Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Comprehension of complex sentences conjoined with "before" and "after" Doke, Wendy Lynne 1982

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-UBC_1982_A6_7 D85.pdf [ 5.5MB ]
Metadata
JSON: 831-1.0095546.json
JSON-LD: 831-1.0095546-ld.json
RDF/XML (Pretty): 831-1.0095546-rdf.xml
RDF/JSON: 831-1.0095546-rdf.json
Turtle: 831-1.0095546-turtle.txt
N-Triples: 831-1.0095546-rdf-ntriples.txt
Original Record: 831-1.0095546-source.json
Full Text
831-1.0095546-fulltext.txt
Citation
831-1.0095546.ris

Full Text

COMPREHENSION OF COMPLEX SENTENCES CONJOINED WITH "BEFORE" AND "AFTER" B by WENDY LYNNE DUKE •A., The U n i v e r s i t y of B r i t i s h Columbia, 1979 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n THE FACULTY OF GRADUATE STUDIES School of Audiology and Speech Sciences We accept t h i s t h e s i s as conforming to the r e q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA A u g u s t 1982 c j Wendy L y n n e Duke, 1982 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of /hvr/rVJ/dcy y b<^J. S^e^cik -JoWrcyC The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date g,.pt4/ / /^ cPA i i ABSTRACT The purpose of t h i s i n v e s t i g a t i o n i s to examine the e f f e c t s of c e r t a i n factors on adults auditory comprehension of complex sentences conjoined with bz^ofiz and &{£<LH. The factors investigated are: conjunction choice, order of mention, clause placement, and general-knowledge constraints. The sentences used i n the study f a l l into four syntactic categories (Before-1, Before-2, After-1, After-2) and two semantic categories (c.on&ttia.A-nzd by general knowledge and an-conAtlciZnzd) . Sixteen subjects each pa r t i c i p a t e d i n two tasks designed to e l i c i t varying reaction times. Stimuli consisted of 160 pre-recorded sentences describing 40 se-quences of two events, with corresponding s l i d e i l l u s t r a t i o n s . Reaction time to task s t i m u l i was recorded to the nearest hundredth of a second. Square roots of the reaction times were subjected to analysis of variance. Results indicate that only the placement of the main clause produces a s i g n i f i c a n t e f f e c t on subjects* responses, thus lending support to a growing body of data which suggests that the main clause holds a priveleged p o s i t i o n i n the comprehension of complex sentences. Results are discussed with respect to experimental design, previous research and theories of sentence comprehension. i i i TABLE OF CONTENTS EMI ABSTRACT n LIST OF TABLES v i LIST OF FIGURES v i i i ACKNOWLEDGEMENT j _ x CHAPTER ONE : A Review of the Lit e r a t u r e 1 Introduction 1 Studies Using Adult Subjects .... 3 Studies with Children as Subjects 8 Other Related Studies 21 On the Role of Semantic Constrained i n the Comprehension of Active and Passive Sentences 21 Studies Using Aphasic Subjects 26 Summary ^ Statement of the Problem 2 9 CHAPTER TWO Method 3 1 Overview 31 Preparation of Sentence Stimuli 32 General Control of Variables Across the Sentences 3 2 S p e c i f i c Considerations Applied to the Cross-matching Procedure 34 The necessity for one actor .... 33 Rejection of ce r t a i n clauses 33 Procedure for cross-matching .... .... .... 34 Rejection of c e r t a i n sentences 35 i v P a 9 e Summary of Sentence Construction Procedures • • • • 39 The Resulting Sentence Stimuli ., 39 Di v i s i o n of Sentences into Four Testing Subsets 42 Semi-Randomization of Sentence Order within the Subsets 43 Assignment of Required Response Values to Sentences 43 Preparation of Vi s u a l Stimuli .... 44 Constraints Applied to the Vi s u a l Stimuli 44 Preparation of Taped Stimuli 47 S p l i c i n g of the Taped Stimuli 47 Equipment 49 Stimulus Presentation and Subject Response .... 49 Temporal Relation of Slid e and Sentence Stimuli 53 F i r s t possible r e l a t i o n 54 Second possible r e l a t i o n 54 Third possible r e l a t i o n 55 Subjects 55 Balancing Subjects, Tasks, and Subsets 56 Procedure • 58 Instructions 60 Verbatim Instructions 60 Obtaining the most comfortable l i s t e n i n g l e v e l 60 S p e c i f i c instructions : Ta^sk 1 60 Sp e c i f i c instructions : Task 2 .... 60 General in s t r u c t i o n s : Task 1 and Task 2 61 Reiteration of instructions : Task 1 61 Reiteration of instructions : Task 2 62 A Comment on these Instructions 62 V Page CHAPTER THREE : Treatment of Data and Results 64 Overview 64 Treatment of Data P r i o r to Analysis 6 4 S u i t a b i l i t y of Data 66 Measurement Procedure 67 Transformation of Data 69 Analysis 69 Task 1 72 Analysis IA : Syntactic factor s i g n i f i e s .... Conjunction choice 7 2 Analysis IB : Syntactic factor s i g n i f i e s .... Order-of-mention 74 Analysis IC : Syntactic factor s i g n i f i e s .... Clause placement 75 Task 2 .. 7 7 Analysis 2A : Syntactic factor s i g n i f i e s .... Conjunction choice 78 Analysis 2B : Syntactic factor s i g n i f i e s .... Order-of-mention 80 Analysis 2C : Syntactic factor s i g n i f i e s .... Clause of Results 82 Summary of Results 8 5 CHAPTER FOUR : Discussion 87 Discussion of Results 87 Inter-subject Differences 87 True/False Differences 88 Constrained/Unconstrained Differences 89 Order-of-mention Differences 91 Clause Placement Differences 9 3 Before/After Differences 96 A Comment on the Relative Importance of Semantic and Syntactic Factors 97 Possible Sources of Experimental Error Physical and Mechanical Sources of Error 99 Sources of Error i n Data C o l l e c t i o n and Analysis 100 A Problem with the Experimental Design 101 CHAPTER FIVE : Conclusions 103 BIBLIOGRAPHY 107 v i LIST OF TABLES Table I : Table II : Table III : Table IV : Table V Table VI : Table VII : Table VIII : Table IX : Table X : Table XI : Table XII : Table XIII : Table XIV : Before/After Complex Sentence Types and the Factors which Distinguish Them Sentences Used i n t h i s Investigation The Eight C l a s s i f i c a t i o n s of Sentences Under Investigation — • • • • Counterbalancing of Tasks, Subsets, and Subjects Factors Distinguishing the Eight Sentence Types Results of Analysis IA : Analysis of Variance Table Results of Analysis IB : Analysis of Variance Table Results of Analysis IC : Analysis of Variance Table Frequenceis, Means, and Standard Deviations for Analysis IC : Syntactic Factor Page 2 . 40 41 57 70 73 75 76 77 Results of Analysis 2A : Analysis of Variance Table •••• Frequencies, Means and Standard Deviations for Analysis 2A : Truth Value Factor Results of Analysis 2B : Analysis of Variance Table Frequencies, Means and Standard Deviations f o r Analysis 2B : Truth Values Factor Results of Analysis 2C : Analysis of Variance Table 79 80 81 82 83 V l l P a 9 e Table XV : Frequencies, Means and Standard Deviations for Analysis 2C : Syntactic Factor 8 4 Table XVI : Frequencies, Means and Standard Deviations for Analysis 2C : Truth Value Factor 8 5 / v i i i LIST OF FIGURES pag_e Figure 1 : Block Diagram of Experimental Apparatus ../ 50 Figure 2 : Physical Set-up of Experimental Apparatus 51 ix ACKNOWLEDGEMENTS I would l i k e to thank the following people to whom I am indebted for t h e i r valuable contributions to t h i s project'. Carolyn Johnson, for supplying references and for her help-f u l reading of my manuscript; Marguerite Drummond for her care-f u l art work; Andr£ Cassty, for h i s design and construction of equipment, and for h i s sketches which comprise Figures 1 and 2. Andre-Pierre Bengeural, for technical advice; Malcolm Greig, for s t a t i s t i c a l counselling; Susan Lee, for f a m i l i a r i z i n g me with the basics of computer operation; Nancy Burns, for advice concerning points of s t y l e and for encouragement throughout. Marshall Chasin and Harold Janzen, f o r valuable comments and advice on many aspects of t h i s project; Judy Lapadat, for assistance with s t i m u l i preparation and for much appreciated support at every stage; the 16 women and men who volunteered as subjects. I am e s p e c i a l l y indebted to my advisor, John Gi l b e r t , for sponsoring t h i s project and for meticulously reading my o r i g i n a l manuscript. Very special thanks go to Hilda Duke, Leonore O'Neill and Sandy Duke, f o r t h e i r unflagging support and encouragement; and for nurturing my c u r i o s i t y , I am forever g r a t e f u l to the l a t e Joseph H, Duke. 1 CHAPTER ONE A REVIEW OF THE LITERATURE Introduction A number of studies i n the past two decades have i n -vestigated the comprehension and/or production of complex sentences conjoined with bn^onz and The following survey w i l l focus on comprehension data. As may be seen i n Table I, four sentences with the same operational meaning can be produced by manipulating c e r t a i n factors within sentences of t h i s kind. The three factors which are opposed within the sentences are: 1) Conjunction Choice - sentences 1 and 2 are con-joined with bz&oiz, whereas sentences 3 and 4 are conjoined with altzK. 2) Clause Placement - sentences 1 and 3 have an i n i t i a l main clause, whereas sentences 2 and 4 have a f i n a l main clause. 3) Order of Mention - for sentences 1 and 4, the order of mention of events corresponds to the order •ci 2 of occurrence of events (OME=OOE). That i s , the order i n which the two events are stated i s the same as the order i n which the two events occurred. In contrast, temporal ordering of the events i s not preserved i n sentences 2 and 3. Table I fizI'A&tz* Complex Sentence Types and the Factors Which Distinguish Them Sentence Type Factors Examples Order of Mention Conjunction Choice Clause Placement 0ME=00Ea Before Main Clause F i r s t 1. She packed the suitcase before she washed the dishes. + + + 2. Before she washed the dishes, she packed the s u i t -case . + -3. She washed the — - + dishes a f t e r she packed the s u i t -case. 4. Af t e r she packed + the suitcase, she washed the dishes. aOME=OOE i s t o be r e a d : O r d e r o f m e n t i o n o f e v e n t s e q u a l s o r d e r o f o c c u r r e n c e o f e v e n t s . 