- Email: [email protected]
Toward a Model of Macrostructure Search
DfSCOURSE PROCESSING A . Hammer and W. Kintsch (eds.) 0North-Holland Publishing Company, 1982
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
Ulrich Glowalla & Hans Colonius I n s t i t u t f u r Psychologie Technische Universitat Braunschweig Federal Republic of Germany I t has been argued t h a t t h e macrostructural organization of a t e x t i s c e n t r a l t o comprehension and r e t e n t i o n of discourse. I n a recogn i t i o n experiment, some aspects of t h e macros t r u c t u r a l representation have been inve,stigated. The r e s u l t s give some preliminary support t o the Roter Faden model (Glowalla, 1981). I n a d d i t i o n , a q u a n t i t a t i v e model capturing some f e a t u r e s of macrostructure search i s outlined. INTRODUCTION
I t has been s t r e s s e d by a number of people t h a t t h e process of condensing the f u l l meaning of a t e x t i n t o i t s g i s t i s central t o t h e comprehension and r e t e n t i o n of discourse ( e . g . , Kintsch and van Dijk, 1978). The main goal of t h i s paper i s t o i n v e s t i g a t e t h i s process f u r t h e r . I t has become convenient t o use the term macrostructure f o r such global descriptions of the semantic s t r u c t u r e o f a t e x t . In c o n t r a s t , t h e term microstructure has been introduced f o r t h e s t r u c t u r e of individual t e x t sentences and t h e i r r e l a t i o n s . To consider a piece of discourse coherent, two conditions must h o l d : i t s respective sentences should be connected a t t h e microstructure level and these sentences should be organized in l a r g e r conceptual u n i t s a t the macrostructure l e v e l . T h u s , a s has been argued elsewhere ( e . g . , Sanford and Garrod, 1981), only models f o r t e x t processing incorporating b o t h of these aspects may a t t a i n t h e o r e t i c a l a s well a s empirical signif icance. I n f a c t , q u i t e a few models f o r t e x t processing incorporating micro- and macro-level d e s c r i p t i o n s have already been proposed (Glowalla, 1981; Graess e r , 1981; Kintsch and van Dijk, 1978; Schank a n d Abelson, 1977). However, the models d i f f e r with respect t o t h e format of t h e microstructure a n d / o r the macrostructure. As t o t h e l a t t e r , several conceptions a r e under d i s cussion. The e x i s t i n g a l t e r n a t i v e s may roughly be c l a s s i f i e d as follows: The micropropositions of a t e x t a r e transformed i n t o a set of macropropos i t i o n s by so-called macro-operators. Recursive a p p l i c a t i o n o f these macrooperators leads t o macrostructural descriptions a t d i f f e r e n t hierarchical l e v e l s . The r e s u l t i n g macrostructure represents t h e g i s t of t h e t e x t . This concept of macrostructure has been developed by Kintsch and van Dijk (1978). A1 t e r n a t i v e l y , t h e macrostructure o f a t e x t may be conceived of as a hierarchical tree s t r u c t u r e consisting of nodes representing conceptual u n i t s of a t e x t and l i n k s representing the r e l a t i o n s between conceptual u n i t s . The microstructure u n i t s of a t e x t , i . e . , i t s micropropositions, a r e 111
112
COHERENCE
represented by t h e terminal nodes of t h e t r e e s t r u c t u r e . Internal nodes represent conceptual u n i t s of t h e macrostructure on d i f f e r e n t hierarchical l e v e l s . No s p e c i f i c assumptions a r e made about t h e contents represented by the node. Distances between microstructure u n i t s in t h e t r e e s t r u c t u r e representing t h e macrostructure a r e viewed a s a measure o f t h e i r connectivity, i . e . , whether they belong t o t h e same or d i f f e r e n t macrostructural u n i t s . As a matter of f a c t , s t r u c t u r e diagrams defining t h e c o n s t i t u e n t s t r u c t u r e of a piece of discourse ( e . g . , Rumelhart, 1977) may be i n t e r p r e t e d in t h i s way. This view does n o t imply t h a t people cannot construct macropropos i t i o n s out of a s e t of connected micropropositions, i . e . , t h a t they a r e unable t o produce a summary of t h e main contents of a t e x t . Again, Rumelh a r t (1977) has been a b l e t o p r e d i c t observed summaries q u i t e nicely by postulating additional summarization r u l e s . The only proposition made here i s t h a t we may know too l i t t l e about t h e cognitive operations producing t h e t h e o r e t i c a l macrostructure. We should t h e r e f o r e avoid making too many i n i t i a l assumptions. Moreover, with respect t o t h e memory representation of discourse, assumptions about t h e contents of macrostructure nodes may even prove t o be unnecessary.
I n t h e following we shall present some experimental r e s u l t s about search processes on macrostructures. The data o r i g i n a t e from an experiment t o t e s t some predictions of a recently proposed model f o r t e x t processing called t h e "Roter Faden" (main t r a i n of thought; Glowalla, 1981). I n the subsequent section we shall give an o u t l i n e of a q u a n t i t a t i v e model. A s what follows will be presented in the t h e o r e t i c a l framework of t h e Roter Faden model, we comnence with a short survey of t h e main assumptions of t h i s model. THE ROTER FADEN MODEL
The Roter Faden model has been designed t o capture comprehension and retention of discourse, in which purposive action sequences of human beings a r e described. The model c o n s i s t s of an i n t e r a c t i v e processing system based o n s p e c i f i c assumptions about t h e processing devices involved in the comprehension of discourse, together with a s e t of knowledge s t r u c t u r e s necessary t o f u l f i l the t a s k . T h e model s t a r t s off with t h e question how incoming t e x t information has t o be processed t o produce a memory representation of i t . Clauses a r e taken a s u n i t s in t h i s a n a l y s i s ; a clause i s defined here t o be a proposition containing an a c t i v e o r s t a t i v e verb. A t each processing s t e p , t h e model t r i e s t o find answers t o t h e following questions: 1. What i s t h e functional meaning of an incoming clause? ( A functiona l l y i d e n t i f i e d c l a u s e will be termed a microstructure u n i t . ) 2 . How may t h e present microstructure u n i t be connected t o already encoded u n i t s ? 3 . Which inferences have t o be generated, i f no d i r e c t connection e x i s t s between two successive u n i t s ? 4 . Which expectations may be generated about t h e t e x t information y e t t o come? 5. I s i t possible t o condense already encoded microstructure u n i t s into more global macrostructure u n i t s ? The i n t e r a c t i v e processing system produces answers t o these questions u t i l i z i n g t h e assumed knowledge s t r u c t u r e s . B o t h of these components of t h e Roter Faden model will be outlined in turn. The processing system c o n s i s t s of a c e n t r a l processor c a l l e d CONSTRUCT and four subprocessors termed IOENT, INFER, MACRO, and ERWIN. The central processor executes t h e various
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
113
processing o p e r a t i o n s i n a c t u a l f a c t , u t i l i z i n g t h e i n f o r m a t i o n produced by t h e subprocessors. I t i s assumed t h a t t h e s e subprocessors work independentl y o f each o t h e r and communicate e x c l u s i v e l y v i a t h e c e n t r a l p r o c e s s o r . A l l processing d e v i c e s d i f f e r w i t h r e s p e c t t o two p r o p e r t i e s , namely t h e i r t a s k and t h e i r a b i l i t y t o access t h e assumed knowledge s t r u c t u r e s . The c e n t r a l processor CONSTRUCT a c t u a l l y produces t h e memory r e p r e s e n t a t i o n o f a t e x t by p r o c e s s i n g i t s i n f o r m a t i o n c l a u s e - b y - c l a u s e . Subprocessor I D E N T t r i e s t o i d e n t i f y t h e f u n c t i o n a l meaning o f each incoming c l a u s e . D u r i n g t h i s p r o cess e x p e c t a t i o n s a r e u t i l i z e d t h a t have been d e r i v e d f r o m t e x t i n f o r m a t i o n a l r e a d y encoded. I f no match between t h e p r e s e n t c l a u s e and one o f these e x p e c t a t i o n s i s found, subprocessor INFER becomes a c t i v a t e d . INFER, then, generates c h a i n s c o n s i s t i n g o f one o r more i n f e r e n c e s i n o r d e r t o connect t h e new p i e c e o f i n f o r m a t i o n t o t h e s t r u c t u r e a l r e a d y e x i s t i n g . A t each p r o c e s s i n g s t e p , subprocessor MACRO checks f o r t h e p o s s i b i l i t y o f condensing a s e t o f c o h e r e n t m i c r o s t r u c t u r a l u n i t s i n t o more g l o b a l macros t r u c t u r a l u n i t s and ERWIN, f i n a l l y , generates e x p e c t a t i o n s a b o u t t h e t e x t i n f o r m a t i o n y e t t o come. Up t o now, a l l p r o c e s s i n g d e v i c e s a r e r e a l i z e d by d e t a i l e d f l o w c h a r t s m o d e l l i n g t h e comprehension o f t e x t i n f o r m a t i o n a c c o r d i n g t o t h e R o t e r Faden model. Moreover, t h e p r o c e s s i n g d e v i c e s a r e p u r e l y f u n c t i o n a l i n nature. Thus, s p e c i f i c w o r l d knowledge i s n o t c o n s i d e r e d , f o r example, whether a s p e c i f i c method encountered by t h e main c h a r a c t e r o f an a c t i o n sequence i s s u i t a b l e t o r e a c h a c e r t a i n g o a l . I n a d d i t i o n , no c a p a c i t y l i m i t a t i o n s o f w o r k i n g memory a r e c o n s i d e r e d , as has r e c e n t l y been done by K i n t s c h and van D i j k (1978). However, t o implement t h e p r o c e s s i n g model as a computer program, assumptions a b o u t b o t h o f t h e s e a s p e c t s would have t o be i n c o r p o r a t e d . N e v e r t h e l e s s , by g i v i n g a d e t a i l e d account o f t h e opera t i o n s performed d u r i n g t h e n e x t comprehension process, t h e p r o c e s s i n g dev i c e s a l l o w f o r s u c c e s s i v e e x p e r i m e n t a l t e s t i n g . I n a c t u a l f a c t , one i m p o r t a n t f e a t u r e o f t h e model, namely t h e process o f g e n e r a t i n g e x p e c t a t i o n s has a l r e a d y been i n v e s t i g a t e d and l e d t o s u p p o r t i v e e v i d e n c e f o r t h e Roter Faden model ( P o h l , t h i s volume). D u r i n g t h e comprehension process t h e p r o c e s s i n g system i s assumed t o have access t o d i f f e r e n t know1 edge s t r u c t u r e . These knowledge s t r u c t u r e s cons i s t o f t h r e e a c t i o n schemata and a s e t o f t e n c o n f i g u r a l r u l e s . The f o r mer s p e c i f y o u r a b s t r a c t knowledge about t h e t y p i c a l s t r u c t u r e o f a c t i o n sequences and t h e l a t t e r d e f i n e o u r knowledge about p e r m i t t e d d e v i a t i o n s from t h e t y p i c a l s t r u c t u r e . To a p p l y t h e s e knowledge s t r u c t u r e s t o t h e comprehension process, i t must be shown which u n i t s o f a t e x t may be ident i f i e d w i t h u n i t s o f t h e knowledge s t r u c t u r e s . I n t h e R o t e r Faden model i t i s assumed t h a t c l a u s e s t a k e n as u n i t s o f t e x t i n f o r m a t i o n belong t o e i t h e r o f t h r e e b a s i c c a t e g o r i e s . These c a t e g o r i e s correspond t o t h e d i f f e r e n t f u n c t i o n a l meanings o f c e r t a i n c l a u s e s w i t h r e s p e c t t o t h e s t r u c t u r a l p r o p e r t i e s o f a t e x t . The c a t e g o r i e s a r e termed I n t e n t i o n ( g o a l s , conceptions, wishes), A c t i o n ( s t a t e changes), and S t a t e ( c o n d i t i o n s o f t h e w o r l d ) . I t i s assumed t h a t t h e f u n c t i o n a l meaning o f a c l a u s e may be d e r i v e d from t h e semantic i m p o r t o f i t unambiguosly and i n d e p e n d e n t l y o f s u r r o u n d i n g c l a u ses. M i c r o s t r u c t u r a l u n i t s b e l o n g i n g t o t h e same c a t e g o r y w i l l be f u r t h e r d i f f e r e n t i a t e d w i t h r e s p e c t t o t h e a c t i o n schema o f which t h e y a r e a cons t i t u e n t and w i t h r e s p e c t t o whether t h e main c h a r a c t e r i s i n v o l v e d . These s u b c a t e g o r i e s a r e d e t a i l e d below: 1. I n t e n t i o n 1.1 Goal ( d e s i r e f o r t h e o c c u r r e n c e o f a c e r t a i n e v e n t o r s t a t e ) 1 . 2 Conception ( f o r m a t i o n o f a PLAN) 1.3 S e l e c t ( s e l e c t i o n o f a c e r t a i n method)
