec_num": null |
| }, |
| { |
| "text": "The Nature of D i s c o u r s e C o n t e n t . , , . , , , . . , . . , , .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "-1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "1,1,1 B a r c l a y , B r a n s f o r d and FrAnks , , , . , . , , . , , , , .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "-1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "1 -1 -2 F r e d e r i k s e r l . . . a ., . , . , , . , , , , , . . . . . . . , . , . , , . , , . , . , , ,", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "-1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "G r i m e s and Meyer , . , , , , , , , , , . , . , , . , . , 1 2 . 4 K i n t s c h a n d v a n D i j k , . , , , , .a . . . . , . ,", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.2.3", |
| "sec_num": null |
| }, |
| { |
| "text": "1 -3 Some C o n s i d e r a t i o n s R e g a r d i n g Memory E x p e r i m e n t s . , , .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.2.3", |
| "sec_num": null |
| }, |
| { |
| "text": "-. ,", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.2.3", |
| "sec_num": null |
| }, |
| { |
| "text": "D i s c u s s i o n and Conclusians -. , -. , . . . . , . . . . . . .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.4", |
| "sec_num": null |
| }, |
| { |
| "text": "T h e Form bf C o n n e c t e d Discourse , , , . , . , ' , . . , . , , , ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ",", |
| "sec_num": "2" |
| }, |
| { |
| "text": "The Organization o f World Knowledge , . . . , a , , . , . , , a", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "3-2", |
| "sec_num": null |
| }, |
| { |
| "text": "S c h a n k a n d A b e l s o n , , , , , , , . . , . . . . . . , . The n a t u r e ( g e n e r a l c h a r a c t e r i s t i c s a n d f e a t u r e s ) of t h e m e a n i n g o f c o n n e c t e d t e x t w i l l b e e x a m i n e d f i r s t . T h e p r i n c i p l e f o c u s of t h e s e studies i s t o compare t h e u n d e r s t a n d i n g of a t e x t t o t h e a c t u a l t e x t a n d attempt t o c h a r n c t e f i z e t h e t y p e s of differences t h a t a r e f o u n d .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "3.201", |
| "sec_num": null |
| }, |
| { |
| "text": "Harrl~ly, Brainsford And Franks - i n t o t h e s e m a n t i c r e p r c s i -n t n t i c r l~ of t l~c b s t o r y a s i t i s u t~d e r s t o o d , and that t h e context can i n f l u~n c c l t l~c n i i ) o u~~t of i n f c rcnciny, t h a t i s done d u r i n g Page 9", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ". . 1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "I\\", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ". . 1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "F r e d e r i k s e n u s e s a s i m p l e , i n \u00a3 ormally def incd s e m a n t i c rep r e s e n t a t i o n i n the above s t u d i e s . s u b j e c t , r a t h e r than t h e a g e n t , was t h e most l i k e l y t o be kecalled. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ". . 1", |
| "sec_num": "1" |
| }, |
| { |
| "text": "Thorndyke - Sentences l i k e these are imbedded i n a meaningful t e x t , and a r e followed l a t e r i n t h e t e F t by a c o n t i n u a t i o n s e n t e n c e which r e f e r e n c e s t h e p r e v i o u s l y mentioned r e l a t i o n s h i p . T h i s c o n t i n u a t i o n s e n t e n c e might be n e u t r a l l i k e which s t r o n g l y suggests t h e i n f e r e n c e H e is f a t from e a t i n g Erench f r i e s .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.1.4", |
| "sec_num": null |
| }, |
| { |
| "text": "A d d i t i o n a l i n v", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.1.4", |
| "sec_num": null |
| }, |
| { |
| "text": "Thorndyke uses r e c o g n i t i o n tests a f t e r t h e p r e s e n t a t i o n of t h e story t o compare i n f e r e n c e s t h a t have been r e i n f o r c e d by a c o n t i n u a t i o n s e n t e n c e w i t h neutral and i n a p p rop r i a t e i n f e r e n c e s . H e found t h e r e i n f o r c e d i n f e r e n c e s much more l i k e l y t o be r e c o g n i z e d a s p a r t of She t e x t t h a n t h e n e u t r a l i n f e r e n c e s , w h i l e r e c o g n i t i o n of i n a p p r o p r i a t e i n f e r e n c~s was v e r y u n l i k e l y .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "1.1.4", |
| "sec_num": null |
| }, |
| { |
| "text": "a r e made and s t o r e d as p a r t of the u n d e r s t a n d i n g o f a t e x t , a more i m p o r t a n t implicaOion e x r s t s . T h i s i s that t h e r o l e of i n f e r e n c l ng i s t o a i d i n t h e 4 n t e g r e t i o t . r of new i n f o r m a t i o n i n t o tble l a r g e r framewi3tk of the u n d e r s t a n d i n g o,f a t e x t when n o a p p r o p r i a t e u n d e r s t a n d i n g could be o b t a i n e d from only the explicitly s t a t e d i n f o r m a t i o n . a r e a d e q u a t e f o r o n l y a r e s t r i c t e d c l a s s of t e x t s . The f o l l o w i n g s t u d i e r all p ropose some type of m u l t i -l e v e l r e p r e s e n t a t i o n o f c o n t e n t . Thc second c o q d i t i o n is t h a t t h e sequence of macro-st r u c t u r e s r e p r e s e n t i n g a coherent t e x t base must itself be coherent.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Thorndyke s u g g e s t s t h a t a l t h o u g h t h i s e v i d e n c e i n d i c a t e s t h a t inferences", |
| "sec_num": null |
| }, |
| { |
| "text": "R e c a l l i n g t h e d e f i n i t i o n of koherence, t h i s implies t h a t no macro-structure f r e q u e n t l y involved the p r c s c n t a t i o n of m a t e r i a l t h a t was t o t a l l y new and one must be very c a u t i o u s i n g e n e r a l i z i n g from t h e results of t e x t recall.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Thorndyke s u g g e s t s t h a t a l t h o u g h t h i s e v i d e n c e i n d i c a t e s t h a t inferences", |
| "sec_num": null |
| }, |
| { |
| "text": "which", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Describe events occurring t o t h e s e c h a r a c t e r s l e a d i n g t o or from", |
| "sec_num": "4." |
| }, |
| { |
| "text": "experiments due t o t h e many i n t e w c t i o n s between t h e s u b j e c t ' s knowledge, the e x p e r i m e n t a l s e t t i n g , t h e type and c o n t e n t of t h e t e x t and t h e e x t r a -e x p e r i m e n t a l effects ( o r lack of them) i n long term experiments. the h i g h e r -l e v e l o r g a n i z a t i o n f i r s t , and u s e s t h e s e s t r u c t u r e s t o a c c e s s more from t h e g e n e r a l understanding of t e x t . One o t h e r inadequacy o f t h e model i s w i l l be p r e s e n t e d q u i t e b r i e f l y , s i n c e t h e p r e s e n t a t i o n i s intended t o s u g g e s t", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Describe events occurring t o t h e s e c h a r a c t e r s l e a d i n g t o or from", |
| "sec_num": "4." |
| }, |
| { |
| "text": "that", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D i s c u s s i o n And Conclusions", |
| "sec_num": "1.4" |
| }, |
| { |
| "text": "t h e kinds of r e s u l t s t h a t have been obtained, r a t h e r t h a n t h e d e t a i l s of t h e s e conclusions. It i s necessary t o d e f i n e t h e s e ideas i n a computational way t h a t can be i n t e g r a t e d i n t o t h e models of t e x t understanding, and some r e c e n t work h a s begun t h i s task. All of t h e s e a n a l y s e s r e l y on t h e i n t u i t i v e understanding of t h e a n a l y z e r ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D i s c u s s i o n And Conclusions", |
| "sec_num": "1.4" |
| }, |
| { |
| "text": "Mandler and Johnson ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "Mandler And Johnson", |
| "sec_num": "2.4" |
| }, |
| { |
| "text": "Omissions and v i o l a t i o n s of the i d e a l s t r u c t u r e w i l l be r e f l e c t e d by ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "6.", |
| "sec_num": null |
| }, |
| { |
| "text": "The p r i n c i p a l hypotheses of t h e s e i n v e s t i g a t o r s i s t h a t t h e r e is are very high l e v e l s t r u c t u r e s which o r g a n i z e c o n t e n t s t r u c t u r e s i n t h e r e p r e s e n t a t i o n Two s e q u e n t i a l EVENTS may be r e l a t e d by e i t h e r CAUSE o r ALLOW). Mandler and Johnson, and t o some e x t e n t Thorndyke, ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "D i s c u s s i o n And Conclusions", |
| "sec_num": "2.7" |
| }, |
| { |
| "text": "Page 47", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "goal, i t is s t i l l important t o r e t a i n i t as part of the goal of understanding t e x t , and t o r e a l i z e t h a t t h e r i c h n e s s of human u s e of language w i l l be only p a r t i a l l y accounted f o r without t h i s component", |
| "sec_num": null |
| }, |
| { |
| "text": "T h i s section w i l l survey computational work on t h e process of text understanding.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "Computational models w i l l be seen t o be of v a l u e i n a t least two ways. FLrst, many of t h e f e a t u r e s postulated i n o t h e r models w i l l be found i n t h e s e computational models, frequently motivated primarily by computational concerns. This tends t o support these hypotheses. Secondly, John drinks a l o t of coffee. John married a blonde.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "Both s e n t e n c e s r e f e r t o John, but do n o t form a coherent discourse.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "Furthermore, r e f e r e n t i a l i d e n t i t y i s unnecessary, Consider t h e two sentences:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "There has been a drought. People are s t a r v i n g , Today was J a c k ' s b i r t h d a y . Penny and J a n e t went t o t h e s t o r e . They were going t o get p r e s e n t s . J a n e t d e c i d e d t o g e t a top. \"Don't do t h a t \" Penny said. \"Jack h a s a top. H e w i l l make you t a k e i t bock.\"", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "These", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "The \"it\" i n t h e l a s t l i n e i s normally understood 2. Enablement inferences suggest s t a t e s t h a t were n e c e s s a r i l y t me input: Mary gave John a book, inference: Mary had t h e book just before she gave it t o John. i n f e r e n c e : P e t e probably has a g a l l bladder. S u r e l y , the t r a n s f e r d f p o s s e s s i a n i a t h e meaning of t h e s e n t e n c e , n o t j u . The former must always be t r u e when a n act occurs. RNCs, however, tare normally t r u e , b u t may be v i o l a t e d w i t h o u t c r e a t i n g an anomalous s i t u a t i o n .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "COMPUTATIONAL MODELS OF TEXT UNDERSTANDING", |
| "sec_num": "3.0" |
| }, |
| { |
| "text": "John was working i n h i s yard. Mowing t h e lawn h a s as ANCs such t h i n g s a s t h e e x i s t e n c e of John, the lawn, a mower, e t c .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "For example", |
| "sec_num": null |
| }, |
| { |
| "text": "and as RNCs t h a t i t is good weather and t h e grass needs mowing. The f i n a l sentence of t h e s t o r y i l l u s t r a t e s t h e p o i n t about normality considerations. To e x p l a i n t h e tomato sauce, one must i n f e r t h a t John g o t i t an h i s toe, b u t t h i s seems l i k e an a b n o m l i n f e r e n c e t o make w i t h no o t h e r explanation. However, i f t h e paragraph was preceded by t h e sentences:", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "For example", |
| "sec_num": null |
| }, |
| { |
| "text": "John was e a t i n g p i z z a o u t s i d e on his lawn, He n o t i c e d t h a t the grass was v e r y long and and h e g o t out his mower.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "For example", |
| "sec_num": null |
| }, |
| { |
| "text": "Then t h e final sentence seems much more reasonable.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "For example", |
| "sec_num": null |
| }, |
| { |
| "text": "One would i n f e r t h a t p i z z a has tomato sauce, and John dropped some on t h e lawn and l a t e r stepped i n ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "For example", |
| "sec_num": null |
| }, |
| { |
| "text": "A s t o r y i s t h e j o i n i n g of causal c h a i n s t h a t crilminote i n the \"point\" o f t h e s t o r y . O t h e r p a t h s are of less i n t e r e s t .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ".", |
| "sec_num": "4" |
| }, |
| { |
| "text": "The O r g a n i z a t i o n O f World Knowledge 2. T h e r e i s some t y p e of h i e r ; r c h i c a l o r m u l t i -l e v e l o r g a n i z a t i o n ( s ) o f t h e u n d e r s t a n d i n g of a t e x t .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": ". 2", |
| "sec_num": "3" |
| }, |
| { |
| "text": "3. Organized, p r e -e x i 5 t e n t knowledge i s requlred t o a c h i e v e this type of understanding. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "O t h e r Work", |
| "sec_num": "3.3" |
| }, |
| { |
| "text": "t h e e n t i r e text was r e c a l l e d , t h e r e p r o d u c t i o n of t h e p a r t i c u l a r p a r a g r a p h wa8 much smaller and l e s s a c c u r a t e t h a n i n t h e f i r s t two c a s e s . T h e i r c o n c l u s i o n s from these r e s u l t s axe t h a t t h e r e c a l l o f s m a l l amounts of t e x t i s a d i f f e r e n t p r o c e s s from t h e r e c a l l of a l o n g t e x t , t h a t r e c a l l o f a l o n g t e x t relies . on t h e m a c r o -s t r u c t u r e of t h e t e x t a s a means of o r g a n i z i n g t h e t e x t , t h a t t h e m i c r o -s t r u c t u r e i s f o r g o t t e n much more e a s i l y and t h a t summarizing i s based on t h e macro-stcructuxe. F i n a l l y , some e x p e r i m e n t s were done i n which i n c o r r e c t summaries w e r e p r e s e n t e d p r i o r t o t h e p r e s e n t a t i o n of t h e s t o r y . These were found t o have p r a c t i c a l l y no i n f l u e n c e on t h e f i n a l u n d e r s t a n d i n g o' f t h e s t o r y , and t h e s u b j e c t s were u n a b l e t o a c c u r a t e l y r e c a l l t h e i n c o r r e c t summaries.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "t h a t the meaning of a text d i f f e r s from t h e meaning of i t s i n d i v i d u a l s e n t e n c e s , i n t h a t p r i o r knovledge i s u s e d t o i n f e r implicit i n f o r m a t i o n during comprehension, i s g e n e r a l l y accepted. T h e i d e a t h a t t h i s meaning must be w e l l s t r u c t u r e d , o r coherent, i s a l s o expressed d i r e c t l y o r i n d i r e c t l y by most of t h e r e s e a r c h e r s . However, e x p l i c i t disagreement e x i s t s on both the n a t u r e of content s t r u c t u r e , and i t s r e l a t i o n s h i p t o r e c a l l .Crothers and", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "t e x t . So, f o r example, S p i r o ' s s u b j e c t s i n t e g r a t e d i n f o r m a t i o n from t h e s t o r i e s t h e y r e a d i n t o t h e i r g e n e r a l e x p e c t a t i o n s a b o u t l o v e and c o u r t s h i p , and l o s t t h e s p e c i f i c s o f t h e t p x t , even r e c a l l i n g i t w i t h r a d i c a l mods f i , c a t i o n s T h i s p x o c e s s would e x p l a i n t h e a l m o s t t o t a l l o s s of d e t a i l i n f o r m a t i o n with t i m e , a s w e l l a s t h e i n t e r a c t i o n s and d i s t o r t i o n s caused by p r e v i o u s knowledge.", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "t h a t he developed , he designated a function. These f u n c t i o n s a c t l i k e meta-actions i n t h a t t h e y d e s c r i b e major e v e n t s o r event complexes t h a t were", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| }, |
| { |
| "text": "t o suspension of the restaurant script f o r a sub-story, o r t o its abandonment, Furthermore, simultaneous,, independent s c r i p t s are possible. The s i t u a t i o n of eating i n a dining c a r involves both t h e t r a i n s c r i p t and the restaurant s c r i p t . Finally, the activation and t e r m i n a t i o n of scripts i s a very complex p roblem. F o r example, consider t h e sequence: John went t o a restaurant, A f t e r e a t i n g l o b s t e r , he bought a watch. Does t h i s describe a restaurant event followed by a purchasing event, o r t h e more unusual case of buying a watch i n a restaurant? Either i s possible, and the story i n which t h i s s,equence i s found would determine the most probable i n t e r p r e t a t i o n .", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "", |
| "sec_num": null |
| } |
| ], |
| "back_matter": [ |
| { |
| "text": "c a u o a l i t y , time ordering and s p a t i a l r e l a t i o n s are rep r e s e n t e d by arcs between t h e M o d a l i t y nodes of e v e n t s . P h i l l i p s d e f i n e s a n o t h e r t y p e of r e l a t i o n a h i p which h e d e s i g n a t e s t h e M e t a l i n g u a l o r g a n i z a t i o n of knowledge. ", |
| "cite_spans": [], |
| "ref_spans": [], |
| "eq_spans": [], |
| "section": "annex", |
| "sec_num": null |
| } |
| ], |
| "bib_entries": { |
| "BIBREF2": { |
| "ref_id": "b2", |
| "title": "The Role of Compreh'ension i n Remembering Sentences. F. C. 1932. Remembering: -A Study -i n Experimental -and S o c i a l Psychology", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [ |
| "R" |
| ], |
| "last": "Barclay", |
| "suffix": "" |
| } |
| ], |
| "year": 1973, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Barclay, J. R. 1973. The Role of Compreh'ension i n Remembering Sentences. F. C. 1932. Remembering: -A Study -i n Experimental -and S o c i a l Psychology, London: Cambridge U n i v e r s i t y Press.", |
| "links": null |
| }, |
| "BIBREF3": { |
| "ref_id": "b3", |
| "title": "A Tagmemic Approach t o Paragraph Analysis", |
| "authors": [ |
| { |
| "first": "A", |
| "middle": [ |
| "L" |
| ], |
| "last": "Becber", |
| "suffix": "" |
| } |
| ], |
| "year": 1965, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Becber, A. L. 1965. A Tagmemic Approach t o Paragraph Analysis.", |
| "links": null |
| }, |
| "BIBREF5": { |
| "ref_id": "b5", |
| "title": "On a Condition of t h e Coherence of Text", |
| "authors": [ |
| { |
| "first": "I", |
| "middle": [], |
| "last": "Bellert", |
| "suffix": "" |
| } |
| ], |
| "year": 1970, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bellert, I. 1970. On a Condition of t h e Coherence of Text.", |
| "links": null |
| }, |
| "BIBREF7": { |
| "ref_id": "b7", |
| "title": "Represent ation -and Understanding", |
| "authors": [ |
| { |
| "first": "D", |
| "middle": [ |
| "G" |
| ], |
| "last": "Bobrow", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bobrow, D. G, and C o l l i n s , A. (eds.) 1975, Represent ation -and Understanding. New York: Academic Press.", |
| "links": null |
| }, |
| "BIBREF8": { |
| "ref_id": "b8", |
| "title": "G U~; A Frame-Driven Dialogue System. A r t i f i c i a l I.ntellip,ence", |
| "authors": [ |
| { |
| "first": "D", |
| "middle": [ |
| "G" |
| ], |
| "last": "Bobxow", |
| "suffix": "" |
| }, |
| { |
| "first": "R", |
| "middle": [ |
| "M" |
| ], |
| "last": "Kaplan", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Kay", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "Norman", |
| "suffix": "" |
| }, |
| { |
| "first": ".", |
| "middle": [ |
| "H" |
| ], |
| "last": "Thompson", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Winograd", |
| "suffix": "" |
| } |
| ], |
| "year": 1977, |
| "venue": "", |
| "volume": "8", |
| "issue": "", |
| "pages": "155--173", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bobxow, D. G., Kaplan, R. M., Kay, M a , Norman, D. A @ , Thompson. H. and Winograd, T o .1977. G U~; A Frame-Driven Dialogue System. A r t i f i c i a l I.ntellip,ence, 8, 2 , pp. 155-173.", |
| "links": null |
| }, |
| "BIBREF9": { |
| "ref_id": "b9", |
| "title": "Some P r i n c i p l e s of Memory Schemata. i n Bobrow and C o l l i n s", |
| "authors": [ |
| { |
| "first": "D", |
| "middle": [ |
| "G" |
| ], |
| "last": "Bobrow", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [ |
| "A" |
| ], |
| "last": "Norman", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "75", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Bobrow, D. G. and Norman, D. A. 1975, Some P r i n c i p l e s of Memory Schemata. i n Bobrow and C o l l i n s [75].", |
| "links": null |
| }, |
| "BIBREF10": { |
| "ref_id": "b10", |
| "title": "An Overview of KRL", |
| "authors": [ |
| { |
| "first": "D", |
| "middle": [ |
| "G" |
| ], |
| "last": "Bobrow", |
| "suffix": "" |
| }, |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Winograd", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [ |
| "D" |
| ], |
| "last": "Bransford", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [ |
| "R" |
| ], |
| "last": "Barclay", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Franks", |
| "suffix": "" |
| } |
| ], |
| "year": 1972, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "3--46", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "@ Bobrow, D. G. and Winograd, T. 1977. An Overview of KRL, A m w l e d g e , 1, 1, pp. 3-46, Bransford, J. D , , Barclay, J. R. and Franks, J. J, 1972. Sentence Memory: A", |
| "links": null |
| }, |
| "BIBREF12": { |
| "ref_id": "b12", |
| "title": "T s : A Review. _ -Cognition", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Eransford", |
| "suffix": "" |
| }, |
| { |
| "first": "D", |
| "middle": [], |
| "last": "", |
| "suffix": "" |
| }, |
| { |
| "first": "F", |
| "middle": [], |
| "last": "R A N K S", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [ |
| "J" |
| ], |
| "last": "", |
| "suffix": "" |
| } |
| ], |
| "year": 1972, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "211--249", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Eransford, J o D, and F r a n k s , J. J. 1972, T s : A Review. _ - Cognition, . 1, 2 , pp. 211-249.", |
| "links": null |
| }, |
| "BIBREF15": { |
| "ref_id": "b15", |
| "title": "F i r s t Technical Report, C o n t r a s t i v e L i n g u i s t i c s P r o j e c t . Department of L i n g u i s t i c s", |
| "authors": [ |
| { |
| "first": "W", |
| "middle": [], |
| "last": "Chafe", |
| "suffix": "" |
| } |
| ], |
| "year": 1972, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chafe W. 1972. F i r s t Technical Report, C o n t r a s t i v e L i n g u i s t i c s P r o j e c t . Department of L i n g u i s t i c s , U n i v e r s i t y o f C a l i f o r n i a st Berkeley.", |
| "links": null |
| }, |
| "BIBREF16": { |
| "ref_id": "b16", |
| "title": "Some Thoughts on Schemata", |
| "authors": [ |
| { |
| "first": "", |
| "middle": [], |
| "last": "Chafe", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "Nash-Webber", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Chafe, W i 1975. Some Thoughts on Schemata. i n Schank . , . and Nash-Webber [75]", |
| "links": null |
| }, |
| "BIBREF20": { |
| "ref_id": "b20", |
| "title": "M s . Maloprop, A Language Compreh", |
| "authors": [ |
| { |