3 The preceding factors have been most consistently investigated as to th e i r role i n the comprehension of complex sentences conjoined with btfiotLZ and a^-te-t. These factors are, therefore, basic to any discussion concerning comprehension studies of such sentences. A number of additional factors have also been investigated and w i l l be elaborated upon where relevant i n the ensuing survey. Studies Using Adult Subjects Although most of the research concerning comprehension of be^o-te. and dittn. i n complex sentences have involved young children as subjects, some of the e a r l i e r investigations examined comprehension i n adult subjects. Clark and Clark (1968) investigated adults' memory fo r such sentences. The purpose of t h e i r study was to determine whether, i n r e c a l l i n g complex sentences, subjects were more l i k e l y to remember ce r t a i n semantic d i s t i n c t i o n s or to remember transformational markers. They proposed that a semantic explanation of memory would be supported by a response bias for sentences i n which the order of occurrence of events was preserved (as i n sentences 1 and 4 i n Table I ) . In contrast, a transforma-t i o n a l complexity model of memory would be supported by a response bias for sentences i n which the subordinate clause was i n f i n a l p o s i t i o n , assuming that sentences with a pre-posed subordinate clause are transformationally more complex than sentences beginning with a main clause. Clark and Clark constructed 72 complex sentences for th e i r investigation. In addition to the four sentence types given i n Table I, they included sentences of the forms "S^ and then and "Sj but f i r s t , S^". These s t i m u l i were each paired with a noun cue, and were presented on IBM cards i n groups of s i x . Subjects were permitted to study each noun cue and sentence for 10 seconds. After completing each set of six, the subjects were presented with a second set of cards, which contained only the noun cues. They were instructed to write the appropriate sentence, verbatim, beside each noun cue. The Clarks found two response biases for verbatim r e c a l l ; one, for sentences i n which temporal order of events was pre-served, and one for sentences beginning with a main clause. They also determined that subjects better remembered the under-ly i n g sense of sentences preserving temporal order. Clark and Clark interpreted these r e s u l t s as supportive of a "semantic explanation of memory as the most general explanation", (p. 130). In discussion, they account for t h e i r findings i n terms of marked and unmarked semantic d i s t i n c t i o n s . Smith and McMahon (1970) investigated adults' comprehen-sion of such complex sentences i n a vari e t y of tasks. These experiments were a l l performed using a reaction time paradigm. For each experiment, sentences of the types depicted i n Table I were projected on s l i d e s . Following the presentation of each s l i d e , subjects were required to supply the appropriate 5 clause i n answer either to the question "What happened f i r s t ? " or to the question "What happened second?". The three experi-ments d i f f e r e d with respect to the temporal r e l a t i o n of the prompt question and the sentence s t i m u l i . In experiment one, the prompt question ("What happened first/second?") was e x p l i c i t l y stated p r i o r to the presentation of each sentence. For experiments two and three, the prompt question was pre-sented after the stimulus sentence. An i n t e r f e r i n g task (oral reading of a three d i g i t number) was presented between the s t i m u l i and prompt f o r experiment three. In each ex-periment, both the error rate and response latency times for each sentence type were analyzed. For experiments two and three, mean inspection time (the time which the subjects chose to study the sentence) was also analyzed. Smith and McMahon found no evidence that sentences i n which temporal order was preserved were comprehended more e a s i l y . One measure (inspection time) was, i n f a c t , consist-ently longer for these sentences. Latency time re s u l t s for order of mention varied between experiments. They did f i n d , however, that the event which was asserted to have happened f i r s t was more accessible, regardless of sentence type. As regards befjo-te and OL^tnn differences, the only s i g n i f i c a n t difference found was i n experiments two and three, where shorter inspection times were found for bzfiofio. sentences. Smith and McMahon found consistent s i g n i f i c a n t r e s u l t s f o r main clauses versus subordinate clauses. In each ex-periment, i t took subjects longer to supply the subordinate clause than the main clause as an answer. The error rate of 6 subordinate clause answers was also s i g n i f i c a n t l y greater. In general, Smith and McMahon*s r e s u l t s indicate that informa-t i o n i n the main clause i s more r e a d i l y available and more e a s i l y interpreted than information i n the subordinate clause. Because these r e s u l t s seemed to contradict those found by the Clarks (as regards order-of-mention d i f f e r e n c e s ) , Smith and McMahon rep l i c a t e d the Clarks* study and found th e i r r e s u l t s to be r e l i a b l e . Smith and McMahon suggested that the differences they observed might be due to differences i n the processes of memory versus comprehension, and proposed that "order of mention does not have much of an influence on understanding, but has an i n t e r f e r i n g e f f e c t on memory" (p. 269). Smith and McMahon indicated that many of the i r r e s u l t s varied remarkably i n r e l a t i v e size depending on the exact nature of the task. Discussing t h i s with respect to implica-tions for a theory of comprehension, they claimed that t h e i r r e s u l t s support Bever's (1970) theory of a single process, multiple output comprehension process. They suggested that, i n terms of Bever's theory, the differences i n r e s u l t s between d i f f e r e n t tasks can be explained "by postulating that various tasks require d i f f e r e n t l e v e l s of abstraction or ... deeper e x i t points i n a sub-routine" (p. 271). Bever, i n 1970, published a paper i n which he proposed several organizing p r i n c i p l e s which he suggested might be basic to the comprehension of complex sentence, and i n which 7 he discussed the studies of Smith and McMahon and the Clarks. Bever summarized these p r i n c i p l e s as follows: A. In comprehension, the assertion of a sentence i s the basis; the presupposition i s organized as psychologically sub-s i d i a r y to i t . B. In the comprehension of ordered events we organize the r e l a t i o n s by s t a r t i n g with the f i r s t event, organizing the other events as subsidiary to the f i r s t . C. CztzsislA pasii.bu& tenporal order preferably corresponds to the actual order of the clauses. (Bever, 1970, p. 286). With respect to previous studies, Bever's p r i n c i p l e A provides a semantic explanation for the primacy of the main clause, as found by Smith and McMahon. Likewise, p r i n c i p l e s B and C predict the memory re s u l t s which the Clarks found. In addition, Bever pointed out that, when combined, p r i n c i p l e s A and B predict that bz^ofit sentences w i l l " be easier to under-stand than a^tzfi sentences. That i s , assertion and order are confounded i n the sense that " i n be^o^e sentences, the asser-ti o n (in the main clause) also describes the f i r s t event, while i n a^ttn. sentences i t i s the presupposition (in the subordinate clause) that describes the f i r s t event" (p. 287). Bever suggested that these p r i n c i p l e s may play d i f f e r e n t r e l a t i v e roles i n memory and comprehension. That i s , depend-ing on the task, one p r i n c i p l e may play a reduced role r e l a t i v e to another. If d i f f e r e n t organizing p r i n c i p l e s are dominant among d i f f e r e n t tasks, then equivalent r e s u l t s would not be expected from one task to another. 8 Studies With Children As Subjects In 1 9 7 1 , Eve Clark published the r e s u l t s of a study i n which she investigated young children's a c q u i s i t i o n of bz&ofiz and d^tzh. i n sentences l i k e those i n Table I. Clark chose these r e l a t i o n a l terms for study f o r two reasons. F i r s t , i n a previous study (Clark, 1 9 6 9 , 1 9 7 0 ) of young Children's spontaneous use of conjunctions, she had noticed that c h i l d -ren (age 3 ; 0 - 3 ; 6 ) generally described temporally related events by use of an order-of-mention strategy; they talked about ordered events i n the same order as they had occurred. Clark postulated, therefore, that young childre n who do not understand bz^onz and afitzn. might s i m i l a r l y r e l y on an order-of-mention strategy i n int e r p r e t i n g such sentences. Second, recent evidence had suggested that young children display an asymmetry i n the a c q u i s i t i o n of c e r t a i n pairs of r e l a t i o n a l words, acquiring the p o s i t i v e member of such pairs prior to the negative member (Donaldson and Wales, 1 9 7 0 ) . In order to investigate children's comprehension of bz^ofiz and a^tzh. i n complex sentences, Clark constructed 3 2 sentences, eight each of the four sentence types shown i n Table I. Forty children, ranging i n age from 3 ; 0 - 5 ; 0 , particiapted i n the study. They were divided into four age groups of ha l f - y e a r l y i n t e r v a l s . The chi l d r e n were asked to carry out in s t r u c t i o n s , based on these sentences. The examiner read a sentence (e.g. "The boy kicked the rock aft e r he patted the dog") and the c h i l d acted out the sentence with appropriate toys. (A second task, was designed to investigate children's production of bz^oH-Z and d^tzn.. This w i l l not be discussed i n d e t a i l here as i t has no d i r e c t bearing on the comprehension of these terms, nor on the methodology de-veloped by Clark to investigate comprehension. Generally, the r e s u l t s of the production task were consistent with those of the comprehension task). Errors were analyzed with res-pect both to the c h i l d ' s age group and to the sentence type. Only reversal errors were analyzed as there were no omission errors. Clark obtained several i n t e r e s t i n g r e s u l t s . F i r s t , as predicted, the younger children i n the study, who did not seem to f u l l y understand bz^onz and CL^tzn, appeared to follow an order-of-mention strategy i n in t e r p r e t i n g these sentences. That i s , they interpreted the f i r s t event mentioned i n the sentence as that event which had occurred f i r s t , r e s u l t i n g i n superior performance for Before-2 and After-1 sentences (sentence types 1 and 4 i n Table I ) . The childr e n i n Clark's study also appeared to go through predictable stages i n t h e i r a c q u i s i t i o n of the terms bzionz and aitzfi, i n which bzioiz was generally acquired at an e a r l i e r age. Clark proposed 4 stages of a c q u i s i t i o n of these terms: 1) Children understand neither term and r e l y on an order-of-mention strategy. 2) Children understand bzhonz, but r e l y on an order-of-mention strategy to i n t e r p r e t a.{tzK. 3) Children t r e a t afittin. as though i t means bz^otie.. 4) Children i n t e r p r e t both terms c o r r e c t l y . Clark accounted for her r e s u l t s by reference to her "semantic feature" hypothesis. B r i e f l y , t h i s hypothesis states that chil d r e n learn the meanings of words by acquiring semantic components one at a time, i n a h i e r a r c h i c a l fashion, from the superordinate features on down. Thus, at one stage, childre n w i l l confuse antonyms (such as bzion.z and afitzi) because they only know the superordinate features which are common to both words. The s p e c i f i c semantic feature matrices which Clark proposed for these terms are as follows: bz^ofizt + Time d^ttm + Time - Simultaneous - Simultaneous + Pr i o r - Prior Clark noted that the po s i t i v e member of the pair w i l l be acquired f i r s t because i t i s l i n g u i s t i c a l l y unmarked. In th i s case, bz{oh.z i s considered by l i n g u i s t s to be the posi -t i v e , or unmarked member of the r e l a t i o n a l pair (Leech, 1970). Barrie-Blackley (1973) conducted a study which also investigated children's understanding of bzhonz and a.