114
COHERENCE
2. A c t i o n 2.1 Do ( a p p l i c a t i o n o f a s e l e c t e d method) 2.2 Event o r O b s t a c l e (an o b j e c t i v e o c c u r r e n c e o r an a c t i o n by another character) 3. S t a t e 3 . 1 p o s i t i v e o r n e g a t i v e Outcome ( o f an ACTION) 3.2 p o s i t i v e o r n e g a t i v e Consequence ( o f a PLAN o r an EPISODE) 3.3 S e t t i n g o r O b s t a c l e (a c o n d i t i o n o f t h e w o r l d e x i s t i n g indep e n d e n t l y o f a main c h a r a c t e r ' s ACTION) A l l I n t e n t i o n - s u b c a t e g o r i e s , t h e ACTION-subcategory Do, and t h e S t a t e s u b c a t e g o r i e s Outcome and Consequence i n v o l v e t h e main c h a r a c t e r . A l l these s u b c a t e g o r i e s form a c o n s t i t u e n t o f one o f t h e a c t i o n schemata t h a t w i l l be d e s c r i b e d i n t h e n e x t s e c t i o n . Whether an Outcome o r a Consequence i s pos i t i v e o r n e g a t i v e and whether an Event o r a S e t t i n g i s an O b s t a c l e i s det e r m i n e d r e l a t i v e t o t h e c u r r e n t I n t e n t i o n o f t h e main c h a r a c t e r . The a c t i o n schemata assumed by t h e R o t e r Faden model a r e termed EPISODE, PLAN, and ACTION; t h e y a r e g i v e n i n t h e f o r m o f r e w r i t e r u l e s :
+
[~~~~~N,i 1 +
1. EPISODE
=
Goal
2. PLAN
=
Conception +
Consequence
p ACTIONS
and / o r q PLANS
+
Consequence
+
( w i t h (p+q)>2; p.qelNo) 3. ACTION
= Select
+
Do + Outcome
Note t h a t t h e s e schemata a l l o w f o r r e c u r s i v e a p p l i c a t i o n . The a c t i o n schemata express o u r a b s t r a c t knowledge a b o u t t h e p u r p o s i v e b e h a v i o r of human beings. They a r e assumed t o g u i d e b e h a v i o r as w e l l as t h e comprehension of d i s c o u r s e t h a t d e s c r i b e s g o a l - d i r e c t e d human a c t i v i t i e s . Note t h a t t h e cons t i t u e n t s o f t h e s e schemata c o v e r most o f t h e m i c r o s t r u c t u r a l u n i t s i n t r o duced i n t h e p r e v i o u s s e c t i o n ; o n l y S e t t i n g , Event, and O b s t a c l e do n o t o c c u r and an e v a l u a t i o n o f u n i t s Outcome and Consequence i s m i s s i n g . T h i s i s m o t i v a t e d by t h e f a c t t h a t t h e s e elements a r e n o t c o n s i d e r e d p r o p e r t i e s o f o u r a b s t r a c t knowledge a b o u t t h e t y p i c a l s t r u c t u r e o f a c t i o n sequences. N e v e r t h e l e s s , one can say t h a t access t o t h e s e schemata would g r e a t l y fac i l i a t e t h e t e x t comprehension process: T h i s would a l l o w f o r g e n e r a t i n g i n f e r e n c e s about i n f o r m a t i o n n o t e x p l i c i t l y s t a t e d i n a t e x t as w e l l as e x p e c t a t i o n s a b o u t t e x t i n f o r m a t i o n y e t t o come. Moreover, s i n g l e m i c r o s t r u c t u r a l u n i t s may be conceived o f as b e l o n g i n g t o l a r g e r s t r u c t u r a l u n i t s o f a t e x t and m i g h t t h e r e f o r e be summarized a c c o r d i n g t o t h e s e a c t i o n schemata. T h u s , t h e a c t i o n schemata correspond t o t h e m a c r o s t r u c t u r a l u n i t s o f t h e R o t e r Faden model. There i s a problem, however: The a c t i o n sequences r e a l i z e d i n many pieces o f d i s c o u r s e d i s p l a y c o n s i d e r a b l e d e v i a t i o n s from t h e t y p i c a l f o r m expressed by t h e schemata. C e r t a i n s t a t e s o f t h e w o r l d o r events may d e t a i n t h e main c h a r a c t e r from p u r s u i n g h i s g o a l s . Moreover, t h e a c t i o n schemata do n o t c o n t a i n any i n f o r m a t i o n a b o u t t h e p o s s i b l e c o n n e c t i o n s between s u c c e s s i v e m i c r o s t r u c t u r a l u n i t s , f o r example, under which c o n d i t i o n s i t i s
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
115
p o s s i b l e t h a t t h e Outcome o f a c e r t a i n ACTION i s succeeded by t h e Concepti o n o f a new PLAN. I t i s assumed t h a t t h e s e c o n d i t i o n s c o n s i s t i n c o n f i g u r a l r e s t r i c t i o n s based on s u r r o u n d i n g m i c r o s t r u c t u r a l u n i t s . I n t h e R o t e r Faden model t h e c o n d i t i o n s f o r c o n n e c t i n g s u c c e s s i v e m i c r o s t r u c t u r a l u n i t s a r e r e a l i z e d by c o n f i g u r a l r u l e s . These r u l e s d e f i n e which two f u n c t i o n a l u n i t s may f o l l o w one a n o t h e r i n t h e m i c r o s t r u c t u r e o f a t e x t . A s i m p l i f i e d v e r s i o n o f one o f t h e r u l e s w i l l serve as an i l l u s t r a t i o n . Note t h a t t h e m i c r o s t r u c t u r a l u n i t s i n q u e s t i o n a r e under1 i n e d : Obstacle p
-f
( a f t e r t h e s e l e c t i o n o f a method t o a c h i e v e goal r )
L
Goal r + 1: Remove o b s t a c l e p ( t 1) S e l e c t a l t e r n a t i v e method f o r goal r N e g a t i v e Consequence f o r goal r ( - 1 )
T h i s r u l e says t h a t whenever a n O b s t a c l e b l o c k s t h e a p p l i c a t i o n (Do) o f a s e l e c t e d method, t h e main c h a r a c t e r has e x a c t l y t h r e e s t r a t e g i e s t o choose from: To s e t h i m s e l f t h e t o remove t h e o b s t a c l e , o r t o S e l e c t an a l t e r n a t i v e method t h a t w i l l p r e v e n t him f r o m e n c o u n t e r i n g t h e experienced o b s t a c l e , o r f i n a l l y t o g i v e up t h e i n t e n d e d g o a l ( N e g a t i v e Consequence). The numbers i n parantheses i n d i c a t e t h a t t h e c o u r s e o f a c t i o n goes i n t o more d e t a i l o r o n t o a more g l o b a l l e v e l . T h i s i s i n d i c a t e d i n t h e memory r e p r e s e n t a t i o n by changes i n t h e h i e r a r c h y l e v e l , i . e . , one l e v e l down ( + 1) o r one l e v e l up ( - 1 ) .
Goal
Both o f t h e s e knowledge s t r u c t u r e s , namely a c t i o n schemata and c o n f i g u r a l r u l e s a r e u t i l i z e d by t h e p r o c e s s i n g system o f t h e R o t e r Faden model. As an example, t h e f i n a l r e p r e s e n t a t i o n o f one o f t h e t e x t s used i n a r e c o g n i t i o n experiment t o t e s t t h e R o t e r Faden model i s shown i n F i g u r e 1. The symbol s t r i n g connected by d o t t e d l i n e s and r u n n i n g a c r o s s d i f f e r e n t h i e r a r c h i c a l l e v e l s r e p r e s e n t s t h e R o t e r Faden, o r m a i n t r a i n o f thought, o f t h a t s p e c i f i c t e x t i n t e r r u p t e d by dead-end s t r u c t u r e s ( u n s u c c e s s f u l a c t i o n sequences). Both c o n s i s t o f e x p l i c i t , i . e . , a c t u a l l y s t a t e d i n t h e t e x t , o r i m p l i c i t , i.e., i n f e r r e d , f u n c t i o n a l u n i t s , together w i t h t h e i r respective c o n n e c t i o n s a c c o r d i n g t o t h e c o n f i g u r a l r u l e s d e f i n e d i n t h e knowledge s t r u c t u r e . T h i s s t r i n g o f connected f u n c t i o n a l u n i t s i s e q u i v a l e n t t o t h e m i c r o s t r u c t u r e o f t h e t e x t . The c i r c l e d symbols E , P , and H r e p r e s e n t t h e t h r e e d i f f e r e n t t y p e s o f m a c r o p r o p o s i t i o n s a c c o r d i n g t o a c t i o n schemata, a g a i n d e f i n e d i n t h e knowledge s t r u c t u r e . These m a c r o p r o p o s i t i o n s , t o gether w i t h t h e i r connecting continuous l i n e s , represent t h e t h e o r e t i c a l m a c r o s t r u c t u r e . T h i s concept o f a m a c r o s t r u c t u r e r e p r e s e n t e d by nodes and e x t r a c o n n e c t i o n s between c o h e r e n t f u n c t i o n a l u n i t s o f t h e m i c r o s t r u c t u r e w i l l be f u r t h e r i n v e s t i g a t e d l a t e r on. B u t f i r s t , we s h a l l g i v e an o u t l i n e o f a n experiment t h a t was conducted t o t e s t some p r o p e r t i e s o f t h e memory r e p r e s e n t a t i o n produced by t h e R o t e r Faden model.
AN EXPERIMENTAL TEST OF THE ROTER FADEN MODEL As has a l r e a d y been i n d i c a t e d , t h e model a l l o w s f o r d i s c r i m i n a t i o n of t h e Roter Faden and dead-ends o f a t e x t on t h e one hand, and more o r l e s s comp l e x u n i t s i n t h e m a c r o s t r u c t u r e on t h e o t h e r . These two a r e , t o o u r knowledge, t h e o n l y v a r i a b l e s f o r which c o n s i s t e n t e m p i r i c a l r e s u l t s have been o b t a i n e d i n experiments designed t o t e s t s t r u c t u r a l p r o p e r t i e s o f memory r e p r e s e n t a t i o n s f o r p i e c e s o f d i s c o u r s e . The f i r s t v a r i a b l e has proved t o be e f f e c t i v e i n s t u d i e s conducted by B l a c k (1978), G l o w a l l a (1981a), and Omanson (1979), t h e second v a r i a b l e i n a s t u d y by B l a c k and Bower (1979).
0
0
I m W
m
z
0
m
35
4.45
Figure 1 R e p r e s e n t a t i o n Graph o f a T e x t According t o t h e R o t e r Faden Model The Numbers R e f e r t o t h e M i c r o p r o p o s i t i o n s o f t h e T e x t . E x p l a n a t i o n o f Symbols: E: EPISODE; P: PLAN; H: ACTION; I I , I , A c t: i o n w i t h Pos.Outcome, Neg.Outcome, o r Obstacle; , rn : Pos. o r Neg.Outcome;O , : Pos. o r Neg.Consequence; A , A : S e t t i n g / E v e n t o r Obs t a c l e ; 0 : Goal; D : Conception; : Dead-End S t r u c t u r e s .