| "first": "E", |
| "middle": [], |
| "last": "Charniak", |
| "suffix": "" |
| } |
| ], |
| "year": 1977, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Charniak, E. 1977. M s . Maloprop, A Language Compreh.-nsion Program.", |
| "links": null |
| }, |
| "BIBREF22": { |
| "ref_id": "b22", |
| "title": "A P a r t i a l Grammar f o r Eskimo F o l k t a l e s", |
| "authors": [], |
| "year": null, |
| "venue": "American Anthropologist", |
| "volume": "73", |
| "issue": "", |
| "pages": "645--662", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "A P a r t i a l Grammar f o r Eskimo F o l k t a l e s . American Anthropologist, 73, pp. 645-662,", |
| "links": null |
| }, |
| "BIBREF24": { |
| "ref_id": "b24", |
| "title": "John Wiley and sons. jk, T. A. 1972. Some Aspects --of Text Grammars. The Hague: Mouton. van D i jk, T. A. 1974. Models of Macro-stmctures. , T. A. 1975b. Discourse Meaning and Memory", |
| "authors": [ |
| { |
| "first": "Languafie", |
| "middle": [], |
| "last": "Comprehension", |
| "suffix": "" |
| }, |
| { |
| "first": "-", |
| "middle": [], |
| "last": "-And T H E N -Of Knowledge", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Languafie Comprehension --and t h e n -of Knowledge. New York: John Wiley and sons. jk, T. A. 1972. Some Aspects --of Text Grammars. The Hague: Mouton. van D i jk, T. A. 1974. Models of Macro-stmctures. , T. A. 1975b. Discourse Meaning and Memory. t o a p p e a r i n J o u r n d of -Reading Behavior.", |
| "links": null |
| }, |
| "BIBREF25": { |
| "ref_id": "b25", |
| "title": "Macro-structures and Cognition", |
| "authors": [ |
| { |
| "first": "T", |
| "middle": [], |
| "last": "Van D I J K", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [], |
| "last": "", |
| "suffix": "" |
| } |
| ], |
| "year": 1976, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "van D i j k , T, A. 1976. Macro-structures and Cognition.", |
| "links": null |
| }, |
| "BIBREF27": { |
| "ref_id": "b27", |
| "title": "A Computational Approach -t o -Discourse A n a l y s i s", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Hobbs", |
| "suffix": "" |
| } |
| ], |
| "year": 1976, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Hobbs, J o 1976. A Computational Approach -t o -Discourse A n a l y s i s . Research r e p o r t 76-2, Department of Computer S c i e n c e , City College, CUNY.", |
| "links": null |
| }, |
| "BIBREF34": { |
| "ref_id": "b34", |
| "title": "Narrative Analysis", |
| "authors": [ |
| { |
| "first": "W", |
| "middle": [], |
| "last": "Labov", |
| "suffix": "" |
| }, |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Waletzky", |
| "suffix": "" |
| } |
| ], |
| "year": 1967, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Labov, W. and Waletzky, J. 1967. Narrative Analysis:", |
| "links": null |
| }, |
| "BIBREF35": { |
| "ref_id": "b35", |
| "title": "Remembrance of Things Parsed", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [ |
| "M" |
| ], |
| "last": "Handler", |
| "suffix": "" |
| }, |
| { |
| "first": "N", |
| "middle": [ |
| "S" |
| ], |
| "last": "Johnson", |
| "suffix": "" |
| } |
| ], |
| "year": 1977, |
| "venue": "", |
| "volume": "9", |
| "issue": "", |
| "pages": "111--151", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Handler, J. M. and Johnson, N. S. 1977. Remembrance of Things Parsed: , 9, 1, pp. 111-151.", |
| "links": null |
| }, |
| "BIBREF36": { |
| "ref_id": "b36", |
| "title": "The Metanovel", |
| "authors": [ |
| { |
| "first": "J", |
| "middle": [], |
| "last": "Meehan", |
| "suffix": "" |
| } |
| ], |
| "year": 1976, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Meehan, J. 1976. -The Metanovel: --", |
| "links": null |
| }, |
| "BIBREF37": { |
| "ref_id": "b37", |
| "title": "The O on Recall", |
| "authors": [ |
| { |
| "first": "B", |
| "middle": [ |
| "J" |
| ], |
| "last": "Meyer", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Meyer, B. J. 1975. -.-The O on Recall, Amsterdam: North Holland.", |
| "links": null |
| }, |
| "BIBREF38": { |
| "ref_id": "b38", |
| "title": "A Framework f o r Representing Knowledge", |
| "authors": [ |
| { |
| "first": "M", |
| "middle": [], |
| "last": "Minsky", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Minsky, M. 1975. A Framework f o r Representing Knowledge, i n Winston, P.", |
| "links": null |
| }, |
| "BIBREF39": { |
| "ref_id": "b39", |
| "title": "The Psycholo_sy of Computer Vision", |
| "authors": [], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "The Psycholo_sy of Computer Vision, New York: McGraw-Hill.", |
| "links": null |
| }, |
| "BIBREF41": { |
| "ref_id": "b41", |
| "title": "Computer Understanding of P h y s i c s Problems S", |
| "authors": [ |
| { |
| "first": "G", |
| "middle": [], |
| "last": "Novak", |
| "suffix": "" |
| } |
| ], |
| "year": 1976, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Novak, G. 1976. Computer Understanding of P h y s i c s Problems S t a t e d i n N a t u r a l", |
| "links": null |
| }, |
| "BIBREF43": { |
| "ref_id": "b43", |
| "title": "Conceptual Memory. i n Schank [75a", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [ |
| "J" |
| ], |
| "last": "Rieger", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rieger, C. J. 1975a. Conceptual Memory. i n Schank [75a].", |
| "links": null |
| }, |
| "BIBREF44": { |
| "ref_id": "b44", |
| "title": "Conceptual Overlays: A Schank, R. and A b c l s o n , R. 1977. -Scripts. P l a n s , Goals and Understanding", |
| "authors": [ |
| { |
| "first": "C", |
| "middle": [ |
| "J" |
| ], |
| "last": "Rieger", |
| "suffix": "" |
| } |
| ], |
| "year": null, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Rieger, C. J. 197Sb. Conceptual Overlays: A Schank, R. and A b c l s o n , R. 1977. -Scripts. P l a n s , Goals and Understanding. . : 1,awrence' Erlbaum Associates.", |
| "links": null |
| }, |
| "BIBREF45": { |
| "ref_id": "b45", |
| "title": "T h e o r e L i c a l I s s u e s in N a t u r a l L a n~u a~e F r o c~e s s i n~. A s s o c i~t", |
| "authors": [], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Schank, R. and ?ash-IJebber, B. (eds.) 1975. T h e o r e L i c a l I s s u e s in N a t u r a l L a n~u a~e F r o c~e s s i n~. A s s o c i~t i o n f o r C c h p u t a t i o n a l L i n g u i s t i c s .", |
| "links": null |
| }, |
| "BIBREF46": { |
| "ref_id": "b46", |
| "title": "SAM: A S t o r y Understander", |
| "authors": [ |
| { |
| "first": "R", |
| "middle": [], |
| "last": "Schank", |
| "suffix": "" |
| }, |
| { |
| "first": "A", |
| "middle": [ |
| "I" |
| ], |
| "last": "E And T H E Yale", |
| "suffix": "" |
| }, |
| { |
| "first": "", |
| "middle": [], |
| "last": "P R O J E C T", |
| "suffix": "" |
| } |
| ], |
| "year": 1975, |
| "venue": "", |
| "volume": "", |
| "issue": "", |
| "pages": "", |
| "other_ids": {}, |
| "num": null, |
| "urls": [], |
| "raw_text": "Schank, R e and t h e Yale A. I. P r o j e c t . 1975. SAM: A S t o r y Understander. Research r e p o r t -no-43, Department; of Computer S c i e n c e , Yale ' U n i v e r s i t y .", |
| "links": null |
| } |
| }, |
| "ref_entries": { |
| "FIGREF0": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "2 C r o t h e r s . , , . . . . , , ." |
| }, |
| "FIGREF1": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t r u c t u r a l Analysis of Text -, , . , , -. , , . . . , . , d l e r a n d J o h n s o n -, , , , , , -, , . . . . , * . , . ." |
| }, |
| "FIGREF2": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "c u s s i o n and C o n c l u s i o n s . , . , . , , . . . , . , . , , , 3, C o m p u t a t i o f i a l Models of Text U n d e r s t a n d i n g . . , , . . , 3.1 T h e N e c e s s i t y o f World Knowledge rn , , . , . , . . . * , , ," |
| }, |
| "FIGREF3": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "e r w o r k . . -. I . , , . , , --~-~, -~, INTRODUCTION The goal of t h i s s t u d y i s t o e x a m i n e work that has s o m e t h i n g t o offer toward t h e c o n s t r u c t i o n of a c o m p u t a b l e model o f t e x t u n d e r s t a n d h g , a n d t h e r e f o r e t o w a r d c o g n i t i v e models i n g e n e n i l . T h e reason for the c r i t e r i o n of u i r e r n e v t makes vague g c n e r n l i z a t i o t~s i~n p o s i s i b l e ( o r a t l e a s t more d i f f i c u l t ) , and f o r c e s e x a c t s p e c i f i c a t i o n o f p r o c c s s c s being h y p o t h e s i z e d -(However, t h i s c r i t e r i o n i s f r e q u e n t l y d i f f i c u l t t o a p p l y , and i n some i n s t a n c e s t h e a u t h o r s d e c i s i o n was u n d o u b t e d l y a s u b j e c t i v c one. )I t s h o u l d be clearly p o i r l t c d o u t that n o a t t e m p t i s made h e r e t o dealw i t h t h e s e m a n t i c s of i n d i v i d u a l s e n t e n c e s .A f a m i l i a r i t y w i t h r e c e n t work i n s e m a n t i c r e p r e s e n t a t i a n i nc o g n i t i v e p s y c h o l o g y , l i n g u i s t i c s a n d a r t i f i c i a l i n t e l l i g e n c e i s assumed (see Norman a n d R u m e l h a r t [75], Schank a n d C o l b y [73] a n d S t e i n b e r g and Jakbbovifis [ 7 1 1 ) A l t h o u g h many p r o b l e m s i n r e p r e s e n t i n g t h e meaning o f i n d i v i d u a l s e n t e n c e s remain u n s o l v e d , t h i s study f o c u s e s an t h o s e a s p e c t s of mcanlng t h a t are conveyed o n l y by g r o u p s of c o n n e c t e d s e n t e n c e st e x t s . A d d i t i o n a l l y , o n l y work t h a t a t t e m p t s t o deal w i t h t h e s e m a n t i c s o r u n d e r s t a n d i n g o f t e x t s , a s opposed t o s t a t i s t i c a l o r s y n t a c t i c a n a l y s i s , i s c o n s i d e r e d . T h i s f o c u s on t e x t has also led t o t h e o m i s s i o n of s t u d i e s o f n o n -t e x t u a l mpmory and o f c o m p u t a t i o n a l s y s t e m s t h a t u n d e r s t a n d and solve p r o b l e m s s t a t e d i n E n g l i s h o r c a r r y on d i a l o g u e . T h i s o m i s s i o n i s n o t meant. t o imply t h a t work i n these areas h a s n o r e l e v a n the d i f f c r e n t focus and a d d i t i o n a l c o n s t r a i n t s of t h e s e s t u d i e s make s u c h i m p l i c a t i o n s d i f f i c u l t t o i s o l a t e . The s t u d y i s d i v i d e d i n t o t h r e e p a r t s , which o c c a s i o n a l l y o v e r l a p , The f i r s t s e c t i o n deals w i t h t h e c o n t e r l t of c o n i~c c t e d t e x tt h a t i s , what e x a c t l y d o e s t e x t commtlnicate. The work i n t h i s s e c t i o n i s p r i m a r i l y t h a t o f e x p e r i m e n t a l p s y c h o l o g i~t s i n t e r e s t e d i n memory and r e c a l l -The s e c o n d s e c t i o n d e a l s w i t h t h e _ s t r u c r u r q o f t e x t , a p a r t from i t s s p e c i f i c c o n t e n t , T h i s work h a s b e e n d o n e mairlly by c u l t u r a l a n t l~r o p o l o g i s t s and t h o s e i n t e r e s t e d i n t h e t h e o r y of l i t e r a t u r c , b u t i s a l s o b e i n g i n v e s t i g a t e d by e x p e r i m e n t a l p s y c h o l o g i s t s . T h e t h i r d s e c t i o n d i s c u s s e s c o m p u t a t i o n a l models prop-osed by c o m p u t a t i o n a l l i n g u i s t s , w h i c h i n c l u d e a t t e m p t s t o implement i n c o m p u t e r p r o g r a m s some o f t h e p r o c e s s e s d i s c u s s e d i n t h e o t h e r s e c t i o n s -Each s e c t i o n i s c o n c l u d e d w i t h a d l s c~~s s i o n of t h e emerging model of t e x t u n d e r s t a n d i n g . F i n a l l y , some d i r e c t i o n s f o r n e e d e d r e s e a r c h are s u g g e s t e d . I t i s hoped t h a t t h i s k i n d o f i n t e r d i s c i p l i n a r y s u r v e y w i l l c a l l a t t e n t i o n o f w o r k e r s 1 1 2 o n e area t o work i n o t h e r a r e a s a d d r e s s i n g t h e same p r o b l e m s , T h i s t y p e o f communication i s v a l u a b l e i n two d i s t i n c t ways. When s i m i l a r c o n c l u s i o n s are r e a c h e d i n d i f f e r e n t d i s c i p l i n e s f r e q u e n t l y u s i n g d i f f e r e n t m e t h o d s a n d w i t h somewhat d i f f e r e n t g o a l s , t h e s e c o n c l u s i a n s must b e g i v e n s p e c i a l c r e d e n c e , On t h e o t h e r h a n d , i t i s sometimks t h e c a s e t h a t one d i s c i p l i n e c o m p l e t e l y o v e r l o o k s a p r o b l e m o r some a s p e c t o f it d u e to t h e i r own i n t e r e s t s and b i a s e s . Such o m i s s i o i~s n e e d t o be b r o u g h t i n t o s h a r p f o c u s . CONTENT OF CONNECTED DISCOURSE T h i s s c c t f o n w i l l p r e s e n t a b r i e f s u r v e y of the r e c e n t work of some e x p e r i m e n t a l psychologist,^ c o n c e r n e d with t h c p r o c e s s of u n d e r s t a n d i n g c o r l n c c t c d d i s c o u r s e . Thc d e t a i l s o f t h e i r c x p e r i~n~n t s w i l l not be p r e s e n t e d , b u t t h e t y p e s o f c x p e r i r n e a l s and g e n e r a l c o n c l u s i o n s w i l l be d i s c u s s e d , a s w e l l a s any madcls p r o p o s e d by t h e i n v e s t i g a t o r s . T h~~s e r e s u l t s seem t o cotlve r g e t o w a r d a s i n g l e model, i n s p i t e of a p p a r e n t c o n t r a d i c t i o n s , P s y c h o l o g i c a l work on t e x t u n d e r s t a n d i n g i s r e l e v a n t t o c o m p u t a t i o n a l m o d e l s b e c a u s e humans are s t i l l t h e o n l y good tcxt u n d c r s t n n d c r s . Also, although some a s p e c t s of p s y c h o l o g i c a l r e s e a r c h , such a s f o r g e t t i n g , a r e n o t currently i n c l u d e d i n a n y s e r i o u s way i n c o m p u t a t i o n a l m o d e l s , i t w o u l d a p p e a r t h a t such models h a v e a v e r y u s e f u l r o l e t o p l a y i n p o i n t i n g o u t where significant d e t a i l s of t h e p r o c e s s of t e x t u n d e r s t a n d i n g a r e b e i n g h a n d l e d i n t u i t i v e l y b y t h e t h e o r i s t , a n d h a v e n o t r e a l l y b e e n d e f i n e d i n t h e t l l e o r y , 1.1 The N a t u r e O f D i s c o u r s e C o n t e n t" |
| }, |
| "FIGREF4": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "n i m tl\\r~t the itrformnt ion used t o construct the selnantic d t~*~-l p t [ o t~ is not wholly contained i n the sentences, but that an underatander lr is ~rrtsvious knowlcdgc. t o a great e x t e n t . They do not make any d e turt lcs r a s t c d hoaidc u f l o a t i n g log, and a f i s h swam beneath t 111-in. 3. 'l'1rrt.c t u r t l m r~8 t~d on a floating log, and a f i s h swam beneath ttiem.Thtplryhi c;\\ t s i t t i n t ion d c~c r i b o d by t h e accond sentctlcc i s essentially t thrm\" is rcplaccd u i e h \"Jt\". T h i s situation d i f f e r s fromt I),\\ t (it-st.-rl \\red i n t lrc f i r s t sent crlct', howcve r. Whcn s u b jc c t s we re p r e s e n t e d w t r 11 .I hcnt c~~c i \\ i n whicll tllc p ronor11\\ substi rut ion had bcrn made, bhosc who had II(*,II tl tlrc* f l r r t scbtrt cl~cc-were able t o make thc d i s t i n c t i o n , whf Ic those who I \\\\vnrtf I r;ccot\\d mnctc no di~tinction. The explanation offered i s that r ; \\~l l jr'ct rz i~r;r\\cl thc2 t spntinl kt.rowlcdgc t o crclate s i t u a t i o t~ d c s c r i p r i o n s , and when two d i f f e r e n t s e n t e n c e s p r o d u c e t h e same d e s c r i p t i o n , i t is d i f f i c u l t t o d i s t i n g u i s h them. S i m i l a r i n v e s t i g a t i o n s u s e d a s e q u e n c e o f s e n t e n c e s , e a c h comparing two o b j e c t s from a set o f f i v e a c c o r d i n g t o some d i m e n s i o n (e.g. p o s i t j o n , h e i g h t , s p e e d , w e i g h t ) . S u b j e c t s who knew t h a t a s i n g l e s i t u a t i o n was b e i n g described had a b e t t e r memory f o r t h e s i t u a t i o n , b u t were less a b l e t o d i s t i n g u i s h s e n t e n c e s t h e y had h e a r d from t h o s e i m p l i e d by t h e d e s c r i b e d s i t u a t i o n . Once again, t h e c o n c l u s i o n i s t h a t a n o n -l i n g u i s t i c r e p r e s e n t a t i o n i s c r e a t e d when a m e a n i n g f u l t e x t i s u n d e r s t o o d . A f i n a l group o f s t u d i e s i n v o l v e d d e s c r i p t i v e t e x t s t h a t were ambiguous o r d i f f i c u l t t o comprehend w i t h o u t a n a p p r o p r i a t e c o n t e x t . The context m i g h t be a p i c t u r e of t h e d e s c r i b e d s i t u a t i o n o r a m e a n i n g f u l t i t l e f o r t h e t e x t . I n e v e r y case t h e r e s u l t s i n d i c a t e d t h e u n d e r s t a n d e r a t t e m p t e d t o b u i l d a d e s c r i p t i o n of t h e whole s i t u a t i o n , and when t h i s was v e r y d i f f i c u l t o r i m p o s s i b l e , comp r e h e n s i o n was low. ( R e l a t e d s t u d i e s by Anderson, R e y n o l d s , S c h a l l e r t and C o e t z [ 7 6 ] and S c a l l e r t [76] u t i l i z e t e x t s , e a c h w i t h two t o t a l l y d i f f e r e n t meanings. Evidence is o b t a i n e d t h a t one o r t h e o t h e r a l t e r n a t i v e i s a l m o s t a l w a y s chosen f o r t h e e n t i r e t e x t . The e f f e c t of s u b j e c t s backgrounds o n i n t e r p r e t a t i o n s of these ambiguous t e x t s I s also Gtudied, and a c o r r e l a t i o n between background and c h o s e n i n t e r p r e t a t i o n i s claimed. ) The p r i n c i p a l c o n c l u s i o n of t h e s e s t u d i e s is t h a t t h e p r o c e s s of u n d e r s t a n d i n g t e x t i n v o l v e s c o n s t r u c t i n g a c o n s i s t e n t u n i f i e d meaning, wlli ch i s n o t s i m p l y t h e s e t of i n d i v i d u a l s e n t e n e e meanings. I t d i f f c r s i n t h h t a d d i t i o n a l knowledge and orga11ir.n t i 0 1 1 i . ; Int rod~iccd by tire uilderstander. [75p, 7Sb1 c o n t i n~~c s t h i s l i n e of t h o u g h t , t r e a t i n g the u n d c r q t a n d i n g of n disccl\\irsc as the. scmarlt i c k n o w l c d g s that is acquired i n l i s t e n i n g t o t h e disco\\rrsc. T h i r; k n o w l c .~l~~,~~ f o r~c q u t r c d by p r o c e s s e s t h a t u t i l i z e p r i o r knowlcdgc, cont c x t , c t c. nnd i n c l u d c s i n f a r m a t i o n which i s i n f e r r e d a s well AS t h a t e x p l i c i t l y p r c s~b~~t c t l . F r e d c u i k s e n also argues thaf: t h e p r o c e s s e s i n v o l v e d i n u n d c r s t n n d i n p , a d i s c o u r s e a r e d i r e q t l y r e l a t e d t o t h e p r o c e s s l i m i t n t i o r~s of t h c P L ' O C C S S O~. The two p r o c e s s e s that h e discusses are overgcnera1i;cat i o n and i n f e r t >~~c r . O v c . r g c n c i r n l i z e t i o n i s t h e d i s c a r d i n g o f d e t a i l i n f o r m a t i o n r e s u l t i n g i n n more p,th1'icr~1 c o n c e p t . 'tie suggests t h a t o v e r g e n e r a 1 , i z a t i o n r e d u c c s t h e amcwnt of i n f o r n l o t i o n t o be u n d e r s t o o d , thus r e d u c i n g t h e processing Zond on the U I I~C ~-s t n i \\ d e Inferred i n f o r m a t i o n i s i n f a n n a t i o n which i s assunlcd t o hc. t even i f it-i s not e x p l i c i t l y present. T h i s p r o c e s s i s clhimcd t o reduce t h e p r u c c s s i n g l o a d by e l i m i n a t i n g t h e n e c e s s i t y of c o m p l e t e l y a n d c rst lrndiny, t\\vtary sctlteilce. FredarAksen p e r f o r m e d n e u t r a l r e c a l l c x p c r i m c n t s nnd r o c t~l l c x l~r r i n~~n t s i n wh.lch t h e c o n t e x tt h e t a s k assigwcd t o t h e subjc.ct p r l o r t o p l i -: ; c~i t n t i o n of t h e storywas changed,. H i s c o n c l u s f o n s a r c that. ovc r g c n c r ; t l i z ,~r 1011 rind i n f e r e n c e s a r e i n c o r p o r a t e d" |
| }, |
| "FIGREF5": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t c o n s i s t s of s e t I n c l u s i o n , i d e n t i t y , and l o g i c a l i m p l i c a t i o n r e l a t i o n s h i p s t h a t hold among t h e c o n c e p t s of the t e x t , . He i n d i c a t e s t h e need f o r & d e t a i l e d model of t h e s c i w n t i c r e p r e s e n t a t i o n , and i n o t h e r work, F r e d e r i k s e n [75cJ begin's t o clef i n e such a r e p r e s e n t a t i o n a l scheme. H e p r o p c s e s t h e u s e of two networks: a s e m a n t i c network and a l o g i c a l network. The semantfc network c o m a i n s t h e rep r e s e n t a t i o n of i n d i v i d u a l p r o p o s i t i o n s w h i l e t h e l o g i c a l network i s composed o f r e l a t i o n s t h a t h o l d between p r o p o s i t i o n s t h a t e x i s t i n t h e semantic network. ~r e d e r i k s e n ' s system i s probably t h e most e l a b o r a t e y e t p r o p o s e d u s i n g t h e b a s i c i d e a s o f case grammar r e p r e s e n t e d by s e m a n t i c networks. He begins w i t h t h e fundamental i d e a s of o b j e c t and a c t i o n h i e r a r c h i e s , a l t h o u g h h i s are q u i t e large end c o n t a i n numerous d i s t i n c t i o n s . He p r o p o s e s a system t h a t d i s t i n g u i s h e s s i x t e e n v e r b , c a s e s , and a l a r g e numhc r of r e l a t i o n s s p e c i f y i n g s t a t e s , q u a n t i f i c a t i o n l o c a t i~n , manner, time, o r d e r , p r o x i m i t y , t e n s e and aspect. H i s l o g i c a l network c o n s i s t s of r e l a t i o n s drawn from p r o p o s i t i o n a l l o g i c p l u s c a u s a l i t y , and h e proposes several modal o p e r a t o r s . The complexity of t h i s p r o p o s a l g i v e s t h e impression of r e p r e s e n t a t i o n a l power, b u t i t seems t h a t F r e d e r i k s e n i s s t i l l p r i m a r i l y o r i e n t e d toward t h e r e p r e q e n t a t i o n of 1 i n d i y i d u a l p r o p o s i t i a n s . Although he shows t h e r e p r e s e n t a t i o n of s i m p l e time and c a u s a l l y ordered a c t i o n s sequences, h e h a s not y e t demonstrated t h e adequacy of h i s system f o r t e x t , i n gerlerdl. And, h e h a s u t i l i z e d a number of elements of q u a n t i f i e d modal l o g i c w i t h o u t d e m o n s t r a t i n g t h e i r u s e f u l n e s s for modelling human u n d e r s t a n d i n g of d'iscoursc. H i s d i s c u s s i o n s d o r a i s e a number o f i n t e r e s t i n g q u e s t i o n s about r e p r e s e n t a t i o n , and c e r t a i n l y d e s e r v e c o n s i d e r a t i o n i n t h e development of any s e t of s e m a n t i c t y p e s and r e l a t~o n s . 1 1 3 K i n t s c h -K i n t s c h [74] C reviewed by van D i jk [75bJ ) a t t e m p t s t o s e t f o r t h a f a i r l y c o m p l e t e t h e o r y of langudge u n d e r s t a n d i n g . K i n t s c h ' s b a s i c rep r e s e n t a t i o n a l u n i t i s t h e p r o p o s i t i o n , and h e i n c l u d e s d i s c u s & i o n of t h e u s u a l d i f f i c u l t problems o f d e f i n i t e n e s s ( i n c l u d i n g g e n e r i c v e r s u s s p e c i f i c d i s t i n c t i o n s ) o f noun p h r a s e s , q u a n t i f i c a t i o n , m o d a l i t y , i m p l i c e t i a n and p r e s u p p o s i t i o n , l o c a t i o n , time and t e n s e . He p r o p o s e s a t e x t b a s e which u n d e r l i e s d i s c o u r s e s , b u t h i s r e v i s e d and e l a b o r a t e d t h o u g h t on t h i s i s d i s c u s s e d i n S e c t i o n 1.2. ,Me, d e s c r i b e s a model of d i s c o u r s e p r o c e s s i n g i n which h e makes a d i s t i n c t i o n between e p i s o d i c and s e m a n t i c memory, a r g u e s t h a t t h e p r o c e s s i n g o p e r a t i o n s must b e w e l l d e f i n e d ( h e s u g g e s t s p a t t e r n matching and completion, a b s t r a c t i o n a n d g e n e r a t i o n ) and a r g u e s t h a t r e c a l l i s e s s e n t i a l l y a d i f f e r e n t p m c e s s from r e c o g n i t i o n i n t h a t r e c a l l r e q u i r e s i n p u t o r g a n i z a t i o n w h i l e r e c o g n i t i o n does n o t , K i n t s c h r e p o r t s a number of e x p e r i m e n t s u n d e r t a k e n t o t e s t a s p e c t s of h;i$ theory. One c o n c l u s i o n t h a t he r e a c h e s i s t h a t t h e semantic r e p r e s e n t a t i o n i s p a r t i a l l y independent of t h e a c t u a l i n p u t s e n t e n c e s . One k i n d o f experiment performed t o test t h i s was t h e performance of a cammsn task by s u b j e c t s who had read s u b s t a n t i a l l y t h e same m a t e r i a l b u t i n forms o t v a r y i n g comnlexity. In o t h e r e x p e r i m e n t s h e n o t e s t h a t s e n t e n c e s c o n t a i n i n g m u l t i p l e p r o p o s i t i o n s a r e much l e s s l i k e l y t o be r e c a l l e d (completnly) t h a n are s i n g l e p r o p o s i t i b n s e n t e n c e s . T h i s a l s o s u p p o r t s t h e i d e a of an u n d c t l y i n g r e p r e s e n t a t i c n . A r e l a t e d conclusion i s t h a t t h e agent of a p r o p o s i t i o n i s t h e most l i k e l y case t o be r e c a l l e d . However, Kintsch d i d f i n d t h a t i n p a s s i v e s e n t e n c e s , t h e" |