{tzh., as well as unti.1, i n simi l a r complex sentences. A comprehension task si m i l a r to Clark's was used to investigate comprehension of these terms for 30 subjects, age 6 years. Unlike Clark, Barrie-Blackley found that her subjects made more reversal errors for bzfioKZ sentences than for afitzA. sentences (74% and 26% of the incorrect responses, respectively) . Barrie-Blackley offered no explanation for t h i s apparent contradiction, and concluded that s i x year olds have not yet mastered the ad-joining of clauses with temporal l i n k s . Although some sub-sequent studies have f a i l e d to f i n d superior performance for be^cte sentences, Barrie-Blackley's r e s u l t of fewer errors for a^tzn sentences has not been r e p l i c a t e d i n any of the l i t e r a t u r e reviewed here, f o r English complex sentences. Clark's study generated considerable research on the com-prehension of these temporal terms; much discussion centred on the differences i n r e s u l t s obtained by Clark and Amidon and Carey. In th e i r 1972 study, Amidon and Carey investigated children's a b i l i t y to perform temporally ordered commmands under a v a r i e t y of conditions. The sentences used i n t h e i r study were a l l i n the imperative (e.g. "Before you move a red plane, move a blue plane"). The 50 subjects i n t h i s study, age 5;4-6;3, were somewhat older than Clarks. These subjects were divided i n t o 5 groups, each of which performed the tasks under d i f f e r e n t t r a i n i n g conditions. One group received feedback as to the correctness of t h e i r responses, one received intonational emphasis on the conjunction i n the command, and one received both feedback and intonational emphasis. In addition, two control groups were included, neither of which received feedback or emphasis. The second control group, however, were trained with i-in.it and la&t instead of betfo/ie and a^te.*.. After the t r a i n i n g sessions, they were given a post-test, during which they received 1 2 neither intonational emphasis nor feedback. The post-test r e s u l t s for the various groups were compared. Amidon and Carey obtained several i n t e r e s t i n g r e s u l t s . F i r s t , the second control group demonstrated very l i t t l e d i f f i c u l t y i n performing tasks with hi.h.i>t and Itx&t, indicating that any d i f f i c u l t i e s with fae^o-te and a^tzh. could not be attributed s o l e l y to d i f f i c u l t y performing ordered tasks. Intonational emphasis was not found to f a c i l i t a t e performance of the task. Feedback, however, was found to greatly reduce errors. Those childre n who did not receive feedback, and subsequently made more errors on the post-test, were found to commit errors which were most often omissions of the sub-ordinate clause, regardless of sentence type. For reversal errors only, there was a nonsignificant trend towards fewer errors when the temporal order of the events was preserved i n the sentence. Contrary to Clark, therefore, Amidon and Carey did not f i n d order-of-mention to be a dominant strategy. Differences for bz^0K.z and a^tzu also f a i l e d to reach s i g n i f i c a n c e , although there was a trend towards fewer over-a l l errors on bz^ofiz sentences. Amidon and Carey pointed out that t h e i r r e s u l t s provided an i n t e r p r e t i v e d i f f i c u l t y for Clark's semantic feature hypothesis. Since children i n t h e i r study experienced much more d i f f i c u l t y with bzfioAz and a^izi than with ^Ifi&t and la&t, the authors suggested that the problem l i e s not i n acq u i s i t i o n of semantic features, but i n handling the additional syntactic 13 complexity of a subordinate clause. Amidon and Carey concluded that these r e s u l t s support Smith and McMahon's contention that information i n main clauses i s more s a l i e n t than information i n subordinate clauses. They noted that, since the childr e n receiving feedback i n t h e i r study appeared to a l t e r t h e i r strategy of attending p r e f e r e n t i a l l y to the main clause, t h e i r study lent support to Bever's (1970) notion that " l i s t e n e r ' s habits of organizing r e l a -tions between clauses may change with age and experimental i n s t r u c t i o n " (p. 422). Helen Johnson, i n 19 75, r e p l i c a t e d both the Clark and Amidon and Carey studies. She found the seemingly contradic-tory r e s u l t s from both studies to be r e l i a b l e . As the same subjects (18 preschool c h i l d r e n , age 4;2-5;2) participated i n a l l of Johnson's task, t h i s would suggest that the d i f f e r e n t r e s u l t s must be due to the nature of the tasks, rather than to random differences i n the populations used. Johnson's study was an attempt to delineate the r e l a t i v e importance of order-of-mention and main-subordinate r e l a t i o n strategies i n children's comprehension of be^o/ie and a £ t e . f i . In f a c t , Johnson found evidence for both strategies; i n d i f f e r e n t situations i t appeared that one strategy was more dominant than the other. In the task modelled after Clark's study, reversal errors predominated and an error analysis revealed that c h i l d r e n seemed to be using an order-of-mention strategy. When the children were given imperative commands, as i n the Amidon and Carey study, errors tended towards 14 omissions of one clause, rather than reversals. Consistent with Amidon and Carey, the subordinate clause was most often omitted. Johnson's explanation f o r her r e s u l t s was that the strategies used by children to in t e r p r e t these sentences may be sentence-form s p e c i f i c . That i s , for declarative sentence, as used by Clark, the ch i l d r e n appeared to follow an order-of-mention strategy, whereas for imperative sentence, the children appeared to follow a strategy of attending p r e f e r e n t i a l l y to the main clause. Johnson did not, however, agree with Amidon and Carey that t h i s r e s u l t supports the view that information i n the main clause i s more e a s i l y interpreted. Rather, Johnson suggested that (for imperative sentences) children may not be aware that the subordinate clause i s also part of the command. She claimed that for a sentence such as Eat youfi chzme. bufioie. you dfitnk you.fi jutce., only the main clause i s a d i r e c t command. The chi l d r e n , therefore, may not have had d i f f i -c u l t y understanding a subordinate clause per se, but simply f a i l e d to understand, without c o r r e c t i v e feedback, that both clauses of an imperative sentence should be out. Pamela Coker (1978) examined children's comprehension of be^ote and afite.fi used both as prepositions and as subordinat-ing conjunctions. The subjects i n her study were 60 kinder-gartners and 60 f i r s t - g r a d e r s , ranging i n age from 5;3 to 7;7. The complex sentence task was sim i l a r to Clarks and u t i l i z e d s i milar sentences. Coker's r e s u l t s indicated that children 15 acquire bzfiotiz and afitzfi f i r s t as prepositions and l a t e r as subordinating conjunctions. On one prepositional task, i n which childre n were presented with a sequence of three pictures and then asked "What did I show you before/after the x?", the c h i l d r e n demonstrated use of a strategy whereby they responded with the next-event-in-time. This strategy, which Coker f e l t i s somewhat analagous to an order-of-mention strategy i n complex sentences, resulted i n superior per-formance fo r a^tci questions. For the subordinate clause task, children were found to u t i l i z e e i t h e r an order-of-mention strategy or a main-clause-f i r s t strategy. Coker pointed out that t h i s l a t t e r strategy, of d i r e c t i n g attention toward the main clause i n the sense of acting that clause out f i r s t , w i l l r e s u l t i n superior per-formance on be^Cie sentences, since the main clause always expresses the f i r s t event i n be^o-te sentences. For t h i s reason, c h i l d r e n using t h i s strategy showed the same response pattern as those i d e n t i f i e d by Clark as treating d^tuK. to mean bejjo^e. Coker argued, however, that t h i s response pattern i s better explained as a syntactic strategy of p r e f e r e n t i a l l y attending to the main clause f o r three reasons. F i r s t , the children who responded i n t h i s manner for the subordinate clause task showed no evidence of treating afit&n. to mean be-tjo^e i n the prepositional tasks; i n f a c t , many of these children, i . e . those using a next-event-in-time strategy, showed superior performance with a£te.n. used as a preposition, 16 Second, t h i s explanation i s compatible with the re s u l t s of previous research by Amidon and Carey and Bever. Third, Coker c i t e d evidence from a number of sources (Coker and Legum, 1975; Coots, 1976; E i l e r s , O i l e r and E l l i n g t o n , 1974; and Glusksberg, Hay and Danks, 1976) i n which p o l a r i t y differences were not found for other polar opposite pairs, thus bringing into question Clark's contention that the po s i t i v e member of a polar pair i s acquired e a r l i e r . Coker also discussed the matter of order-of-mention versus main-clause strategies. In Coker's study, both strategies were evident (order-of-mention being more common), whereas Amidon and Carey had found much more evidence of a main-clause strategy than was apparent i n either Coker's or Clark's studies. In an e f f o r t to reconcile t h i s apparent contradiction, Coker suggested, as did Johnson before her, that these strategies are task s p e c i f i c . Unlike Johnson, however, she did not a t t r i b u t e the difference to imperative versus declarative sentences, but rather to a difference i n how the c h i l d i s cued to attend to the test sentences. She argued that when the c h i l d i s cued to pay attention to both clauses, an order-of-mention strategy w i l l be more dominant and the main-clause strategy w i l l surface only i n terms of better performance for bz^oKz sentences. When, however, the c h i l d i s not cued to attend to both clauses (as i n Amidon and Carey's non-feedback group) the main-clause strategy w i l l dominate and w i l l be evident i n omissions of the subordinate clause. 