+
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
117
Comparable p r e d i c t i o n s a r e a l s o made by Schank and A b e l s o n ' s (1977) "Conceptual Dependency / Know1 edge S t r u c t u r e " computer s i m u l a t i o n model. Tests of o t h e r s t r u c t u r a l p r o p e r t i e s o f memory r e p r e s e n t a t i o n s o f d i s c o u r s e l i k e membership o f d i f f e r e n t e p i s o d i c c a t e g o r i e s o r p o s i t i o n i n t h e t e x t s t r u c t u r e h i e r a r c h y - have l e d t o i n c o n s i s t e n t r e s u l t s (Omanson, 1979; Trabasso, 1980). Nevertheless, i t i s n o t c l e a r which o f t h e two v a r i a b l e s i s t h e more e f f e c t i v e . The reason f o r t h i s l i e s i n t h e f a c t t h a t i n experiments i n v e s t i g a t i n g one o f t h e s e v a r i a b l e s , p o s s i b l e e f f e c t s o f t h e o t h e r v a r i a b l e have not been c o n t r o l l e d f o r . I n t h e f o l l o w i n g r e c o g n i t i o n experiment, t h e s e two v a r i a b l e s have been v a r i e d i n d e p e n d e n t l y . I n t h i s experiment s u b j e c t s were t e s t e d i n t h r e e d i f f e r e n t s e s s i o n s w i t h e x a c t l y t h e same procedure. Each s e s s i o n s t a r t e d w i t h a s t u d y p e r i o d i n which two t e x t s were presented a c o u s t i c a l l y t o t h e s u b j e c t s . D u r i n g each s e s s i o n , a d i f f e r e n t p a i r o f t e x t s was used. A f t e r an i n t e r v e n i n g p e r c e p t i o n t a s k , t h e s u b j e c t s were t e s t e d on r e t e n t i o n o f t h e p r e v i o u s l y l e a r n e d t e x t p a i r i n t h e f o l l o w i n g way: On each t r i a l two sentences stemming f r o m t h e t e x t p a i r were p r e sented v i s u a l l y . S u b j e c t s had t o d e c i d e whether b o t h sentences o r i g i n a t e d from t h e same o r d i f f e r e n t t e x t s . Percentages o f c o r r e c t answers and r e sponse l a t e n c i e s were measured. A l l probe sentences o r i g i n a t e d from t h r e e d i f f e r e n t t y p e s o f m a c r o s t r u c t u r a l u n i t s , namely a PLAN d o m i n a t i n g an EPISODE, a PLAN c o n s i s t i n g o f s e v e r a l ACTIONS, and f i n a l l y a s i n g l e ACTION. F o r each o f t h e s e s t r u c t u r e t y p e s t h e r e e x i s t e d two comparable u n i t s , one belonging t o t h e R o t e r Faden and t h e o t h e r b e i n g a dead-end. As a n example, look a t t h e t e x t r e p r e s e n t a t i o n shown i n F i g u r e 1. The m a c r o s t r u c t u r a l u n i t s PLAN w i t h embedded EPISODE c o n s i s t o f t h e m i c r o s t r u c t u r e u n i t s 30 t o 38 and 39 t o 47, r e s p e c t i v e l y , t h e f i r s t b e i n g t h e dead-end s t r u c t u r e and the second b e i n g t h e R o t e r Faden s t r u c t u r e . B o t h sentences of a p o s i t i v e probe stemmed from t h e same m a c r o s t r u c t u r a l u n i t and b o t h e i t h e r belonged t o t h e R o t e r Faden s t r u c t u r e o r t o t h e c o r r e s p o n d i n g dead-end s t r u c t u r e . This i s a l s o t r u e f o r n e g a t i v e probes w i t h one i m p o r t a n t e x c e p t i o n : One sentence o r i g i n a t e d f r o m t h e f i r s t t e x t and one from t h e second t e s t . The observed r e a c t i o n t i m e s have been analysed by s e v e r a l ANOVAs; t h e main r e s u l t s a r e as f o l l o w s : Ifa p i e c e o f i n f o r m a t i o n belongs t o t h e R o t e r Faden, i t r e c e i v e s some s o r t o f s t r u c t u r a l b e n e f i t i n s o f a r as a s t r o n g decay i n a v a i l a b i l i t y i s compensated f o r when i t o r i g i n a t e s f r o m a l e s s i m p o r t a n t m a c r o s t r u c t u r a l u n i t . I f t h e p i e c e o f i n f o r m a t i o n stems from a dead-end s t r u c t u r e , t h e decay i n a v a i l a b i l i t y w i l l be i n t e n s i f i e d . These r e s u l t s l e d t o t h e c o n c l u s i o n t h a t any model must account f o r t h e e f f e c t s of b o t h o f t h e s e p r o p e r t i e s . The s u b s t a n t i a l l y s t r o n g e r e f f e c t s o f membership i n more o r l e s s complex m a c r o s t r u c t u r e u n i t s l e d t o t h e f u r t h e r conc l u s i o n t h a t t h e m a c r o s t r u c t u r a l o r g a n i z a t i o n o f t e x t s should be i n v e s t i gated more t h o r o u g h l y i n t h e f u t u r e . F o r d e t a i l s o f t h i s experiment see Glowalla (1981). PRELIMINARY RESULTS ON MACROSTRUCTURAL ORGANIZATION
AS a m a t t e r o f f a c t , p a r t of t h e d a t a o f t h e experiment j u s t d e s c r i b e d a l l o w s f o r some p r e l i m i n a r y i n s i g h t s i n t o t h e s e phenomena. D u r i n g t h e stepwise c o n s t r u c t i o n of t h e memory r e p r e s e n t a t i o n f o r a g i v e n t e x t according t o t h e R o t e r Faden model, m a c r o s t r u c t u r a l c o n n e c t i o n s between coherent m i c r o s t r u c t u r a l u n i t s ( t h o s e b e l o n g i n g t o t h e same m a c r o s t r u c t u r a1 u n i t ) a r e e s t a b l i s h e d v i a a node r e p r e s e n t i n g t h a t p i e c e o f t h e macros t r u c t u r e as a whole. P a i r s o f m i c r o s t r u c t u r a l u n i t s d i f f e r w i t h r e s p e c t
118
COHERENCE
t o t h e i r d i s t a n c e i n t h e m a c r o s t r u c t u r e . The d i s t a n c e i s expressed by t h e number o f l i n k s on t h e p a t h between two r e s p e c t i v e u n i t s . I f , f o r example, two m i c r o s t r u c t u r a l u n i t s a r e dominated by t h e same m a c r o s t r u c t u r e node H, t h e i r d i s t a n c e i s 2, i f t h e y a r e dominated by t h e same node P b u t belong t o d i f f e r e n t A C T I O N S , t h e i r d i s t a n c e i s 4. I t may be asked whether d i f f e r e n t response l a t e n c i e s a r e t o be observed f o r probes h a v i n g d i f f e r e n t d i s t a n c e s i n t h e m a c r o s t r u c t u r e . Assuming equal search r a t e s f o r a l l l i n k s i n t h e m a c r o s t r u c t u r e , search t i m e s s h o u l d i n c r e a s e w i t h i n c r e a s i n g distance. T h i s p r e d i c t i o n can be t e s t e d by i n s p e c t i o n o f p a r t o f t h e d a t a f r o m t h e d e s c r i b e d r e c o g n i t i o n experiment. I n s t r u c t u r e t y p e 1 (PLAN w i t h an embedded EPISODE) t h e r e a r e probes w i t h d i s t a n c e s o f 2, 4, o r 6 l i n k s , r e s p e c t i v e l y . I f one t a k e s t h e d a t a o f a l l s i x t e x t s t h a t o r i g i n a t e from s t r u c t u r e t y p e 1 and computes mean r e a c t i o n t i m e s f o r t h e t h r e e d i s t i n c t c a t e g o r i e s s e p a r a t e l y , t h e r e s u l t s a r e as f o l l o w s : 2443 ms, 2546 ms, and 2742 ms f o r probes w i t h d i s t a n c e s 2, 4, and 6 l i n k s . The observed d i f f e rences of 103 ms and 196 ms a r e s t a t i s t i c a l l y r e l i a b l e ( F 2 , a s = 68.28; p < . 0 0 1 ) . S i m i l a r r e a c t i o n t i m e d i f f e r e n c e s can be o b t a i n e d i f t h i s a n a l y s i s i s performed f o r each t e x t s e p a r a t e l y , b u t f o r two o u t o f s i x t e x t s t h e s e d i f f e r e n c e s a r e n o t s i g n i f i c a n t . T h i s r e s u l t may be t a k e n as s u p p o r t i n g evidence f o r t h e m a c r o s t r u c t u r e assumptions o f t h e R o t e r Faden model. However, one can o b j e c t t h a t t h e v a r i a b l e d i s t a n c e s i n t h e m a c r o s t r u c t u r e a r e confounded w i t h a n o t h e r v a r i a b l e , namely t h e number o f c l a u s e s between t h e two sentences o f a probe i n t h e s u r f a c e s t r u c t u r e o f a g i v e n t e x t . I n f a c t , t h i s v a r i a b l e should a f f e c t response l a t e n c i e s , s i n c e t h e mean number o f coherence r e l a t i o n s between two sentences i s i n v e r s e l y r e l a t e d t o t h e number o f i n t e r v e n i n g sentences ( C l a r k and Sengul, 1979; Lesgold, Roth and C u r t i s , 1979). Whether t h i s p r o p e r t y had any i n f l u e n c e on t h e observed response t i m e s has been i n v e s t i g a t e d by a s e p a r a t e a n a l y s i s i n v o l v i n g probes o f s t r u c t u r e t y p e 2 (PLAN c o n s i s t i n g o f s e v e r a l ACTIONS). The sentences o f t h e s e probes possess t h e same d i s t a n c e of 4 l i n k s i n t h e macros t r u c t u r e t o g e t h e r w i t h 0, 1, o r 2 i n t e r v e n i n g sentences. Over a l l s i x t e x t s used i n t h e experiment, mean r e a c t i o n t i m e s o f 2716 ms, 2811 ms, and 2861 ms were observed f o r probes b e l o n g i n g t o t h e s e d i f f e r e n t c a t e g o r i e s . The r e a c t i o n t i m e d i f f e r e n c e s o f 95 ms and 50 ms a r e s t a t i s t i c a l l y s i g n i f i c a n t ( F Z , ~= ~7.99; p < . 0 1 ) . W i t h m i n o r q u a l i f i c a t i o n s s i m i l a r d i f f e r e n c e s were observed f o r e v e r y s i n g l e t e x t . As i t stands, t h e d i s t a n c e e f f e c t m i g h t be e x p l a i n e d by t h e number o f i n t e r v e n i n g sentences as w e l l . To c l a r i f y whether t h e observed d i f f e r e n c e s i n response l a t e n c i e s a r e a f f e c t e d by b o t h o f t h e s e v a r i a b l e s o r depend e x c l u s i v e l y on t h e number o f i n t e r v e n i n g sentences w i t h r e s p e c t t o t h e sentences o f a probe, a t h i r d a n a l y s i s was conducted. A subset o f probes b e l o n g i n g t o s t r u c t u r e t y p e 1 proved t o be s u i t a b l e f o r such an a n a l y s i s . The sentences o f t h e s e probes a r e separated by two i n t e r v e n i n g sentences i n t h e s u r f a c e s t r u c t u r e o f t h e r e s p e c t i v e t e x t and connected v i a 4 o r 6 l i n k s i n t h e m a c r o s t r u c t u r e . C o n s i d e r i n g such probes o f a l l s i x t e x t s t o g e t h e r , mean r e a c t i o n t i m e s o f 2606 ms and 2822 ms were observed. The d i f f e r e n c e o f 216 ms i s s t a t i s t i c a l l y s i g n i f i c a n t (F1,+$ = 23.26; p < .001). Again, s i m i l a r r e s u l t s a r e o b t a i n e d , i f s e p a r a t e ANOVAs a r e conducted f o r s i n g l e t e x t s . A j o i n t consideration o f a l l t h r e e analyses l e d t o t h e f o l l o w i n g c o n c l u s i o n : Both v a r i a b l e s , namely t h e d i s t a n c e i n t h e macros t r u c t u r e as w e l l as t h e number o f i n t e r v e n i n g sentences as a measure of
TOWARD A MODEL OF MACROSTRUCTURE SEARCH coreferentiality,
119
a f f e c t t h e u n d e r l y i n g response process.
A MODEL FOR MACROSTRUCTURE SEARCH I n t h i s s e c t i o n , we g i v e a n o u t l i n e o f a q u a n t i t a t i v e model f o r t h e search processes hypothesized i n t h e above experiment. I d e a l l y , one would expect such a model t o account f o r a l l main f e a t u r e s o f t h e d a t a , t h a t i s , t h e dependence o f r e a c t i o n t i m e s and p r o p o r t i o n of c o r r e c t answers on t h e v a r i o u s probe t y p e s . I t should be c l e a r f r o m t h e o u t s e t t h a t what f o l l o w s i s o n l y meant as a s t a r t i n g p o i n t toward such a model. Moreover, t h e model b u i l d i n g h e r e i s a post-hoc e n t e r p r i s e and t h u s , no model t e s t i n g i s y e t involved. I t i s common i n r e a c t i o n t i m e - s t u d i e s o f l o n g - t e r m memory search t o conc e i v e o f t o t a l r e a c t i o n t i m e (RT) as an a d d i t i v e c o m p o s i t i o n o f two p a r t s : a search t i m e ( S ) depending on t h e e x p e r i m e n t a l c o n d i t i o n , and a r e s i d u a l t i m e ( R ) c o n s i s t i n g o f encoding and motor components t h a t a r e c o n s t a n t over e x p e r i m e n t a l c o n d i t i o n s ( c f . P a c h e l l a , 1 9 7 4 ) :
RT=S+R An i m p o r t a n t p r e s u p p o s i t i o n i n t h e f o l l o w i n g argument i s t h a t t h e S component o f t h e r e a c t i o n t i m e i n t h e above experiment can be t a k e n as generated by a search process o f t h e m a c r o s t r u c t u r e t h a t resembles t h e spreading a c t i v a t i o n assumptions used i n semantic network models ( c f . King and Anderson, 1976). A f i n d i n g s u p p o r t i n g t h i s v i e w i s t h e afore-mentioned i n c r e a s e i n mean r e a c t i o n t i m e as a f u n c t i o n o f t h e number o f l i n k s separ a t i n g two probe sentences i n t h e m a c r o s t r u c t u r e graph r e p r e s e n t a t i o n . Bef o r e i n t r o d u c i n g some s p e c i f i c a s p e c t s o f t h i s search process, two remarks a r e a p p o s i t e . F i r s t l y , w h i l e o u r d a t a a n a l y s i s suggested an e f f e c t o f t h e number o f m i c r o s t r u c t u r e l i n k s between t h e probe sentences on RT, t h i s has n o t y e t been i n c o r p o r a t e d i n t o t h e model assumptions, s i n c e t h e experiment was n o t e s p e c i a l l y designed t o e x p l o r e t h e n a t u r e o f t h i s e f f e c t . The obvious a l t e r n a t i v e s t o handle t h i s would be e i t h e r t o p o s t u l a t e a p p r o p r i a t e search processes on t h e m i c r o s t r u c t u r e l i n k s o r t o l e t some parameters o f t h e m a c r o s t r u c t u r e search depend on t h e m i c r o s t r u c t u r e d i s t a n c e between t h e two probe sentences. Secondly, a s i m i l a r p o i n t should be made f o r t h e e f f e c t o f a n o t h e r v a r i a b l e , t h e m a c r o s t r u c t u r e t y p e which t h e probe sentences f o r m p a r t o f . One way t o t a k e t h i s e f f e c t i n t o a c c o u n t would be t o break down t h e search t i m e ( S ) i n t o a n o t h e r two a d d i t i v e subcomponents, one being t h e a c t u a l t i m e t o a c t i v a t e a l l necessary l i n k s , t h e o t h e r r e p r e s e n t i n g t h e t i m e t o g a i n access t o t h e m i c r o s t r u c t u r e r e p r e s e n t a t i o n o f t h e probes, where t h e l a t t e r t i m e component may d i f f e r f r o m one m a c r o s t r u c t u r e t y p e t o t h e n e x t . The f o l l o w i n g o n l y r e f e r s t o t h e a c t i v a t i o n search p r o cess i t s e l f : ( a ) each node and l i n k e i t h e r i s i n t h e a c t i v e s t a t e o r n o t ; ( b ) i f a l i n k i s i n t h e a c t i v e s t a t e , t h e n t h e two nodes connected by t h e l i n k are also active; ( c ) t h e t i m e f o r a c t i v a t i o n t o spread from an a c t i v e node t o an i n c i d e n t l i n k i s a n e x p o n e n t i a l l y d i s t r i b u t e d random v a r i a b l e w i t h a r a t e i n v e r s e l y r e l a t e d t o t h e t o t a l number o f l i n k s i n c i d e n t t o t h a t node; ( d ) a c t i v a t i o n spreads i n d e p e n d e n t l y f r o m b o t h probe nodes and t h e s u b j e c t g i v e s an answer whenever a p a t h c o n n e c t i n g b o t h probe nodes i s a c t iva t e d .