| }, |
| "FIGREF6": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "final conclusion conce ms t h e presence of i n f e r r e d i n \u00a3 ormation i n memory. K i n t s c h found t h a t immediate r e c a l l showed some d i f f e r e n c e between i m p l i c i t and e x p l i c i t p r o p o s i t i o n s , b u t t h a t a f t e r twenty minutes o r more the d i f f e r e n c e had disappeared and t h e two were i n d i s t i n g u i s h a b l e t o t h e s u b j e e t . An i n t e r e s t i n g side result was t h e d i s t i n c t d i f f e r e n c e i n r e a d i n g and r e c o g n i t i o n times f o r argumentative versus d e s c r i p t i v e discourse. The argumentative d i s c o u r s e required s i g n i f i c a n t l y more time i n both tasks.. K i n t s c h also i n c l u d e s r e a c t i o n time experiments t o study p r o p o s i t i o n r e t r i e v a l , deterruin-ation of t h e t r u t h o r f a l s i t y of g e n e r a l p r o p o s i t i o n s and p r o c e s s i n g of complex l e x i c a l i t e m s b u t t h e s e w i l l not be discussed." |
| }, |
| "FIGREF7": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "e s t i g a t i o n b on t h e r o l e of i n f e r e n c e i n u n d e r s t a n d i n g t e x t h a v e been c a r r i e d o u t by Thorndyke [76]. H e u s e s compound s e n t e n c e s a s s e r t i n g a c a u s a l connection which i s not f a m i l i a r o r obvious. For example: The hamburger chain owner was a\u00a3 r a i d t h a t h i s l o v e f o x f rench f r i e s would ruin h i s marriage." |
| }, |
| "FIGREF8": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "He d e c i d e d t o see a m a r r i a g e c o u n s e l o r i n o r d e r t o save h i s marriage. o r i t might encourage one p a r t i c u l a r e q l a n a t c r r y inference l i k e He d e c i d e d t o j o i n weight w a t c h e r s i n o r d e r t o s a v e h i s marri ~g e ." |
| }, |
| "FIGREF9": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "The S t r u c t u r e O f D i s c o u r s e C o n t e n t A t t e m p t s t o c h a r a c t e r i z e t h e c o n t e n t of t e x t n e c e s s a r i l y lead t o a s t r u c t u r i n g o f t h a t c o n t e n t . Simple one-dimensional r e p r e s e n t a t i o n s i n which p r o p o s i t i o n s a r e connected o n l y by c o -r e f e r e n c e o r by time o r c a u s a l o r d e r i n g Page 13" |
| }, |
| "FIGREF10": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t seems a p p r o p r i a t e t o b e g i n a d i s c u s s i o n of memory s t r u c t u r e w i t h B a r t l e t t [32] s i n c e h e i s o f t e n r e f e r e n c e d as t h e o r i g i n a t o r o f s e v e r a l cu rrently p o p u l a r hypotheses. I t s h o u l d be remembered t h a t Bart1 et t was concerned with memory i n g e n e r a l and t h e phenomena a s s o c i a t e d w i t h r e c a l l , and did, n o t r e s t r i c t h i m s e l f t o t h e s t u d y o f d i s c~u r s e . His p x i n c i p l e c o n c l u s i o n s were t h a t memories a r e n o t s t o r e d i n i s o l a t e d , s t a t i c u n i t s , and t h a t exact r e c a l l is very rare. \\In f a c t , he o f t e n found c a s e s of g r o s s d i s t o r t i o n s i n the r e c a l l o f h i s s u b j e c t s . H e s u g g e s t s t h a t i n s t e a d , memory i s composed of a number of a c t i v e , o r g a n i z e d masses of past r e a c t i o n s and experiences, which he d e s i g n a t e s schemata, and a s m a l l amount o f o u t s t a n d i n g d e t a i l -Remembering i s s e e n a s a c o n s t r u c t i v e process s t r o n g l y a f f e c t e d by memories o t h e r t h a n t h e one b e i n g r e t r i e v e d . H i s ideas of o r g a n i z a t i o n beyond t h e s e n t e n c e , of s t o r a g e o f something other t h a n a c t u a l i n p u t s e n t e n c e s , and of a c t i v e p r o c e s s e s t h a t modify t e x t p r i o r t o i t s r e p r o d u c t i o n i n r e c a l l a r e a l l currently e n joying w j d e a c c e p t a n c e . s concerned w i t h t h e r e c a l l of short, e x p o s i t o r y p a r a g r a p h s containin!:. m a t e r i a l not l i k e l y t o be f a m i l i a r t o his s u b j e c t s , He p r e s e n t s r e s u l t s obtained w i tll p a r a g r a p h s a b o u t hebulae. He p r o p o s e s t h e e x i s t e n c e of a semantic r e p r e s e n t a t i o n u n d e r l y i n g any p a r t i c u l a r d i s c o u r s e which c o n t a i n s t h e rncal~lng of t h e d i s c o u r g e , but, does not r e f l e c t t h e d e t a i l s of the s u r f a c e form of the t e y t . Thus, a s i n g l e s e m a n t i c r e p r e s e n t~t i o n might u n d e r l i e numerbus a c t u a l d i s c o u r s e s . H i s s e m a n t i c r e p r e s e n t a t i o n assumes a c o n c e p t u a l taxonomy showing t h e r e l a t i o n s h i p of each known c o n c e p t t o i t s s u p e r o r d i n a t e concept (i.e. the f a m i l i a r semantic h i e r a r c h y ) , b u t h e does not p r o v i d e d e t a i l s on t h i s . H e also p r o p o s e s an a d d i t k o n a l set of h i e r a r c h i e s showing t h e r e l a t i o n s h i p s of t h e c o n c e p t s i n a p a r t i c u l a r d i s c o u r s e ." |
| }, |
| "FIGREF11": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "h e f o l l o w i n g : means t h a t a n e b u l a i s e i t h e r seen o r i t i s n o t s e e n , where a l l of t h e c o n c e p t s a r e d e f i n e d on t h e c o n c e p t u a l graph. C r o t h e r s does n o t c a r e f u l l y d e f i n e h i s n o t a t i o n , s o i t i s ~o t c l e a r e x a c t l y what may o r may n o t o c c u r a t a node, P r i m a r i l y , c o n c e p t s (words) or c o n n e c t i v e s are used (e.g. NEBULA. I S AND, OR, WHY). S i n c e hc d e a l s o n l y w i t h e x p o s i t o r y , d e s c r i p t i v e material h e f i n d s no need t o r e p r e s e n t a c t i o n s o r t i m e , C r o t h e r s performed r e c a l l e x p e r i m e n t s on two d i f f e r e n t v e r s i o n s of t h e same m a t e r i a l , and arrives a t t h r e e c o n c l u s i b n s ." |
| }, |
| "FIGREF12": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "i r s t , he c o n c l u d e s t h a t t h e s u r f a c e p a r a g r a p h * s n o t a Page 15 -fieant f a c t o r i n r e c a l l , t h u s a r g u i n g f o r a surface-independent s e m a n t i c repnsentation f o r t h e paragraphs. H i slast two c o n c l u s i o n s are n e g a t i v e xejecting hypotheses he had p r e v i o u s l y suggested. The f i r s t of t h e s e i s t h a t superordinate nodes i n t h e meaning o f t h e paragraph w i l l be r e c a l l e d more frequently than s u b o r d i n a t e nodes. The second r e j e c t e d h y p o t h e s i s i s t h a t i a f o r a a t i o n n o t d i r e c t l y connected t o t h e most s u p e r o r d i n a t e pode o f t h e paragraph I& lese l i k e l y t o be r e c a l l e d t h a n i n f o r m a t i o n t h a t i s connected. Since h i s r e s u l t s do n o t s u p p o r t e i t h e r of t h e s e hypotheses, C r o t h e r s s u g g e s t s tbat other v a r i a b l e a , such as frequency o f concepts, are probably a l s o of irportance. h a s proposed a n extended c a s e grammr supplemented w i t h what he calls rhetorical p r e d i c a t e s , which are h i g h e r -o r d e r p r e d i c a t e s t h a t t a k e other p r o p o s i t i o n s as t h e i r arguments. H e s u b d i v i d e s r h e t o r i c a l p r e d i c a t e s i n t o t h r e e groups: p a r a t a c t i c , h y p o t a c t i c and n e u t r a l , P a r a t a e t i c p r e d i c a t e s always t a k e arguments of e q u a l weight (i. e. no argument i s s u b o r d i n a t e t o any other argument). H y p o t a c t i c p r e d i c a t e s have one dominant argument, t o which t h e o t h e r s are ,subordinate. N e u t r a l p r e d i c a t e s may b e used a s e i t h e r p a r a t a c t i c o r h y p o t a c t i c p r e d i c a t e s . The following i s a l i s t o f some of t h e r h e t o r i c a l p r e d i c a t e s proposed by G r i m e s , and t h e i r b a s i c meaning: P a r a t a c t i c p r e d i c r t e s restatement of the domiriafit p r o p o s i t i o n Gives more s p e c i f i c i n f o m a t i o n a b o u t a g e n e r a l dominant p r o p o s i t i o n Gives an a b s t r a c t e k p l a n a t i o n for t h e s p e c i f i c dominant p r o p o s i t i o n Gives an analogy t o s u p p o r t the dominant p r o p o s i t i o n L i s t of elements related i n some u n s p e c i f i e d way C a u s a l i t y , an a n t e c e d e n t and consequent Meyer [75] u s e s Grimes' r h e t o r i c a l p r e d i c a t e s and most of his casegrammar t o create analyses of paragraphs s h e uses i n r e c a l l experiments. F o r example: e s s e n t i a l s of an analysis of a paragraph which stated t h a t the need t o generate e l e c t r i c power while protecting the environment and of the s p e c i f i c s e t of resources t o the generaJ idea. eyer's analyses of expository paragraphs using t h i s scheme always follow the author's organizational st rvcture and always result i n a purely h i e r a r c h i c a l structure like those depicted above Meyer a l s o recognizes class of sentences that can occur i n expository paragraphs, but do not contribute any Page 18 content t o the passage. S h e c a l l s these> s i g n a l l i n g because t h e i r f u n c t i o n is. t o e x p l i c k t l y i n d i c a t e t h e s t r u c t u r e of t h e passage. She n o t e s f o u r t y p e s of s i g n a l l i n s 1. ~~e c i f i c a t i b n of t h e s t r u c t u r e (e. g. \"Two o p t i o n s exist.'') 2. P r e m a t u r e l y r e v e a l e d i n f o r m a t i o n abstracted f r o m t h e remainder of t h e p a s s a g e ( e m go \"The a l t e r n a t i v e s are s o l a r energy, n u c l e a r energy, geothermal energy aad l a s e r fusion energy. \") 3. Summary s t a t e m e n t s t t 4. P o i n t e r words ( e . g . u n f o r t u n a t e l y \" , \"an impdrtant p o i n t is\") Meyer conducted r e c a l l e x p e r i m e n t s v a r y i n g t h e p o s i t i o n of c e r t a i n material i n t h e content hierarchy of p a s s a g e s , and i n c l u d i n g o r o m i t t i n g certain s ~g n a l l i n g i n \u00a3 o m a t ion. u c t u r a l l y higher i n t h e c o n t e n t h i e r a r c h y i s remembered s u b s t a n t i a l l y b e t t e r than i d e n t i c a l i n f o r m a t i o n p l a c e d l o w e r i n the h i e r a r c h y in a n o t h e r passage. She a l s o n o t e s t h a t passages w i t h i d e n t i c a l s t r u c t u r e b u t t o t a l l y d i f f e r e n t c o n t e n t e x h i b i t very s i m i l a r patterns of r e c a l l a t the h i g h e r l e v e $ s . However, a t t h e l o w e r l e v e l s , t h e p a t t e r n of r e c a l l v a r i e d , i n d i c a t i n g c o n t e n t dependence. I n regard t o s i g n a l l i n g , s h e concludes t h a t i thas v e r y l i t t l e ' e f f e c t a t t h e t o p l e v e l , b u t seems b e n e f i c i a l a t t h e middle" |
| }, |
| "FIGREF13": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t o a t h e o r e t i c a l e x p l a n a t i o n of why t h e highest i d e a s a r e r e c a l l e d b e s t , Meyer s u g g e s t s t h r e e proposals and p o i n t s o u t the weaknesses of each. F i r s t , lower p r o p o s i t i o n s might be subsumed by h i e h e r ones w i t h the passage of time. Immediate r e c a l l experiments s h w t h a t t h e phenomenon occurs even w i t h o u t a time lapse, however. Secondly, a t i s p o s s i b l e t h a t a l l p r o p o s i t i o n s are stored b u t t h a t r e t r i e v a l i s e a s i e r a t the h i g h e r levels. She c r i t i c i z e s t h i s proposal because even cued r e c a l l experiments were unable t o r e t r i e v e t h e lower propositions. F i n a l l y , perhaps only h i g h e r i d e a s are ever stored. But h e r r e s u l t s show t h a t t h e lose of lower l e v e l p r o p o s i t i o n s begins immediately, but continues w i t h time a t a more r a p i d rate than f o r t h e h i g h e r l e v e l propositions. Meyer s u g g e s t s t h a t probably some combination of these processes i s occurring, a s well a s o t h e r processes s e n s i t i v e t o t h e s t r u c t u r e of t h e passage as a whole. 1 . 2 4 Kintsch And Van Dijk -Kintsch and van Dl jk (van D i jk [74,7Sa,76], van D i jk and Kintsch [forthcoming] and Kintsch and van Di jk [76]) p r e s e n t a model f o r t h e organization of discourse as whole, and a number of (sometimes informal) experiments attempting t o v a l i d a t e t h e model. Beginning a t the lowest level, a discourse r e p r e s e n t a t i o n c o n s i s t s of a s e t of propositions. A d i s t i n c t i o n i s made between two d i f f e r e n t t y p e s of r e p r e s e n t a t i o n , callCd t e x t b a s e s , a s t o whether a l l implied information i s made e x p l i c i t o r w h e t h e r i t i s l e f t i m p l i c i t . The notion of coherence i s introduced, which i s t h e p r o p e r t y t h a t d i s t i n g u i s h e s a discourse from a random s e t of sentences, R e f e r e n t i a l i d e n t i t y has been auggqsted a s a m a j o r t e s t of coherence, b u t i t i e n o t an a d e q u a t e d e f i n i t i o n ( s e e d i s c u e s i o n o f coherence and Bellert's p r o p o s a l s i n Section 3). K i n t s c h and van D i j k a r g u e t h a t coherence i s more accuratelycaptured b y the requirement t h a t each p r o p o s i t i o n of t h e t e x t base b e c o n n e c t e d w i t h one o r more p r e c e d i n g p r o p o s i t i o n s . P r o p o s i t i o n s are connected i f one is a c o n d i t i o n f o r the o t h e r , with t h e s t r e n g t h of t h e connectfon ranging from p o s a i b l e t o necessary. Thus, a n e x p l i c i t t e x t b a s e i s one c o n t a i n i n g a l l of t h e p r o p o s i t i o n s n e c e s s a r y f o r coherence, w h i l e an i m p l i c i t t e x t base i s one w i t h some of t h e s e p r o p o s i t i n n s d e l e t e d . The d e l e t e d p r o p o s i t i o n s are t h o s e t h a t can be assumed known o r which normally would be i n f~r r e d by t h e unde rs tande r. D i f f e r e n t t y p e s o f d i s c o u r s e would have d i f f e r e n t rules g o v e r n i n g t h e d e l e t i o n of p r o p o s i t i o n s . F o r example, casual conve r g a t i o n would a l l o w more d e l e t i o n thari c a r e f u l argumentation. The t e x t b a s e i s o r g a n i z e d h i e r a r c h i c a l l y u n d e r macro-s t r u c t u r e s , which are higher b r d e r p r o p o s i t i o n s . Macro-st ructures may be related t o t h e i r p r o p o s i t i o n a l arguments by a number of macro-rules. F o u r of t h e s e rules are suggested. =The f i r s t , i n f o r m a t i o n r e d u c t i o n , i s a rule of g e n e r a l i z a t i o n which would e x p l a i n the e x i s t e n c e of t h e macro-structure John i s ill, which had a s i t s arguments p r o p o s i t i o n s l i k e J o h n h a s a feve r. John has t h e f l u . The second rule, d e l e t i o n , would e x p l a i n t h e r e l a t i o n s h i p between P e t e r saw a b a l l . and i t s s u b o r d i n a t e p r o p o s i t i o n s P e t e r saw a b a l l . The b a l l wSas blue. The t h i r d r u l e , i n t e g r a t i o n , combines a c e n t r a l e v e n t w i t h i t s normal p r e -c o n d i t i o n s , r e s u l t s and component a c t i o n s . Thus John went t o P a r i s . might have as i t s arguments John took a cab T O t h e s t a t i o n . H e bought t i c k e t s . H e went t o P a r i s . The f o u r t h rule, c o n s t r u c t i o n , explains t h e r e l a t i o n s h i p between a complex f a c t and i t s component p a r t s . F o r example, t h e m a c r o -s t r u c t u r e P e t e r b u i l t a house, might have as i t s arguments t h e e v e n t sequence P e t e r l a i d a foundation. P e t e r b u i l t walls. P e t e r b u i l t a roof. I n g e n e r a l , two c o n d i t i o n s h o l d f u r a l l macro-structures. F i r s t , the Page 22 macro-structure must b e semant l c a l l y implied by i t s micro-structure (i.e. i t s p r o p o s i t i o n a l arguments). T h e e x a c t meaning of t h e term i m p l i c a t i o n is not c l e a r l y s t a t e d by Kintsch and van D l j k , They sometimes r e f e r t o i t as e n t a i l m e n t and t r e a t i t a s a formal l o g i c a l r e l a t i o n , b u t a t o t h e r times say t h a t i t i s not l o g i c a l i n the s t r i c t sense. I n any c a s e , i t i s c l e a r t h a t t h e r e are semantic rules o r s t r u c t u r e s t h a t allow c r e a t i o n of macro-~tructures such a s those above." |
| }, |
| "FIGREF14": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "may d e l e t e information contained i n i t s m i c r o -s t r u c t u r e which fs a c~ndition f o r a n o t h e r macro-structure An important consequence of t h i s i s t h a t t h e macro-structures 06. a ttext, taken b y themselves, form a coherent summary of t h a t t e x t . A f i n a l component of a tept r e p r e s e n t a t i o n i s the s p e c i f i c a t i o n of t h e \u00a3om o r s t r u c t u r e of t h e d i s c o u r s e , b u t d i s c u s s i o n of this p r o p o s a l w i l l be postponed u n t i l S e c t i o n 2. T h e i r i n v e s t i g a t i o n s a r e p r i m a r t l y w i t h n a r r a t i v e t e x t , and they c a l l t h e n a r r a t i v e s t r u c t u r e a schema. K i n t s c h and v a n 4 i jk shggcst t h a t the concept of -n a r r a t i v e st-ructure coabined with the i d e a of macro-structures leads them t o the f o l l a w i m a r i l y w h a t i s s t o r e d when a t e x t i s understood. R e c a l l u s e s macro-structures a s a s t a r t i n g p o i n t i n r e t r i e v a l , and summaries d i r e c t l y r e f l e c t t h e nlacro-st r u c t u r e s . Secondly, s i n c e macro-st ructures a r e e s s e n t i a l t o comprehension, they must be c o n s t r u c t e d a t t h e time of reading. Page 23 F i n a l l y , a narrative schema is necessary for the organization of t h e t e x t representation. K i n t s c l l and van D i j k have done a number of summary, recall and o t h e r experiments t o test t h e s e hypotheses. Most u s e a 1600 word t e x t from -The ------..--Decamcrot~. I n r e c a l l experiments, the propositions corresponding t o the macro-st ructures were found t o b e recalled most o f t e n , and were very u n l i k e l y tod i s a p p e a r i n delayed r e c a l l (nine days l a t e r ). Xn contrastGive major goals of characters" |
| }, |
| "FIGREF15": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "r o b a b l e c o n s e q u e n c e s of an a c t i o n o r a n e v e n t6. M e t a -n a r r a t i v e s t a t e m e n t s by t h e a u t h o rSummarizing e x p e r i m e n t s showed t h b t summaries of t h e s t o r y *wet.e very much l i k e d e l a y e d recall.When a summary was w r i t t e n a f t e r p r e s e n t a t i o n of eachsuccessive p a r t of t h e s t o r y , f o l l o w e d i m m e d i a t e l y by a complete summary, p r o p o s i t i o n s were i n c l u d e d Fn t h e p a r t i a l summaries t h a t were o m i t t e d i n t h e f i n a l summary. A f l n a l group of r e c a l l e x p e r i m e n t s compared r e c a l l of a 70 word p a r a g r a p h i n i s o l a t i o n , t o cued r e c a l l o f t h e same p a r a g r a p h imbedded i n ran 85O word t e x t . S u r p r i s i n g l y , t h e r e c a l l s were a l m o s t i d e n t i c a Some Conside r a t i on's Regarding Memory Experiments B e f o r e a t t e m p t i q g t o d e s c r i b e an i n f o r m a l madel based upon these h y p o t h e s e s and e x p e r i m e n t a l r e s u l t a , some g e n e r a l criticisms o f memory e x p e r i m e n t s exp r e s s e d by S p i r o [ 751 d e s e r v e c o n s i d e r a t i o n . spire's p r i n c i p l e argument i s t h a t r e c a l l c o n s i s t s of a c t i v e r e c o n s t r u c t i o n p r o c e s s e s , r a t h e r t h a n p a s s i v e p r o c e s s e s which merely reproduce t h a t which was s t o r e d st the time of comprehension. H i s g e n e r a l p o s i t i o n i s t h u s much l i k e t h a t of B a r t l e t t , and h e Observes t h a t almo3t a l l e x p e r i m e n t e t s s i n c e Bartlett have f a i l e d t o r e p l i c a t e h i s f i n d i n g s of s i g n i f i c a n t e r r o r s i n r e c a l l . As a r e s u l t , p s y c h o l o g i s t s have tended t o concent rate on t h e process of u n d e r s t a n d i n g , o r c o n s t r u c t i o n , and h a v e t r e a t e d r e c a l l a s somethihg o f a s i m p l e r e t r i e v a l process. S p i r o a r g u e s t h a t f o r r e c o n s t r u c t i v e e r r o r s t o o c c u r , i t i s n e c e s s a r y t h a t t h e i n p u t b e u n d e r s t o o d i n terms o f some p r e e x i s t i n g schema. Any t e x t t h a t r e s u l t s i n t h e c r e a t i o n of a new achema would n o t b e s u b j e c t t o t h e same k i n d of e f f e c t s of p r e v i o u s knowledge. The i n t e r f e r e n c e caused by t h e p r e -e x i s t e n c e o f schemata i s even more pronounced i f o t h e r u s e s of t h e schema are made between t h e time of comprehension and t h e time of r e c a l l . S p i r o a l s o p o i n t s o u t t h a t any e x p e r i m e n t a l s u b j e c t would be u n l i k e l y t o i n t e g r a t e material read o r h e a r d i n an experiment i n t o h i s g e n e r a l knowledge s i n c e i t s t r u t h and s o u r c e a r e unknom. T h i s coupled w i t h t h e d e s i r e t o perform w e l l , could e a s i l y r e s u l t i d a c o n s i d e r a b l y modified p r o c e s s of u n d e r s t a n d i n g t e x t i n e x p e r i m e n t a l s e t t i n g s . . S p i r o performed a number of r e c a l l e x p e r i m e n t s which s u p p o r t h i s h y p o t h e s i s . He u s e d text a b o u t i n t e r p e r s o n a l r e l a t i o n s h i p s and i n s t r u c t e d t31e s u b j e c t s t h a t t h e e x p e r i m e n t was concerned w i t h t h e i r r e a c t i o n s t o t h e i n c i d e n t s d e s c r i b e d , t h u s maximally Page 26 i n v o l v i n g t h e i r p r e -e x i s t i n g s t r u e r u r e a i n t h e understanding process. S p i r o found s u b s t a n t i a l e r r o r s i n t h e subjects r e c a l , l~, which h e e x p l a i n e as the e f f e c t of i n t e r a c t i o n w i ti1 prc-cxisting s t r u c t u r e s f o r i n t e r p e r s o n a l r e l a t i o n s h i p s . (Kintsch and van D i j k [75] found similar r e s u l t s u s i n g s t o r i e s from t h e Bible.) S p i r o c o n c l r~d e s t h a t most t e x t r e c a l l experiments have not shown t h e s e e f f e c t s because they d i d n o t meet t h e r e q u i r e d c o n d i t i o n s , b u t" |