17 This i s a very i n t e r e s t i n g argument. Not only does i t allow for an integration of those r e s u l t s discussed by Coker, but i t also seems to provide a l o g i c a l r e c o n c i l i a t i o n for the differences i n r e s u l t s found by the Clarks, as opposed to Smith and McMahon, when working with adult subjects. The Clarks' verbatim r e c a l l task, which resulted i n an order-of-mention response preference, forced the subjects to attend to both clauses. Conversely, Smith and McMahon*s instructions to sv£> jects to respond to a question with the appropriate clause, may well not have forced the subject to attend to both clauses; as Coker's argument would predict, the primary role of the main clause was evident i n t h i s study. A further problem associated with Clark's semantic feature hypothesis was suggested by Lynne Feagans (1980) Feagans investigated children's comprehension of terms related to the concepts of order, duration and simultaneity i n complex sentences, as expressed by the words bzfiofiz, afitzA, &<lncz and unt<LZ. Feagans c i t e d evidence from philosophical analysis (van Fraasen, 1970) which has suggested that order i s a more basic concept than simultaneity or duration. Feagans also referred to Piaget's (1966) claim that children acquire a sense of temporal order p r i o r to a sense of duration, pointing out some older language a c q u i s i t i o n data (Ames, 1964; Stern, 1962) which tend to support t h i s view, The problem which this notion represents for Clark's hypoyhesis r e l a t e s to the s p e c i f i c semantic feature matrices which Clark devised i n order to account for her bz{on.z/a{tzn. data i n terms of her semantic feature hypothesis, i n these matrices, the feature "Simultaneous" i s represented as superordinate to the feature " P r i o r " . Since one of the p r i n c i p l e s of the semantic feature hypothesis i s that general features are acquired p r i o r to more s p e c i f i c ones, any evidence demonstrating that terms expressing order are acquired at an e a r l i e r age than those expressing simultaneity would c a l l Clark's hypothesis into question. Feagans examined t h i s problem with 60 children, ages 3, 5 and 7 years. S i m i l a r l y to the Clark study, childr e n were asked to act out sentences presented to them by the ex-aminer. As predicted by Feagans, the r e s u l t s indicated that bz{oiz and a{tzh. (the temporal order terms) were generally comprehended at an e a r l i e r age than itncz and unttl. Feagans found that, even at seven years of age, the children did not respond above chance l e v e l for durational and simultaneity terms. This r e s u l t casts doubt on eith e r the above p r i n c i p l e of the semantic feature hypothesis, or, at the very l e a s t , on the s p e c i f i c a t i o n s of how t h i s hypothesis rel a t e s to the a c q u i s i t i o n of bz{onz and a. {tin, as outlined by Clark, (1970). In 1977, French and Brown published a paper investigating the r o l e of semantic constraints on children's comprehension of bz{on.z and a{tin i n complex sentences, thus adding another dimension to t h i s problem. They contended that Clark's semantic feature hypothesis f a i l s to account for the role 19 played by supportive context i n language a c q u i s i t i o n . A l -though t h i s question had not previously been examined with s p e c i f i c reference to the a c q u i s i t i o n of be.fion.z and afitzti, French and Brown reported research which indicates that children are better able to understand r e l a t i o n s i n active and passive sentences when contextual semantic constraints are provided (Bever, 1970; Olson and Nickerson, 1977; S i n c l a i r - d e Zwart, 1969). In addition, they pointed out that Brown (1976) found that young children's memory for ordered sequences improved when the sequences were meaningfully ordered. In order to examine the r o l e of contextual support i n the a c q u i s i t i o n of bzfioKZ and afitzti, French and Brown con-structed a number of two-event sentences i n which events were meaningfully ordered; that i s , events i n the sentences bore a pr e d i c t i b l e temporal r e l a t i o n s h i p to each other (e.g."Raggedy Ann f i l l s the b o t t l e before she feeds the baby"). The com-ponent clauses of these l o g i c a l l y constrained sentences were then randomly cross-matched with each other i n order to pro-duce an equal number of sentences i n which the events were a r b i t r a r i l y ordered. Forty children, ranging i n age from 3;5-5;l, were asked to act out these sentences, i n a procedure sim i l a r to that used by Clark. French and Brown found that performance was markedly superior for the l o g i c a l sequences. They concluded that t h e i r r e s u l t s support the importance of context i n language development i n general, and, s p e c i f i c a l l y that a c q u i s i t i o n 20 o f bn^oh-Z and a£to.A. i s f a c i l i t a t e d b y s i t u a t i o n s w h i c h p r o v i d e c o n t e x t u a l s u p p o r t . O t h e r i n t e r e s t i n g r e s u l t s were o b t a i n e d i n t h i s s t u d y . F r e n c h a nd Brown r e p o r t e d a n o n s i g n i f i c a n t t e n d e n c y f o r f e w e r r e v e r s a l e r r o r s on s e n t e n c e s w h i c h p r e s e r v e t h e o r d e r o f m e n t i o n o f e v e n t s . I n an a n a l y s i s o f number c o r r e c t , no d i f f e r e n c e was f o u n d f o r t h e be^oA-e/d^tih. v a r i a b l e . An i n t e r e s t i n g p a t t e r n was n o t e d f o r o m i s s i o n e r r o r s . F o r a r b i t r a r y s e n t e n c e s , o m i s s i o n e r r o r s t e n d e d t o be o m i s s i o n s o f t h e s u b o r d i n a t e c l a u s e , a s f o u n d b y Amidon and C a r e y . F o r l o g i c a l s e n t e n c e s , however, e r r o r s t e n d e d t o be a s s o c i a t e d w i t h t h e s e c o n d e v e n t . F r e n c h and Brown p o s t u l a t e d t h a t " g i v e n t h e s e m a n t i c c o n s t r a i n t s o f t h e l o g i c a l s e n t e n c e s , t h e m a i n c l a u s e l o s e s i t s p r i v i l e g e d p o s i t i o n " ( p . 2 5 3 ) . K a v a n a u g h (1979) a l s o i n v e s t i g a t e d t h e r o l e o f l o g i c a l c o n s t r a i n t s o n c h i l d r e n ' s u n d e r s t a n d i n g o f bzioKZ and a^tzn. i n c o m p l e x s e n t e n c e s . I n a p r o c e d u r e s i m i l a r t o F r e n c h and Browns, Kavanaugh p r e s e n t e d 30 c h i l d r e n w i t h s e n t e n c e s i n w h i c h t h e e v e n t s were e i t h e r l o g i c a l l y o r a r b i t r a r i l y o r d e r e d . T h e s e c h i l d r e n were d i v i d e d i n t o two age g r o u p s , (3;6-4;2 and 4; 3 - 5 ; 0 ) . C o n s i s t e n t w i t h t h e r e s u l t s o f F r e n c h and Brown, Kavanaugh f o u n d s u p e r i o r p e r f o r m a n c e f o r l o g i c a l l y o r d e r e d s e n t e n c e s , when t h e t o t a l number o f e r r o r s was a n a l y z e d . On f u r t h e r a n a l y s i s , t h i s f i n d i n g was m a i n t a i n e d f o r r e v e r s a l e r r o r s , b u t n o t f o r o m i s s i o n e r r o r s . K a v anaugh s u g g e s t e d t h a t t h i s r e s u l t was due t o t h e f a c t t h a t o l d e r c h i l d r e n 21 commit few omission errors; these children, he f e l t , w i l l be more l i k e l y to reverse a sequence of events i n preference to omitting one event, when encountering d i f f i c u l t y . One f i n a l point raised by Kavanaugh bears mention. Although the role of semantic constraints and no n l i n g u i s t i c strategies i n language a c q u i s i t i o n was not considered when Clark o r i g i n a l l y proposed the semantic feature hypothesis (as French and Brown have pointed out), Kavanaugh claimed "that more recent formulations of semantic feature theory (Clark, 1973, 1975, 1977; Clark and Garnica, 1974) are not imcompatible with the demonstrated e f f e c t s of constrained sentences" (p. 357). Other Related Studies On the Role of Semantic Constraints i n the Comprehension  of Active and Passive Sentences As French and Brown noted, although no previous research was concerned with the role of semantic constraints on the comprehension of bzionz and afitzA., a number of studies ex-amined t h i s question with regard to a c q u i s i t i o n of passive sentences. French and Brown referred only to studies which investigated the e f f e c t s of such constraints for young children's a c q u i s i t i o n of passive sentences. Two i n f l u e n t i a l studies which examined t h i s e f f e c t for adults' comprehension of passives should be mentioned. The following i s not intended to be a comprehensive review of the l i t e r a t u r e pertaining to r e v e r s i b i l i t y i n passive sentences; rather, i t i s offered to demonstrate findings when a question s i m i l a r to the one of in t e r e s t here (namely, the role of semantic constraints i n the comprehension of bzfiofiz and afitzH.) has been investigated. Slobin (1966) investigated comprehension of active and passive sentences i n both children and adults using a picture-v e r i f i c a t i o n procedure. He used sentences of two semantic types - re v e r s i b l e and nonreversible. " R e v e r s i b i l i t y " i s used to indicate whether the actor and r e c i p i e n t of an action could l o g i c a l l y exchange roles within the sentence. Examples of r e v e r s i b l e and nonreversible sentences are "The g i r l chased the boy" and "The g i r l watered the flowers", respectively. Sxiteen subjects i n each of f i v e age groups (kindergarten, grades 2, 4 and 6, and adults) took part i n th i s experiment. These subjects were evenly divided with respect to sex. Reaction time for p i c t u r e - v e r i f i c a t i o n , i n response to an illuminated picture immediately following the verbal presenta-t i o n of a sentence, was measured. Slobin found that, (at a l l ages), passive sentences took more time to evaluate than d i d actives. The e f f e c t of non-r e v e r s i b i l i t y was to decrease p i c t u r e - v e r i f i c a t i o n time. This e f f e c t was e s p e c i a l l y f a c i l i t a t i v e for passive sentences. That i s , "making sentences nonreversible l a r g e l y washed out the difference i n syntactic complexity between active and passive sentences", (p. 219). While v a l i d for a l l age groups, 23 t h i s e f f e c t was most pronounced for the younger subjects. Slobin suggested that n o n r e v e r s i b i l i t y i s more f a c i l i t a t i v e for passive sentences because a l i s t e n e r does not have the d i f f i c u l t y ' of deciding which party i s the subject - only one choice i s possible. Forster and Olbrei (1973) also studied l i s t e n e r s ' responses to re v e r s i b l e and nonreversible active and passive sentences. The question which Forster and Olbrei asked was subtly d i f f e r e n t from that asked by Slobin. Rather than examining how long subjects took to evaluate these sentences, they were interested i n determining whether "the component of sentence processing d i r e c t l y attributable to syntactic processing depends c r i t i c a l l y on c e r t a i n semantic properties of the sentence" (p.319). They maintain that two opposing views,the " i n t e r a c t i v e " and "constancy" hypotheses, are held on t h i s question. The " i n t e r a c t i v e " hypothesis, as they used t h i s term, refers to the view that feedback from the semantic l e v e l of processing af f e c t s syntactic decision making. The re s u l t s obtained by Slobin, they claimed, are often c i t e d as evidence for t h i s hypothesis. The alternative view, that "the component of t o t a l processing time d i r e c t l y a t t r i b u t a b l e to syntactic processing remains constant despite v a r i a t i o n i n meaning", i s referred to as the "constancy" hypothesis. In order to evaluate these two opposing views experimentally, Forster and Olbrei chose a technique modified from the decision latency procedure which had been used by Rubinstein, G a