120
COHERENCE
The e x p o n e n t i a l d i s t r i b u t i o n i s used m a i n l y f o r mathematical t r a c t a b i l i t y : 1 ) t h e minimum o f a f i n i t e number o f independent e x p o n e n t i a l l y d i s t r i b u t e d random v a r i a b l e s w i t h r a t e s r . i s a g a i n e x p o n e n t i a l l y d i s t r i b u t e d w i t h r a t e J C r.
j
J '
and 2) t h e p r o b a b i l i t y o f t h e i - t h v a r i a b l e b e i n g t h e minimum i s ri/C r . J J
.
The assumption t h a t t h e r a t e o f a c t i v a t i o n s p r e a d i n g from a node towards an i n c i d e n t l i n k depends on t h e t o t a l number o f l i n k s i n c i d e n t w i t h t h a t node i s m o t i v a t e d by t h e s o - c a l l e d f a n - e f f e c t i n semantic network r e s e a r c h ( c f . K i n g and Anderson, 1976). Given ( a ) - ( d ) i t i s a r o u t i n e m a t t e r t o compute t h e expected search t i m e s f o r any m a c r o s t r u c t u r e c o n f i g u r a t i o n . L e t u s i l l u s t r a t e t h i s w i t h an example o f a m a c r o s t r u c t u r e grapii segment ( v i s . F i g u r e 2 ) . Suppose t h e m i c r o s t r u c t u r e nodes l a b e l e d 2 and 4 a r e a c t i v a t e d by t h e probe sentences. I n o r d e r t o compute t h e expected t i m e t h e p a t h a-b-c-d from 2 t o 4 needs t o be a c t i v a t e d , we have t o c o n s i d e r a l l p o s s i b l e ways t h i s a c t i v a t i o n may come about. Suppose t h e o r d e r o f a c t i v a t i o n o f t h e l i n k s i s a, b, d, c. I f r i s t h e parameter o f t h e e x p o n e n t i a l d i s t r i b u t i o n , t h e expected t i m e f o r a t o be a c t i v a t e d f r o m 2 i s l / r ; f o r b , t h e t i m e t o be a c t i v a t e d i s 3 / r , since there a r e t h r e e l i n k s i n c i d e n t w i t h H2 not y e t a c t i v e ; then f o r d , t h e t i m e i s a g a i n l / r , w h i l e f o r e , we have 1 / 2 r , s i n c e a c t i v a t i o n may spread f r o m b o t h E and H 4 . Thus t h e expected t i m e f o r t h e o r d e r a,b,d,c i s 5 . 5 / r . The p r o b a b i l i t y o f t h i s o r d e r o c c u r r i n g i s 1 / 2 ( a " w i n n i n g " a g a i n s t d ) t i m e s 1/4 ( b " w i n n i n g " a g a i n s t d) t i m e s 1 / 2 ( d " w i n n i n g " against e) = 1/16. I n o r d e r t o g e t t h e expected t o t a l a c t i v a t i o n t i m e we have t o sum over the expected a c t i v a t i o n t i m e s o f a l l p o s s i b l e o r d e r s weighted by t h e i r r e s p e c t i v e p r o b a b i l i t i e s o f occurrence. I t i s n o t d i f f i c u l t t o show t h a t f o r a p a t h o f n l i n k s t h e r e a r e 2"-' d i f f e r e n t p o s s i b l e a c t i v a t i o n o r d e r s . However, i n most cases Thus, c o m p u t a t i o n l o a d i n c r e a s e s q u i c k l y w i t h n n i s o n l y o f moderate s i z e and, i n any event, t h e t a s k s h o u l d be amenable t o programming.
.
I n t h e above example, t h e expected t o t a l t i m e t o a c t i v a t e t h e 4 l i n k s conn e c t i n g t h e t e s t probes 2 and 4 can be shown t o be about 5 / r by e v a l u a t i n g a l l 8 d i f f e r e n t a c t i v a t i o n o r d e r s . The c o r r e s p o n d i n g t i m e f o r t e s t probes 1 and 2 w i t h o n l y 2 c o n n e c t i n g l i n k s i s 1 . 7 5 / r . I t can be shown more gen e r a l l y t h a t , no m a t t e r what t h e s p e c i f i c c o n f i g u r a t i o n , t h e expected t o t a l t i m e i s an i n c r e a s i n g f u n c t i o n o f t h e number o f l i n k s c o n s t i t u t i n g t h e p a t h c o n n e c t i n g t h e two t e s t probes. As i t stands, t h e model a l l o w s some f a i r l y d e f i n i t e p r e d i c t i o n s t o be made. I n a d d i t i o n t o t h e m o n o t o n i c i t y p r o p e r t y j u s t mentioned, one would expect equal t o t a l a c t i v a t i o n t i m e s - a p a r t from s t a t i s t i c a l v a r i a b i l i t y - f o r a number o f t e s t probe p a i r s due t o an i n h e r e n t symnetry i n t h e macrostruct u r e graph, f o r example f o r t h e p a i r s 1 - 3, 2 - 3 , 1 - 4, and 2 - 4 . I t s h o u l d be p o i n t e d o u t , however, t h a t any c o n c l u s i o n drawn f r o m a t e s t o f t h e s e p r e d i c t i o n s i s o n l y v a l i d i f t h e i n f l u e n c e o f t h e o t h e r f a c t o r s on RT d i s c u s s e d above has been accounted f o r a p p r o p r i a t e l y .
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
121
Figure 2 Segment o f a M a c r o s t r u c t u r e Tree The Numbers r e f e r t o M i c r o s t r u c t u r e , t h e C a p i t a l L e t t e r s t o M a c r o s t r u c t u r e Nodes.
CONCLUSION
I t has been argued t h a t t h e m a c r o s t r u c t u r e o f a t e x t can be r e p r e s e n t e d by a graph, d i v i d i n g c o h e r e n t m i c r o s t r u c t u r a l u n i t s i n t o c l u s t e r s a t d i f f e r e n t h i e r a r c h i c a l l e v e l s . The q u a n t i t a t i v e model o u t l i n e d i n t h e p r e v i o u s s e c t i o n makes some d e f i n i t e p r e d i c t i o n s about search processes on m a c r o s t r u c t u r e graphs. Most i m p o r t a n t l y , t h e t i m e t o v e r i f y a probe should be a n i n c r e a s i n g f u n c t i o n of t h e number o f l i n k s c o n s t i t u t i n g t h e p a t h c o n n e c t i n g t h e two sentences o f t h e p r o b e i n t h e m a c r o s t r u c t u r e . The observed response l a t e n c i e s i n t h e r e p o r t e d r e c o g n i t i o n experiment a r e i n accordance w i t h t h i s p r e d i c t i o n . However, as has a l r e a d y been p o i n t e d o u t , t h e experiment was n o t e s p e c i a l l y designed t o i n v e s t i g a t e search processes on macros t r u c t u r e s . Thus, no r e a s o n a b l e t e s t o f t h e proposed model c o u l d be p e r formed. We s h a l l t h e r e f o r e c o n c l u d e w i t h some s u g g e s t i o n s a b o u t t h e d e s i g n o f such a n experiment. R e t a i n i n g t h e r e c o g n i t i o n t e c h n i q u e presented, one e f f e c t i v e m o d i f i c a t i o n c o u l d c o n s i s t o f t h e a d d i t i o n a l employment o f probes w i t h sentences stemming f r o m d i f f e r e n t s t r u c t u r e t y p e s . I n an experiment i n v e s t i g a t i n g the j o i n t representation o f several p a r t i a l l y coherent t e x t s (Glowalla, 1980), r e l i a b l e r e a c t i o n - t i m e d i f f e r e n c e s between probes f r o m t h e same o r d i f f e r e n t s t r u c t u r e t y p e s have been observed. By a l t e r i n g t h e d i s t a n c e o f
122
COHERENCE
m a c r o s t r u c t u r e u n i t s o u t o f w h i c h t h e two sentences o f a probe a r e taken s i m u l t a n e o u s l y , a s t r o n g t e s t o f t h e proposed q u a n t i t a t i v e model should be p o s s i b l e . Another m o d i f i c a t i o n m i g h t c o n s i s t i n a v a r i a t i o n o f t h e pres e n t a t i o n o r d e r o f probe sentences. T h i s v a r i a b l e has p r o v e n t o be h i g h l y e f f e c t i v e i n r e s e a r c h on a c t i v a t i o n o f semantic networks ( G l o w a l l a , Schulze and Wender, 1980). We assume t h a t an a l t e r a t i o n o f p r e s e n t a t i o n o r d e r s h o u l d f u n d a m e n t a l l y i n f l u e n c e t h e e f f e c t s o f t h e number o f m i c r o s t r u c t u r e l i n k s ( a s a measure o f c o r e f e r e n t i a l i t y ) on RT. I n t h e experiment reported here, t h e o r d e r o f probe sentences corresponded t o t h e i r appearance i n the s u r f a c e s t r u c t u r e o f a t e x t . The o p p o s i t e o r d e r should r e s u l t i n a dec r e a s e o f c o r e f e r e n t i a l i t y e f f e c t s on RT. The u s e o f t h e s e two and o t h e r s i m i l a r m o d i f i c a t i o n s i n experiments on m a c r o s t r u c t u r e search should lead t o a deeper u n d e r s t a n d i n g o f t h e n a t u r e of m a c r o s t r u c t u r e s and t h e i r applic a t i o n d u r i n g the processing o f discourse. REFERENCES B l a c k , J.B. 1978.
Story memory s t r u c t u r e s . PhD Thesis, S t a n f o r d U n i v e r s i t y ,
Episodes as chunks i n n a r r a t i v e memory. B l a c k , J.B., and Bower, G.H. Journal of Verbal Learning and Verbal aehuvior, 1979, 18, 309 - 318. I n search o f r e f e r e n t s f o r nouns and proC l a r k , H.H., and Sengul, C.J. nouns. Memory and Cognition, 1979, 7 , 35 - 41. Der Rote. Fuden - e i n handlungstheoretisches Model1 zur TextGlowalla, U . verarbeitung. D i s s e r t a t i o n , U n i v e r s i t a t Braunschweig, 1981. Reproduzieren und Zusammenfassen von Texten. Sprache und G l o w a l l a , U. Kognition, i n p r e s s .
Probleme b e h Enkodieren und Dekodieren mehrerer ProblemGlowalla, U . ZBsegeschichten. Unpublished m a n u s c r i p t , U n i v e r s i t a t Braunschweig, 1980. The a c t i v a t i o n o f sentenG l o w a l l a , U . , Schulze, H.H., and Wender, K.F. ces i n semantic networks. I n F. K l i x and J . Hoffmann (Eds.), Cognition and Memory. Amsterdam: N o r t h H o l l a n d , 1980. Graesser, A.C. 1981.
Prose comprehension beyond t h e word. New York: S p r i n g e r ,
King, D., and Anderson, J.R. Long-term memory r e t r i e v a l : An i n t e r s e c t i n g a c t i v a t i o n process. Journal o f Verbal Learning and Verbal Behavior, 1976, 15, 587 - 605. K i n t s c h , W., and van D i j k , T.A. Toward a model of d i s c o u r s e comprehensi o n and p r o d u c t i o n . Psychological Review, 1978, 85, 363 - 394. Lesgold, A.M., Roth, S.F., and C u r t i s , M.E. Foregrounding e f f e c t s i n d i s c o u r s e comprehension. Journal of VerbaZ Learning and VerbaZ Behavior, 1979, 18, 291 - 308. Omanson, R.C. The narrative anaZysis. PhO Thesis, U n i v e r s i t y o f Minnesota, Minnesota, 1979.
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
? 23
Pohl, R.F. Acceptability of s t o r y continuations. I n A. Flammer, and W . Kintsch ( E d s . ) , T e x t processing. Amsterdam: North Holland, in press. Pachella, R.G. T h e i n t e r p r e t a t i o n of reaction time i n information-processing research. In B.H. Kantowitz ( E d . ) , Human information processing: t u t o r i a l s i n performance and cognition. H i l l s d a l e , N.J.: Erlbaum, 1974. Rumelhart, D . E . Understanding and summarizing brief s t o r i e s . I n D . LaBerge, and S.J. Samuels ( E d s . ) , Basic processes i n reading. H i l l s d a l e , N.J.: Erlbaum, 1977. Sanford, A.J., and Garrod, S.C. Wiley, 1981.
Understanding w r i t t e n language. New York:
Schank, R . C . , and Abelson, R.P. S c i p t s , pZans, goals, and understanding. H i l l s d a l e , N.J.: Erlbaum, 1977. Trabasso, T. Relevance perception i n t a t conrprehensioiz. Paper presented a t t h e Department o f Psychology, Universitat Braunschweig, 1980.