| }, |
| "FIGREF16": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "was kept i s o l a t e d by t h e s u b j e c t . These c r i t i c i s m s not only are r e l e v a n t t o the q u e s t i o n of c o n s t r u c t i o n v e r s u s r e c o n s t r u c t i o n , b u t a l s o t o i n f e r r i n g d i s c o u r s e organf z a t i o n f rofn r e c a l l experiments, and i n d i c a t e that" |
| }, |
| "FIGREF17": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "g r e a t deal of agreement i s seen in t h e p r e c e d i n g s t u d i e s . The idea knowledge i s b u i l t , o r t h e a u t h o r ' s s t r u c t i l r e used. K i n t s c h and . van D i j k e s s e n t i a l l y u s e b o t h a p p r o a c h e s s i n c e the t i m e o r d e r i n g i n n a r r a t j v e i s a u t h o r d e f i n e d ( a l t h o u g h t h e c o n c e r u c t u r e s which h i e r a r c h i c a l l y o r g a n i z e t h e t e x t are p r e -e x i s t e n t -Meyer, K i n t s c h and van D i j k a l l a g r e e t h a t p r o p o s i t i o n s higher i n t h e s t r u c t u r e a r e more l i k e l y t o be r e c a l l e d , b u t C r o r h e r s e v i d e n c e d i d n o t s u p p o r t t h i s c o n c l u s i o n . Clearly, t h i s h y p o t h e s i s , i f true, could be confirmed o n l y i f i t i s t e s t e d a g a i n s t t h e a p p r o p r i a t e c o n t e n t s t r u c t u r e . S p i r o ' s comments r e g a r d i n g t h e u s e ( p r l a a h of use) of p r e -e x i s t i n g s t r u c t u r e s are r e l e v a n t , I n t h e case of u n f a m i l i a r e x p o s i t o r y material, s u c h as that used by C r o t h e r s and Meyer, i t seems r e a s o n a b l e to suppose t h a t t h e a u t h o r s o r g a n i z a t i o n would b e used t o o r g a n i z e t h e s e m a n t i c r e p r e s e n t a t i o n d u e t o the l a c k ( o r n,on-use) o f any a p p r o p r i a t e p r e -e x i s t i n g c o n t e n t s t r u c t u r e s . T h u s Meyer's r e s u l t s suppart t h e importance of p r o p o s i t i o n a l l e v e l , and C r o t h e r a r e s u l t s do n o t -Ylowevev, a f a m i l i a r s t r u c t u r e , l i k e an a c t i o n sequence, i s u n d e r s t p o d in t h e . same way r e g a r d l e s s of p r e s e n t a t i o n s e t t i n g . Thus, Kintsch and van D i jk found t h a t t h e r o l e o f a p r o p o s i t i o n i n t h e p r e -e x i s t i n g s t r u c t u r e ( t h e m a c r o -s t r u c t u r e s ) was primary i n d e t e r m i n i n g i t s i m p o r t a n c e , It s h o u l d be n o t e d t h a t even Meyer found v a r i a t i o n s i n t h e r e c a l l of m i d c l e v e l p r o p o s i t i o n s t h a t s h e c o u l d n o t e x p l a i n by r e f e r e n c e t o t h e a u t h o r ' s o r g a n i z a t i o n . I t i s p l a u s i b l e t o assume t h a t t h e s e v a r i a t i o n s were c a u s e d by i n d i v i d u a l d i f f e r e n c e s i n p r i o r knowledge, r e s u l t i n g i n d i f f e r i n g c o n t e n t s t r u c t u r e s ." |
| }, |
| "FIGREF18": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "In l i g h t of t h t s , i t seems p l a u s i b l e t o p r o p o s e a g e n e r a l model t h a t d i s t i n g u i s h e s t h e f o l l o w i n g component p r o c e s s e s b f t e x t u n d e r s t a a d i n g ( r e a l i z i n g t h a t l i k e almost a l l d i s t i n c t i o n s , such as s y n t a x versus s e m a n t i c q , t h e y are r e a l l y ~n a d e q u a t e , b u t u s e f u l , o v e r s i m p l i f i c a t i o n s ) . F i r s t t h e r e i s t h e p r o c e s s of comprehensioh of t h e s u r f a c e s t r u c t u r e which b u i l d s a r e p r e s e n t a t i o n t h a t i s independent of t h e s u r f a c e form o f t h e t e x t , T h i s r e p r e s e n t a t $ o n c o n t a i n s as much i n f e r r e d i n f o r m a t i o n a s i s n e c e s s a r y t o meet some minimal l e v e l o f u n d e r s t a n d a b i l i t y , o r coherence. T h i s r e p r e s e n t a t i o n i s o r g a n i z e d h i e r a r c h i c a l l y by h i g h e r l e v e l c o n t e n t st r u c t u r e s g h i c h summarize o r g e n e r a l i z e o v e r a l a r g e r amount o f d e t a i l e d i n \u00a3 o n n a t i o n . These h i g h e r l e v e l s t r u c t u r e s would n o r m a l l y b e p r e -e x i s t i n g ( i . e. known t o t h e u n d e r s t a n d e r ) a l t h o u g h a t e x t c o u l d r e s u l t i n t h e c r e a t i o n of new s t r u c t u r e s . D i f f e r e n t t y p e s of t e x t s c o u l d c e r t a i n l y d i f f e r i n t h e complexity of each o f t h e s e tasks, e x p l a i n i n g t h e v a r i a t i o n , i n d i f f i c u l t y of u n d e r s t a n d i n g . The p r o c e s s ( e s ) e x p l a i n i n g f o r g e t t f n g , and t h e i n t e g r a t i o n of new i n f o r m a t i o n i n t o p r i o r knowledge, o p e r a t e s on t h i s t e x t representat-ion. T h i s p r o c e s s i s p o o r l y u n d e r s t o o d , and t h e r e a r e no w e l l d e f i n e d h y p o t h e s e s a b o u t i t s o p e r a t i o n . I t a p p e a r s t o o p e r a t e p r i m a r i l y on t h e c o n t e n t s t r u c t u r e of a F i n a l l y , r e c a l l i s a retrieval p r o c e s s which o p e r a t e s on t h e c u r r e n t form of t h e rep resentation. If t h a t form i s r e l a t i v e l y unchanged s i n c e comprehension, r e c a l l may be almost d i s t o r t i o n -f ree. But i f s i g n i f i c a n t changes have occurred, r e c a l l w i l l be d i s t o r t e d . A s suggested by Kintsch and van D i j k , summary i s seen as recall t o a c o n t r o l l e d depth. R e t r i e v a l accesses" |
| }, |
| "FIGREF19": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "d e t a i l e d information. A s r e t r i e v e d information i s generated a s output t e x t , a test of coherence would be made. I f something h a s been f o r g o t t e n o r changed s o t h a t a necessqry explanation or l i n k i s missing, t h e r e t r i e v a l p r o c e s s would supply a probable p i e c e of information from t h e g e n e r a l p r e -e x i s t i n g s t r u c t u r e s . This, along w i t h modifications from i n t e g r a t i o n i n t o p r i o r knowledge, accounts f o r t h e r e c o n s t r u c t i v e aspect of recall-T h i s g e n e r a l model does not account f o r the r e c a l l o f very s p e c i f i c s u r f a c e d e t a i l s (e.g. words and c o n s t r u c t i o n s used i n t h e i n p u t t e x t ) b u t as suggested by Kintsch's work, i t seems b e s t t o treat t h i s as a p r o c e s s s e p a r a t e" |
| }, |
| "FIGREF20": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "i t s p e c i f i e s complete comprehension of a l l i n p u t p r o p o s i t i o n s , i n c l u d i n g assignment t o an a p p r o p r i a t e r o l e i n a h i g h e r l e v e l s t r u c t u r e . Frederiksen s u g g e s t s t h a t o v e r g e n e r a l i z a t i o n i s a l o s s of d e t a i l which reduces t h e processing load. The i d e a t h a t p a r t i a l processing i s sometimes done on i n p u t information i s a l s o suggested by Bobrow and Norman [75]. They d e s c r i b e such processes as resource-limited. This kind of processing seems t o be i n d i c a t e d by f a c t s such as l o s s of d e t a i l i n immediate r e c a l l and t h e s u b j e c t i v e impression of p a r t i a l comprehension i n a number of s i t u a t i o n s (such a s reading verg r a p i d l y , reading while p a r t i a l l y d i s t r a c t e d o r reading very complex material). Wowever, w h i l e t h e idea of d i s c a r d i n g comp rellended i n \u00a3 ormation* could be described c o m p u t a t i o n a l l y , s e l e c t i v e comprehension i a a p r o c e s s t h a t r e q u i t e s f u r t h e r i n v e s t i g a t i o n . One of t h e p r i n c i p l e c o n t r i b u t i o n s of computational i n v e s t i g a t i o n s w i t h t h i s sort of model w i l l be s p e c i f i c a t i o n of t h e p r o c e s s e s of c o m p r e h e n~i o n , which can only be done by s p e c i f y i n g t h e n a t u r e of p r e -e x i s t i n g memory structures. Meyer's work i s an e x a m p l e of t h e u s e of undefined b a s i c s t m c t u r e s . She c r i t i c i z e s C r o t h e r a f o r r e q u i r i n g two g r a p h s t o represent a t e x t one of which i a h i s Eundame~ital h i e r a r c h y of kndwn concepts. She, i n s t e a d , simply doesn't d e f i n e any of h e r p r e d i c a t e s first o r h i g h e r o r d e r , a p p a r e n t l y t a k i n g them as p r i m i t i v e . Assuming that one a t t e m p t e d t o d e f i n e even t h e argument r o l e s f~r t h e r h e t o r i c a l predicates, enormous d i f f i c u l t y would be encountered. Advocates of case systems have always had d i f f i c u l t y d e f i n i n g t h e e x a c t r o l e of c a s e s , o r d e l i m i t i n g t h e c l a s s of e n t i t i e s t h a t c o u l d f i l l a case role. C o n s i d e r how much more difficult i t would be t o d e f i n e what p a i r s of t h i n g s could e x i s t i n t h e RESPONSE o F COVARIANCE r e l a t i o n s , o r e x a c t l y what the c h a r a c t e x i s t i c s of t h e s e arguments are. K i n t s c h and van Di jk a t t e m p t t o s e t o u t much more w e l l d e f i n e d s t r u c t u r e s , b u t e v e n t h e y have l i t t l e t o say about t h e s e t of s t r u c t u r e s t h a t m u s t e x i s t , what i n f o r m a t i o n they must c o n t a i n and how t h e camprehension a l g o r i t h m s u s e t h e t e x t and t h e s t r u c t u r e s t o b u i l d h i g h e r -l e v e l s t r u c t u r e s . T h e s e c t i o n on c o m p u t a t i o n a l models w i l l c o n s i d e r a l g o r i t h m i c attempts a t t e x t understanding, and should c l e a r l y i n d i c a t e t h e complexity of these issues. Page 31 2.0 THE FORM OF CONNECTED DISCOURSE The general model of t e x t understanding suggested i n t h e l a s t s e c t i o n d e s c r i b e s t h e s t r u c t u r e d c o n t e n t of t e x t r e p r e s e n t a t i o n . This s e c a a n c o n t a i n s work i n v e s t i g a t i n g a similar, y e t d i s t i n c t , t y p e of s t r u c t u r ethe a b s t r a c t , underlying form of a t e x t . (Not everyone would accept t h i s d i s t i a c t i o n , b u t i t seems analogous t o t h e d i s t i n c t i o n between t h e s y n t a c t i c structure of sentences and t h e semantic s t r u c t u r e o f t h e i r underlying meaning.) Much o f t h i s work has been done by c u l t u r a l a n t h x o p o l o p i s t s or by t h o s e i n t e r e s t e d i n l i t e r a r y a n a l y s i s . L i t t l e of t h e t r a d i t i o n a l work ha8 been done w i t h c o g n i t i v e o r computational models i n mind, and hence t e n d s t o r e l y on i n t u i t i v e understanding o f t h e terms. T h i s less computational work" |
| }, |
| "FIGREF21": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "2.1 S t r u c t u r a l Analysis O f TextMost modern work i n t h e s t r u c t u r a l a n a l y s i s of t e x t s h a s r o o t s i n t h e work of t h e Russian s t r u c t u r a l i s t , Propp [68]. Propp was concerned w i t h t h e form of Russian f o l k t a l e s . and developed a method f o r d e s c r i b i n g t h es i m i l a r i t i e s he found i n a corpus of 100 f o l k t a l e s . The major s t r u c t u r a l u n i t r e p e a t e d l y found i n t h e t a l e s . F o r example, A b s e n t a t i o n i s t h e f u n c t i o n t h a t d e s c r i b e s t h e s i t u a t i o n i n which one ~f t h e members of t h e f a m i l y a b s e n t s h i m s e l f from home, and T r i c k e r y is t h e f u n c t i o n which d e s c r i b e s a v i l l a i n ' s a t t e m pt t o d e c e i v e h i s v i c t i m i n o r d e r t o t a k e p o s s e s s i o n o f him o r h i s belongings. Propp fourid t h i r t y -o n e such f u n c t i o n s t o be adequate t o d e s c r i b e t h e e v e n t s i n t h e t a l e s t h a t h e s t u d i e d . I m p o r t a n t l y , t h e o r d e r i n g of t h e f u n c t i o n s was found t o be fixed. Some functicrns a r e o p t i o n a l , b u t i f they am p r e s e n t t h e i r p o s i t i o n i s f i x e d with r e s p e c t t o t h e o t h e r ftlnctions. All of the f u n c t i o n s are d e f i n e d i n terns of a c t o r s t h a t p a r t i c i p a t e i n t h e s i t u a t i o n s they d e s c r i b e . P.ropp d i d some a n a l y s i s of t h e r e s t r i c t i o n s on t h e assignment of v a r i o u s c h a r a c t e r s t o t h e s e f u n c t i o n a l r o l e s . He n o t e s t h a t t h e aame c h a r a c t e r i s l i k e l y t o play a f i x e d set of r o l e s , which h e d e s i g n a t e s a s p h e r e of a c t i o n . These spheres of a c t i o n i n c l u d e s u c h f a m i l i a r s t o r y r o l e s as hero, v i l l a i n and f a l s e hero. F i n a l l y , Propp n o t e s t h a t a s i n g l e f o l k t a l e may b e composed of a sequence of b a s i c t a l e s . He d e s i g n a t e s t h e b a s i c t a l e a I1 move\". Thus, a f o l k t a l e ,may be a one-move o r a multi-move t a l e , with each move conforming t o t h e d e f i n i t i o n of a t a l e . These r u l e s c a p t u r e t h e s i m i l a r i t y of t h e f o l k t a l e s which could d i f f e r widely i n t h e details o f c h a r a c t e r s , s e t t i n g , s p e c i f i c a c t i o n s , e t c . I t Is c l e a r t h a t Propp's r u l e s could be expressed formally, and t h a t i s what Klein e t a1 f ? 4 j d i d by w r i t i n g a computer program t o g e n e r a t e t h e f u n c t i o n sequences of a number of f o l k t a l e s , using Prupp's d e f i n i t i o n s . The i d e a s of Propp have been r e f i n e d and g e n e r a l i z e d by Barthes, Bremond, Greimas and Todorov ( t h e i r work is not generally a v a i l a b l e i n English b u t i s reviewed i n van D i jk [ 7 2 ] and Weinolci (72) ). T h e i r work has been d i r e c t e d toward the r e d e f i n i t i o n of Propp's f u n c t i o n s as p r o p o s i t i o n s , and of actor8 as case r e l a t i o n s of t h e s e p r o p o s i t i o n s . They have a l s o attempted t o generalize the f u n c t i o n s by making them less s p e c i f i c , and t o apply them t o o t h e r forms d f t e x t s . However, Hendricks [ 7 2 ] demonst rates the f l e x i b i l i t y of Propp's o r i g i n a l f u n c t i o n s by u s i n g them t o a n a l y z e p a r t of t h e Bradbury novel Something Wicked T h i s Way Comes. Chafe [72, 751 d i s c u s s e s the p r o c e s s of \" v e r b a l i z a t i o u \" , by which he means t h e t r a n s l a t i o n of non-verbal knowledge i n t o v e r b a l o u t p u t . One o f t h e proposed component processes of v e r b a l i z a t i o n is b r e a k i n g the knowledge i n t o smaller chunks according t o p a t t e r n s , which a r e c a l l e d \" s~h e m a t a '~. These schemata a r e grammar l i k e d e s c r i p t i o n s of p o s s i b l e v e r b a l sequences. Thus, he n o t e s t h a t a certain class of s t o r i e s w i l l always be of t h e form: PLOT + MORAL Chafe a l s o mentions such schemata a s TRICK and VISIT as o c c u r r i n g i n t h e stories he h a s analyzed. He e x p l i c i t l y s t a t e s t h a t , a t t h i s time, schematic analysis of a story is done by \"imagination abd i n t u i t i o n \" . He discusses aspects of t h e schematic analysis of s e v e r a l simple t a l e s , b u t these are mostly i l l u s t r a t i v e rather than complete, i n any sense. Colby [73] a n a l y z e s Eskimo f o l k t a l e s , i d e n t i f y i n g \"eidons\", which are similar t o Propp's f u n c t i o n s , and producing a grammar capable of g e n e r a t i n g some of t h e s e t a l e s . Colby observes t h a t most n a t i v e s p e a k e r s are probably n o t very familiar with story grammars of t h i s t y p e (or t h e knowledge they r e p r e s e n t ) , s i n c e only a f e w i n d i v i d u a l s a r e a b l e t o g e n e r a t e such t a l e s Nonetheless, i t seems reasonable t o assume t h a t t h e same t y p e of s t r u c t u r e can Page 34 be found in more c o n v e n t i o n a l s t o r i e s Although most work i n t h i s a r e a has involved a n a l y s i s of some form of t r a d i t i o n a l n a r r a t i v e , Labov and Waletzky [67] p r e s e n t a n a l y s e s o f oral v p r s i o n s of p e r s o n a l experiences. Even i n t h i s i n f o r m a l type of n a r r a t i v e , t h e y f i n d f i v e r e g u l a r components: an o r i e n t a t i o n , a complication, an e v a l u a t i o n , t h e r e s o l u t i o n and a coda o r moral, The l a s t two components are o p t i o n a l . A t o t a l l y d i f f e r e n t t y p e of material i s examined by Becker [ 6 5 ] who d i s c u s s e s p a t t e r n s i n e x p o s i t o r y paragraphs." |
| }, |
| "FIGREF22": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t o grasp t h e meaning of t h e s t r u c t u r e s from i n f o m a l d e s c r i p t i o n s and t o f i n dt h e s e s t r u c t u r e s i n t h e t e x t being analyzed. For t h i s reason, t h e p a r t i c u l a r s t r u c t u r e s t h a t have been suggested are of l e s s i n t e r e s t than i s t h e general r e s u l t t h a t q u i t e r e g u l a r h i g h l e v e l p a t t e r n s are found i n t e x t s of many t y p e s , which may be c h a r a c t e r i z e d independently of t h e s p e c i f i c t e x t u a l c o n t e n t .2.2 Kintsch And Van DijkI n a d d i t i o n t o t h e c o n t e n t macro-structures ( d i s c u s s e d i n s e c t i o n 1.2.4) K i n t s c h and van D i jk [ 7 5 ] and van D i jk [75a, 761 d i s c u s s even h i g h e r l e v e l \"super-structures\", used t o o r g a n i z e t h e content of a t e x t according t o t h e t y p e of t h e t e x t , which a r e d e r i v e d from t h e work i n s t r u c t u r a l a n a l y s i s of t e x t s . They l i m i t t h e i r work t o narratives, which they d e f i n e a s a s p e c i f i c t y p e of a c t i o n d i s c o u r s e . Narrative c a t e g o r i e s such as Episode S e t t i n g , Complication, etc. are s u p e r s t r u c t u r e s under which are organized either more n a r r a t i v e c a t e g o r i e s o r t h e c o n t e n t macro-structures of t h e text. Van Dijk [75aJ p r e s e n t s a f a i r l y complete example. I t i n c l u d e s a 1600 word t e x t from T h e Decameron, a p r o p o s i t i o n a l a n a l y s i s o f t h e s t o r y , t h e content macro-st r u c t u r e s of t h e s t o r y and t h e n a r r a t i v e c a t e g o r i e s under which t h e macro-st N C t u r e s are organized. The following a b b r e v i a t e d fragment should give an i n d i c a t i o n of t h e kind of a n a l y s i s being proposed. L l l l l l l l l l l ----l o ---*~~ Landolfo l o s e s h i s f o r t u n e i ( d e t a i l p r o p o s i t i o n s about t h i s l o s s ) Some expetiments were done t o t e s t t h e role of t h e understander's n o t i o n o f n a r r a t i v e s t r u c t u r e . An Apache myth and three Decameron s t o r i e s , which were of comparable d i f f i c u l t y a t t h e s e n t e n c e l e v e l , were presented. They d i f f e r e d i n t h a t t h e Apache myth had a n o r g a n i z a t i o n n o t f a m i l i a r t o t h e s u b j e c t s . There was much g r e a t e r v a r i e t y i n the p r o p o s i t i o n s d i f f e r e n t p e o p l e Page 36 used in tllc recall' of the myth than i n the o t h e r three stories. Another' experiment comparcd recalls of s u b j e c t s who had read a normal story to those of s u b j e c t s who had rend scrambled v e r s i o n s of t h e same s t o r y . The recalls were i n d i s t i n g u i s h a b l e by judges. Both of t h e~e groups of expeiiments i n d i c a t e the a c t i v e role o f one's e x p e c t a t i o n s about t h e f o m of a discourse in arganizir~g the r e p r e s e n t a t i o n of t h a t discourse, When this is impossible, as in the Apache myth, recall is less organized and more random. 2 . 3 Rumelhart Rumelhart ( 7 4 , 751 p r o p o s e s a story grammar t h a t i s capable of producing structures s i m i l a r t o t h o s e described by Kintsch and van Dijk (section 2.2) H i s grammar contains rules like the following: 1. STORY -> SETTING + EPISODE 2. EPISODE -> EVENT + REACTION 3. REACTlON -> INTERNAL-RESPONSE + OVERT-RESPONSE U s i n g these rules, he d e s c r i b e s the s y n t a c t i c s t r u c t u r e of simple s t o r i e s . The following fragment i l l u s t r s t e s the t y p e s of structures derived (ignoring t h e p a r e n t h c ? s i z c d p r e d i c a t e s f o r now): o p o s i t i o n s d e s c r i b i n g ~a r g i e ' s I b a l l o o n b e i n g popped ) Margie i s s a d M a r g i e cries A s s o c i a t e d w i t h most s y n t a c r t c s t o r y rules are s e m a n t i c rules which are used t o g e n e r a t e a c o r r e s p o n d i n g s e m a n t i c s t r u c t u r e f o r t h e s t o r y , The semantic rules for t h e s y n t a c t i c rules g i v e n above are: 1, ALLOW (SETTING, EPISODE) 2, INITIATE (EVENT, REACTION) where t h e s e m a n t i c p r e d i c a t e s are i n t e n d e d t o mean what i s suggested by their names (e.g. t h e SETTING ALLOWS t h e EPISODE t o o c c u r ) . The f i g u r e g i v e n above c o n t a i n s t h e a p p r o p r i a t e s e m a n t i c p r e d i c a t e , i n p a r e n t h e s e s , a t e a c h node i n the s y n t a c t i c s t r u c t u r e . (Rumelhart uses a c o m p l e t e l y s e p a r a t e s e m a n t i c s t r u c t u r e , ) Rumelhart d i g c u s s e s some elementary summarization r u l e s that operate on t h e semantic s t r u c t u r e o f a s t o r y . Some i l l u s t r a t i v e r u l e s are: 1. MOTIVATE ( t h o u g h t , r e s p o n s e ) => response 2. INITIATE (X, Y) => Y 3 . INITIATE (X, Y) => Y ( b e c a u s e 1 when 1 after) X These rules could produce t h e e i t h e r o f t h e f o l l o w i n g summaries ( f o r t h e EPISODE i n t h e s t o r y fragment g i v e n above): 1. Margie c r i e d . 2, Margie c r i e d b e c a u s e h e r b a l l o o n had popped, Rume.lhart8s i d e a s are i n t e n d e d t o be p r i m a r i l y i l l u s t r a t i v e (\"a t e n t a t i v e b e g i n n i n g of a t h e o r y u ) , and d e a l w i t h very s i m p l i f i e d s t o r i e s . They do argue c l e a r l y f o r t h e n o t i o n of b o t h a s y n t a c t i c and a s e m a n t i c structure f o r t h e r e p r e s e n t a t i o n of a s t o r y ' s meaning. The s y n t a c t i c s t r u c t u r e a l l o w s p r o d u c t i o n of t h e s e m a n t i c s t u c t u r e , which i s j u s t i f i e d by i t s u s e f u l n e s s i n p roducing summaries." |