r f i e l d and M i l l i k a n (1970) to examine word 24 recognition. Forster and Olbrei asked t h e i r adult subjects to decide whether v i s u a l l y presented strings of words con-s t i t u t e d meaningful sentences. Distractors used i n the experiment equalled the number of well-formed sentences and were of two types; semantically anomalous and grammatically i l l - f o r m e d . Decision latency for t h i s task was measured and analyzed. A second procedure used to study the e f f e c t s of r e v e r s i b i l i t y followed the rapid s e r i a l v i s u a l presentation (RSVP) technique, previously used by Forster and Ryder (1971). In t h i s procedure, each word of a sentence or of an anomalous s t r i n g , i s presented i n d i v i d u a l l y i n quick, over-lapping succession (the trace of one word does not completely fade u n t i l the following word i s projected). The subject i s asked, a f t e r a l l the words of a given s t r i n g have been thus presented, to r e c a l l the sequence of words. I t i s assumed that "the presentation rate i s slow enough to permit each word to be i d e n t i f i e d but too f a s t to allow each word to be separately encoded into memory" unless "the subject i s able to organize the input meaningfully" (p. 339). The r e s u l t s from both tasks, i n the opinion of Forster and O l b r e i , f a i l e d to provide support for the i n t e r a c t i o n hypothesis. Ambiguous r e s u l t s were obtained when rev e r s i b l e versus nonreversible responses were analyzed. In one experiment, there was a nonsignificant trend towards shorter response latencies f o r r e v e r s i b l e sentences. In another experiment, a marginally s i g n i f i c a n t r e s u l t indicated per-formance for nonreversible sentences. The c r u c i a l r e s u l t , 25 however, was that for each experiment, s i g n i f i c a n t differences were found between responses to active and passive sentences, regardless of whether the sentences were re v e r s i b l e . That i s , unlike Slobin, Forster and Olbrei f a i l e d to show that n o n r e v e r s i b i l i t y allows for passive sentences to be processed with approximately the same ease as active sentences. Forster and Olbrei concluded that t h e i r r e s u l t s supported the constancy hypothesis. On the basis of t h e i r d e f i n i t i o n of t h i s hypothesis, and i f one accepts the assumptions inherent i n t h e i r tasks, t h i s may be a v a l i d conclusion. However, i t i s important to bear i n mind that the tasks used by Forster and Olbrei to investigate re-v e r s i b i l i t y were very d i f f e r e n t from those used by Slobin. For example, they did not use verbal s t i m u l i , and at no point did they d i r e c t l y evaluate t h e i r subjects' comprehension of the s t i m u l i sentences. Although i n t e r e s t i n g , conclusions reached by Forster and Olb r e i may be considered to have questionable relevance to the problem of sentence comprehen-sion. The preceding two studies demonstrated very d i f f e r e n t procedural approaches and t h e o r e t i c a l biases for examining the r o l e of r e v e r s i b i l i t y i n the comprehension of active and passive sentences. Although these studies do not d i r e c t l y r e l a t e to the comprehension of be.£oie. and a^tzn. i n complex sentences, they provide a framework upon which to develop a methodology for investigating the r o l e of sim i l a r semantic constraints i n the comprehension of these sentences. Studies Using Aphasic Subjects Recently, some research has been conducted with aphasic subjects, examining t h e i r comprehension of bzfioiz and a{tzh.. Although conclusions based on such studies of comprehension in a language disordered population are not d i r e c t l y r e l a t a b l e to comprehension by normal adult subjects, i t i s i n t e r e s t i n g to note what trends have been observed i n t h i s population. Sasanuma and Kamio (1976) conducted a study using 57 aphasic subjects, a l l of whom were native speakers of Japanese. The subjects were asked to perform a series of commands i n -volving complex sentences with bz&ofiz and a^tzn. (As the st i m u l i sentences were a l l i n Japanese, they do not d i r e c t l y correspond to those outlined i n Table I. They were of four si m i l a r sentence types; h a l f of the sentences contained bzfioAz versus a^tzh., and half had the subordinate clause preposed.) A l l of the s t i m u l i sentences were presented verbally, and a l l used the verb touch, both reversal and item errors were analyzed. Sasanuma and Kamio found that these subjects made f a r more reversal errors on b z ^ o i z than a^tzn. sentences. Item errors appeared to be d i s t r i b u t e d equally among the sentence types. Subjects made three times more errors on bz^onz sentences than they did on a f t e r sentences. No c o r r e l a t i o n 27 was found between c l i n i c a l syndrome of aphasia and error pattern. In a li m i t e d follow-up study involving 9 of the 57 subjects, Sasanuma and Kamio noted that reversal errors appeared more r e s i s t a n t to recovery than did item errors. Sasanuma and Kamio pointed out that, i n Japanese, order-of-mention and conjunction choice are confounded i n these sentences. That i s , the temporal order of events i s always preserved i n a f t e r sentences and i s always reversed i n bzfiotiz sentences. For t h i s reason, the authors f e l t that i t was not possible to t e l l whether the dominant response pattern they obtained was due to subjects u t i l i z i n g an order-of-mention strategy or to subjects overgeneralizing the meaning of A n s e l l and Flowers (1982) investigated t h i s question using English-speaking aphasic subjects. Their 12 subjects were a l l considered to have r e l a t i v e l y preserved auditory comprehension, as indicated by testing with the Boston Vtagnoittc kpha&ta Examtnatton (Goodglass and Kaplan, 1972) and the Shontznzd Vzulon ofi thz Takzn Jz&t (Derenzi and F a g l i o n i , 1978). Test sentences, written i n the imperative, were varied along the same dimensions as sentences shown i n Table I. Two sets of sentences were constructed. The set considered to be more complex s p e c i f i e d the shape of the object to be manipulated. Examples of the sentences used are "Touch the yellow one before touching the green one" and 28 "Touch the red square a f t e r touching the green c i r c l e " . The sentences were recorded with normal intonation and were presented to the subjects 11 seconds apart. Errors were analyzed along these parameters: adverb choice, adverbial clause placement, and coincidence of order of mention with order of occurrence of events. Ansel1 and Flowers found that only adverb choice was s i g n i f i c a n t i n aphasics" comprehension of these complex sentences, and that t h i s e f f e c t was only s i g n i f i c a n t when re-versal-of-order errors were analyzed. In contrast to Sasanuma and Kamio, they found s i g n i f i c a n t l y fewer errors on b<L{ofio. sentences. Ans e l l and Flowers pointed out that aphasics' error patterns are not s i m i l a r to those of the children studied by Clark, since no use of an order-of-mention strategy was evident. They claimed that t h e i r r e s u l t s do not support the regression hypothesis (Jackobson, 1968) , which contends that language d i s s o l u t i o n i n aphasia r e f l e c t s language a c q u i s i t i o n i n children. Summary Studies have been conducted to investigate comprehen-sion of bifioKQ. and afitiK i n complex sentences by children, adults and aphasics. In general, studies involving adults have u t i l i z e d written s t i m u l i , whereas studies involving children and asphasics have used verbal s t i m u l i . Many super-f i c i a l contradictions are reported i n the l i t e r a t u r e . Use of an order-of-mention strategy has appeared dominant i n many studies (Clark and Clark, 1968; Clark, 1971). Primacy of the main clause i s indicated i n other studies (Amidon and Carey, 1972; Smith and McMahon, 1970). S t i l l others have suggested that these two strategies may not be mutually exclusive; rather, one may become more dominant than the other given a s p e c i f i c task (Coker, 1979; Johnson, 1975). The influence of bzfioAz/'afitzn. differences on the comprehension of the sentence types depicted i n Table I show equivocable r e s u l t s . In those studies where a p r e f e r e n t i a l response to bzfiotiz sentences has been demonstrated, i t has been argued that t h i s r e s u l t may be associated with bzfionz as the unmarked member of the r e l a t i o n a l pair (Clark, 1971), or, c o n t r a r i l y , that t h i s r e s u l t i s evidence of a less dominant form of a main-clause strategy for comprehension of such sentences (Coker, 1979). Although most of the research reported above has not been concerned with the r o l e of semantic constraints i n the comprehension of complex sentences of the foregoing variety, French and Brown (1977) and Kavanaugh (1979) have examined t h i s e f f e c t with young children and have found that semantic constraints appear to f a c i l i t a t e a c q u i s i t i o n of bzfionz and afitzA. Statement of the Problem A number of gaps e x i s t i n our current knowledge of the comprehension of bzfioKz and afitzn i n complex sentences. Th 30 present study was undertaken to determine: 1) The auditory comprehension of sentence types shown i n Table I, by adult subjects, (previous studies with adults haved used only written s t i m u l i ) . 2) The role of semantic constraints i n adults' comprehension of sentence types shown i n Table I. S p e c i f i c a l l y , the following four n u l l hypotheses were posed: 1) Adults do not demonstrate d i f f e r e n t response patterns to sentences which are constrained by general knowledge versus those that are uncon-strained. 2) Adults do not demonstrate d i f f e r e n t response patterns to be^cte. and afitzti sentences. 3) Adults do not demonstrate d i f f e r e n t response patterns to sentences i n which order-of-mention and order of occurrence of events correspond, versus sentences i n which t h i s order i s not preserved. 4) Adults do not demonstrate d i f f e r e n t response patterns to sentences i n which the subordinate clause i s preposed versus those sentences which have an i n i t i a l main clause. In developing the experimental procedures and hypotheses, a long term goal was to devise a t e s t which might l a t e r prove useful for investigating sentence comprehension strategies used by aphasic subjects. 