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
Ulrich Glowalla & Hans Colonius I n s t i t u t f u r Psychologie Technische Universitat Braunschweig Federal Republic of Germany I t has been argued t h a t t h e macrostructural organization of a t e x t i s c e n t r a l t o comprehension and r e t e n t i o n of discourse. I n a recogn i t i o n experiment, some aspects of t h e macros t r u c t u r a l representation have been inve,stigated. The r e s u l t s give some preliminary support t o the Roter Faden model (Glowalla, 1981). I n a d d i t i o n , a q u a n t i t a t i v e model capturing some f e a t u r e s of macrostructure search i s outlined. INTRODUCTION
I t has been s t r e s s e d by a number of people t h a t t h e process of condensing the f u l l meaning of a t e x t i n t o i t s g i s t i s central t o t h e comprehension and r e t e n t i o n of discourse ( e . g . , Kintsch and van Dijk, 1978). The main goal of t h i s paper i s t o i n v e s t i g a t e t h i s process f u r t h e r . I t has become convenient t o use the term macrostructure f o r such global descriptions of the semantic s t r u c t u r e o f a t e x t . In c o n t r a s t , t h e term microstructure has been introduced f o r t h e s t r u c t u r e of individual t e x t sentences and t h e i r r e l a t i o n s . To consider a piece of discourse coherent, two conditions must h o l d : i t s respective sentences should be connected a t t h e microstructure level and these sentences should be organized in l a r g e r conceptual u n i t s a t the macrostructure l e v e l . T h u s , a s has been argued elsewhere ( e . g . , Sanford and Garrod, 1981), only models f o r t e x t processing incorporating b o t h of these aspects may a t t a i n t h e o r e t i c a l a s well a s empirical signif icance. I n f a c t , q u i t e a few models f o r t e x t processing incorporating micro- and macro-level d e s c r i p t i o n s have already been proposed (Glowalla, 1981; Graess e r , 1981; Kintsch and van Dijk, 1978; Schank a n d Abelson, 1977). However, the models d i f f e r with respect t o t h e format of t h e microstructure a n d / o r the macrostructure. As t o t h e l a t t e r , several conceptions a r e under d i s cussion. The e x i s t i n g a l t e r n a t i v e s may roughly be c l a s s i f i e d as follows: The micropropositions of a t e x t a r e transformed i n t o a set of macropropos i t i o n s by so-called macro-operators. Recursive a p p l i c a t i o n o f these macrooperators leads t o macrostructural descriptions a t d i f f e r e n t hierarchical l e v e l s . The r e s u l t i n g macrostructure represents t h e g i s t of t h e t e x t . This concept of macrostructure has been developed by Kintsch and van Dijk (1978). A1 t e r n a t i v e l y , t h e macrostructure o f a t e x t may be conceived of as a hierarchical tree s t r u c t u r e consisting of nodes representing conceptual u n i t s of a t e x t and l i n k s representing the r e l a t i o n s between conceptual u n i t s . The microstructure u n i t s of a t e x t , i . e . , i t s micropropositions, a r e 111
112
COHERENCE
represented by t h e terminal nodes of t h e t r e e s t r u c t u r e . Internal nodes represent conceptual u n i t s of t h e macrostructure on d i f f e r e n t hierarchical l e v e l s . No s p e c i f i c assumptions a r e made about t h e contents represented by the node. Distances between microstructure u n i t s in t h e t r e e s t r u c t u r e representing t h e macrostructure a r e viewed a s a measure o f t h e i r connectivity, i . e . , whether they belong t o t h e same or d i f f e r e n t macrostructural u n i t s . As a matter of f a c t , s t r u c t u r e diagrams defining t h e c o n s t i t u e n t s t r u c t u r e of a piece of discourse ( e . g . , Rumelhart, 1977) may be i n t e r p r e t e d in t h i s way. This view does n o t imply t h a t people cannot construct macropropos i t i o n s out of a s e t of connected micropropositions, i . e . , t h a t they a r e unable t o produce a summary of t h e main contents of a t e x t . Again, Rumelh a r t (1977) has been a b l e t o p r e d i c t observed summaries q u i t e nicely by postulating additional summarization r u l e s . The only proposition made here i s t h a t we may know too l i t t l e about t h e cognitive operations producing t h e t h e o r e t i c a l macrostructure. We should t h e r e f o r e avoid making too many i n i t i a l assumptions. Moreover, with respect t o t h e memory representation of discourse, assumptions about t h e contents of macrostructure nodes may even prove t o be unnecessary.
I n t h e following we shall present some experimental r e s u l t s about search processes on macrostructures. The data o r i g i n a t e from an experiment t o t e s t some predictions of a recently proposed model f o r t e x t processing called t h e "Roter Faden" (main t r a i n of thought; Glowalla, 1981). I n the subsequent section we shall give an o u t l i n e of a q u a n t i t a t i v e model. A s what follows will be presented in the t h e o r e t i c a l framework of t h e Roter Faden model, we comnence with a short survey of t h e main assumptions of t h i s model. THE ROTER FADEN MODEL
The Roter Faden model has been designed t o capture comprehension and retention of discourse, in which purposive action sequences of human beings a r e described. The model c o n s i s t s of an i n t e r a c t i v e processing system based o n s p e c i f i c assumptions about t h e processing devices involved in the comprehension of discourse, together with a s e t of knowledge s t r u c t u r e s necessary t o f u l f i l the t a s k . T h e model s t a r t s off with t h e question how incoming t e x t information has t o be processed t o produce a memory representation of i t . Clauses a r e taken a s u n i t s in t h i s a n a l y s i s ; a clause i s defined here t o be a proposition containing an a c t i v e o r s t a t i v e verb. A t each processing s t e p , t h e model t r i e s t o find answers t o t h e following questions: 1. What i s t h e functional meaning of an incoming clause? ( A functiona l l y i d e n t i f i e d c l a u s e will be termed a microstructure u n i t . ) 2 . How may t h e present microstructure u n i t be connected t o already encoded u n i t s ? 3 . Which inferences have t o be generated, i f no d i r e c t connection e x i s t s between two successive u n i t s ? 4 . Which expectations may be generated about t h e t e x t information y e t t o come? 5. I s i t possible t o condense already encoded microstructure u n i t s into more global macrostructure u n i t s ? The i n t e r a c t i v e processing system produces answers t o these questions u t i l i z i n g t h e assumed knowledge s t r u c t u r e s . B o t h of these components of t h e Roter Faden model will be outlined in turn. The processing system c o n s i s t s of a c e n t r a l processor c a l l e d CONSTRUCT and four subprocessors termed IOENT, INFER, MACRO, and ERWIN. The central processor executes t h e various
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
113
processing o p e r a t i o n s i n a c t u a l f a c t , u t i l i z i n g t h e i n f o r m a t i o n produced by t h e subprocessors. I t i s assumed t h a t t h e s e subprocessors work independentl y o f each o t h e r and communicate e x c l u s i v e l y v i a t h e c e n t r a l p r o c e s s o r . A l l processing d e v i c e s d i f f e r w i t h r e s p e c t t o two p r o p e r t i e s , namely t h e i r t a s k and t h e i r a b i l i t y t o access t h e assumed knowledge s t r u c t u r e s . The c e n t r a l processor CONSTRUCT a c t u a l l y produces t h e memory r e p r e s e n t a t i o n o f a t e x t by p r o c e s s i n g i t s i n f o r m a t i o n c l a u s e - b y - c l a u s e . Subprocessor I D E N T t r i e s t o i d e n t i f y t h e f u n c t i o n a l meaning o f each incoming c l a u s e . D u r i n g t h i s p r o cess e x p e c t a t i o n s a r e u t i l i z e d t h a t have been d e r i v e d f r o m t e x t i n f o r m a t i o n a l r e a d y encoded. I f no match between t h e p r e s e n t c l a u s e and one o f these e x p e c t a t i o n s i s found, subprocessor INFER becomes a c t i v a t e d . INFER, then, generates c h a i n s c o n s i s t i n g o f one o r more i n f e r e n c e s i n o r d e r t o connect t h e new p i e c e o f i n f o r m a t i o n t o t h e s t r u c t u r e a l r e a d y e x i s t i n g . A t each p r o c e s s i n g s t e p , subprocessor MACRO checks f o r t h e p o s s i b i l i t y o f condensing a s e t o f c o h e r e n t m i c r o s t r u c t u r a l u n i t s i n t o more g l o b a l macros t r u c t u r a l u n i t s and ERWIN, f i n a l l y , generates e x p e c t a t i o n s a b o u t t h e t e x t i n f o r m a t i o n y e t t o come. Up t o now, a l l p r o c e s s i n g d e v i c e s a r e r e a l i z e d by d e t a i l e d f l o w c h a r t s m o d e l l i n g t h e comprehension o f t e x t i n f o r m a t i o n a c c o r d i n g t o t h e R o t e r Faden model. Moreover, t h e p r o c e s s i n g d e v i c e s a r e p u r e l y f u n c t i o n a l i n nature. Thus, s p e c i f i c w o r l d knowledge i s n o t c o n s i d e r e d , f o r example, whether a s p e c i f i c method encountered by t h e main c h a r a c t e r o f an a c t i o n sequence i s s u i t a b l e t o r e a c h a c e r t a i n g o a l . I n a d d i t i o n , no c a p a c i t y l i m i t a t i o n s o f w o r k i n g memory a r e c o n s i d e r e d , as has r e c e n t l y been done by K i n t s c h and van D i j k (1978). However, t o implement t h e p r o c e s s i n g model as a computer program, assumptions a b o u t b o t h o f t h e s e a s p e c t s would have t o be i n c o r p o r a t e d . N e v e r t h e l e s s , by g i v i n g a d e t a i l e d account o f t h e opera t i o n s performed d u r i n g t h e n e x t comprehension process, t h e p r o c e s s i n g dev i c e s a l l o w f o r s u c c e s s i v e e x p e r i m e n t a l t e s t i n g . I n a c t u a l f a c t , one i m p o r t a n t f e a t u r e o f t h e model, namely t h e process o f g e n e r a t i n g e x p e c t a t i o n s has a l r e a d y been i n v e s t i g a t e d and l e d t o s u p p o r t i v e e v i d e n c e f o r t h e Roter Faden model ( P o h l , t h i s volume). D u r i n g t h e comprehension process t h e p r o c e s s i n g system i s assumed t o have access t o d i f f e r e n t know1 edge s t r u c t u r e . These knowledge s t r u c t u r e s cons i s t o f t h r e e a c t i o n schemata and a s e t o f t e n c o n f i g u r a l r u l e s . The f o r mer s p e c i f y o u r a b s t r a c t knowledge about t h e t y p i c a l s t r u c t u r e o f a c t i o n sequences and t h e l a t t e r d e f i n e o u r knowledge about p e r m i t t e d d e v i a t i o n s from t h e t y p i c a l s t r u c t u r e . To a p p l y t h e s e knowledge s t r u c t u r e s t o t h e comprehension process, i t must be shown which u n i t s o f a t e x t may be ident i f i e d w i t h u n i t s o f t h e knowledge s t r u c t u r e s . I n t h e R o t e r Faden model i t i s assumed t h a t c l a u s e s t a k e n as u n i t s o f t e x t i n f o r m a t i o n belong t o e i t h e r o f t h r e e b a s i c c a t e g o r i e s . These c a t e g o r i e s correspond t o t h e d i f f e r e n t f u n c t i o n a l meanings o f c e r t a i n c l a u s e s w i t h r e s p e c t t o t h e s t r u c t u r a l p r o p e r t i e s o f a t e x t . The c a t e g o r i e s a r e termed I n t e n t i o n ( g o a l s , conceptions, wishes), A c t i o n ( s t a t e changes), and S t a t e ( c o n d i t i o n s o f t h e w o r l d ) . I t i s assumed t h a t t h e f u n c t i o n a l meaning o f a c l a u s e may be d e r i v e d from t h e semantic i m p o r t o f i t unambiguosly and i n d e p e n d e n t l y o f s u r r o u n d i n g c l a u ses. M i c r o s t r u c t u r a l u n i t s b e l o n g i n g t o t h e same c a t e g o r y w i l l be f u r t h e r d i f f e r e n t i a t e d w i t h r e s p e c t t o t h e a c t i o n schema o f which t h e y a r e a cons t i t u e n t and w i t h r e s p e c t t o whether t h e main c h a r a c t e r i s i n v o l v e d . These s u b c a t e g o r i e s a r e d e t a i l e d below: 1. I n t e n t i o n 1.1 Goal ( d e s i r e f o r t h e o c c u r r e n c e o f a c e r t a i n e v e n t o r s t a t e ) 1 . 2 Conception ( f o r m a t i o n o f a PLAN) 1.3 S e l e c t ( s e l e c t i o n o f a c e r t a i n method)