| }, |
| "FIGREF23": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "[ 7 7 ] d i s c u s s an extended form o f ~u m e l h a r t ' s s t o r y grammar, which they propose a s a b a s i s f o r t e x t r e c a l l s t u d i e s . The extended grammar a l l o w s m u l t i p l e e p i s o d e s and does n o t u t i l i z e t h e a d d i t i o n a l s t r u c t u r e c o n t a i n i n g t h e semantic i n t e r p r e t a t i o n which Rumelhart suggests. Mandler and Johnson f e e l t h a t t h i s second s t r u c t u r e i s unwieldy and f r e q u e n t l y redundaat., Conjunction, time-ordering and c a u s a l i t y are i n c l u d e d a s categories i n t h e grammar. They observe t h a t r e s t r i c t i o n s imposed on a story by this t y p e of grammar i n c l u d e a l l o w i n g only a s i n g l e p r o t a g o n i s t per e p i s o d e , and p r o h i b i t i n g t h e r e a l i z a t i o n of h i g h e r l e v e l nodes i n t h e actual text. They demonstrate t h e u t i l i t y of t h e i r grammar on s e v e r a l s t o r i e s , i n c l u d i n g \"The War of t h e Ghosts\". They a l s o n o t e t h a t a t r a n s f o r m a t i o n a l component remains t o be developed which would p r o v i d e f o r d e l e t i o n s and r e o r d e r i n g s of t h e i d e a l story s t r u c t u r e -They d i s c u s s a set of p r e d i c t i o n s f o r r e c a l l s t u d i e s , which are suggested by t h i s d e f i n i t i o n of s t o r y . s t r u c t u r e . These i n c l u d e : 1. Accuracy w i l l be b e t t e r as the s t o r y i s complete and o r d e r e d as d e f i n e d by t h e i d e a l s t r u c t u r e -2, E l a b o r a t i o n s w i l l be p o o r l y r e c a l l e d . 3. O p t i o n a l nodes' r e a l i z a t i o n s w i l l be less w e l l r e c a l l e d m 4. C a u s a l i t y w i l l be better r e c a l l e d t h a n simple time-ordering. 5. I n v e r s i o n s should be fewer a s t h e story i s c l o s e r t o t h e i d e a l s t r u c t u r e -" |
| }, |
| "FIGREF24": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "a d d i t i o n s and d i s t o r t ions. Page 40 U n f o r t u n a t e l y , t h e only experimental work reported i s a c o w r i s o n of recalls of s t u d e n t s from f i r s t grade, f o u r t h g r a d e and u n i v e r s i t y l e v e l s , on a set of s t o r i e s which are very n e a r t h e i d e a l s t r u c t u r e . Hopefully, t h i s i s only t h e beginning s t a g e of v a l i d a t i o n of t h e proposed model* An i n t e r e s t i n g r e s u l t from t h i s experinlent i s t h e p r o b a b i l i t y of r e c a l l o f a p r o p o s i t i o n as a f u n c t i o n of i t s r o l e F o r u n i v e r s i t y s t u d e n t s , S e t t i n g s ( i n t r o d u c t i o n of c h a r a c t e r s ) and Beginnings ( i n i t i a t i o n of event sequences) (which t o g e t h e r formed Goal P a t h s ) were n e x t best r e c a l l e d * Reactions (mental e v e n t s ) and Endings (which were u s u a l l y either very p r e d i c t a b l e o r o m i t t e d i n t h e stories used) were least w e l l r e c a l l e d . 'These r e s u l t s a r e i n harmony w i t h t h e r e s u l t s reported by K i n t s c h and van D i j k (which are d i s c u s s e d i n S e c t i o n 1.2.4). 2.5 Thorndyke Thorndyke [771 attempts t o v a l i d a t e the b a s i c n o t i o n of u n d e r l y i h g text form as a n e c e s s a r y p a r t of a t e x t understanding. Re applies a s t o r y s t r u c t u r e grammar, which is n e a r l y i d e n t i c a l t o t h a t proposed by Rumelhart, t o two s t o r i e s , \"Ci r c l e I s l a n d \" (analyzed by F r e d e r i k s e n [75b] ) and \"The Old Farmer\" ( g i v e n by Rumelhart [74J), each pf which r e s u l t s i n a s t r a i g h t f o r w a r d h i e r a r c h i c a l s t r u c t u r e . The i n i t i a l rule of t h e grammar i s STORY -> SETTING + THEME t PLOT + RESOLUTION Thorndyke wishes t o t e s t the e f f e c t s of v i o l a t i n g this rule. He does s o by performing r e c a l l , summary and r e c o g n i t i o n experiments on f o u r v a r i -a t i o n s of t h e s t o r i e s . Thesc a r e a normal Form of t h e s t o r y , a story w i t h the Theme moved t o t h e end, a s t o r y with t h e Theme omitted and a descriptive version which omits c a u s a l and temporal c o n t i n u i t y . using only s t a t i v e o r single a c t i o n sentences. I n comp r e h e n s i b i l i t y judgements and r e c a l l t e s t s , the normal form s t o r y was b e s t , t h e Theme-after form n e x t , with t h e last two cases more dependent on t h e p a r t i c u l a r story. Recalls of t h e Theme-after passages alao showed a s t r o n g tendency t o r e l o c a t e t h e Theme t o i t s normal p o s i t i o n , near t h e beginning of t h e s t o r y . There was a l s o a tendency i n t h e more s t r u c t u r e d passages f o r h i g h e r l e v e l p r o p o s i t i o n s t o have a h i g h e r p r o b a b i l i t y of r e c a l l . Summarizations showed t h a t , of t h e r e c a l l e d p r o p o s i t i o n s , the h i g h e r l e v e l ones were much more l i k e l y t o be included a s part of the summary. Summaries of t h e d e s c r i p t i v e p r e s e n t a t i o n s yielded a wider selection of proposi,tions. I n recognition tests, t h e more s t r u c t u r e d passages produced more recognition e r r o r s when t h e t e s t p r o p o s i t i o n was c o n s i s t e n t w i t h t h e s t o r y , while t h e less s t r u c t u r e d passages produced more a c c u r a t e recognition. I n attempts t o s e p a r a t e t h e e f f e c t s of content and s t r u c t u r e , Thorndyke t e s t e d t h e f o u r s t o r i e s obtained by using each of t h e o r i g i n a l p l o t s w i t h two c h a r a c t e r / o b j e c t sets. He found t h a t the Farmer p l o t was always more comprehensible. He a l s o found t h a t p r e s e n t a t i o n of a second s t o r y using t h e same p l o t s t r u c t u r e improved l a t e r r e c a l l of t h e f i r s t s t o r y . Thorndyke concludes t h a t t e x t s t r u c t u r e s are b a s i c t o comprehension, although some are i n h e r e n t l y more complex than others. The d i s t i n c t i o n between s t r u c t u r e and content i s supported by t h e p o s i t i v e e f f e c t of repeated s t r u c t u r e p r i o r t o r e c a l l . And f i n a l l y , t h e s t r u c t u r a l p o s i t i o n of a p r o p o s i t i o n has s i g n i f i c a n t in\u00a3 h e n c e on both r e c a l l end summarization. 2.6 Text Grammar \"Text grammar\" i s the d e s i g n a t i o n used f o r a n o t h e r c u r r e n t l y a c t i v e line of i n v e s t i g a t i o n i n t o t h e s t r u c t u r e of t e x t . The fundamental i d e a s of t e x t grammar are discussed by van Dijk [ 7 2 ] , Ihwe [ 7 2 ] , Kummer [ 7 2 ] , Petofi [72] and P e t o f i and R i e s e r [73]. (Unfortunately, a g r e a t p o r t i o n of t h e work i s not a v a i l a b l e i n English.) The p r i n c i p l e concern i s t h e formulation of g e n e r a t i v e d e s c r i p t i o n s of texts, r a t h e r than i n d i v i d u a l s e n t e n c e s , and i s u s u a l l y approached u s i n g a model s i m i l a r t o t h a t of t h e generative semantics school of l i n g u i s t i c s . Van Dijk s u g g e s t s t h e following f i v e components of a text grammar: 1. Semantic formation r u l e s f o r t h e meanings of t e x t s , as a whole.2. Transformations which o p e r a t e on this t e x t meaning.3. Transformations which produce a sequence of s e n t e n t i a l semantic rep r e s e n t a t i o n s from t h e t e x t meaning, 4. Transformations which produce q sequence of s e n t e n t i a l s y n t a c t i c r e p r e s e n t a t i o n s (including l e x i c a l i terns) from tho sequence of s e n t e n t i a l semantic rep resent a t ions.5. Rules p a i r i n g s y n t a c t i c r e p r e s e n t a t i o n s with morphonologicalr e p r e s e n t a t i o n s . would operate i n t h e o r d e r given t o generate a t e x t * He cqhasires this is only the o u t l i n e of a theory, and t h a t most of t h e hard dttdls rerain t o be specified. A s i g n i f i c a n t amount of work has been done onccqoaents 1 and 2, formall,y defining represenmtions of text. This has in&ded n u m e m s examinations of t h e use of some extended p r e d i c a t e c a l c u l u s u a basic representation (primarily addressing t h e meaning of i n d i v i d u a l satemces and objects) but has a l s o l e d t o psychologically motivated imestiwtioas (as discussed i n Section 1.2.4) and t o studies i n t h e general stacture of textual types (as discussed i n Section 2.2). Discussions of capmmnts 3 and 4 have been mostly d e s c r i p t i v e , i n d i c a t i n g phenomena t h a t r s t be accounted f o~ without a c t u a l l y defining them. Many of t h e s e phenomena seem best characterized as operatdons operating on two o r more underlying propositions-These phenomena include anaphora (including pronominalization md article s e l e c t i o n ) , sentence stress, c o n t r a s t , use of c l a u s a l caajtanctions, and verb t e n s e determination. O f course, c e r t a i n l i t e r a r y e t i t i o n over prooariaanzation omission of causal connect ion, repet i t i o n of c e r t a i n #.tcnce forms, etc. Operations which occur on a s i n g l e p r o p o s i t i o n include: . a m t i c t r a n s f o n a t i o n s , such as p e r s o n i f i c a t i o n and t h e use of metaphor; w m c t i c traasformations, such a s inversions and o t h e r s t y l i s t i c operations;.rd pbaaological transformations, such a s those producing rhyme, a l l i t e r a t i o n .od meter.The set of r e a l i z a t i o n r u l e s used t o produce a p a r t i c u l a r t e x t fm its generated underlying s t r u c t u r e is assumed t o be l i mi t e d by t h e type of tart that being generated (e.g. mystery, n a r r a t i v e , etc.), which is Lpdicated by some underlying a b s t r a c t type marker. The s e l e c t e d r u l e s w i l l also determine t h e t e x t u a l c h a r a c t e r i s t i c s u s u a l l y r e f e r r e d t o as s t y l e o r literary merit. I t seems t h a t any complete t e x t u n d e r s t a n d i n g model must have a set of rules r e l a t i n g t h e u n d e r l y i n g s t r u c t u r e of t h e texti n d i v i d u a l and groups of p r o p o s i t i o n s t o g e t h e r w i t h g l o b a l a s p e c t s of t h e t e x tt o i t s surface r e a l i z a t i o n . Although u s i n g a g e n e r a t i v e model, t e x t grammarians are concerned with i d e n t i f y i n g and c h a r a c t e r i z i n g these r e l a t i o n s h i p s , and the r e s u l t s of t h e i r i n v e s t i g a t i o n s s h o u l d be of d e f i n i t e v a l u e i n d e s c r i b i n g t e x t u n d e r s t a n d i n g , once t h e i r r e s e a r c h has reached t h e p o i n t of c l e a r l y d e f i n i n g t h e n a t u r e of t h e s e s u r f a c e phenomena i n terms of t h e meaning of t e x t ." |
| }, |
| "FIGREF25": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "\\of a t e x t . T h i s s e c t i o n was begun w i t h t h e warning t h a t t h e d i s t i n c t i o nbetween t h e s e two d i f f e r e n t t y p e s of s t r u c t u r e s might be d i f f i c u l t t o define.I t seems c l e a r t h a t although each p r o p o s a l t h a t h a s been considered claims t o be concerned w i t h t h i s high l e v e l s t r u c t u r e , some confusion e x i s t s . The i n t e r a c t i o n s between t h e a b s t r a c t s t r u c t u r e s and t h e content s t r u c t u r e s aren o t c h a r a c t e r i z e d a t a l l . Thus, Kintsch and van Dijk do not d e s c r i b e t h e r e s t r i c t i o n s on what t y p e of e v e n t s may r e a l i z e a CON.PLICATION category. S i m i l a r l y , Rumelhart does not e x p l a i n how t h e c o r r e c t semantic r u l e would be s e l e c t e d when a choice i s p o s s i b l e (e.g." |
| }, |
| "FIGREF26": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "create a d d i t i o n a l c o n f u s i o n by mixing s y n t a c t i c c a t e g o r i e s w i t h categories t h a t would u s u a l l y b e c o n s i d e r e d semantic. Thus, c o n j u n c t i o n time o r d c d n g and c a u s a l i t y may a p p e a r i n t h e \" s y n t a c t i c \" s t r u c t u r e of a story. Although t h i s d i f f i c u l t y i n d i s t i n g u i s h i n g between t h e two t y p e s o f structure. suggest t h a t t h e r e is no d i s t i n c t i o n . t h e e v i d e n c e p r o v i d e d by t h e exirrtenca of d e s c r i b a b l e c l a e s e a of t e x t s seems t o i n d i o a t e t h a t some t y p e of h i g h level s t r u c t u r e s d o e x i s t . It remains t o c l e a r l y d e f i n e t h e n a t u r e of thasa s t r u c k u r e s , and t o e x p l a i n t h e i r r e l a t i o n s h i p w i t h c o n t e n t s t r u c t u r e s , without r e p e a t i n g t h e fallacy of c r e a t i n g a m u l t i t u d e of s u b c a t e g o r i e s (e.g. w n y d i f f e r e n t EVENT s u b t y p e s ) which a r e a l l e g e d t o be p u r e l y s y n t a c t i c c a t e g o r i e s . These s t r u c t u r e s f i t n e a t l y i n t o t h e model d i s c u s s e d i n t h e c o n c l u s i o n of t h e c o n t e n t s e c t i o n ( s e c t i o n 1.4). They are t h e h i g h e s t l e v e l o f s t r u c t u r e , u n d e r which c o n t e n t s t r u c t u r e s are organized. T h i s is c o n s i s t e n t w i t h the h y p o t h e s i s of K i n t s c h and van D i j k ( d i s c u s s e d i n s e c t i o n 1 -2 * 4 ) t h a t the s t r u c t u r a l d e s c r i p t i o n of a t e x t is one of t h e i m p o r t a n t a s p e c t s o f its r e p r e s e n t a t i o n . They s p e c i f i c a l l y s u g g e s t t h a t t h e s e st m c t u r e s organize m a c r o -s t r u c t u r e s , which p r o v i d e t h e c o n t e n t o r g a n i z a t i o n of t h e t e x t . T h i s i a q u i t e ~0 n S i S t e h t w i t h t h e r e s u l t s of s t r u c t u r a l a n a l y s i s which, as H e n d r i c k c [731 p o i n t s o u t , o r d i n a r i l y u s e s a summary o r s y n o p s i s , n o t t h e a c t u a l t e x t . as a b a s i s f o r a n a l y s i s . (Recall t h a t m a c r o -s t r u c t u r e s are proposed ae t h e b a s i s f o r summary g e n e r a t f o n ) . O f c o u r s e very simple o r v e r y n a v e l t e x t a might not have any structure of t h i s t y p e , scnce no a p p r o p r i a t e s t r u c t u r e would be pre-exis'tent. O f c o u r s e , t h e problem of l e a r n i n g these s t r u c t u r e is as d i f f i c u l t t o explain ( i f not more d i g f i c u l t ) a s t h e l e a r n i n g of c o n t e n t struct-uree. Such acquisition i s simply not w e l l understood. The work of t h e i c aspects of t e x t 0 may seem a d i s t a n t" |
| }, |
| "FIGREF27": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "many p o i n t s passed over very quickly by o t h e r workers are seen t o o f f e r formidable problems when one attempts t o set out a f u l l computational description. The discovery of inadequate d e s c r i p t i o n i s e s s e n t i a l i n develeping sound models. \"Computational\" w i l l be taken t o mean any model t h a t i s a c t u a l l y programmable o r t h a t i s formally defined. The work s e l e c t e d for discussion i n t h i s s e c t i o n i s generally t r e a t e d i n more d e t a i l than t h a t i n previous sectibns. The reason f o r t h i s is t h a t i f t h e complexity of rcomputationally specifying c e r t a i n r e p r e s e n t a t i o n s and processes is t o be made clear, i t i s impossible t o t r e a t these matters cursorily. This, i n t u r n , has n e c e s s i t a t e d g r e a t e r selectivity on t h e part of t h e author i n choosing work t h a t seems t o b e s t convey c e r t a i n types of problems. B e l l e r t 1701 attempts t o define the notion of t h e coherence of a t e x t , and i n doing s o suggests some of the same conclusions reached by computational l i n g u i s t s i n their t e x t understanding work. Coherence r e f e r s t o t h e p r o p e r t y of a set of u t t e r a n c e s which make i t a connected d i s c o u r s e r a t h e r t h a n a random c o l l e c t ion of utterances. R e f e r e n t i a l i d e n t i t y has of t e n been suggested a s an i n d i c a t i o n of coherence, but i t is c l e a r l y i n s u f f i c i e n t . For example:" |
| }, |
| "FIGREF28": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "are coherent without any e x p l i c i t r e f e r e n t i a l i d e n t i t y .Bellert d e f i h e s a coherent t e x t as one i n which t h e semantic i n t e r p r e t a t i o n of each sentence i s dependent on t h e semantic i n t e r p r e t a t i o n of the preceding sentences.The semantic i n t e r p r e t a t i o n of a sentence i a defined as t h e set of conclusions t h a t can be i n f e r r e d from t h a t sentence. She s u g g e s t s t h a t t h e r e are two t y p e s of conclusions t h a t may be drawn:1. those drawn only from knowledge of t h e language2. t h o s e drawn from knowledge of t h e worldBoth t y p e s of conclusions a r e a b s o l u t e l y necessary i n understanding t e x t and may be a p p r o p r i a t e l y drawn when t h e coherence of t h e t e x t requires. She concludes t h a t \"an u t t e r a n c e h a s meaning only i n t h e e n t i r e context and through o u r knowledge of t h e world\".The r e s t of the work discussed i n t h i s s e c t i o n wd 11 s t r o n g l y r e i n f o r c e these conclusions.The d i s c u s s i o n i s d i v i d e d i n t o two sections. The f i r s t c o n t a i n s e a r l i e r work t h a t s u p p o r t s the claim t h a t world knowledge is e s s e n t i a l t o t h e understanding of t e x t . The c l o s e l y r e l a t e d problem of making i m p l i c i t l y conveyed information e x p l i c i t i s a l s o a p r i n c i p l e concern. The second s e c t i o n Page 49 describes later work in whlch t h e o r~a n i z a t i o n of world knowledge is recognized as a c r i t i c a l question for t e x t understanding. work [72, 74, 761 is probably the f i r s t attempt t o s e t o u t in a well defined fashion the dimension of information processing that must be carried on i n the understanding of a story. The tarn \"understanding\" i s necessarily vague, but Charniak suggests an intuitive definition. Consider the following story fragment: Fred was goipg t o the s t o r e . Today was Jack's birthday and Fred was going t o get a present. It should be clear t h a t i f a human had read and u n d e r s t o o d t h i s fragment that he would be able to answer such questions as a semantic representation of t h i s fragment (i.e. an understanding of i t ) should e x p l i c i t l y contain t h e answers to ordinary questions such as these. Note the i m p o r t a n t p o i n t that this type of u n d e r s t a n d i n g could be a t t a i n e d only through the u s e of general knowledge of t h e world, a l o n g w i t h t h e e x p l i c i t s t a t e m e n t s of t h e t e x t . Knowledge required t o answer t h e above q u e s t i o n s would i n c l u d e such f a c t s as: 1. A p e r s o n h a v i n g a b i r t h d a y i s likely t a receive p r e s e n t s . 2. P r e s e n t s are of t e n bought a t s t o r e s . So Charniak's g o a l becomes o u t l i b i n g a n answer t o t h e q u e s t i o n of h w common sense knowledge may be i n c o r p o r a t e d i n t o t h e p r o c e s s of understanding natural language. A c l o s e l y r e l a t e d g o a l i s t h e d e t e r m i n a t i o n of how much knowledge of this t y p e i s r e q u i r e d by t h e basic problems of n a t u r a l language, such as the r e s o l u t i o n of pronominal r e f e r e n c e , Charniak b r e a k s t h e problem of p r o c e s s i n g n a t u r a l language i n t o two parts. The f i r~l t p a r t i s the t r a n s l a t i o n of n a t u r a l l'anguage i n t o a farm t h a t i s convenient f o r use i n making deductions. T h i s i n t e r n a l r e p r e s e n t a t i o n i s l i k e t h e u n d e r s t a n d i n g t h a t a person would be capable of o b t a i n i n g without cqntext. The i n t e r n a l r e p r e s e n t a t i o n i s a c a n o n i c a l p r o p o s i t i o n a l form, Thus, e i t h e r of t h e s e n t e n c e s Jack caught a cold. Jack came down w i t h a cold, might be represented by t h e p r o p o s i t i o n (BECOME-SICK-WITH JACK COLD) which r e p r e s e n t s the e x p l i c i t meaning o f e i t h c r of t h e sentences. The second p a r t of t h e problem i s what Charniak c a l l s Deep Semantic Processing (DSP). T h i s i s t h e p r o c e s s i n g t h a t makes e x p l i c i t t h e i m p l i c i t i n f o r m a t i o n conveyed by the story. Charniak e l e c t s t o examine only DSP. The f u n c t i o n of t h e system t h a t Charniak p r o p o s e s i s t o read a s t o r y which has a l r e a d y been t r a n s l a t e d i n t o i n t e r n a l r e p r e s e n t a t i o n and o u t p u t a data b a s e of p r o p o s i t i o n s t h a t e x p l i c i t l y represent a l l of t h e i n f o r m a t i o n conveyed by t h e s t o r y . In o r d e r t o do t h i s , t h e system uses commdn sense knowledge t o make i m p l i c i t i n f o r m a t i o n e x p l i c i t -T h i s knowledge i s coded i n t o t h e system i n i t i a l l y , and i s n o t learned o r modified, Charniak's g o a l o f u n d e r s t a n d i n g s t o r i e s l e a d s d i r e c t l y t o t h e q u e s t i o n s of what k i n d of i m p l i c i t information i s conveyed, what k i n d s of common sense knowledge are r e q u i r e d t o e x p l i c a t e t h a t i n f o r m a t i o n , how should this common s e n s e knowledge be r e p r e s e n t e d and when and how i s i t used. I n answering the q u e s t i o n about t h e t y p e s of i m p l i c i t in\u00a3 ormation Charni ak d i s c o v e r e d t h a t r e s o l u t i o n of pronominal r e f e r e n c e f r e q u e n t l y r e q u i red common s e n s e knowledge. C o n s i d e r Charniak's most d i s c u s s e d example [Charniak, 741 :" |
| }, |