31 CHAPTER TWO METHOD Overview The comprehension of sentences expressing temporally related events was examined i n two separate tasks; a re-action time paradigm was used for both tasks. For the ^LKbt task, each stimulus item consisted of two s l i d e s and one pre-recorded sentence, which were pre-sented simultaneously. Subjects were required to press a button i n d i c a t i n g which s l i d e most appropriately depicted the accompanying sentence. (One s l i d e depicted the two events i n the order stated i n the sentence; the other showed the two events i n the reverse order). The time required for t h i s response was measured to the nearest one-hundredth of a second. The izcond task was s i m i l a r to the f i r s t , except that only one s l i d e was presented with each sentence. The sub-jects* task i n t h i s case was to decide whether or not the s l i d e c o r r e c t l y depicted the accompanying sentence. (Some of these s l i d e s represented the correct order of events; others showed the reverse order). Response time was again measured to the nearest hundredth of a second. The same subjects, s l i d e s and sentence s t i m u l i were used for both tasks. Preparation of Sentence Stimuli Forty core sentences were constructed to serve as s t i m u l i . (As each sentence had four paraphrastic v a r i a t i o n s , corresponding to the four sentence types of Table I, t h i s re-sulted i n a t o t a l of 160 sentence s t i m u l i ) . Twenty of the for t y basic sentences were c l a s s i f i e d as con.&tn.cxlnzd by general knowledge (C sentences) and 20 were c l a s s i f i e d as u.vi-con6tn.a<Lne.d (U sentences) . General Control of Variables Across the Sentences In constructing these sentences, c e r t a i n variables were cont r o l l e d . The actions described by the sentences had to be re a d i l y picturable, unambiguous, and common enough that i t could be1 s a f e l y assumed that both the events and the vocabu-l a r y describing them would be known by the subjects. A l l clauses comprising the sentences had to be of r e l a t i v e l y equi-valent syntactic structure, thus, core sentences were of the form "She verbed (± p a r t i c l e ) a r t i c l e noun". To further ensure that unequivalent syntactic complexity of the t e s t sentences did not confound results, the 40 clauses which comprised the 20 C sentences were cross-matched with each other to construct the 20 U sentences. This procedure en-sured that l e s s frequent vocabulary items or events evenly 33 distribued among the two c l a s s i f i c a t i o n s of sentences (U and C). This procedure ensured that the two groups of sentences varied only i n semantic complexity. S p e c i f i c Considerations Applied to the Cross-Matching  Procedure The necessity for one actor Although cross-matching the clauses ensured syntactic equivalence f o r the two sentence c l a s s i f i c a t i o n s , t h i s pro-cedure raised other problems which required attention. One such problem was that the same actor had to be the subject of a l l of the clauses comprosing the C sentences. Otherwise, when these clauses were cross-matched to construct the U sentences, these r e s u l t i n g U sentences would have been seman-t i c a l l y more complex i n the sense that they would have had two actors as opposed to the U sentences' one. For t h i s reason, one young woman was depicted as the actor of a l l the clauses. To r e t a i n syntactic equivalence, t h i s woman was always referred to as "she". Rejection of c e r t a i n clauses A second problem involved the r e j e c t i o n of some clauses considered by the experimenter to predispose a l i s t e n e r to inter p r e t such clauses as either f i r s t or second events. That i s , since cross-matching dictated that a l l clauses had to par t i c i p a t e not only i n sentences which were constrained by general knowledge, but also i n sentences which were not so constrained, such a bias made them inappropriate. For example, the sentence She. l i t thi aZaAm bifioAi &hi wint to iZnp was considered as a possible sentence s t i m u l i . I t was rejected, however, since the author believed that the clause, "she went to sleep", could not be r e a d i l y cross-matched with any other clause, without the resultant sentence biasing the l i s t e n e r to inte r p r e t that clause as the second event. In other words, a sentence containing the clause, ihi toint to iZnp, would l i k e l y be constrained by general knowledge regardless of what other clause i t was matched with, since people normally sleep aft e r they have pa r t i c i p a t e d i n other events; a t r u l y a r b i t r a r y or unconstrained sentence containing that clause would,therefore, be d i f f i c u l t to formulate. Procedure for cross-matching As described above, every attempt was made to take note of, and r e j e c t , any clause which i n t r i n s i c a l l y was l i k e l y to be interpreted as occurring i n a p a r t i c u l a r order r e l a t i v e to any other clause with which i t might be combined. I t i s possible, however, that some clauses i n the C sentences had such a subtle bias i n t h i s d i r e c t i o n that they were undetected, and hence, not rejected. In order to minimize any a f f e c t such clauses may have had on the resultant U sentences, a s p e c i f i c procedure was followed during cross-matching. The f o r t y clauses comprising the twenty C sentences were l a b e l l e d as either 1 or 2, corresponding to whether they were the f i r s t or second events i n the C sentences. They were then cross-matched i n fiouA ways to construct the U sentences: - f i v e U sentences were composed of 2 f i r s t event C clauses, - f i v e were composed of 2 second event C clauses, - f i v e were composed of f i r s t event C clauses followed by a second event C clause, - f i v e were composed of a second event C clause followed by a f i r s t event C clause. This procedure minimized one subtle source of biasing which may have otherwise resulted i n U sentences which were not wholly unconstrained. Rejection of c e r t a i n sentences Although t h i s cross-matching procedure was considered necessary, i t l i m i t e d the number of possible combinations of C clauses which could be used to generate U sentences. None-theless, every attempt was made to ensure that the r e s u l t i n g U sentences met three further constraints. The &<Lh.6t i s ob-vious; i . e . , that cross-matched clauses did not, by chance, r e s u l t i n sentences which were constrained by general knowledge. Two such precautions against t h i s have already been described (e.g., r e j e c t i o n of ce r t a i n clauses and pro-cedure for cross-matching). However, i t was s t i l l possible that, had the clauses been randomly matched from t h i s point on, that such randomly matched clauses might have resulted i n constrained sentences, simply by coincidence. For example, the sentence She. took a &houozK balonz iho. wtnt to woik, could have resulted from cross-matching, at t h i s point. This sentence would have been rejected, however, since these two 36 events could be expected to occur i n that order; the new sentence i s not the unconstrained one that was aimed for, but yet another constrained sentence. A Atcond constraint concerns the reverse of the previous s i t u a t i o n ; i . e . , clauses could not be randomly matched i n such a way as to r e s u l t i n sentences which were contrary to an expected order of events. In other words, a sentence could not be one describing a highly u n l i k e l y sequence of events. For example, the sentence Skz got on tho. bu6 be^o-te 4fie packzd hzn. Aui-tccue., (which was a possible consequence of cross-matching), would not simply be unconstrained but would be describing events i n an order reverse to expectation. As u n l i k e l y sentences are, i n the author's opinion, quite d i f f e r e n t from sentences i n which the temporal r e l a t i o n of two events i s not predictable (unconstrained), such sentences would be rejected. Stated otherwise, both of these f i r s t two constraints dealt with the concept of r e v e r s i b i l i t y . In order to be acceptable, the U sentences had to be judged as equally l i k e l y regardless of how t h e i r component clauses were ordered. A thltid constraint also dealt with implausible sentences. Although the second constraint dealt with sentences which were u n l i k e l y i n the sense of describing events which occur i n an unexpected order, t h i s i s not, of course, the only manner i n which a sentence can be considered u n l i k e l y . A sentence can be judged as implausible because i t describes two events which one cannot imagine as being related, or for many other reasons. Since the purpose of t h i s i n v e s t i g a t i o n 37 was to study how subjects react to "Unconstrained" versus "Constrained" sentences, and not "unlikely" versus " l i k e l y " sentences, a " p l a u s i b i l i t y " c r i t e r i a was used. This pZau&tbtZtty c r i t e r i a was that the U sentences had to be judged as plausible sentences describing events which did not require unreasonable stretches of the l i s t e n e r s ' imagination. This was by far the most d i f f i c u l t constraint to meet, since determining the p l a u s i b i l i t y of a sentence, out of context, i s a very nebulous task. Although no pro-cedure used to judge such a subjective concept as " p l a u s i b i -l i t y " can be t o t a l l y adequate, i t was necessary to deal with t h i s problem i n some way. Therefore, a U ie.nte.nce. wa& ne.-jzctzd tfi t t wai judgzd to be, AtgntfitcantZy Ze.66 pZauitbZz/ than the. two C &e.nte.nce.i finom whtch tt& component cZau6e.4> we.ne. dnauon. Three judges were used f o r the task of deter-mining p l a u s i b i l i t y ; i f one of them considered the sentence to be u n l i k e l y , i t was rejected. The key word here, of course, i s " s i g n i f i c a n t " , which allowed for some difference i n judged p l a u s i b i l i t y between the constrained and uncon-strained sentences. This was necessary, however, since i t i s only reasonable that a sentence composed of two clauses so related as to be expected to occur i n a p a r t i c u l a r temporal r e l a t i o n to each other (C sentences) would be somewhat more plau s i b l e than sentences composed of clauses which are not so related; that i s , the p l a u s i b i l i t y of C sentences i s obvious. Yet the uncomfortable question remains of how s i g n i f i c a n t a difference judged p l a u s i b i l i t y i s considered too s i g n i f i c a n t . At the very l e a s t , however, outlandish sentences have been eliminated by t h i s judging procedure. Hopefully, those sentences which remain are indeed merely unconstrained and not implausible. Some constraints p a r t i c u l a r to the U sentences have been discussed. A fiousith constraint was s p e c i f i c to the C sentences. Although the C sentences had to be, by d e f i n i -t i o n , ones i n which the clauses occurred i n an expected temporal r e l a t i o n to each other, such sentences i n which the reverse order of events was t o t a l l y impossible were rejected. In other words, the purpose of the C sentences was to suggest a p a r t i c u l a r order of events to the l i s t e n e r , not to s t r i c t l y impose one. I t i s possible, (although the author knows of no evidence to deny or confirm t h i s ) , that d i f f e r e n t processes are at play i n comprehending sentences i n which one in t e r p r e t a t i o n i s more l i k e l y than another, as opposed to sentences i n which, owing to content alone and regardless of syntax, one i n t e r p r e t a t i o n i s demanded since the other i s impossible. Therefore, i n order to maintain some degree of homogeneity among the C sentences, a