114
COHERENCE
2. A c t i o n 2.1 Do ( a p p l i c a t i o n o f a s e l e c t e d method) 2.2 Event o r O b s t a c l e (an o b j e c t i v e o c c u r r e n c e o r an a c t i o n by another character) 3. S t a t e 3 . 1 p o s i t i v e o r n e g a t i v e Outcome ( o f an ACTION) 3.2 p o s i t i v e o r n e g a t i v e Consequence ( o f a PLAN o r an EPISODE) 3.3 S e t t i n g o r O b s t a c l e (a c o n d i t i o n o f t h e w o r l d e x i s t i n g indep e n d e n t l y o f a main c h a r a c t e r ' s ACTION) A l l I n t e n t i o n - s u b c a t e g o r i e s , t h e ACTION-subcategory Do, and t h e S t a t e s u b c a t e g o r i e s Outcome and Consequence i n v o l v e t h e main c h a r a c t e r . A l l these s u b c a t e g o r i e s form a c o n s t i t u e n t o f one o f t h e a c t i o n schemata t h a t w i l l be d e s c r i b e d i n t h e n e x t s e c t i o n . Whether an Outcome o r a Consequence i s pos i t i v e o r n e g a t i v e and whether an Event o r a S e t t i n g i s an O b s t a c l e i s det e r m i n e d r e l a t i v e t o t h e c u r r e n t I n t e n t i o n o f t h e main c h a r a c t e r . The a c t i o n schemata assumed by t h e R o t e r Faden model a r e termed EPISODE, PLAN, and ACTION; t h e y a r e g i v e n i n t h e f o r m o f r e w r i t e r u l e s :
+
[~~~~~N,i 1 +
1. EPISODE
=
Goal
2. PLAN
=
Conception +
Consequence
p ACTIONS
and / o r q PLANS
+
Consequence
+
( w i t h (p+q)>2; p.qelNo) 3. ACTION
= Select
+
Do + Outcome
Note t h a t t h e s e schemata a l l o w f o r r e c u r s i v e a p p l i c a t i o n . The a c t i o n schemata express o u r a b s t r a c t knowledge a b o u t t h e p u r p o s i v e b e h a v i o r of human beings. They a r e assumed t o g u i d e b e h a v i o r as w e l l as t h e comprehension of d i s c o u r s e t h a t d e s c r i b e s g o a l - d i r e c t e d human a c t i v i t i e s . Note t h a t t h e cons t i t u e n t s o f t h e s e schemata c o v e r most o f t h e m i c r o s t r u c t u r a l u n i t s i n t r o duced i n t h e p r e v i o u s s e c t i o n ; o n l y S e t t i n g , Event, and O b s t a c l e do n o t o c c u r and an e v a l u a t i o n o f u n i t s Outcome and Consequence i s m i s s i n g . T h i s i s m o t i v a t e d by t h e f a c t t h a t t h e s e elements a r e n o t c o n s i d e r e d p r o p e r t i e s o f o u r a b s t r a c t knowledge a b o u t t h e t y p i c a l s t r u c t u r e o f a c t i o n sequences. N e v e r t h e l e s s , one can say t h a t access t o t h e s e schemata would g r e a t l y fac i l i a t e t h e t e x t comprehension process: T h i s would a l l o w f o r g e n e r a t i n g i n f e r e n c e s about i n f o r m a t i o n n o t e x p l i c i t l y s t a t e d i n a t e x t as w e l l as e x p e c t a t i o n s a b o u t t e x t i n f o r m a t i o n y e t t o come. Moreover, s i n g l e m i c r o s t r u c t u r a l u n i t s may be conceived o f as b e l o n g i n g t o l a r g e r s t r u c t u r a l u n i t s o f a t e x t and m i g h t t h e r e f o r e be summarized a c c o r d i n g t o t h e s e a c t i o n schemata. T h u s , t h e a c t i o n schemata correspond t o t h e m a c r o s t r u c t u r a l u n i t s o f t h e R o t e r Faden model. There i s a problem, however: The a c t i o n sequences r e a l i z e d i n many pieces o f d i s c o u r s e d i s p l a y c o n s i d e r a b l e d e v i a t i o n s from t h e t y p i c a l f o r m expressed by t h e schemata. C e r t a i n s t a t e s o f t h e w o r l d o r events may d e t a i n t h e main c h a r a c t e r from p u r s u i n g h i s g o a l s . Moreover, t h e a c t i o n schemata do n o t c o n t a i n any i n f o r m a t i o n a b o u t t h e p o s s i b l e c o n n e c t i o n s between s u c c e s s i v e m i c r o s t r u c t u r a l u n i t s , f o r example, under which c o n d i t i o n s i t i s
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
115
p o s s i b l e t h a t t h e Outcome o f a c e r t a i n ACTION i s succeeded by t h e Concepti o n o f a new PLAN. I t i s assumed t h a t t h e s e c o n d i t i o n s c o n s i s t i n c o n f i g u r a l r e s t r i c t i o n s based on s u r r o u n d i n g m i c r o s t r u c t u r a l u n i t s . I n t h e R o t e r Faden model t h e c o n d i t i o n s f o r c o n n e c t i n g s u c c e s s i v e m i c r o s t r u c t u r a l u n i t s a r e r e a l i z e d by c o n f i g u r a l r u l e s . These r u l e s d e f i n e which two f u n c t i o n a l u n i t s may f o l l o w one a n o t h e r i n t h e m i c r o s t r u c t u r e o f a t e x t . A s i m p l i f i e d v e r s i o n o f one o f t h e r u l e s w i l l serve as an i l l u s t r a t i o n . Note t h a t t h e m i c r o s t r u c t u r a l u n i t s i n q u e s t i o n a r e under1 i n e d : Obstacle p
-f
( a f t e r t h e s e l e c t i o n o f a method t o a c h i e v e goal r )
L
Goal r + 1: Remove o b s t a c l e p ( t 1) S e l e c t a l t e r n a t i v e method f o r goal r N e g a t i v e Consequence f o r goal r ( - 1 )
T h i s r u l e says t h a t whenever a n O b s t a c l e b l o c k s t h e a p p l i c a t i o n (Do) o f a s e l e c t e d method, t h e main c h a r a c t e r has e x a c t l y t h r e e s t r a t e g i e s t o choose from: To s e t h i m s e l f t h e t o remove t h e o b s t a c l e , o r t o S e l e c t an a l t e r n a t i v e method t h a t w i l l p r e v e n t him f r o m e n c o u n t e r i n g t h e experienced o b s t a c l e , o r f i n a l l y t o g i v e up t h e i n t e n d e d g o a l ( N e g a t i v e Consequence). The numbers i n parantheses i n d i c a t e t h a t t h e c o u r s e o f a c t i o n goes i n t o more d e t a i l o r o n t o a more g l o b a l l e v e l . T h i s i s i n d i c a t e d i n t h e memory r e p r e s e n t a t i o n by changes i n t h e h i e r a r c h y l e v e l , i . e . , one l e v e l down ( + 1) o r one l e v e l up ( - 1 ) .
Goal
Both o f t h e s e knowledge s t r u c t u r e s , namely a c t i o n schemata and c o n f i g u r a l r u l e s a r e u t i l i z e d by t h e p r o c e s s i n g system o f t h e R o t e r Faden model. As an example, t h e f i n a l r e p r e s e n t a t i o n o f one o f t h e t e x t s used i n a r e c o g n i t i o n experiment t o t e s t t h e R o t e r Faden model i s shown i n F i g u r e 1. The symbol s t r i n g connected by d o t t e d l i n e s and r u n n i n g a c r o s s d i f f e r e n t h i e r a r c h i c a l l e v e l s r e p r e s e n t s t h e R o t e r Faden, o r m a i n t r a i n o f thought, o f t h a t s p e c i f i c t e x t i n t e r r u p t e d by dead-end s t r u c t u r e s ( u n s u c c e s s f u l a c t i o n sequences). Both c o n s i s t o f e x p l i c i t , i . e . , a c t u a l l y s t a t e d i n t h e t e x t , o r i m p l i c i t , i.e., i n f e r r e d , f u n c t i o n a l u n i t s , together w i t h t h e i r respective c o n n e c t i o n s a c c o r d i n g t o t h e c o n f i g u r a l r u l e s d e f i n e d i n t h e knowledge s t r u c t u r e . T h i s s t r i n g o f connected f u n c t i o n a l u n i t s i s e q u i v a l e n t t o t h e m i c r o s t r u c t u r e o f t h e t e x t . The c i r c l e d symbols E , P , and H r e p r e s e n t t h e t h r e e d i f f e r e n t t y p e s o f m a c r o p r o p o s i t i o n s a c c o r d i n g t o a c t i o n schemata, a g a i n d e f i n e d i n t h e knowledge s t r u c t u r e . These m a c r o p r o p o s i t i o n s , t o gether w i t h t h e i r connecting continuous l i n e s , represent t h e t h e o r e t i c a l m a c r o s t r u c t u r e . T h i s concept o f a m a c r o s t r u c t u r e r e p r e s e n t e d by nodes and e x t r a c o n n e c t i o n s between c o h e r e n t f u n c t i o n a l u n i t s o f t h e m i c r o s t r u c t u r e w i l l be f u r t h e r i n v e s t i g a t e d l a t e r on. B u t f i r s t , we s h a l l g i v e an o u t l i n e o f a n experiment t h a t was conducted t o t e s t some p r o p e r t i e s o f t h e memory r e p r e s e n t a t i o n produced by t h e R o t e r Faden model.
AN EXPERIMENTAL TEST OF THE ROTER FADEN MODEL As has a l r e a d y been i n d i c a t e d , t h e model a l l o w s f o r d i s c r i m i n a t i o n of t h e Roter Faden and dead-ends o f a t e x t on t h e one hand, and more o r l e s s comp l e x u n i t s i n t h e m a c r o s t r u c t u r e on t h e o t h e r . These two a r e , t o o u r knowledge, t h e o n l y v a r i a b l e s f o r which c o n s i s t e n t e m p i r i c a l r e s u l t s have been o b t a i n e d i n experiments designed t o t e s t s t r u c t u r a l p r o p e r t i e s o f memory r e p r e s e n t a t i o n s f o r p i e c e s o f d i s c o u r s e . The f i r s t v a r i a b l e has proved t o be e f f e c t i v e i n s t u d i e s conducted by B l a c k (1978), G l o w a l l a (1981a), and Omanson (1979), t h e second v a r i a b l e i n a s t u d y by B l a c k and Bower (1979).
0
0
I m W
m
z
0
m
35
4.45
Figure 1 R e p r e s e n t a t i o n Graph o f a T e x t According t o t h e R o t e r Faden Model The Numbers R e f e r t o t h e M i c r o p r o p o s i t i o n s o f t h e T e x t . E x p l a n a t i o n o f Symbols: E: EPISODE; P: PLAN; H: ACTION; I I , I , A c t: i o n w i t h Pos.Outcome, Neg.Outcome, o r Obstacle; , rn : Pos. o r Neg.Outcome;O , : Pos. o r Neg.Consequence; A , A : S e t t i n g / E v e n t o r Obs t a c l e ; 0 : Goal; D : Conception; : Dead-End S t r u c t u r e s .