| "FIGREF29": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "t o refer t o t h e t o p t h a t Penny would buy, b u t any p u r e l y s y n t a c t i c p r o c e d u r e s would p r o b a b l y s e l e c t t h e t o p t h a t Jack a l r e a d y h a s , on t h e b a s i s of recency. The c o r r e c t c h o i c e of r e f e r e n t s seems t o b e based on t h e common s e n s e knowledge that i f a p r e s e n t i s purchased and g i v e n , i t may be r e t u r n e d o r exchanged. The f a c t t h a t if a person r e c e i v e s a d u p l i c a t e present h e may not w i s h t o keep i t i s also r e l e v a n t . T h i s r e a l i z a t i o n l e d Charniak t o c o n c e n t r a t e on t h e problem of pronominal r e f e r e n c e .I n s t a n c e s of r e f e r e n c e are c l e a r l y r e c o g n i z a b l e i n t h e i n p u t , and t h e need t o u t i l i z e common s e n s e knowledge seems f u l l y p r e s e n t .Charniak d e c i d e d t o r e p r e s e n t common sense knowledge as i n f e r e n c e s . That i s , i m p l i c i t knowledge i s made e~p l i c i t by having r u l e s which i n f e r the i m p l i c i t f a c t upon p r e s e n t a t i o n of the n e c e s s a r y e x p l i c i t information. Por example, i f one knows t h a t i t i s r a i n i n g , t h e i n f e r e n c e t h a t anyone o u t s i d e w i l l g e t w e t i s v a l i d , The p r e s e n t a t i o n of the s e n t e n c e John went o u t s i d e .is s u f f i c i e n t t o trigger t h e common sense i n f e r e n c e J o h n g o t w e t , i n a s i t u a t i o n i n which i t i s known t o be r a i n i n g .Charniak's chaice of in\u00a3 e r e n c e s as a r e p r e s e n t a t i o n was probably s t r o n g l y i n f l u e n c e d by t h e a v a i l a b i l i t y of MICRO-PLANNER. ( F o r a d i s c u g s i o n of MICRO-PLANNER, see Winograd [74] and Charniak 1761) Before examining h i s i n f e r e n c e s i n more detail, one o t h e r q u e s t i o n must b e c o n s i d e r e d . S i n c e an i n f e r e n c e i s drawn a t some p a r t i c u l a r t i m e , t h e q u e s t i o n of when i n f e r e n c e s should be made arises. Three possible answers t o t h e q u e s t i o n a r e s u g g e s t e d : 1. Make n o i n f e r e n c e s until t h e s t o r y i s a c c e s s e d ( e . g . t o answer a q u e s t i o n a b o u t i t , summarize a p a r t of i t , e t c , ) Page 53 2. Make i n f e r e n c e s as t h e story i s read, b u t only t h o s e t h a t are necessary t o s o l v e some p a r t i c u l a r problem (e. g. t o solve p roblems of ambiguity o r r e f e r e n c e ) 3. Make non-problem i n f e r e n c e s as t h e s t o r y i s read, making e x p l i c i t aa much information as p o s s i b l e Charniak r e j e c t s t h e f i r s t p o s s i b i l i t y because i f i t is necessary t o make dedu.cti0ns from a l l p r e v i o u s p r o p o s i t i o n s when accessing some p r o p o s i t i o n , there i s no t h e o r e t i c a l difference from making them as t h e story is read, And he argues thqt t h e r e is no way t o be s u r e t h a t you have made t h e c o r r e c t and necessary i n f e r e n c e s without examining all p r e v i o u s propositions. An argument a g a i n s t both t h e f i r s t and second p o s s i b i l i t i e s i s t h a t t h e meaning of any p r o p o s i t i o n may be c o n t e x t -s e n s i t i v e Consider a story i n which Janet wants t o t r a d e w i t h Jack f o r h i s p a i n t s . The sentence\"Those p a i n t s make your airplane look funny\" J a n e t s a i d .should probably be understood as p a r t of a bargaining s t r a t e g y and n o t an expression' of Janet's true f e e l i n g s . S i n c e no e x p l i c i t problem i s presented by this sentence, no i n f e r e n c i n g would be done, and hence t h e correct understanding would be missed.So Charniak s e l e c t s t h e t h i r d p o s s i b ' i l i t y , making i n f e r e n c e s whenever p o s~i b l e as t h e story i s read. Charniak implements i n f e r e n c e s e s s e n t i a l l y a s MICRO-PLANNER antecedent theorems. H e d i s t j n g u i s h e s two types of common sense i n f e r e n c e s which he d e s i g n a t e s base rout i n e s and demons. Base r o u t i n e s rep r e s e n t knowledge t h a t should always be a v a t l a b l e and does not need t o be t r i g g e r e d by a s p e c i f i c c o n t e x t * F o r example, knowledge about t rsding s h o u l d always be a v a i l a b l e ro t h a t any s t a t e m e n t about a t r a d e t h a t h a s occurred w i l l cause t h e ibfarencer t h a t t h e ownership r e l a t i o n s have been reversed. Base routincar are implemented a s a n t e c e d e n t theorems matched a g a i n s t each i n p u t proporition. Some i n f e r e n c e s are n o t always a p p r o p r i a t e , such as the p r e v i o u s example about g e t t i n g wet i n t h e rain. The i n f e r e n c e would be i n c o r r e c t i f made in r non-raining s i t u a t i o n . T h i s t y p e of inference, c a l l e d a demon, is implementad a s a n t e c e d e n t theorems that are not always a c t i v e . Demons are activated by b a s e r o u t i n e s . Thua, t h e r e would be a base r o u t i n e about r a i n i n g which could make immediate i n f e r e n c e s and a c t i v a t e demons a p p r o p r i a t e t o t h e c o n t e x t , r a i n . The p r e v i o u s i n f e r e n c e rule uould be a demon of this type. Charniak r e f e r s t o the set of p r o p o s i t i o n s which match a base r o u t i n e ' s p a t t e r n as its t o p i c concept. He notes t h a t a t o p i c concept may occur e i t h e r b e f o r e o r a f t e r a p r o p o s i t i o n which would match one of t h e t o p i c concept'^ demons. Consider t h e two sequences: 1, It was raining. J a c k was o u t s i d e . 2. Jack was o u t s i d e . I t was r a i n i n g . I n t h e f i r s t sequence, t h e r a i n b a s e r o u t i n e i s matched which a c t i v a t e s t h e outside-implies-wet demon. The demon matches the next p r o p o s i t i o n causing t h e i n f e r e n c e t h a t Jack i s wet. Charniak c a l l s t h i s looking f o m a r d , and i t i s handled c o r r e c t l y by antecedent theorems. I n t h e second sequence, however, the demon i s n o t a c t i v a t e d u n t i l a f t e r t h e p r o p o s i t i o n s o the i n f e r e n c e i s Page 55 mirsed. T h i s i s c a l l e d looking backward. Charniak proposes a n e x t e n a i o n t o a n t e c e d e n t theorems s o t h a t when a demon is a c t i v a t e d , the d a t a base i s m a r c h e d f o r matches. Then, i n t h e second sequence, the i n f e r e n c e would be made when t h e t h e demon was a c t i v a t e d . A f i n a l o b s e r v a t i o n about demons concerns d e a c t i v a t i o n . Obviously, i f demons are c o n t e x t dependent, they must be d e a c t i v a t e d when t h e c o n t e x t i s no longer preeent. T h i~ i s i l l u s t r a t e d by It was raining. When i t q u i t , J a c k went o u t s i d e . It should n o t be i n f e r r e d t h a t J a c k g o t w e t . Although t h i a simple example could probably be handled by t h e rain base r o u t i n e , i n g e n e r a l t h e problem i s q u i t e d i f f i c u l t and Charniak offers no real s o l u t i o n s . He assumes t h a t d e a c t i v a t i n g demons a f t e r some f i x e d number of i n t e r m e d i a t e p r o p o s i t i o n s would be a s a t i s f a c t o r y f i r s t approximation. Charniak's model i n c l u d e s two o t h e r components. Be s u g g e s t s a bookkeeping component t o keep t h e data base u p d a t e d and t o n s i s t e n t . When inferences are made, they w i l l f r e q u e n t l y r e p l a c e o t h e r a s s e r t i o n s p r e v i o u s l y true. For example i f J a c k i a i n s i d e , t h e n goes o u t s i d e , bookkeeping would mark the o r i g i n a l f a c t as no l o n g e r t r u e , b u t would keep i t f o r h i s t o r i c a lpurposes (such a s answering t h e q u e s t i o n , \"Was Jack inside?\").The f o u r t h component i s made up of f a c t f i n d e r s . These are necessary as a r e s u l t of Charniak's d e c i s i o n t o make a l l p o s s i b l e i n \u00a3 e r e n c e s which are exp resaed as demons as the s t o r y i s read. C l e a r l y , many possible i n f e r e n c e s should n o t be made t o avoid clogging t h e system with a huge number of assertions. To a v o i d t h i s , t h e s e unnecessary i n f e r e n c e s w i l l n o t be r e a l i z e d a s dcmons. I n f e r e n c e s such a s t h e f a c t s t h a t John i s i n h i s house, h i s neighbor!~ood, c t c , i f h e is i n h i s k i t c h e n f a l l i n t o t h i s category. Some s i t u a t i o n s may r e q u i r e t h e a v a i l a b i l i t y o f t h e s e f a c t so The sequence Jack was i n t h e house. L a t e r he was i n t h e k i t c h e n . should n o t be i n t e r p r e t e d as a change i n l o c a t i o n , from t h e h o u s e t o the k i t c h e n . F a c t f i n d e r s a r e implemented as MICRO-PLANNER consctluent theorems, w i t h p a t t e r n s which a r e matched a g a i n s t d e s i r e d goals. They t h u s make c e r t a i n i n f o r m a t i o n a v a i l a b l e through d e d u c t i o n t h a t i s not s u f f i c i e n t l y important t o i n f e r a s soon as p o s s i b l e . Fact finders a r e always a v a i l a b l e , n o t a c t i v a t e d and d e a c t i v a t e d l i k e demons. Charniak's complete model i s d e p i c t e d below. The model shows t h a t t h e i n f e r e n c e s a r e t r e a t e d ;Like a d d i t i o n a l p r o p o s i t i o n s and a r e t h u s subject t o In some ways Rieger's work [ 7 4 . 751 seems c l o s e l y r e l a t e d t o Charniak's, b u t i t i s not c l e a r t h a t t h i s i s completely true. Ricger's s y s t e m a c c e p t s s e n t e n c e s as i n p u t which a r e a l r e a d y analyzed i n t o t h e i r semantic r e p r e s e n t a t i o n s , and makes e x p l i c i t a d d i t i o n a l i n f o r m a t i o n t h a t he claims i s Page 57 i q l i c i t l y in the input. R e designates these additional pieces of information inferences, and agrees with Charniak t h a t they should b e made whenever possible Rieger discusses t h e use of i n \u00a3 erences i n handling problems such as reference, but the bulk of h i s work i s s p e c i f i c a t i o n of t h e inferences, themselves, and i t i s here t h a t t h e main value of his work is found. Rieger Sdentlfies sixteen classes of inferences which will be discussed i n t h r e e b r o d categories (not Rieger's categories). are inferences which are concerned with the causal camectiopg between states and a c t s (Rieger's treatment resembles t h e use i n robotics wo* ( e . 8 . Fikes and Nilsson [71] ) of preconditionsstates that be true for an a c t t o occurand postconditionsstates t h a t result fror m act occurring). Given t h a t a s t a t e or act i s t r u e or has occurred, rbrt inferences may be made? 1 . Causative inferences suggest the likely cause input: Mary has t h e diamond ring. inference: Someone must have given or s o l d the r i n g to Mary." |
| }, |
| "FIGREF30": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "suggest results t h a t f ollswed input: Mae gave John the book. inference: John h a s the book, 4. Missing enablement inferences explain why something cannot occur input: Mary couldn't see t h e horses finish. inference: Something must be blocking her view. Page 58 5 . I n t e r v e n t i o n i n f e r e n c e s e x p l a i n how something may be stopped o r p revent ed input: Mary was h i c t i n g John l j i t h a bat. i n f e r e n c e : Taking t h e b a t away from Mary would stop her. s i n g I n f o r m a t i o n -The second c a t e g o r y of in\u00a3 e rences concerns s u p p l y i n g common knowladgo a b o u t f a m i l i a r o b j e c t s o r a c t i o n s t h a t is n o t in t h e input. 1, S p e c i f i c a t i o n i n f e r e n c e s f i l l i n missing p a r t s i n p u t : J o h n h i t Mary. i n f e r e n c e : John used h i s hand t o h i t Mary, 2. F u n c t i o n i n f e r e n c e s supply t h e normal r o l e of o b j e c t s i n p u t : J o h n got a book. i n f e r e n c e : John will read tbe book. 3. Normatiwe i n f e r e n c e s supply information a b o u t what f s normally t r u e i n p u t : Pete i s a, human." |
| }, |
| "FIGREF31": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "d u r a t i o n i n f e r e n c e s s u g g e s t how long some s t a t e w i l l p e r s i s t i n p u t : J o h n s t a r t e d e a t i n g a t 6 : 0 0 . i n f e r e n c e : fie is probably still eating a t 6:15.5 . Feature inferences connect features of objects w i t h t h e o b j e c t si n p u t : F r e d wagged h i s t a i l . Xnference: Fred is a non-human animal-6. s i t u a t i o n i n f e r e n c e s supply o t h e r l i k e l y aspects of a s f t u a t i o n i n p u t : M a r y is going t o a masquerade.i n f e r e n c e : Mary i s probabky wearing a costume*3.1.2.3 Motivation Atid Rnwledse -The third category of lnfarsoces concerns human mrivationr and knowledge, and their relation t o one's actions.1. Motivation inferences suggest reasons for an actor to do eo~vlthing input: John h i t Mary.infetenee John wanted Mary t o be hurt.2. Actioa-pwdiction infbmncao suggert a poseible courae of actian from a pamon's wants input: John wants some n a i l a . inaereacer John is likely t o go t o a hardware store.3. Enablement-p r e d i c t i o n inferences suggest reasons for s pamon bringing about a certain e t a t e input: Maryput on hot glasses. inferencd: Mary probably wants t o look a t e m t h i n g . 4, Utterance-intent inferences supply information intended, but not actually ~t a t e d input: Mary couldn't jump the fence. inference: Mary wanted t o jump t h e fence. 5 . Knwledge propagation inferences predict whet else a patson would know i f given that he knows certain t h i n g s input: B i l l knew that Mary hit. John with a bat. inference: R i l l knew that John had been hurt. One point that ehould be c l a r i f i e d i s that the use of the tern inference to designate the i m p l i c i t information a sentence conveys i e p a r t i d l l y misleedcng. What i s i m p l i c i t , and thus misht need to be made explicit, i s directly a function of the seaant i c representation. Consider t h e previous examp le Mary gave John the book. I f t h e fact t h a t John p o s s e s s e s t h e book a f t e r being g i v e n i t i s an i n f e r e n c e , I t 1s not c l e a r what meaning t h e r e p r e s e n t a t i o n of t h e sentence captures." |
| }, |
| "FIGREF32": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "s t an i m p l i c i t a d d i t i o n . A t the o t h e r extreme, t h e example John h i t Mary. s u r e l y does q o t n e c e s s a r i l y mean t h a t J o h n wanted t o h u r t Mary. T h i s i s c l e a r l y conveyed i m p l i c i t l y , i f a t a l l . It i s very d i f f i c u l t t o e v a l u a t e t h e completeness o r adequacy of such a l a r g e set of i n f e r e n c e types, b u t R i e g e r c e r t a i n l y s u g g e s t s how large the class of i n f o r m a t i o n i s t h a t can be i m p l i c i t l y conveyed by t e x t and, therefore which might need t o be made e x p l i c i t d u r i n g the p r o c e s s i n g of that t e x t . Any proposed scheme f o r r e p r e s e n t i n g and understanding t e x t s h o u l d be examined t o see how each of t h e s e t y p e s of information i s s t o r e d and i n what way t h e i n f o r m a t i o n i s a s s e r t e d t o be p a r t of t h e meaning of t h e input. model of s t o r y comprehension p r i m a r i l y discgsses c o n n e~t t o t~a between p r o p o s i t i o n s of a s t o r y that a l t e r t h e sequence of e v e n t s involved i n u n d e r s t a n d i n g t h e s t o n . Far example, t h e understanding of some s e n t e n c e of a s t o r y . may be c o r r e c t because a demon had been p r e v i o u s l y a c t i v a t e d , thus a l l o w i n g t h e c o r r e c t i n t e r p r e t a t i o n of t h a t sentence. Charniak dges not say much about e x p l i c i t l y rep r e s e a t i n g t h e connection between t h e p r o p o s i t i o n r e s p o n s i b l e f o r a c t i v a t i n g t h e demon, and t h e c o r r e c t l y understood Page 61 proposition. However, s i n c e h e mainly i n v e s t i g a t e s reference, o b t a i n i n g t h e c o r r e c t r e f e r e n t i s an e x p l i c i t r e p r e s e n t a t i o n of t h e understanding. Clearly, many kinds of connections might e x i s t between p r o p o s i t i o n s of a connected discourse. Schank ( [73b], Schank and Abelson [77] ) investigates c a u s a l i t y as one of t h e primary discourse connectives. Schank's proposals are discussed i n terms of his Conceptual Dependegcy theory [Schank, 73a], but are largely independent of it. The term conceptualization i s used t o r e f e r t o e i t h e r a p r o p o s i t i o n (using one of t h e s m a l l s e t of p r i m i t i v e p r e d i c a t e s o r a c t s ) o r t o a s t a t e ( i . e . an o b j e c t having some value f o r some attribute). Schank notes t h a t t h e sentence John cried because Mary s a i d h e was ugly. a s s e r t s a connection between Mary said John was ugly. and John cried. But c l o s e r consideration reveals t h a t some of the links i n this connection are unspecified, John must have found out t h a t Mary had s a i d i t and t h i s knowledge must have made h i m unhappy, which was manifested in his crying. I n o r d e r t o be a b l e to recognize instances of missing l i n k s and t o be a b l e t o supply them, Schank develops a c l a s s i f i c a t i o n of t y p e s of causality and characterizes their form and meaning. ( I t should be c l e a r t h a t here, as i n a l l Conceptual Dependency, there i s no simple mapping from English t o t h e Page, 62 underlying r e p r e s e n t a t i o n , ) T h e f i r s t type o f c a u s a t i o n i s a r e s u l t i n which a n a c t may bring about a change of s t a t e , F o r example John went t o New York, -RESULT-> John i s i n New York, The second t y p e of c a u s a t i o n i s enablement T h i s is t h e s i t u a t i o n i n which a change of s t a t e b r i n g s about t h e c o n d i t i o n s necessary f o r some act t o occur, John has a ring, -ENABLE-> John gave i t t o Mary. The t h i r d c a u s a l t y p e , i n i t i a t i o n , i s t h e r e l a t i o n s h i p between any c o n c e p t u a l i z a t i o n and the a c t bf someone t h i n k i n g a b o u t t h a t thing. John -INITIATE-> I t h i n k about B i l l . (John reminds m e of B i l l . ) The fourth and f i n a l t y p e is reason c a u s a t i o n , T h i s i s t h e r e l a t i o n s h i p t h a t holds between t h e a c t of d e c i d i n g t o do something and a c t u a l l y doing that thing. John decided t o l e a v e town. -RFASON-> John l e f t town, Schank also s u g g e s t s the utility o f s non-specific caugal connection, which h e designates \"lead-t 0\". This would be used t o r e p resent c a u s a l l y connected e v e n t s i n s i t u a t i o n s where t h e s p e c i f i c t y p e of c a u s a l i t y o r t h e s p e c i f i c c h a i n of causes i s unknown, Page 63 Although Schnnk explores t h e purely s y n t a c t i c expansion of English c a u s a l s i n t o h i s formal c a u s a l s , h e concludes t h a t t h e s y n t a c t i c approach i s t o o limited. Consider t h e sentence The hurricane caused my depression. T r e a t i n g t h i s a s an i n s t a n c e of i n i t i a t i o n c a u s a t i o n ( t h e h u r r i c a n e initiated depressed t h i n k i n g ) makes i t causal-syntact i c a l l y c o r r e c t , b u t t h e r e i e c l e a r l y something m i s s i n g t h a t e x p l a i n s what l e d t o t h e depressian. Schank s u g g e s t s t h a t o t h e r knowledge i s used t o f i n d the connection, and t h i s suggestion seems t o indicate a f a r more g e n e r a l approach t o understanding English e x p r e s s i o n s of c a u s a l i t y t h a n any s y n t a c t i c approach, Schank s u g g e s t s t h a t a p e r s o n would, i f p o s s i b l e , f i n d a reasonable connection such as t h a t t h e h u r r i c a n e probably blew down t h e speaker's house which caused him t o l o s e money which caused him t o become depressed, I n order t o accomplish t h i s e l a b o r a t i o n a person would have t o u s e a g r e a t d e a l of world knowledge. Schank p r o p o s e s t h a t a number of d i f f e r e n t k i n d s of knowledge are involved here. F i r s t , t h e r e are axioms about t h e way people f e e l , such a s A l . Bad results can cause depression. Associated with an axiom would be a number of more i d i o s y n c r a t i c b e l i e f s , such as IB1. Less money i s a bad r e s u l t . F i n a l l y , general knowledge of t h e world would a l s o be necessary. T h i s would necessarily include f a c t s auch as WK1. O b j e c t s can have monetary value. WK2. Changes i n an o b j e c t ' s p h y s i c a l s t a t e can cause a change i n i t s worth* A p e r s o n making the above connections would essentially be engaged i n problem solving. Given the f i n a l s t a t e , d e p r e s s i o o , he would f i n d en axiom which e x p l a i n s i t then f i n d a n i d i o s y n c r a t i c b e l i e f which would meet the e x i m ' s p r e c o n d i t i o n s (bad r e s u l t s ) , t h e n u s e world knowledge t o e s t a b l i s h t h a t a n e g a t i v e change of s t a t e could be t h e real c u l p r i t , then n o t i c e t h a t a h u r r i c a n e i s able t o destroy o b j e c t s . The real p o i n t o f t h i s discussion seems t o be that much a d d i t i o n a l knowledge i s r e q u i r e d t o e s t a b l i s h conrlections t h a t p e o p l e are able t o (and do) make when understanding connected sentences. This i s an a f f i r m a t i o n of Bellert's hypothesis, and con\u00a3 irms Charniak's c o n c l u s i o n s , b u t i s reached a f t e r work on a d i f f e r e n t t e x t understanding p roblem, Schank [75b] a t t e m p t s t o use t h e s e types of causality as a b a s i s f o r d e f i n i n g t h e semantic rep r e s e n t a t i o n of a paragraph. lie a r g u e s t h a t t h e c o l l e c t e d rep r e s e n t a t i o n s of t h e i n d i v i d u a l s e n t e n c e s of t h e paragraph do not form a r e p r e s e n t a t i o n o f t h e paragraph a s a whole. Much addttinnal implicit i n f o r m a t i o n could be made e x p l i c i t , b u t Schank suggests t h a t u n d i r e c t e d e x p l i c a t i o n i s not p l a u s i b l e e He o f f e r s a s o l u t i o n t o t h e problem o f what i n f o r m a t i o n should be made e x p l i c i t by d e f i n i n g a paragraph r e p r e s e n t a t i o n as a c a u s a l l y connected sequence. I n f e r e n c e s a r e l i m i t e d t o t h o s e items of i n f o r m a t i o n r e q u i r e d t~ f i n d t h e c a u s a l connections. The e t a t e a t h a t are u s u a l l y true when a n a c t o c c u r s a r e c a l l e d t h e neceasaty c o n d i t i o n 8 of t h a t act. These states would be ~o n n e c t c d t o t h e a c t by an e n a b l i n g c a u s a t i o n . Schank d i v i d e s necessary c o n d i t i o n s i n t o two types: absolutely necessary c o n d i t i o n s and r e a s o n a b l e necessary c o n d i t i o n s (ANCs and RNCs)" |
| }, |