l l of these sentences had to be ones that could possibly occur i n the reverse order but would be u n l i k e l y to be heard i n that form. 39 Summary of Sentence Construction Procedures To b r i e f l y r e c a p i t u l a t e : 20 sentences were constructed i n which the temporal order of events described by them were constrained by general knowledge (C sentences). Both these sentences and the 40 clauses of which they were composed met a number of constraints which have been outlined above. The same 40 clauses were then cross-matched, using a s p e c i f i c procedure previously described, i n order to construct 20 sentences i n which the temporal order of events was not con-strained by general knowledge, (U sentences). These sentences also met a number of constraints which have been discussed. o The Resulting Sentence Stimuli As mentioned, these fo r t y basic sentences were a l l of one syntactic type (Before-2 form). Each of these Before-2 sentences were then expanded into t h e i r four syntactic v a r i a -t i o n s , as shown i n Table I, r e s u l t i n g i n 160 sentences (four syntactic v a r i a t i o n s each of 40 paraphrastically d i f f e r e n t sentences). The 40 core sentences are reproduced i n Table I I . Two of these sentences are shown i n a l l of t h e i r syntactic forms (Table I I I ) . As the complete set of 160 sentences can be subdivided into four syntactic classes and two semantic classes, t h i s r e s u l t s i n eight c l a s s i f i c a t i o n s of sentences, of which there are 20 each, i n t o t a l . Table III shows an example of each of these sentence types. Table II a Sentences Used i n t h i s Investigation Sentence Constrained Sentences  1. She ate breakfast before she went to work. 2. She packed the suitcase before she got on the plane. 3. She ate dinner before she washed the dishes. 4. She ate the steak before she ate dessert. 5. She turned on the light before she read the book. 6. She took a shower before she dried her hair. 7. She dialled the number before she talked on the phone, 8. She washed her face before she put on her make-up. 9. She bought groceries before she cooked dinner. 10. She bought some paint before she painted the furniture. 11. She bought the cake before she had the party. 12. She addressed the envelope before she mailed the letter. 13. She counted her change before she got on the bus. 14. She stood in line before she watched the movie. 15. She struck the match before she smoked the cigarette. 16. She beat the eggs before she baked the cookies. 17. She f e l l off the ladder before she went to the hospital. 18. She got in the boat before she caught the fish. 19. She robbed the bank before she hid the money. 20. She skiied the race before she won the medal. Unconstrained Sentences 21. She packed the suitcase before she washed the dishes. 22. She bought the cake before she talked on the phone. 23. She ate dinner before she won the medal. 24. She got in the boat before she ate dessert. 25. She bought groceries before she got on the bus. 26. She smoked the cigarette before she turned on the light. 27. She got on the plane before she ate breakfast. 28. She went to work before she addressed the envelope. 29. She mailed the letter before she counted her change. 30. She had the party before she robbed the bank. 31. She took a shower before she f e l l off the ladder. 32. She skiied the race before she ate the steak. 33. She bought some paint before she dialled the number. 34. She beat the eggs before she washed her face. 35. She stood in line before she struck the match. 36. She read the book before she caught the fish. 37. She went to the hospital before she painted the furniture. 38. She hid the money before she watched the movie. 39. She put on her make-up before she cooked dinner. 40. She dried her hair before she baked the cookies. aThese sentences are depicted in Before-2 form. 4-1 Table III The Eight C l a s s i f i c a t i o n s of Sentences Under Investigation Sentence Type Example Before-2 Constrained Unconstrained Before-1 Constrained Unconstrained After-2 Constrained Unconstrained After-1 Constrained Unconstrained She ate dinner before she washed the dishes. She packed the suitcase before she washed the dishes. Before she washed the dishes, she ate dinner. Before she washed the dishes, she packed the suitcase. She washed the dishes aft e r she ate dinner. She washed the dishes after she packed the suitcase. After she ate dinner, she washed the dishes. After she packed the suitcase, she washed the dishes. 42 D i v i s i o n of Sentences Into Four Testing Subsets Each subject p a r t i c i p a t e d i n two tasks, each involving the above described 160 sentences. In order to minimize the r i s k of l o s i n g l i s t e n e r attention during the experiments, i t was necessary to divide the sentences into t e s t i n g subsets. The 160 sentences, therefore, were divided into four subsets, (a, b, c, and d), which were used independently i n d i f f e r e n t runs of the experiment. These four sets each consisted of 20 constrained (C) and 20 unconstrained (U) sentences. Each of the 4 0 core sentences was represented only once i n one of i t s four syntactic forms, i n each set; each set, therefore, contained 40 sentences each with d i f f e r e n t semantic content. The four syntactic c l a s s i f i c a t i o n s were represented by ten sentences each i n the four subsets. The r e s u l t was four subsets which contained 40 paraphrastically d i f f e r e n t sentences with equal representation by each of the eight sub-c l a s s i f i c a t i o n s of sentence type ( i . e . , f i v e each of U/Before-1, C/Before-1, U/Before-2, C/Before-2, U/After-1, C/After-1, U/After-2, and C/After-2). I t was intended that t h i s procedure should not only balance the sets of s t i m u l i for sentence type, but also that i t would render homogeneous the amount of v i s u a l decoding required by the subjects from subtest to subtest, as each of the 40 sequences of two events would be pictured once and only once i n every subtest. 43 Semi-Randomization of Sentence Order Within the Subsets Once the sentences had been subdivided into subsets a, b, c and d, as described above, they were semi-randomized as to t h e i r order within each subset. One constraint was placed on the semi-randomization of the sentence. I t w i l l be r e c a l l e d that the same 40 clauses were used to construct both the U and C sentences. In each subset, therefore, a given clause would appear twice, accompanied by a d i f f e r e n t clause each time. I t was possible, therefore, that subjects would hear the same clause i n two consecutive sentences, perhaps re-su l t i n g i n some momentary confusion as to whether they were accidently hearing the same sentence twice. For t h i s reason, the semi-randomization of the subsets was constrained such that no clause appeared i n two consecutive sentences. Assignment of Required Response Values to Sentences Once the four subsets had been compiled and semi-randomized, the next variable was dealt with. As indicated previously, f o r Task 1, two s l i d e s were presented and the subject was expected to decide whether the l e f t s l i d e or the r i g h t s l i d e was appropriate. Likewise, for Task 2, the sub-j e c t was presented with one s l i d e and was expected to decide whether the s l i d e corresponded to the sentence or not (yes/no response). Pr i o r to te s t i n g , i t was, therefore, necessary to deter-mine for each sentence i n each subtest, whether the correct s l i d e would be presented on the l e f t or r i g h t (for Task 1) or whether the correct s l i d e or f o i l would be presented (for 44 Task 2 ) . To accomplish t h i s , each stimulus sentence was randomly assigned a number (1 or 2 ) , with the constraint that there would be an equal occurrence of each number i n each sub-set. For Task 1, an assignment of 1 corresponded to the correct s l i d e appearing on the l e f t . For Task 2, an assign-ment of 1 meant that the correct s l i d e and not the f o i l would be presented. Thus, i n each testing subsets, each of the two responses ( l e f t / r i g h t , or yes/no) would be indicated randomly and with equal frequency. Preparation of Vis u a l Stimuli The events described i n the sentences were i l l u s t r a t e d by a professional a r t i s t . Each of the for t y clauses were i l l u s t r a t e d by single l i n e drawings, to scale, and executed to the s a t i s f a c t i o n of both the a r t i s t and the author. These were reproduced, then matched with each of the two other i l l u s t r a t i o n s with which i t belonged, for conversion to standard 35 mm black and white s l i d e s . Each complete s l i d e , therefore, contained two i l l u s t r a t i o n s (corresponding to the two clauses of each sentence) which depicted the events of the sentence from l e f t to r i g h t . The t o t a l s l i d e area was divided into h a l f , with each i l l u s t r a t i o n occupying equal areas on the s l i d e . Constraints Applied to the Vis u a l Stimuli A number of conditions and precautions were observed i n preparing the i l l u s t r a t i o n s . The i l l u s t r a t i o n s were kept as v i s u a l l y simple as possible while s t i l l providing enough con-textual d e t a i l s to provide a c l e a r and unambiguous interpre-t a t i o n ; no unnecessary d e t a i l c l u t t e r e d the i l l u s t r a t i o n s . I t was necessary for each clause to be i l l u s t r a t e d i n -dependently of the other clauses, but, none-the-less, i n such a way as to make them v i s u a l l y compatible with t h e i r sentence-mate clauses. This i s an important point, f o r r e c a l l that each clause was matched with another clause which resulted i n a constrained sequence of events and one clause which resulted i n a more a r b i t r a r y sequence of events. I t was, therefore, not desirable to have a clause i l l u s t r a t e d such that i t could v i s u a l l y be t i e d together more with the other clause when they were matched to form constrained sentences than when i t was matched with a d i f f e r e n t clause to form an unconstrained sentence. I t was necessary that no bias associating the two clauses of a constrained sentence should r e s u l t from the depiction; i f such a bias did a f f e c t the l i s t e n e r , t h i s should r e s u l t from the sentential content and not from how the sentences were i l l u s t r a t e d . This l a s t point affected many de-t a i l s of the i l l u s t r a t i o n s . For example, consider the clause, A he. ate. dz&&e.n.t. This clause was matched with & he. ate. the. Atzak (to form a C sentence) and with &hz got tn the. boat (to form a U sentence - see Table I I ) . Thus, the clause i n question had to be depicted such that the de.6ie.n.t being eaten by the actor was one which could be eaten with equal l i k e l i -hood on a boat or i n a l o c a t i o n where one was also l i k e l y to eat a steak ( i . e . , at a kitchen t a b l e ) . Also, of course, neither the boat nor the kitchen table could a c t u a l l y be 46 depicted i n the i l l u s t r a t i o n of ike. ate. de.