+
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
117
Comparable p r e d i c t i o n s a r e a l s o made by Schank and A b e l s o n ' s (1977) "Conceptual Dependency / Know1 edge S t r u c t u r e " computer s i m u l a t i o n model. Tests of o t h e r s t r u c t u r a l p r o p e r t i e s o f memory r e p r e s e n t a t i o n s o f d i s c o u r s e l i k e membership o f d i f f e r e n t e p i s o d i c c a t e g o r i e s o r p o s i t i o n i n t h e t e x t s t r u c t u r e h i e r a r c h y - have l e d t o i n c o n s i s t e n t r e s u l t s (Omanson, 1979; Trabasso, 1980). Nevertheless, i t i s n o t c l e a r which o f t h e two v a r i a b l e s i s t h e more e f f e c t i v e . The reason f o r t h i s l i e s i n t h e f a c t t h a t i n experiments i n v e s t i g a t i n g one o f t h e s e v a r i a b l e s , p o s s i b l e e f f e c t s o f t h e o t h e r v a r i a b l e have not been c o n t r o l l e d f o r . I n t h e f o l l o w i n g r e c o g n i t i o n experiment, t h e s e two v a r i a b l e s have been v a r i e d i n d e p e n d e n t l y . I n t h i s experiment s u b j e c t s were t e s t e d i n t h r e e d i f f e r e n t s e s s i o n s w i t h e x a c t l y t h e same procedure. Each s e s s i o n s t a r t e d w i t h a s t u d y p e r i o d i n which two t e x t s were presented a c o u s t i c a l l y t o t h e s u b j e c t s . D u r i n g each s e s s i o n , a d i f f e r e n t p a i r o f t e x t s was used. A f t e r an i n t e r v e n i n g p e r c e p t i o n t a s k , t h e s u b j e c t s were t e s t e d on r e t e n t i o n o f t h e p r e v i o u s l y l e a r n e d t e x t p a i r i n t h e f o l l o w i n g way: On each t r i a l two sentences stemming f r o m t h e t e x t p a i r were p r e sented v i s u a l l y . S u b j e c t s had t o d e c i d e whether b o t h sentences o r i g i n a t e d from t h e same o r d i f f e r e n t t e x t s . Percentages o f c o r r e c t answers and r e sponse l a t e n c i e s were measured. A l l probe sentences o r i g i n a t e d from t h r e e d i f f e r e n t t y p e s o f m a c r o s t r u c t u r a l u n i t s , namely a PLAN d o m i n a t i n g an EPISODE, a PLAN c o n s i s t i n g o f s e v e r a l ACTIONS, and f i n a l l y a s i n g l e ACTION. F o r each o f t h e s e s t r u c t u r e t y p e s t h e r e e x i s t e d two comparable u n i t s , one belonging t o t h e R o t e r Faden and t h e o t h e r b e i n g a dead-end. As a n example, look a t t h e t e x t r e p r e s e n t a t i o n shown i n F i g u r e 1. The m a c r o s t r u c t u r a l u n i t s PLAN w i t h embedded EPISODE c o n s i s t o f t h e m i c r o s t r u c t u r e u n i t s 30 t o 38 and 39 t o 47, r e s p e c t i v e l y , t h e f i r s t b e i n g t h e dead-end s t r u c t u r e and the second b e i n g t h e R o t e r Faden s t r u c t u r e . B o t h sentences of a p o s i t i v e probe stemmed from t h e same m a c r o s t r u c t u r a l u n i t and b o t h e i t h e r belonged t o t h e R o t e r Faden s t r u c t u r e o r t o t h e c o r r e s p o n d i n g dead-end s t r u c t u r e . This i s a l s o t r u e f o r n e g a t i v e probes w i t h one i m p o r t a n t e x c e p t i o n : One sentence o r i g i n a t e d f r o m t h e f i r s t t e x t and one from t h e second t e s t . The observed r e a c t i o n t i m e s have been analysed by s e v e r a l ANOVAs; t h e main r e s u l t s a r e as f o l l o w s : Ifa p i e c e o f i n f o r m a t i o n belongs t o t h e R o t e r Faden, i t r e c e i v e s some s o r t o f s t r u c t u r a l b e n e f i t i n s o f a r as a s t r o n g decay i n a v a i l a b i l i t y i s compensated f o r when i t o r i g i n a t e s f r o m a l e s s i m p o r t a n t m a c r o s t r u c t u r a l u n i t . I f t h e p i e c e o f i n f o r m a t i o n stems from a dead-end s t r u c t u r e , t h e decay i n a v a i l a b i l i t y w i l l be i n t e n s i f i e d . These r e s u l t s l e d t o t h e c o n c l u s i o n t h a t any model must account f o r t h e e f f e c t s of b o t h o f t h e s e p r o p e r t i e s . The s u b s t a n t i a l l y s t r o n g e r e f f e c t s o f membership i n more o r l e s s complex m a c r o s t r u c t u r e u n i t s l e d t o t h e f u r t h e r conc l u s i o n t h a t t h e m a c r o s t r u c t u r a l o r g a n i z a t i o n o f t e x t s should be i n v e s t i gated more t h o r o u g h l y i n t h e f u t u r e . F o r d e t a i l s o f t h i s experiment see Glowalla (1981). PRELIMINARY RESULTS ON MACROSTRUCTURAL ORGANIZATION
AS a m a t t e r o f f a c t , p a r t of t h e d a t a o f t h e experiment j u s t d e s c r i b e d a l l o w s f o r some p r e l i m i n a r y i n s i g h t s i n t o t h e s e phenomena. D u r i n g t h e stepwise c o n s t r u c t i o n of t h e memory r e p r e s e n t a t i o n f o r a g i v e n t e x t according t o t h e R o t e r Faden model, m a c r o s t r u c t u r a l c o n n e c t i o n s between coherent m i c r o s t r u c t u r a l u n i t s ( t h o s e b e l o n g i n g t o t h e same m a c r o s t r u c t u r a1 u n i t ) a r e e s t a b l i s h e d v i a a node r e p r e s e n t i n g t h a t p i e c e o f t h e macros t r u c t u r e as a whole. P a i r s o f m i c r o s t r u c t u r a l u n i t s d i f f e r w i t h r e s p e c t
118
COHERENCE
t o t h e i r d i s t a n c e i n t h e m a c r o s t r u c t u r e . The d i s t a n c e i s expressed by t h e number o f l i n k s on t h e p a t h between two r e s p e c t i v e u n i t s . I f , f o r example, two m i c r o s t r u c t u r a l u n i t s a r e dominated by t h e same m a c r o s t r u c t u r e node H, t h e i r d i s t a n c e i s 2, i f t h e y a r e dominated by t h e same node P b u t belong t o d i f f e r e n t A C T I O N S , t h e i r d i s t a n c e i s 4. I t may be asked whether d i f f e r e n t response l a t e n c i e s a r e t o be observed f o r probes h a v i n g d i f f e r e n t d i s t a n c e s i n t h e m a c r o s t r u c t u r e . Assuming equal search r a t e s f o r a l l l i n k s i n t h e m a c r o s t r u c t u r e , search t i m e s s h o u l d i n c r e a s e w i t h i n c r e a s i n g distance. T h i s p r e d i c t i o n can be t e s t e d by i n s p e c t i o n o f p a r t o f t h e d a t a f r o m t h e d e s c r i b e d r e c o g n i t i o n experiment. I n s t r u c t u r e t y p e 1 (PLAN w i t h an embedded EPISODE) t h e r e a r e probes w i t h d i s t a n c e s o f 2, 4, o r 6 l i n k s , r e s p e c t i v e l y . I f one t a k e s t h e d a t a o f a l l s i x t e x t s t h a t o r i g i n a t e from s t r u c t u r e t y p e 1 and computes mean r e a c t i o n t i m e s f o r t h e t h r e e d i s t i n c t c a t e g o r i e s s e p a r a t e l y , t h e r e s u l t s a r e as f o l l o w s : 2443 ms, 2546 ms, and 2742 ms f o r probes w i t h d i s t a n c e s 2, 4, and 6 l i n k s . The observed d i f f e rences of 103 ms and 196 ms a r e s t a t i s t i c a l l y r e l i a b l e ( F 2 , a s = 68.28; p < . 0 0 1 ) . S i m i l a r r e a c t i o n t i m e d i f f e r e n c e s can be o b t a i n e d i f t h i s a n a l y s i s i s performed f o r each t e x t s e p a r a t e l y , b u t f o r two o u t o f s i x t e x t s t h e s e d i f f e r e n c e s a r e n o t s i g n i f i c a n t . T h i s r e s u l t may be t a k e n as s u p p o r t i n g evidence f o r t h e m a c r o s t r u c t u r e assumptions o f t h e R o t e r Faden model. However, one can o b j e c t t h a t t h e v a r i a b l e d i s t a n c e s i n t h e m a c r o s t r u c t u r e a r e confounded w i t h a n o t h e r v a r i a b l e , namely t h e number o f c l a u s e s between t h e two sentences o f a probe i n t h e s u r f a c e s t r u c t u r e o f a g i v e n t e x t . I n f a c t , t h i s v a r i a b l e should a f f e c t response l a t e n c i e s , s i n c e t h e mean number o f coherence r e l a t i o n s between two sentences i s i n v e r s e l y r e l a t e d t o t h e number o f i n t e r v e n i n g sentences ( C l a r k and Sengul, 1979; Lesgold, Roth and C u r t i s , 1979). Whether t h i s p r o p e r t y had any i n f l u e n c e on t h e observed response t i m e s has been i n v e s t i g a t e d by a s e p a r a t e a n a l y s i s i n v o l v i n g probes o f s t r u c t u r e t y p e 2 (PLAN c o n s i s t i n g o f s e v e r a l ACTIONS). The sentences o f t h e s e probes possess t h e same d i s t a n c e of 4 l i n k s i n t h e macros t r u c t u r e t o g e t h e r w i t h 0, 1, o r 2 i n t e r v e n i n g sentences. Over a l l s i x t e x t s used i n t h e experiment, mean r e a c t i o n t i m e s o f 2716 ms, 2811 ms, and 2861 ms were observed f o r probes b e l o n g i n g t o t h e s e d i f f e r e n t c a t e g o r i e s . The r e a c t i o n t i m e d i f f e r e n c e s o f 95 ms and 50 ms a r e s t a t i s t i c a l l y s i g n i f i c a n t ( F Z , ~= ~7.99; p < . 0 1 ) . W i t h m i n o r q u a l i f i c a t i o n s s i m i l a r d i f f e r e n c e s were observed f o r e v e r y s i n g l e t e x t . As i t stands, t h e d i s t a n c e e f f e c t m i g h t be e x p l a i n e d by t h e number o f i n t e r v e n i n g sentences as w e l l . To c l a r i f y whether t h e observed d i f f e r e n c e s i n response l a t e n c i e s a r e a f f e c t e d by b o t h o f t h e s e v a r i a b l e s o r depend e x c l u s i v e l y on t h e number o f i n t e r v e n i n g sentences w i t h r e s p e c t t o t h e sentences o f a probe, a t h i r d a n a l y s i s was conducted. A subset o f probes b e l o n g i n g t o s t r u c t u r e t y p e 1 proved t o be s u i t a b l e f o r such an a n a l y s i s . The sentences o f t h e s e probes a r e separated by two i n t e r v e n i n g sentences i n t h e s u r f a c e s t r u c t u r e o f t h e r e s p e c t i v e t e x t and connected v i a 4 o r 6 l i n k s i n t h e m a c r o s t r u c t u r e . C o n s i d e r i n g such probes o f a l l s i x t e x t s t o g e t h e r , mean r e a c t i o n t i m e s o f 2606 ms and 2822 ms were observed. The d i f f e r e n c e o f 216 ms i s s t a t i s t i c a l l y s i g n i f i c a n t (F1,+$ = 23.26; p < .001). Again, s i m i l a r r e s u l t s a r e o b t a i n e d , i f s e p a r a t e ANOVAs a r e conducted f o r s i n g l e t e x t s . A j o i n t consideration o f a l l t h r e e analyses l e d t o t h e f o l l o w i n g c o n c l u s i o n : Both v a r i a b l e s , namely t h e d i s t a n c e i n t h e macros t r u c t u r e as w e l l as t h e number o f i n t e r v e n i n g sentences as a measure of
TOWARD A MODEL OF MACROSTRUCTURE SEARCH coreferentiality,
119
a f f e c t t h e u n d e r l y i n g response process.
A MODEL FOR MACROSTRUCTURE SEARCH I n t h i s s e c t i o n , we g i v e a n o u t l i n e o f a q u a n t i t a t i v e model f o r t h e search processes hypothesized i n t h e above experiment. I d e a l l y , one would expect such a model t o account f o r a l l main f e a t u r e s o f t h e d a t a , t h a t i s , t h e dependence o f r e a c t i o n t i m e s and p r o p o r t i o n of c o r r e c t answers on t h e v a r i o u s probe t y p e s . I t should be c l e a r f r o m t h e o u t s e t t h a t what f o l l o w s i s o n l y meant as a s t a r t i n g p o i n t toward such a model. Moreover, t h e model b u i l d i n g h e r e i s a post-hoc e n t e r p r i s e and t h u s , no model t e s t i n g i s y e t involved. I t i s common i n r e a c t i o n t i m e - s t u d i e s o f l o n g - t e r m memory search t o conc e i v e o f t o t a l r e a c t i o n t i m e (RT) as an a d d i t i v e c o m p o s i t i o n o f two p a r t s : a search t i m e ( S ) depending on t h e e x p e r i m e n t a l c o n d i t i o n , and a r e s i d u a l t i m e ( R ) c o n s i s t i n g o f encoding and motor components t h a t a r e c o n s t a n t over e x p e r i m e n t a l c o n d i t i o n s ( c f . P a c h e l l a , 1 9 7 4 ) :
RT=S+R An i m p o r t a n t p r e s u p p o s i t i o n i n t h e f o l l o w i n g argument i s t h a t t h e S component o f t h e r e a c t i o n t i m e i n t h e above experiment can be t a k e n as generated by a search process o f t h e m a c r o s t r u c t u r e t h a t resembles t h e spreading a c t i v a t i o n assumptions used i n semantic network models ( c f . King and Anderson, 1976). A f i n d i n g s u p p o r t i n g t h i s v i e w i s t h e afore-mentioned i n c r e a s e i n mean r e a c t i o n t i m e as a f u n c t i o n o f t h e number o f l i n k s separ a t i n g two probe sentences i n t h e m a c r o s t r u c t u r e graph r e p r e s e n t a t i o n . Bef o r e i n t r o d u c i n g some s p e c i f i c a s p e c t s o f t h i s search process, two remarks a r e a p p o s i t e . F i r s t l y , w h i l e o u r d a t a a n a l y s i s suggested an e f f e c t o f t h e number o f m i c r o s t r u c t u r e l i n k s between t h e probe sentences on RT, t h i s has n o t y e t been i n c o r p o r a t e d i n t o t h e model assumptions, s i n c e t h e experiment was n o t e s p e c i a l l y designed t o e x p l o r e t h e n a t u r e o f t h i s e f f e c t . The obvious a l t e r n a t i v e s t o handle t h i s would be e i t h e r t o p o s t u l a t e a p p r o p r i a t e search processes on t h e m i c r o s t r u c t u r e l i n k s o r t o l e t some parameters o f t h e m a c r o s t r u c t u r e search depend on t h e m i c r o s t r u c t u r e d i s t a n c e between t h e two probe sentences. Secondly, a s i m i l a r p o i n t should be made f o r t h e e f f e c t o f a n o t h e r v a r i a b l e , t h e m a c r o s t r u c t u r e t y p e which t h e probe sentences f o r m p a r t o f . One way t o t a k e t h i s e f f e c t i n t o a c c o u n t would be t o break down t h e search t i m e ( S ) i n t o a n o t h e r two a d d i t i v e subcomponents, one being t h e a c t u a l t i m e t o a c t i v a t e a l l necessary l i n k s , t h e o t h e r r e p r e s e n t i n g t h e t i m e t o g a i n access t o t h e m i c r o s t r u c t u r e r e p r e s e n t a t i o n o f t h e probes, where t h e l a t t e r t i m e component may d i f f e r f r o m one m a c r o s t r u c t u r e t y p e t o t h e n e x t . The f o l l o w i n g o n l y r e f e r s t o t h e a c t i v a t i o n search p r o cess i t s e l f : ( a ) each node and l i n k e i t h e r i s i n t h e a c t i v e s t a t e o r n o t ; ( b ) i f a l i n k i s i n t h e a c t i v e s t a t e , t h e n t h e two nodes connected by t h e l i n k are also active; ( c ) t h e t i m e f o r a c t i v a t i o n t o spread from an a c t i v e node t o an i n c i d e n t l i n k i s a n e x p o n e n t i a l l y d i s t r i b u t e d random v a r i a b l e w i t h a r a t e i n v e r s e l y r e l a t e d t o t h e t o t a l number o f l i n k s i n c i d e n t t o t h a t node; ( d ) a c t i v a t i o n spreads i n d e p e n d e n t l y f r o m b o t h probe nodes and t h e s u b j e c t g i v e s an answer whenever a p a t h c o n n e c t i n g b o t h probe nodes i s a c t iva t e d .