| "FIGREF33": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "a8 an ANC (among o t h e r s ) t h a t John h a s a yard. An RNC might be t h a t i t was n i c e weather, b u t t h i s could be v i o l a t e d * I t would only i n d i c a t e something u n u s u a l , and p e r h a p s s i g n i f i c a n t . Schank i l l u s t r a t e s t h e u s e of neceeeary c o n d i t i o n s t o e s t a b l i s h c a u s a l c h a i n s by a n a l y z i n g s e v e r a l s t o r i e s . A b r i e f d e s c r i p t i o n of one of t h e s e should i n d i c a t e h i s methods. H e c o n s i d e r s t h e f o l l o w i n g paragraph: John began t o mow h i s lawn. Suddenly h i s t o e s t a r t e d bleeding, ... When he cleaned o f f h i s t o e , h e d i s c o v e r e d t h a t h e had s t e p p e d i n tomato sauce." |
| }, |
| "FIGREF34": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "These c o n d i t i o n s would be i n f e r r e d when t h e f i r s t s e n t e n c e was processed s i n c e they were not e x p l i c i t l y s t a t e d . I f t h e f i r s t sentence had been preceded by a n e x p l i c i t s t a t e m e n t of e i t h e r o r b o t h of t h e s e RNC6, they would have been connected as e n a b l i n g t h e mowing. (There i s g r e a t d i f f i c u l t y i n f i n d i n g a g e n e r a l scheme which a l l o w s i n f e r r i n g of normal n e c e s s a r y c o n d i t i o n s , b u t p r o h i b i t s i n f e r r i n g unusual ones. I n t h i s case, i t secms reasanable t o infer a l l of them.) The second sentence p r e s e n t s a more d i f f i c u l t problem. Same sequence of i n f e r e n c e s would have t o be made which produces t h e chain: mowing i n v o l v e s t u r n i n g b l a d e s , which could h i t a t o e , which could cause i t t o bleed." |
| }, |
| "FIGREF35": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": ", Schank s t a t e s t h e following c o n d i t i o n s f o r t h e r e p r e s e n t a t i o n of a paragraph: 1. Each i n p u t sentence i s represented. 2. These rep r e s e n t a t i o n s should be conceptually connected, p r i m a r i l y by c a u s a l chains,3. The necessary condl t f o n s f o r every conceptualization must be e x p l i c i t l y represented, and may o r i g i n a t e as i n p u t sentences o r may be i n f e r r e d e i t h e r from p r e v i o u s c o n c e p t u a l i z a t i o n s o r because they are normal." |
| }, |
| "FIGREF36": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "c h a r a c t e r i s t i c common t o a l l of t h e p r e c e d i n g work i s t h a t a great deal of world knowledge i a r e q u i r e d , b u t there i s no c l e a r o r g a n i z a t i o n of t h i s knowledge. I t i s clear t h a t any system with large amounts o f knowledge r e p r e s e n t e d simply as demons o r i n f e r e n c e r u l e a would become bogged dawn searching f o r r e l e v a n t knowledge and would q u i t e probably draw i n c o r r e c t in\u00a3 e r e n c e s because knowledge would be a p p l i e d i n inappropriate contexts (recall Charniak's concern about t h e d e a c t i v a t i o n of demons). Charniak [ 75, 76, 771 himself s u g g e s t s t h a t more organized knowledge would be s u p e r i o r t o the demon approach. These r e a l i z a t i o n s have l e d t o a number of p r o p o s a l s f o r t h e o r g a n i z a t i o n of knowledge. Minsky [75] i a t r o d u c e d t h e term \"frame\" f o r knowledge s t r u c t u r e s , but Winograd I751 Bobrow and Norman [ 7 5 ] ( u s i n g t b term \"schema\") and Rumelhart and Ortbny (771 d i s c u s s c l o s e l y r e l a t e d ideas. Recently, Bobrow and Norman 1771 have d e s c r i b e d a language f o r knowledge r e p r e s e n t a t i o n (Ec2RL) i n which the frame concept p l a y s a c e n t r a l role. These knowledge s t r u c t u r e s are intended t o o r g a n i z e conventional o r e n c y c l o p e d i c kncrwledge, rather than d e f initionsl features or c h a r a c t e r i s t i c s , T h i s knowledge i s described i n terms of roles o r s l o t s which p a r t i c i p a t e i n t h e s i t u a t i o n . These are t h e v a r i a b l e s of t h e frame. Frames must also have t h e a b i l i t y t o r e f e r e n c e o t h e r frames. A d e t a i l e d d i s c u s s i o n o t t h e nature of Page 58 frames i s i n a p p r o p r i a t e h e r e , p r i m a r i l y because i t i s o u t s i d e the scope of t h i s survey, b u t i n p a r t because many of t h e computational q u e s t i o n s regarding frames are n o t completely answered. However, r e c e n t computational work haa begun t h e a t t e m p t t o i n c o r p o r a t e such o r g a n i z a t i o n of knowledge i n t o t e x t u n d e r s t a n d i n g systems. Schank e t a 1 1751 and C u l l i n g f o r d [751) extends his ideas wfth t h e a d d i t i o n of l a r g e , p r e -e x i s t i n g knowledge s t r u c t u r e s t h a t he c a l l s s c r i p t s . H e s a y s t h a t i t i s necessary t o have l a r g e amounts o f specific knowledge about known s i t u a t i o n s , s i n c e otherwise i t i s d i f f i c u l t o r i m p o s s i b l e t o recognize t h e c a u s a l r e l a t i o n s h i p s between events. A s c r i p t is a sequence of a c t i o n s t h a t p r o v i d e knowledge about t h e t y p i c a l occurrence of some s i t u a t i o n . T h e f o l l o w i n g i s a partial d e s c r i p t i o n of t h e s c r i p t for going t o a r e s t a u r a n t : S c r i p t : R e s t a u r a n t Track: Coffee shop Roles: Custome r, W a i t r e s s , Chef, CasKier Props: Tables, Menu, Food, Check, Money Reason: To g e t food t o eat E n t r y c o n d i t i o n s : Customer ( i s hungry; has money) R e s u l t s : Customer ( i s n o t hungry; h a s less money;i s p l e a s e d ) Cashier (has more money) Scene 1: E n t e r i n g Go t o r e s t a u r a n t Find a n ernp ty t a b l e" |
| }, |
| "FIGREF37": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "scenes a r e t h e main e p i s o d e s of t h e e v e n t ' a n d each i s d e f i n e d i n terms of a sequence of s p e c i f i c a c t i o n s . Schank s u g g e s t s t h a t s c r i p t s would need t o be d i v i d e d a t t h e t o p l e v e l i n t o d i f f e r e n t t r a c k s whicih d i s t i n g u i s h t h e sequences f o r d i f f e r e n t t y p e s of restaurants. Th& u s e f u l n e s s of a s c r i p t i s seen by c o n s i d e r i n g t h e sequence J o h n w e n t to a r e s t a u r a n t . H e ordered a hamburger from t h e waitress. H e p a i d and l e f t . Many d e t a i l s of this sequence have been omitted and connections between t h e a c t i o n s would be impossible to e s t a b l i s h wf t h o u t knowledge of what c o n s t i t u t e s g o i n g t o a r e s t a u r a n t . Furthermore, t h e d e f i n i t e r e f e r e n c e t o t h e waitress i s meaningful only because all p a r t i c i p a n t s i n a s c r i p t are a u t o m a t i c a l l y i n t r o d u c e d by any r e f e r e n c e to t h a t s c r i p t ." |
| }, |
| "FIGREF38": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "The s c r i p t also c o n t a i n s t h ec o n d i t i o n s that must be t r u e f o r s u c c e s s f u l e x e c u t i o n of t h e s c r i p t , as w e l l as t h e r e s u l t s of s u c c e s s f u l execution. Each scene h a s one act t h a t i s c o n s i d e r e d t h e e s s e n t i a l a c t of t h a t p a r t of t h e s c r i p t . These acts are c a l l e d t h e M A I N CONceptualizations of a scene. S c r i p t s r e p r e s e n t t h e knowledge about conventional s i t u a t i o w t h a t any r e a d e r would b e assumed t o know.But t h e s e p a r a t e q u e s t i o n of how t o r e p r e s e n t a t e x t which c o n t a i n s s c r i p t usages must be answered. I f t h ercy r e s e n t a t i o n of t h e r e s t a u r a n t s t o r y above wee merely an e l a b o r a t e d c a u s a l c h a i n c o n t a i n i n g many e v e n t s i n \u00a3 e r r e d from t h e r e s t a u r a n t s c r i p t , t h e u n i f y i n g d e s c r i p t i o n t h a t t h e s e e v e n t s t o g e t h e r c o n s t i t u t e going t o a r e s t a u r a n t worrld be l o s t . Schank. proposes t h a t t h a t t h e e l a b o r a t e d c a u s a l c h a i n i s only one l e v e l of t h e s t o r y ' s r e p r e s e n t a t i o nt h e Conceptual Dependency (CD) l e v e l . T h e r e i s a l s o a second l e v e l of r e p r e s e n t a t i o nt h e Knowledge S t r u c t u r e (KS) l e v e l . A t t h e KS l e v e l t h e above s t o r y would have a r e p r e s e n t a t i o n l i k e : S c r i p t s R e s t a u r a n t Custome rcJohn Food=Hambu rge r A d d i t i o n a l l y , t h e s c r i p t r e p r e s e n t a t i o n would be l i n k e d t o each of t h e e v e n t s a t t h e CD l e v e l which was p a r t of t h a t s c r i p t i n s t a n c e . Schank s u g g e s t s t h a t t h e c a u s a l c h a i n a t t h e CD l e v e l should c o n t a i n a l l t h e e v e n t s e x p l i c i t l y mentioned p l u s t h e MAINCONS of any s c e n e t h a t is mentioned. I f any e v e n t i s encountered which i s not a n d e r s t a n d a b l e i n terms of t h e c u r r e n t s c r i p t , t h e e v e n t w i l l be r e p r e s e n t e d a t t h e CD l e v e l , b u t w i l l a l s o be placed on a Wierdlist. F o r example Schank d i s c u s s e s t h e p r o c e s s i n g of a s t o r y i n which John h a s h i s pocket picked w h i l e r i d i n g a subway. L a t e r , he has no money w i t h which t o pay the b i l l a f t e r e a t i n g a t a r e s t a u r a n t -The f i r s t sequence of e v e n t s i s understood i n terms of t h e subway s c r i p t , b u t t h e pocket p i c k i n g e v e n t i s p l a c e d on t h e Wierd l i s t . When a p p l y i n g t h e r e s t a u r a n t G c r i p t , t h e &aability to pay is encountered i n the t e x t and i s inexplicable from t h e r o t m~r m t script. The s c r i p t a p p l i e r asks t h e monitoring program t o look f o r c p r r s~a l occurrences t h a t could r e s u l t i n John having no money. The d t o r finds the pocket picking event, and r e t u r n s t h e informati6n t o t h e scrlpt q p l i e r which uses i t as the required explanation.Of course, the problems i n general are much more d i f f i c u l t than this r i q l e example indicates. For example, an unusual event w i t h i n a s c r i p t could imterrupt the script's normal continuation. Obstacles, such as not being able to get wbat you want a t a r e s t a u r a n t , may cause a l t e r i n g o r abandoning t h e restaurant script. D i s t r a c t i o n s , sueh a s a robbery i n a t e s t a u r a n t may lead 3.2.1 2 Goals And P l a n s -Schankand Abelson (75, 771 ( a l s o see Meehan 1761 and Wilensky [ 7 6 ] ) have r e c e n t l y suggested t h a t a11 connected event sequences are not apprcrp riatoly represented by s c r l p t s . Consider t h e sequence: Johh wanted t o become king-H e went t o g e t some arsenic. It seems that t r e a t i n g t h e poisoning-the-king s i t u a t i o n a s so c o n v e n t i o a r l t h a t a s c r i p t f o r i t e x i s t s i s u n r e a l i s t i c , The s o l u t i o n l i e s i n realixlag t h a t w h i l e s c r i p t s handle w e l l known s i t u a t i o n s , mechanisms rmst e x i s t which are a b l e t o handle novel s i t u a t i o n s using g e n e r a l knowledge. \"Plans\" are suggested as t h e a p p r o p r i a t e mechanism. A p l a n i s a general course of actaon i n t e n d e d t o r e a l i z e some g o a l ( s ) , A p o s s i b l e h i e r a r c h y of high-level g o a l s i s shown i n Figure 3-l(a). Motivated a c t i o n s are always a s s o c i a t e d w i t h the s a t i s f a c t i o n of some high-level goal-Frequently, more s p e c i f i c sub-goals e x i s t which are motivated by a high-level goal. For example, someone may want t o go t o t h e t r a i n s t a t t o n ( i n s t r u m e n t a l g o a l ) s o he can go t o New York (a 1 I s p e c i f i e d g o a l ) s o he can enjoy pleasure\" (high-level g o a l ) . The purpose of c l e a r l y d e f i n i n g a g o a l h i e r a r c h y i s t o enable understanding of s i t u a t i o n s i n which an actor faces goal conflict and elects $0 pursue the h i g h e s t g o a l ( e . g w \"preserve health\" r a t h e r than \"en joy pleasure\"), But general knowledge about Row goals can be achieved must a l s o be a v a i l a b l e . The \"D-goal\" i s proposed as t h e fundamental unit of organization f o r suhh i n f o r m a t i o n . A D-goal i e a p o i n t pf access t o p l a n n i n g information f o r t h e r e a l i z a t i o n of some goal. Foe example, D-KNOW i s t h e D-goal f o r t h e goal o f knowing something, D-CQNT f o r the goal of p h y s i c a l l y c o n t r o Z l i n g something, and D-PROX f o r the goal of being in proximity to something. The D-goals' a s s o c i a t e d knowledge is an o r d e r e d l i s t of \"planboxee\", each of which p t o v i d e s d e t a i l i n f o p t i o n on one method for achieving t h e goal. The o r d e r i n g prwides thb sequence i n which the methods are l i k e l y t o be considered. For example, D-KNOW has an ASK planbox (as a highly likely method) which specifies the a c t u a l act of a s k i n g the appropriate q u e s t i o n , t h e i n t e n d e d r e s u l t ( f e e . g e t t i n g t h e answer), and t h e p r e -c o n d i t i o n s n e c e s s a r y t o successfujlly ask. Some of t h e p r e -c o n d i t i o n s are: communication i s possi\"b1e; t h e p e r s b n being asked knows the answer; the person being asked i s d i s p o s e d t o answer t h e q u e s t i o n , When a planbbx i s ext rernely s p e c i f i c , i t becomes a s c r i p t . Using t h e telephone book is a method o f acquiring certain k i n d s of knowledge that i s so conventional that it is a script, Certain 'recurring sequences of B-goals are called \"named plans\". The named plan, r e p a r a t l o n s a n d d o a c t i o n a p p r o p r i a t e t o x Figure 3 -l ( b ) s h a m the r e l a t i o n s h i p of D-goals, planboxes, s c r i p t s and named p l a n s t o each other and t o t h e g o a l h i e r a r c h y . The r e p r e s e n t a t i o n of t h e p l a n s of a c t o r s i n a t e x t would be a t t h e KS level ( w i t h s c r i p t s ) . The D-goals would be e x p l i c i t l y r e p r e s e n t e d and would be connected t o a c t u a l a c t s a t the CD level which ( a t t e m p t e d t o ) implement t h e p l a n For example, t h e s e n t e n c e : John t r i e d t o f i n d o u t who ate t h e candy. would be repreaentsd arro: (John,?.ate t h e candy)---implementati-on--> DO I Failure <--------------------result------------I That is, the plan of t r y i n g t o know who a t e the candy was implemented B S some u n s p e c i f i e d a c t (DO) which f a i l e d t o achieve t h e p l a n ' s goal. A c t s ( a t t h e CD level) which meet p r e -c o n d i t i o n s of p l a n b o x e s would be l i n k e d t o the D-goal they e n a b l e . A d d i t i o n a l i n f o r m a t i o n i s suggested as b e i n g a p p r o p r i a t e and necessary f o r t h e KS level. T h i s i n c l u d e s , f o r each character, h i s goals, t h e c u r r e n t s t a t u s of each goal, s t r a t e g i e s which c o u l d be used t o a c h i e v e each goal a n d f a c t s (true i n f o r m a t i o n as opposed t o a c t u a l occurrences) relevant t o goal u n d e r s t a n d i n g . T h i s in\u00a3 o n n a t i o n i.s m a i n t a i n e d on \"Coal F a t e Graphs\", w h i c h a l s o c o n t a i n a s s o c i a t i o n s between c h a r a c t e r s and any \"Themes\" (large goal complexes such as BECOMING-RICH) i n which h e i s p a r t i c i p a t i n g . Schapk and Abeleon appear t o be committed t o the development of a very complex system f o r t h e o r g a n i z a t i o n o f knarledga, and t h e r e p r e s e n t a t i o n of such knowledge when i t b c c u r s i n stories. I t s h o u l d be a p p a r e n t t h a t t h e y rare a t t e m p t i n g t o model g o a l related human knowledge and a c t i o n s , since such knowledge and a c t i o n s are common in s t o r i e s . Since t h i s knowledge p n be r e q u i r e d t o u n d e r a t a n d p a r t i c u l a r s t o r i e s i t 1 4 n o t clear a t what p o i n t ( i f any) one ceases t o s t u d y t e x t unde rs t a n d i n g and b e g i n s m o d e l l i n g p e r s o n a l i t y . 3.2.2 P h i l l i p s -P h i l l i p s * [ 7 5 a , 7 5 b ] p r e s e n t s probably t h e most comprehensive c o m p u t a t i o n a l model of t e x t , i n t h a t h e is concerned w i t h t h e r e p r e s e n t a t i o n of a l l types of knowledge and t e x t u a l r e l a t i o r s h i p s . T h i s b r e a d t h i s i n f o r m a t i v e , but n e c e s s q r i l y r e s u l t s i n a l a c k of d e p t h i n some areas. H e presents a r e p r e s e n t a t i o n a l scheme which h e u s e s f o r t h e v a r i~u s r e q u i r e d t y p e s of knowledge, which i n c l u d e world knowledge, l i % g u l s t i r knowledge and t s i s t s of both t h e s t a t i c d a t a s t r u c t u r e s and t h e dynamic p r o c e s s e s which o p e r a t e upon these s t r u c t u r e s . T h e static d a t a s t r u c t u r e i s a f a i r l y c o n v e n t i o n a l b u t very w e l l d e f i n e d , semantic network ( c l o s e l y r e l a t e d t o t h e s u g g e s t i o n s of Hays [73] ) which use nodes f o r e n t i t i e s and e v e n t s , and a r c s for t h e relationships between nodes. The set of hierarchical r e l a t i o n s h i p s between e n t i t i e s ( t h e t a x o n o m i c -s t r u c t u r e ) i s c a l l e d ' t h e p a r a d i g m a t i c s t r u c t u r e . The syntagmat i c s t r u c t u r e is t h e see o f r e l a t i o n s h i p s between e v e n t s and t h e e v e n t p a r t i c i p a n t s ( c a s e o r argument r e l a t i o n s h i p s ) . D i s c o u r s e , o r t e x t , i s r e p r e s e n t e d u s i n g t h e same r e p r e s e n t a t i o n a l scheme, b u t may be c h a r a c t e r i z e d by p r o p e r t i e s n o t applicable t o d i s c o n t i n u o u s knawlddge. The f i r s t of t h e s e p r o p e r t i e s i s connectedness. Two p r o p o s i t i o n s are p a r a d i g m a t i c a l l y connected i f each h a s an argument such t h a t the two arguments have a common p a r a d i g m a t i c s u p e r o r d i n a t e node, o r i f t h e f i r s t is the immediate s u p e r o r d i n a t e of t h e the second. F o r example, \" l i o n s \" and ' I t i g e r s \" b o t h have \"animal\" as s u p e r o r d i n a t e , s o t h e f o l l o w i n g are connected p a r a d i g m a t i c a l l y : L i o n s are indigenous co Africa. T i g e r s have s t r i p e s . And t h e two p r o p o s i t i o n s : Man i s a h u n t i n g animal. Modern man h u n t s f o r s p o r t . are a l s o connected s i n c e t h e second i s a more s p e c i f i c p r o p o s i t i o n (\"modern man\" i s immediately s u b o r d i n a t e t o ''man\") t h a n t h e f i r s t . Two p r o p o s i t i o n s are d i s c u r s i v e l y connected when d i s c u r s i v e r e l a t i o n s (em g. c a u s a l i t y ) e x i s t between them. The second a t t r i b u t e of d i s c o u r s e i s t h e m a t i c i t y . A theme i s a p r e s c r i b e d p a t t e r n ( r e p r e s e n t e d by a M e t a l i n g u a l c o n s t r u c t i o n ) t o which a d i s c o u r s e may conform. A theme may be ~' c o n t e n t i v e \" l i k e \" a c c i d e n t a l drowning\" t l which i s r e p r e s e n t e d a s e x i s t i n g when a person i s caused t o b e i n t h e w a t e r and i s unable t o a c t , t h e combination of which causes him t o drown\". T h i s i s c o n t e n t i v e s i q c e i t s p e c i f i e s both t h e p a r t s ( e v e n t s ) of t h e theme and t h e i r i n t e r r e l a t i o n s h i p s . A 'lnon-contentive\" theme i s one t h a t provides only a s t r u c t u r a l p a t t e r n o r t h e i n t e r r e l a t i o n s h i p s between some u n s p e c i f i e d e n t i t i e s \"clue\" i s defined t o be \"an unobserved act that causes the e x i s t e n c e of something\". The exact events and results are unspecified. A d i s c o u r s e i s thematic i f i t s propoaitiona can be matched t o a theme or h i e r a r c h y o f themes ( i . e . a theme which matches several p r o p o s i t i o n s i n a d i s c o u r s e , such as \"accidental drowning\", may i n turn be a part o f another theme, s u c h as \"tragedyft). P h i l l i p a t h e n d e f i n e s a coherent d i s c o u r s e as one which i s b o t h connected (each proposition i s connected t o a t l e a s t one other p r o p o s i t i o n ) a n d thematic. P h i l l i p s points out t h a t although t h e r e is no simple one-to-one mapping between h i s r e p r e s e n t a t i o n s and t h e c o n s t r u c t s used i n t e x t analyses like those of Grimes ( s e c t i o n 1.2.3) and Propp ( s e c t i o n 1. .2), his structures do provide f o r t h e r e p r e s e n t a t i o n of most of t h e proposed relationships, Some of Grimes r h e t o r i c a l p r e d i c a t e s , such as A t t r i b u t i o n , S p e c i f i c and C o l l e c t i o n , c o r r e s p o n d t o paradigmatic connectedness, while o t h e r s , such as Covariance, correspond t o d i s c u r s i v e copnectedness. S t i l l o t h e r r h e t o r i c a l p r e d i c a t e s , such a s Respqnse and Analogy and Propp's patterns of f u n c t i o n s i n a move (as w e l l a s t h e functions, themselves) correspond t o thematic s t r u c t u r e . Hence, P h i l l i p s claIms t o have presented computational i n t e r p r e t a t i o n s f o r t h e p r i n c i p l e t e x t phenomena." |
| }, |
| "FIGREF39": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "s t o f h i s proposed model of t e x t , P h i l l i p s p r e s e n t s a model of t e x t understanding embodied i n a computer program. H i s program i n p u t s a discourse i n t h e form of parse t r e e s of the sentences and b u i l d s t h e knowledge r e p r e s e n t a t i o n of t h e discourse. An i n t e r e s t i n g use of t h e Metalingual const m c t i on i s i t s function i n r e p l a c i n g non-cognitive s u r f a c e w o r d s by the Page 80 a p p r o p r i a t e c o g n i t i v e s t r u c t u r e ." |
| }, |
| "FIGREF40": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "o r example, t h e p r e p o s i t i o n \"through\" i s replaced by t h e c o g n i t i v e s t tucture f o r \"in-contact-wLth\" i n a s e n t e n c e l i k e :The Abominable Snwman walked t h r o u g h t h e snow.\"fhus, t h e same r e p r e s e n t a t i o n a l mechanism i s used f o r s u c h d i v e r g e n t t y p e s of knowledge a s s y n t a c t i c a l l y -r e l a t e d i n f o r m a t i o n and t h e m a t i c s t r u c t u r e s . Once t h e p a r s e t r e e h a s been c o n v e r t e d i n t o t h e knowledge r e p r e s e n t a t i o n of t h e p r o p o s i t i o n , t h e u n d e r s t a n d i n g p r o c e s s i n v o l v e s two major s t e p s . The f i r s t fs begun by matching t h e i n p u t p r o p o s i t i o n (IP) a g a i n s t t h e e n c y c l o p e d i a t o f i n d a c o r r e s p o n d i n g g e n e r a l i z e d p r o p o s i t i o n (GP). Thus, t h e IP The boat c o n t a i n s H o r a t i o Smith. i s matched t o the GP Things c o n t a i n peaple.N o t i c e t h a t P h i l l i p s u s e s GPs t o c a p t u r e t h e same knowledge t h a t f e a t u r e s and s e l e c t i o n a l r e s t r i c t i o n s c a p t u r e i n many o t h e r systems.So t h e c o r r e c t i n t e r p r e t a t i o n of a GP i s t h a t i t i s a p l a u s i b l e p r o p o s i t i o n , not a n e c e s s a r y t r u t h . Once t h e I P h a s been matched t o a GP, t h r e e a d d i t i o n a l checks are made s t a r t i n g from t h e matched GP. One is t o d e t e r m i n e i f any of th e terms o f t h e p r o p o s i t i o n have M e t a l i n g u a l d e f i n i t i o n s . I f s o , new knowledge c o r r e s p o n d i n & t o t h e d e f i n i t i o n i s added t o t h e d i s c o u r s e r e p r e s e n t a t i o n . F o r example, i f t h e d i s c o u r s e i n c l u d e d t h a t \"John was poisoned\" t h e e l a b o r a t e d i n f o r m a t i o n t h a t \"John i n g e s t e d something t h a t caused him t o be ill\" would be added. A s e c o n d check i s t o see i f t h e matched GP i s d i s c u r s i v e l y connected t o any Page 81 o t h e r GPs. I f s o , new k~lowledge i s added t o t h e d i s c o u r s e which c o r r e s p o n d s t o t h e r e l a t e d GP and t h e d$.ecursive connection. I f t h e rnatche'd GP were \"People are i n j u r e d \" and i t had a causal l i n k t o \"People are u n a b l e t o act\", t h e n t h e IP \"John was i n j u r e d \" would r e s u l t i n t h e a d d i t i o n oE \"John was u n a b l e t o a c t \" w i t h a c a u s a l l i n k from t h e o r i g i n a l I P , I t is by u s i n g t h e s e i n t e r r e l a t i o n s h i p s o f G P s t h a t P h i l l i p s a c c o m p l i s h e s i n f e r e n c e . The t h i r d t e s t i s t o d e t e r m i n e i f t h e matched GP i s a p a r t of any ' c o n t e n t i v e theme. (The e n c y c l o p e d i a ' s GPs have p o i n t e r s t o a l l c o n t e n t i v e themes which contain them.) I f s o t h e theme i s matched a g a i n s t t h e d i s c o u r s e as a whole t o see i f all components are p r e s e n t and a l l c o n s t r a i n t s s a t i s f i e d , I f t h e y are, the theme i s addtid t o t h e d i s c o u r s e rep r e s e n t a t i o n . \"Accidental drowning1' would be added t o a d i s c o u r s e r e p r e s e n t a t i o n which had matched the GPs and c o n n e c t i o n s I' (\"People contact-water\" and \"~e o p l e cannot a c t \" ) cause (\"People drown\") \" w i t h a c o r e f e r e n c e c o n s t r a i n t on \"people\", A f t e r p r o c e s s i n g Ips and a d d i n g t h e r e l a t e d s t r u c t u r e t o t h e d i s c o u r s e r e p r e s e n t a t i o n , t h e second m a j o r s t e p is t e s t i n g t h e d i s c o u r s e for coherence. T h i s i n v o l v e t w o tests. The f i r s t : d e t e r m i n e s i f t h e d i s c o u r s e i s connected. Then, t h e m a t i c i t y , i s tested by checking to see i f a s i n g l e undominated theme h a s been found. I t s h o u l d be n o t e d t h a t non-contenrive themes, s i n c e t h e y have no component a c t i o n s , and t h u s cannot be p o i n t e d t o by GPs, m u s t be tested f o r in a serial f a s h i o n . I f t h e discourse p a s s e s both tests, t h e n i t i s judged coherent. Page 82 To illustrate t h e p r o c e s s , a v e r y b r i e f d e s c r i p t i o n o f the u n d e r s t a n d i n g of a atory w i l l be given. Given the s t o r y : (IP1) A b o a t c o n t a i n s H o r a t i o Smith. ( I P 2 ) The b o a t o v e r t u r n s . ( I P 3 ) H o r a t i o Smith drowns, t h e f o l l o w i n g e v e n t s would o c c u r : 1, I P 1 and I P 2 would be matched t o GPs which a r e p a r t s of a complex e v e n t that has as a c a u s a l r e s u l t , when i n s t a n t i a t e d , t h e added p r o p o s i t i o n s : (AP1) H o r a t i o Smith is i n t h e water. (caused jointly by IPl and IP2) (AP2) H o r a t i o Smith i s injured, (caused j o i n t l y by I P 1 and I P 2 ) 2. The GP c o r r e s p o n d i n g t o AP2 r e s u l t s i n t h e a d d i t i o n o f : (AP3) H o r a t i o Smith cannot a c t -( c a u s e d by AP2) 3. The GPs c o r r e s p o n d i n g t o AP1 and AP3 are part o f a complex event that h a s the c a u s a l result \"People drownt', But this i s t h e GP which m a t c h e s IP3. S o a causal l i n k i s added from AP1 and AP3 t o IP3. 4. The d i s c o u r s e i s t e s t e d f o r c o n n e c t e d n e s s , and passes t h e t e s t . 5. The d i s c o u r s e matches t h e \"accidental drowning\" theme, which i s t h e only theme i t matches, s o i t i s t h e m a t i c , and t h u s c o h e r e n t , Page 83 P h i l l i p s p r i n c i p l e g o a l was a computational model of t e x t * His model of rate that the proposed r e p r e s e n t a t i o n s c o u l d a c t u a l l y be b u i l t from a n i n p u t text. His t e x t model does seem t o present a reasonable representation f o r a number of t e x t phenomena, s e v e r a l of which have not been considered by o t h e r computational models, Paradigmatic c o n n e c t e d n e s s and some t y p e s of thematic structure seem p a r t i c u l a r l y important. Iiowever, s e v e r a l o b j e c t i o n s must be r a i s e d t o his model of t e x t underslanding. The f i r s t o b j e c t i o n i s t h a t his t e s t a for coherence a r e a p p l i e d o n l y t o the complete d i s c o u r s e , and are not formulated i n such a way as t o suggest s t r a t e g i e s t o avoid i n c o r r e c t interpretation of p r o p o s i t i o n s pnd r e l a t i o n s h i p s , I f t h i s were done, t h e s e q u e n t i a l processing of input p t o p o s i t i o n s would c o n t i n u a l l y test f o r coherence, and incoherence would immediately suggest t h a t some m i s i n t e r p r e t a t i o n might have been made, o r some inference omitted. A second o b j e c t i o n i& t h a t t h e d e f i n i t i o n s of connectedness and thematici t y are inadequate. A common s u p e r o r d i n a t e node i s simply not s u f f i c i e n t t o e x p l a i n paradigmatic connectedness. The example of sentences about l i o n s and t i g e r s would be connected only i n some context t h a t explained why t h e s e statements were being made (e, g. \"All* I know about lions and t i g e r s is . . .\"). T h e m a t i c i t y i s defined without respect t o how much of t h e discourse the theme accounts for. A theme cannot account f o r a l l of t h e p r o p o s i t i o n s in a discourse u n l e s s t h e d i s c o u r s e i s very t r i v i a l , and y e t i f a theme matched o n l y t h e f i r s t three propositionS of a one hundred p r o p o s i t i o n d i s c o u r s e , it could h a r d l y be c a l l e d t h e theme of the discourse." |
| }, |
| "FIGREF41": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "number of aspects of t h e understanding processes a r e not convincingly shown t o b e computationally p r a c t i c a l . The problem of avoiding i n c o r r e c t a d d i t i o n s t o t h e d i s c o u r s e r e p r e s e n t a t i o n from a r i c h e n c y c l o p e d i a i s ignored. Each e x a w l e h a s e x a c t l y t h e r i g h t i n f a m a t i o n a v a i l a b l e . Also, i t ie mot c l e a r that: t h e use of g e n e r a l i z e d p r o p o s i t i o n s would work when t h e r e a m ~averrrl .levels of g e n e r a l i z a t i o n p o s s i b l e . F o r example, t h e g p e r e l i z e d p r o p o s i t i o n \"People c o n t a c t w a t e r \" h a s a s s o c i a t e d knowledge, b u t t h e h i g h e r l e v e l p r o p o s i t i o n \"People c o n t a c t t h i n g s t t would a l s o need t o be p r e s e n t w i t h i t s a s s o c i a t e d knowledge. And f i n a l l y , t h e methods of a c c e s s i n g themesby p o i n t e r s t o a l l o c c u r r e n c e s of g e n e r a l i z e d p r o p o s i t i o n s o r by s e r i a l s e a r c hb o t h seem comput a t i o n a l l y unaccep t a b l e . P r o p o s i t i o n s l i k e \"~e o p l e go\" would r e s u l t i n , a c o m b i n a t o r i c e x p l o s i o n of p o s s i b i l i t i e s ." |
| }, |
| "FIGREF42": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "o r a l l o f these r e a s o n s , P h i l l i p s ' u n d e r s t a n d i n g model i s u s e f u l more i n s u g g e s t i n g t h e k i n d s of problems i n v o l v e d , t h a n i n p r o v i d i n g a n a c t u a l model of t e x t understanding." |
| }, |
| "FIGREF43": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "Space l i m i t a t i o n s and t h e narrowly d e f i n e d s c o p e of t h i s sunrey have combined t o e l i m i n a t e c e r t a i n i n t e r e s t i n g work from d e t a i l e d c o n s i d e r a t i o n . W i l k s [ 7 5 , 76, 771 h a s d e s c r i b e d a t e x t u n d e r s t a n d i n g (and t r a n s l a t i o n ) system which u s e s a meaning r e p r e s e n t a t i o n c a l l e d P r e f e r e n c e Semantics. The system n o r m a l l y opercltes i n a b a s i c mode, d e a l i n g with s e n t e n c e s i n d i v i d u a l l y , b u t i s c a p a b l e of e n t e r i n g an extended mode when a r e f e r e n c e problem occurs. An example i s \" i t \" i n J o h n d r a n k t h e whiskey from t h e g l a s s andi t f e l t warm i n h i 3 stoomch. S e v e r a l p r o c e s s e s are i n v o l v e d i n t h e a t t e m p t t o d e t e r m i n e t h e c o r r e c t r e f e r e n t . \"Ext r a c t i o n l ' i s t h e a d d i t i o n of l o g i c a l l y t r u e p r o p o s i t i o n s , only A t t h e l e v e i of t e x t , Wilks d i f f e r s from R i e g e r and Charniak p r i m a r i l y i nhis i n s i s t e n c e upun e n t e r i n g a n extended made o n l y when a problem r e q u i r e s i t .R e f e r e n c e p toblems are t h e only problems he d i s c u s s e s a s t r i g g e r i n g t h i a mode H i s u s e of pseudo-text B i n e s t a b l i s h i n g t e x t u a l c o n n e c t i o n s i s s t i l l t o o b r i e f l y d e s c r i b e d t o c r i t i c a l l y e v a l u a t e . R i e g e r ~( [ 7 5 b , 75c, 76a, 76b1, Rieger and G r i n b e r g [771) h a s d e s c r i b e d a complex system f o r t h e o r g a n i z e d r e p r e s e n t a t i o n of cause and e f f e c t k n w l e d g e ,.a& p l a n s u t i l i z i n g t h i s knowledge.A set of d e c i s i o n trees (\"selection 11 networks\") a r e p o s t u l a t e d which. g i v e n a goal, s e l e c t c e r t a i n common s e n s e a l g o r i t h m s \" capable of r e a l i z i n g t h a t goal. R i e g e r asserts t h a t the same knowledge s t -r u c t u r e s used f o r p l a n n i n g , should b e used f o r u n d e r s t a n d i n g t h e i n t e n t i o n a l a c t s of o t h e r s , To do t h i s i n t e x t , h e r e q u i r e s k n w l e d g e which allows p r e d i c t i o n of g o a l s and a c t i o n s l i k e l y t o be made i n response t o t h e o c c u r r e n c e of some e v e n t o r s t a t e , A subsequent s e n t e n c e i s t e s t e d t o see i f i t confirms any p r e d i c t i o n by being a s t e p i n an expected a c t i o n o r i n an a l g o r i t h m t o achieve a n expected g o a l , T h i s t e s t i n g r e q u i r e s t h a t a l l a l g o r i t h m s be indexed by s t e p . Thus, i t must b~ p o s s i b l e t o f i n d a l l o c c u r r e n c e s of (X GOT0 Y) i n some s e t of a l g o r i t h m s , A c o n f i r m a t i o n i s used t o morecoslfidently p r e d i c t the c o u r s e of a c t i o n b e i n g followed, C l e a r l y , Rieger i s d e a l i n g with the same problems a s Schank and Abelson ( s e c t i o n 3,2.1.2), b u t ha.s not y e t c l e a r l y demonstrated t h e u t i l i t y o r p r a c t i c a l i t y of h i s approach t o t h i s aspect of t e x t understanding. Page 87 Hobbs 176, 771 has d i s c u s s e d a n a l y s e s of v a r i o u s t e x t s based on h i s own s y s t e m of semantics. P o s s i b l e f n t e r s e n t e n t i a l r e l a t i o n s are d e s c r i b e d by p a t t e r na c t i o n p a i r s which, when matched by i n p u t s e n t e n c e s modify t h e t e x t r e p r e s e n t a t i o n tree apprap r i a t e l y .These r e l a t i o n s i n c l u d e causality, time o r d e r i n g , p a r a p h r a s e , examp 1.2, c o n t r a s t , p a r a l l e l c o n s t . r u c t i o n and v l o l a t e d e x p e c t a t i o n . These r e l a t i o n s h i p 8 i n d i c a t e t h a t Hobbs does n o t make a d i s t i n c t i o n between c o n t e n t and form ( a u t h o r imposed) r e l a t i o n s . Thc complexity o f Hobbs' system ( i n t h e l a r g e number of r u l e s and t h e i r i n t e r a c t i o n s ) makes e v a l u a t i o n d i f f i c u l t until h e h a s completed his computer implementation. O t h e r work t h a t s h o u l d be mentioned i n c l u d e s Schmidt ([761, Schmidt and Sridharan [77]), who d i s c u s s e s t h e problem of recognitioh of p l a n s from a c t i o n s , as w e l l as Novak [76] and Bobrow, e t a 1 [77] who both u s e frame l i k e knowledge s t r u c t u r e s i n s p e c i a l i z e d language u n d e r s t a n d i rag systems. 3.4 D i s c u s s i o n And Conclusions What have the discrissed c o m p u t a t i o n a l models added t o t h e p r e v i o u s l y described model of t e x t u n d e r s t a n d i n g ? Charniak, Riegex and W i l k s have demonstrated h o w -i t i s possible t o infer i n f o r m a t i o n o n l y i m p l i c i t i n a t e x t . Schank h a s e s p e c i a l l y examined t h e r i c h n e s s and complexity of c a u s a l c o n n e c t i o n s , which a r e often i m p l i c i t . T h i s c a p a b i l i t y raises many a d d i t i o n a l q u e s t i o n s , however, i n c l u d i n g when, and how many, i n f e r e n c e s s h o u l d be made. Whi le Cllarniak and Rieger s u g g e s t making many i n f e rences whencvv r p o s s i b l e , Wilks argues f o r making them only a s required. N e i t h e r a p p r o a c h h a s y e t b e e n a p p l i e d t o s u f f i c i e n t ly 1 argc t r x t s and k n w l e d g e bases t o convincingly demonstrate i t s v a l i d i t y . Pre-existcnt k n~w l e d g r s t r u c t u r e s h a v e been s u g g e s t e d by Schaak and Philips t o avoid t h e extrc:mc8ly d i f f i c u l t problem of making matiy-step inference chains to e s t a b l i s h i n l p l i c i t c o n~~c c t ions. Although use o f these s t r u c t u r e s has d e m o n s t r a t e d t h e i r a b i l i t y t o meet t h i s goal, as w e l l a s t h e i r u s e f u l n e s s i n generation of summaries a n d p a r a p h r a s e s , problems have appeared. Complex knowledge s t r u c t u r e s and r e a l t e x t s p r e s e n t many d i f f i c u l t . k e a la matching an input p r o y o s l t l o n w i t h an element in the s t r u c t u r e . When t h e match is i m p e r f e c t , orwhen many choices a r e (computationally) p o s g i b l e , i t i s very d i f f i c u l t t o perform t h e required matching correctly, I n a n attempt t o d e a l w i t h this d i f f i c u l t y (as w e l l a s others), and t o r e c o g n i z e t h e f a c t t h a t n o v e l s i t u a t i o n s are a l s o unde rstandable, Schank, Abelson, R i e g e r a n d S c h m i d t have been c o n c e r n e d with r e c o g n i z i n g the m o t i v a t i o n s and intentions o f a c t o r s . T h i s aspect of u n d e r s t a n d i n g i s c l e a r l y i m p o r t a n t , f o r t h e s t a t e d r t 3 a s o n s , b u t t h e approaches d e s c r i b e d have almost c e r t a i n l y b e e n t o o s i m p l i f i c d f o r the u n d e r s t a n d i n g ,of a c t u a l t e x t s . I t Thus *the requi rcmcnt t h a t a n u n d e r s t a r~d i n g of a t e x t c o n t a i n s as much i n f e r r e d i n f o r n i a t~o~~ a s h~! c~s s~i r y t o mqet some m i n i m a l l e v e l of ... coherence\" (section 1 . 4 ) i s w e n t o bc a c o m p u t a t i o n a l l y d i f f i c u l t problem. T h e orgat-lizing c o r~t t~i~t S I 1 1 1 , t t~r l~~; a r c i~s c~f u l i n d e a l i n g w i t h t h i s problem, b u t c r e a t e new problems. Knocll c-dgc) nrltl rr~p r e s r~n t a t i o n of p l a n s and intentions i s introduced t n hnndalr* sor.\\c3 of t h c s r problems. T t s h o u l d be noted t h a t t h e p r e v i o u s l y d j sc~~r;sotl n r o t l c~l t l~~ n o t s p c r l f i c , i l l y d i s t i n g u i s h e d t h i s t y p e of Page 89 knowledge. Coherence has been defined i n tllcse computational models as connectivity e s t a b l i s h e d by inferrencing, matching a knowledge structure or being a step in a plan. The exact d i s t i n c t i o n s between these three types of knowledge are s t i l l blurred, and i t remains t o be demonstrated that these d i s t i n c t i o n s are appropriate and adequate. Only P h i l l i p s and Hobbs have addressed representing the form of t e x t s , bqt these suggestions have not beeh computationally adequate, nor has there been any clear d i s t i n c t i o n between form and content structures. A better defined notion of t e x t form should certainly play a role in the ongoing determinatiion of t e x t u a l coherence t h a t occurs during comp rehension in the t e x t understanding model. With regard t o f o r g e t t i n g , l i t t l e has been added, although h i g h e r l e v e l knowledge structures could be used t o summarile t h e i r more d e t a i l e d components (which could then be \"forgotted'). Schank has a l s o suggested that the least connected propositions i n a representation are t h o s e most likely t o be forgotten. However, these ideas have not been s e r i o u s l y investigated. I n general, computational models understand t e x t s p e r f e c t l y ( i f a t a l l ) , and do not contain any imperfect ret rieval processes. Permanent learning and integration with p r i o r knowledge has not been investigated in these systems, nor has any explanation been offered f o r o c c a s i o n a l r e c a l l of surface t e x t , --These systems generally d o not keep the surface t e x t a t a l l , b u t c o u l d e a s i l y do sv. However, i t would result i n p e r f e c t r e c a l l o f t h i s information. Page 90 The p r i n c i p a l v a l u e of c o m p u t a t i o n a l models has been t h e d e m o n s t r a t i o n of t h e g r e a t d i f f i c u l t y i n a c t u a l l y s p e c i f y i n g comprehension a l g o r i t h m s . I t i s all t o o e a s y , when e x p l a i n i n g h w a p a r t i c u l a r r e s u l t i s a c h i e v e d , t o i g n o r e t h e problem o f a v o i d i n g i n c o r r e c t p a t h s . The f a c t t h a t h i g h e r -l e v e l content st ructures are c o m p u t a t i o n a l l y u a e f u l , and also p s y c h o l o g i c a l l y i n d i c a t e d , i s a n i m p o r t a n t c o n f i r m a t i o n . The i n v e s t i g a t i o n s of more, a c t u a l texts by systems h a v i n g more knowledge available ( n o t j u s t the relevant knowledge) i san i m p o r t a n t s t e p i n validating t h e models b e i n g proposed. l l o w i n g p o i n t s are a c c e p t e d by most researchers concerned with t e x t u n d e r s t a n d i n g :1. Much i n f o r m a t i o n i m p l i c i t i n t e x t s i s explicit i n an u n d e r s t a n d i n g of t h a t t e x t ." |
| }, |
| "FIGREF44": { |
| "num": null, |
| "uris": null, |
| "type_str": "figure", |
| "text": "The s t u d y of knowledge r e p r e s c n t a t i o n , a n d t h e n a t u r e of knowledge s t r u c t u r e s i s a primary concern. Many computational problems have n o t been resolved. A p a r t i c u l a r l y important q u e s t i o n f o r text u n d e r s t a n d i n g c o n c e r n s t h e n e c e s s i t y of redundantly copying g e n e r a l knowledge f o r s p e c i f i c i n s t a n t i a t i o n s . Fahlman(773 d i s c u s s e s t h i s problem i n d e t a i l , and o f f e r s some ways i n which t o a v o i d unnecessary redundancy. The r e l a t i o n s h i p of t h e c o n t e n t and t h e form o f t e x t s n e e d s a d d i t i o n a l c l a r i f i c a t i o n . S i m i l a r l y , t h e r e l a t i o n s h i p o f g e n e r a l c o n t e n t s t r u c t u r e s t o r e p r e s e n t a t i o n s f o r p l a n s and i n t e n t i o n s n e e d s study t o see how d i s t i n c t t h e s e are. And of course, t h e large a r e a s of l e a r n i n g and f o r g e t t i n g a r e i m p o r t a h t , b u t m i s s i n g , components of a t e x t u n d e r s t a n d i n g model. I n a d d i t i o n t o t h e s e open problems, a t t e m p t s t o a p p l y t h e combined insights of t h e d i v e r s e r e s e a r c h p e r s p e c t i v e s t o a c t u a l t e x t s i s a n e c e s s a r y s t e p i n e v a l u a t i n g t h e adequacy o f c u r r e n t t e x t u n d e r s t a n d i n g models. P e t o f i , J. and Rieser,H.-(eds.) 1973, S t u d i e s -i n Text Grammar. Dordrecht: \"Reidel. P h i l l i p s , B. 1975a. Topic -Analysis. unpublished PhD d i s s e r t a t i o n , S t a t e U n i v e r s i t y of New York a t Buffalo. P h i l l i p s , B. 1975b. Judging t h e Coherence of Discourse. i n P r o~e e d i n~s of t h e 13th Annual M e c t i n~ o f t h e A s s o c i a t i o n f o r Computational L j n g u i s t i c s . p~, V. 1968. Morphologycf ---t h e F o l k t a l e . -Austin: U n i v e r s i t y of Texas Press. Rieger, C. J. 1974. Conccjt u a l Memory. u~~p u b l i s h e d PhD d i s s e r t a t i o n , S t a n f o r d U n i v e r s i t y . e r p r e t a t i o n of Sentence Meaning i n Context* i n Advance Papers -of t h e Fourth --I n t e r n a t i o n a l J o i n t conference -on A r t i f i c i a l I n t e l l i g e n c e . R i e g e r , C. J -1 9 7 5~. The Commonsense Algorithm as a Basis f o r Computer Models of Human Memory, I n f e r e n c e , B e l i e f and Contextual Language Comprehension. i n Schank and Nash-Webber [75] . R i e g e r , C. J o 1976a. An O r g a n i z a t i o n of Knowledge f o r Problem Solving and Language Comp rehension. A r t i f i c i a l I n t e l l i g e n c e 7, 2, pp. 89Computation i n C o g n i t i v e Models. T e c h n i c a l r e p o r t no. 459, Department of ~o m p u t e r~c i e n c e U n i v e r s i t y of Maryland. R i e g e r , C. J. and Grinberg, M e 1977. The D e c l a r a t i v e R e p r e s e n t a t i o n and P rocedu raf. Sinlulation a \u00a3 C a u s a l i t y i n P h y s i c a l Mechanisms. i n Proceedings--oft h e F i f t h I n t e r n a t i o n a l J o i n t Conferenceon A r t i f i c i Eo 1975. Notes on a Schema f o r StorPes. i n Bobrow and C o l l i n s [75]. Rumelhart, D. Eo and Ortony, A. 1977. The R e p r e s e n t a t i o n of Knowledge i n Memory. i n Anderson, S p i r o and Montague [77]. S c a l l e r t , D o L. 1976. Improving Memory f o r P m s e : The R e l a t i o n s h i p Between Depth of P r o c e s s i n g and Context. J o u r n a lof Verbal Learning and Verbal Behavior, 1 5 , 6 , pp, 621-632. Schank R. 1973a. I d e n t i f i c a t i o n of C o n c e p t u a l i z a t i o n s Underlying N a t u r t i t u t o p e r g l i S t u d i Semantici e C o g n i t i v i , Castagnola, Switzerland* 1975b. The S t r u c t u r e of Episodes i n Memory. i n Bohrow and C o l l i n s [75] . Schank, R o and Abelson, R , 1975. S c r i p t s , P l a n s and Knowledge. i n Advance Papers of t h e Fourth I n t e r n a t i o n a l ---J o i n t ----Conferenceon A r t i f i c i a l I n t e l l i p , e n c e ." |
| } |
| } |
| } |
| } |