120
COHERENCE
The e x p o n e n t i a l d i s t r i b u t i o n i s used m a i n l y f o r mathematical t r a c t a b i l i t y : 1 ) t h e minimum o f a f i n i t e number o f independent e x p o n e n t i a l l y d i s t r i b u t e d random v a r i a b l e s w i t h r a t e s r . i s a g a i n e x p o n e n t i a l l y d i s t r i b u t e d w i t h r a t e J C r.
j
J '
and 2) t h e p r o b a b i l i t y o f t h e i - t h v a r i a b l e b e i n g t h e minimum i s ri/C r . J J
.
The assumption t h a t t h e r a t e o f a c t i v a t i o n s p r e a d i n g from a node towards an i n c i d e n t l i n k depends on t h e t o t a l number o f l i n k s i n c i d e n t w i t h t h a t node i s m o t i v a t e d by t h e s o - c a l l e d f a n - e f f e c t i n semantic network r e s e a r c h ( c f . K i n g and Anderson, 1976). Given ( a ) - ( d ) i t i s a r o u t i n e m a t t e r t o compute t h e expected search t i m e s f o r any m a c r o s t r u c t u r e c o n f i g u r a t i o n . L e t u s i l l u s t r a t e t h i s w i t h an example o f a m a c r o s t r u c t u r e grapii segment ( v i s . F i g u r e 2 ) . Suppose t h e m i c r o s t r u c t u r e nodes l a b e l e d 2 and 4 a r e a c t i v a t e d by t h e probe sentences. I n o r d e r t o compute t h e expected t i m e t h e p a t h a-b-c-d from 2 t o 4 needs t o be a c t i v a t e d , we have t o c o n s i d e r a l l p o s s i b l e ways t h i s a c t i v a t i o n may come about. Suppose t h e o r d e r o f a c t i v a t i o n o f t h e l i n k s i s a, b, d, c. I f r i s t h e parameter o f t h e e x p o n e n t i a l d i s t r i b u t i o n , t h e expected t i m e f o r a t o be a c t i v a t e d f r o m 2 i s l / r ; f o r b , t h e t i m e t o be a c t i v a t e d i s 3 / r , since there a r e t h r e e l i n k s i n c i d e n t w i t h H2 not y e t a c t i v e ; then f o r d , t h e t i m e i s a g a i n l / r , w h i l e f o r e , we have 1 / 2 r , s i n c e a c t i v a t i o n may spread f r o m b o t h E and H 4 . Thus t h e expected t i m e f o r t h e o r d e r a,b,d,c i s 5 . 5 / r . The p r o b a b i l i t y o f t h i s o r d e r o c c u r r i n g i s 1 / 2 ( a " w i n n i n g " a g a i n s t d ) t i m e s 1/4 ( b " w i n n i n g " a g a i n s t d) t i m e s 1 / 2 ( d " w i n n i n g " against e) = 1/16. I n o r d e r t o g e t t h e expected t o t a l a c t i v a t i o n t i m e we have t o sum over the expected a c t i v a t i o n t i m e s o f a l l p o s s i b l e o r d e r s weighted by t h e i r r e s p e c t i v e p r o b a b i l i t i e s o f occurrence. I t i s n o t d i f f i c u l t t o show t h a t f o r a p a t h o f n l i n k s t h e r e a r e 2"-' d i f f e r e n t p o s s i b l e a c t i v a t i o n o r d e r s . However, i n most cases Thus, c o m p u t a t i o n l o a d i n c r e a s e s q u i c k l y w i t h n n i s o n l y o f moderate s i z e and, i n any event, t h e t a s k s h o u l d be amenable t o programming.
.
I n t h e above example, t h e expected t o t a l t i m e t o a c t i v a t e t h e 4 l i n k s conn e c t i n g t h e t e s t probes 2 and 4 can be shown t o be about 5 / r by e v a l u a t i n g a l l 8 d i f f e r e n t a c t i v a t i o n o r d e r s . The c o r r e s p o n d i n g t i m e f o r t e s t probes 1 and 2 w i t h o n l y 2 c o n n e c t i n g l i n k s i s 1 . 7 5 / r . I t can be shown more gen e r a l l y t h a t , no m a t t e r what t h e s p e c i f i c c o n f i g u r a t i o n , t h e expected t o t a l t i m e i s an i n c r e a s i n g f u n c t i o n o f t h e number o f l i n k s c o n s t i t u t i n g t h e p a t h c o n n e c t i n g t h e two t e s t probes. As i t stands, t h e model a l l o w s some f a i r l y d e f i n i t e p r e d i c t i o n s t o be made. I n a d d i t i o n t o t h e m o n o t o n i c i t y p r o p e r t y j u s t mentioned, one would expect equal t o t a l a c t i v a t i o n t i m e s - a p a r t from s t a t i s t i c a l v a r i a b i l i t y - f o r a number o f t e s t probe p a i r s due t o an i n h e r e n t symnetry i n t h e macrostruct u r e graph, f o r example f o r t h e p a i r s 1 - 3, 2 - 3 , 1 - 4, and 2 - 4 . I t s h o u l d be p o i n t e d o u t , however, t h a t any c o n c l u s i o n drawn f r o m a t e s t o f t h e s e p r e d i c t i o n s i s o n l y v a l i d i f t h e i n f l u e n c e o f t h e o t h e r f a c t o r s on RT d i s c u s s e d above has been accounted f o r a p p r o p r i a t e l y .
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
121
Figure 2 Segment o f a M a c r o s t r u c t u r e Tree The Numbers r e f e r t o M i c r o s t r u c t u r e , t h e C a p i t a l L e t t e r s t o M a c r o s t r u c t u r e Nodes.
CONCLUSION
I t has been argued t h a t t h e m a c r o s t r u c t u r e o f a t e x t can be r e p r e s e n t e d by a graph, d i v i d i n g c o h e r e n t m i c r o s t r u c t u r a l u n i t s i n t o c l u s t e r s a t d i f f e r e n t h i e r a r c h i c a l l e v e l s . The q u a n t i t a t i v e model o u t l i n e d i n t h e p r e v i o u s s e c t i o n makes some d e f i n i t e p r e d i c t i o n s about search processes on m a c r o s t r u c t u r e graphs. Most i m p o r t a n t l y , t h e t i m e t o v e r i f y a probe should be a n i n c r e a s i n g f u n c t i o n of t h e number o f l i n k s c o n s t i t u t i n g t h e p a t h c o n n e c t i n g t h e two sentences o f t h e p r o b e i n t h e m a c r o s t r u c t u r e . The observed response l a t e n c i e s i n t h e r e p o r t e d r e c o g n i t i o n experiment a r e i n accordance w i t h t h i s p r e d i c t i o n . However, as has a l r e a d y been p o i n t e d o u t , t h e experiment was n o t e s p e c i a l l y designed t o i n v e s t i g a t e search processes on macros t r u c t u r e s . Thus, no r e a s o n a b l e t e s t o f t h e proposed model c o u l d be p e r formed. We s h a l l t h e r e f o r e c o n c l u d e w i t h some s u g g e s t i o n s a b o u t t h e d e s i g n o f such a n experiment. R e t a i n i n g t h e r e c o g n i t i o n t e c h n i q u e presented, one e f f e c t i v e m o d i f i c a t i o n c o u l d c o n s i s t o f t h e a d d i t i o n a l employment o f probes w i t h sentences stemming f r o m d i f f e r e n t s t r u c t u r e t y p e s . I n an experiment i n v e s t i g a t i n g the j o i n t representation o f several p a r t i a l l y coherent t e x t s (Glowalla, 1980), r e l i a b l e r e a c t i o n - t i m e d i f f e r e n c e s between probes f r o m t h e same o r d i f f e r e n t s t r u c t u r e t y p e s have been observed. By a l t e r i n g t h e d i s t a n c e o f
122
COHERENCE
m a c r o s t r u c t u r e u n i t s o u t o f w h i c h t h e two sentences o f a probe a r e taken s i m u l t a n e o u s l y , a s t r o n g t e s t o f t h e proposed q u a n t i t a t i v e model should be p o s s i b l e . Another m o d i f i c a t i o n m i g h t c o n s i s t i n a v a r i a t i o n o f t h e pres e n t a t i o n o r d e r o f probe sentences. T h i s v a r i a b l e has p r o v e n t o be h i g h l y e f f e c t i v e i n r e s e a r c h on a c t i v a t i o n o f semantic networks ( G l o w a l l a , Schulze and Wender, 1980). We assume t h a t an a l t e r a t i o n o f p r e s e n t a t i o n o r d e r s h o u l d f u n d a m e n t a l l y i n f l u e n c e t h e e f f e c t s o f t h e number o f m i c r o s t r u c t u r e l i n k s ( a s a measure o f c o r e f e r e n t i a l i t y ) on RT. I n t h e experiment reported here, t h e o r d e r o f probe sentences corresponded t o t h e i r appearance i n the s u r f a c e s t r u c t u r e o f a t e x t . The o p p o s i t e o r d e r should r e s u l t i n a dec r e a s e o f c o r e f e r e n t i a l i t y e f f e c t s on RT. The u s e o f t h e s e two and o t h e r s i m i l a r m o d i f i c a t i o n s i n experiments on m a c r o s t r u c t u r e search should lead t o a deeper u n d e r s t a n d i n g o f t h e n a t u r e of m a c r o s t r u c t u r e s and t h e i r applic a t i o n d u r i n g the processing o f discourse. REFERENCES B l a c k , J.B. 1978.
Story memory s t r u c t u r e s . PhD Thesis, S t a n f o r d U n i v e r s i t y ,
Episodes as chunks i n n a r r a t i v e memory. B l a c k , J.B., and Bower, G.H. Journal of Verbal Learning and Verbal aehuvior, 1979, 18, 309 - 318. I n search o f r e f e r e n t s f o r nouns and proC l a r k , H.H., and Sengul, C.J. nouns. Memory and Cognition, 1979, 7 , 35 - 41. Der Rote. Fuden - e i n handlungstheoretisches Model1 zur TextGlowalla, U . verarbeitung. D i s s e r t a t i o n , U n i v e r s i t a t Braunschweig, 1981. Reproduzieren und Zusammenfassen von Texten. Sprache und G l o w a l l a , U. Kognition, i n p r e s s .
Probleme b e h Enkodieren und Dekodieren mehrerer ProblemGlowalla, U . ZBsegeschichten. Unpublished m a n u s c r i p t , U n i v e r s i t a t Braunschweig, 1980. The a c t i v a t i o n o f sentenG l o w a l l a , U . , Schulze, H.H., and Wender, K.F. ces i n semantic networks. I n F. K l i x and J . Hoffmann (Eds.), Cognition and Memory. Amsterdam: N o r t h H o l l a n d , 1980. Graesser, A.C. 1981.
Prose comprehension beyond t h e word. New York: S p r i n g e r ,
King, D., and Anderson, J.R. Long-term memory r e t r i e v a l : An i n t e r s e c t i n g a c t i v a t i o n process. Journal o f Verbal Learning and Verbal Behavior, 1976, 15, 587 - 605. K i n t s c h , W., and van D i j k , T.A. Toward a model of d i s c o u r s e comprehensi o n and p r o d u c t i o n . Psychological Review, 1978, 85, 363 - 394. Lesgold, A.M., Roth, S.F., and C u r t i s , M.E. Foregrounding e f f e c t s i n d i s c o u r s e comprehension. Journal of VerbaZ Learning and VerbaZ Behavior, 1979, 18, 291 - 308. Omanson, R.C. The narrative anaZysis. PhO Thesis, U n i v e r s i t y o f Minnesota, Minnesota, 1979.
TOWARD A MODEL OF MACROSTRUCTURE SEARCH
? 23
Pohl, R.F. Acceptability of s t o r y continuations. I n A. Flammer, and W . Kintsch ( E d s . ) , T e x t processing. Amsterdam: North Holland, in press. Pachella, R.G. T h e i n t e r p r e t a t i o n of reaction time i n information-processing research. In B.H. Kantowitz ( E d . ) , Human information processing: t u t o r i a l s i n performance and cognition. H i l l s d a l e , N.J.: Erlbaum, 1974. Rumelhart, D . E . Understanding and summarizing brief s t o r i e s . I n D . LaBerge, and S.J. Samuels ( E d s . ) , Basic processes i n reading. H i l l s d a l e , N.J.: Erlbaum, 1977. Sanford, A.J., and Garrod, S.C. Wiley, 1981.
Understanding w r i t t e n language. New York:
Schank, R . C . , and Abelson, R.P. S c i p t s , pZans, goals, and understanding. H i l l s d a l e , N.J.: Erlbaum, 1977. Trabasso, T. Relevance perception i n t a t conrprehensioiz. Paper presented a t t h e Department o f Psychology, Universitat Braunschweig, 1980.