&&zKt, since t h i s would cause i t to be incompatible v i s u a l l y with one of i t s two matching clauses. V i r t u a l l y every i l l u s t r a t i o n was held up to such scrutiny with respect to how well i t depicted the clause and how compatible i t would be with both of i t s sentence-mate clauses. For sim i l a r reasons, wherever l o g i c a l l y possible, the actor was portrayed as wearing the same set of clothes throughout the i l l u s t r a t i o n s . The above point i s not t r i v i a l . I t refers back to the issue raised i n a previous section, concerning the difference between sentences which are unconstrained by general know-ledge versus those that are not simply unconstrained but also implausible. Just as clauses were not merely randomly matched to produce possibly implausible sentences, care was taken to ensure that the s l i d e s which i l l u s t r a t e d these sentences resulted i n equally v i s u a l sequences to accompany both the constrained and unconstrained sentences. Had the clauses not been i l l u s t r a t e d i n t h i s "independent yet v i s u a l l y compatible" manner, the unconstrained sentences might have appeared to be a less l i k e l y sequences of events (visually) than the constrained sentences. One further constraint was placed on these i l l u s t r a t i o n s ; i t was not permissible for them to contain any l i n g u i s t i c content. I t was considered that such an in c l u s i o n might have affected the time needed for v i s u a l processing of i n d i v i d u a l s l i d e s , thus confounding the investigation. Hence, i n some 47 i l l u s t r a t i o n s , such as those taking place i n a hospital or i n a store, printed signs were not considered to be an acceptable way of conveying context. A t o t a l of 80 s l i d e s were thus prepared. Forty s l i d e s depicted a sequence of two events, shown i n a l e f t to r i g h t order which corresponded with the f o r t y s t i m u l i sentences. The remaining 40 s l i d e s served as f o i l s ; they showed the same 40 sequences of two events i n the reverse order. Preparation of Taped Stimuli The 160 sentences were recorded at a normal speaking rate, at a tape recorder speed of lh inches per second. Stressed sections of the sentences peaked at O dB on the VU meter and the microphone was set at a constant distance from the speaker. The tape recorder used was a Revox, Model 77A; recordings were made on acetate audiotape (Ampex 406) . These sentences were recorded on Track 1 of the audiotape; on Track 2 a constant 1000 Hz tone was recorded. The purpose of the tone, which was to t r i g g e r c e r t a i n mechanisms of the reaction-time equipment, w i l l be dealt with more f u l l y l a t e r i n t h i s chapter. S p l i c i n g of the Taped Stimuli Once recorded, the sentences were s p l i c e d , then joined with leader tape to form the four t e s t i n g subsets Ca, b, c 48 and d). The purpose of s p l i c i n g these sentences was not merely organizational. More importantly, the s p l i c e at the beginning of each sentence ensured that the onset of each sentence would be exactly synchronized with the onset of the 1000 Hz tone, on Track 2. I t was highly undesirable, there-fore, that the s p l i c e occur either p r i o r to or aft e r the on-set of the sentence; i f p r i o r , the 1000 Hz tone would be present before the sentence began, thus tri g g e r i n g the equip-ment prematurely; i f a f t e r , the onset of the sentence would be l o s t . The sentences were played on an Otari tape recorder. Once the onset of each sentence had been roughly located at a speed of 7% inches per second, the machine was turned to the ed i t mode. The tape was then slowly advanced and re-tracted over the playback head u n t i l the f i r s t g l o t t a l pulse had been located. This point was marked with a fine point pen. This procedure was repeated, by hand, two times to ascertain that the correct spot on the tape had been located. The loca-tions was then rechecked at lh inches per second, to ensure that the area marked was indeed the onset of speech, and not a preceding throat-clearing or other such noise. (The end of each sentence was s i m i l a r l y located. Since the exact ending of the sentence was not c r u c i a l , the experimental concern here was that the ent i r e sentence be preserved and not sp l i c e d pre-maturely. I t was therefore, permissible, i n t h i s case, to mark the tape f o r s p l i c i n g following the end of voice, rather than exactly at the instant the voice ended.) 49 The tape was then removed from the machine, cut with a razor blade v e r t i c a l l y (to produce a sudden and not gradual onset) at the points marking the onset and end of each sentence, and spliced c a r e f u l l y to white leader tape. Five seconds worth of leader tape was used between each sentence. The r e s u l t of the foregoing procedure was four r e e l s , each containing a testing subset of 40 sentences, joined by leader tape. Each set was i n t e r n a l l y arranged i n the pre-determined experimental order. Equipment A device for measuring response time was designed and b u i l t s p e c i f i c a l l y for t h i s experiment. The major function of t h i s device was to co-ordinate the timing of the various components of the experiment and to provide an accurate means of measuring the subjects' response times to the s t i m u l i . The experimental equipment included the device i t s e l f , plus two s l i d e projectors, a Revox tape recorder, and a subject response box which contained two pushbuttons. Figures 1 and 2 depict the equipment and set-up schematically. Stimulus Presentation and Subject Response As each stimulus sentence was played, the 1000 Hz tone, recorded on Tract 2, of the audiotape, triggered the e l e c t r o -nic switch of the device. The e l e c t r o n i c switch, i n turn, triggered both r e l a y A and the timer of the device. (The Si u H 0) o > Q_D 2 Track Tape Recorder Projector 2 Remote Control & Power Trqck 2 Pulse Remote Control Response Box F i g u r e 1 B l o c k D i a g r a m o f E x p e r i m e n t a l A p p a r a t u s o Tape Recorder OPERATOR Figure 2 Physical Set-up of Experimental Apparatus p u r p o s e o f t h e t i m e r was t o p r o v i d e a means o f a u t o m a t i c a l l y t u r n i n g o f f t h e e q u i p m e n t i n t h e e v e n t o f no r e s p o n s e f r o m t h e s u b j e c t a f t e r an e l a p s e d t i m e o f one m i n u t e f r o m t h e p r e s e n t a -t i o n o f t h e s t i m u u s . ) R e l a y A s i m u l t a n e o u s l y a c t i v a t e d b o t h s l i d e p r o j e c t o r s , t h e d i g i t a l s t o p w a t c h , and a d e l a y e d r e l a y B. T h u s , i t was assumed t h a t t h e o n s e t o f e a c h s e n t e n c e began s i m u l t a n e o u s l y w i t h t h e p r o j e c t i o n o f t h e a c c o m p a n y i n g s l i d e (s) and w i t h t h e o n s e t o f t h e s t o p w a t c h . The s t o p w a t c h m e a s u r e d e l a p s e d t i m e between t h e s t a r t o f t h e s e n t e n c e and t h e s u b j e c t ' s r e s p o n s e , t o t h e n e a r e s t one h u n d r e d t h o f a s e c o n d . The p u r p o s e o f t h e d e l a y e d r e l a y B was t o i n t e r r u p t t h e t r i g g e r i n g s i g n a l g o i n g t o t h e s t o p w a t c h . T h i s i n t e r r u p t i o n was n e c e s s a r y s i n c e t h r e e f u n c t i o n s ( " s t o p " , " s t a r t " , and " r e s e t " ) , were c o n t r o l l e d on one c o n t a c t o f t h e s t o p w a t c h . T h e r e f o r e , o n c e t h e s t o p w a t c h had been t r i g g e r e d t o " s t a r t " , t h i s s i g n a l must be i n t e r r u p t e d i n o r d e r t h a t , a t a l a t d r i n s t a n t , a n o t h e r t r i g g e r i n g i m p u l s e c o u l d g e t t h r o u g h t o t r i g g e r t h e s t o p w a t c h t o " s t o p " . The v o l t a g e n e c e s s a r y t o a c c o m p l i s h t h e " s t o p " i m p u l s e was p r o v i d e d by one c h a n n e l o f t h e s u b j e c t ' s p u s h b u t t o n . The o t h e r c h a n n e l o f t h e p u s h b u t t o n mechanism p r o v i d e d v o l t a g e t o t u r n o f f t h e s l i d e p r o j e c t o r s . T h e r e f o r e , o n c e t h e s u b j e c t had i n d i c a t e d h i s / h e r r e s p o n s e b y p u s h i n g t h e a p p r o p r i a t e p u s h b u t t o n , b o t h t h e p r o j e c t o r s and t h e s t o p w a t c h were t u r n e d o f f . i A t t h i s p o i n t , t h e s t o p w a t c h r e a d o u t d i s p l a y e d t h e e l a p s e d t i m e and i n d i c a t e d w h i c h p u s h b u t t o n h a d been p r e s s e d by t h e s u b j e c t . T h i s d i s p l a y was h e l d u n t i l t h e s t o p w a t c h had b e e n m a n u a l l y r e s e t by p r e s s i n g t h e r e s e t b u t t o n p r o v i d e d on t h e d e v i c e . ( T h e r e was no d a n g e r o f t h e f o l l o w i n g s e n t e n c e b e i n g p l a y e d p r i o r t o t h e s u b j e c t s ' r e s p o n s e , s i n c e t h e Revox a u t o m a t i c a l l y c e a s e d a d v a n c i n g when t h e w h i t e l e a d e r t a p e p a s s e d t h e p l a y b a c k head.) Once t h i s i n f o r m a t i o n had been r e c o r d e d , t h e s l i d e p r o -j e c t o r (s) and t a p e were a d v a n c e d , a n d t h e r e s e t b u t t o n on t h e s t o p w a t c h were d e p r e s s e d , i n p r e p a r a t i o n f o r p r e s e n t a t i o n o f t h e n e x t s t i m u l u s . ( T h i s p r o c e d u r e , i n c l u d i n g r e c o r d i n g r e s p o n s e t i m e , c o u l d be p e r f o r m e d i n a p p r o x i m a t e l y f i v e s e c o n d s , w h i c h was t h e minimum d e l a y between t h e r e s p o n s e t o one s t i m u l u s and p r e s e n t a t i o n o f t h e n e x t . ) S i n c e o n l y one s l i d e p r o j e c t o r was n e c e s s a r y f o r T a s k 2, t h e s e c o n d p r o j e c t o r was d i s c o n n e c t e d i n t h i s c a s e . T e m p o r a l R e l a t i o n o f S l i d e and S e n t e n c e S t i m u l i I t s h o u l d be n o t e d t h a t t h e e q u i p m e n t d e s c r i b e d above p r o v i d e d a means o f e n s u r i n g t h a t t h e s l i d e s t i m u l i and s e n t e n c e s t i m u l i were p r e s e n t e d s i m u l t a n e o u s l y . T h i s was one o f s e v e r a l p o s s i b l e t e m p o r a l r e l a t i o n s w h i c h c o u l d have been imposed on t h e p r e s e n t a t i o n o f s l i d e and s e n t e n c e s t i m u l i ; f o r e xample, i n s t e a d o f b e i n g p r e s e n t e d s i m u l t a n e o u s l y , t h e s l i d e s c o u l d have b e e n p r e s e n t e d e i t h e r p r i o r t o , o r f o l l o w i n g , t h e p r e s e n t a t i o n o f t h e s e n t e n c e s . E a c h o f t h e s e t h r e e p o s s i b l e r e l a t i o n s was c o n s i d e r e d . 54 F i r s t possible r e l a t i o n At l e a s t two disadvantages were moted i n association with presentation of the s l i d e s p r i o r to the onset of the sentences; f i r s t , the processing required of subjects i n order to make a decision would involve the confounding i n -fluence of VSLAUCLI mzmoKy; second, i n the case of s l i d e s de-pi c t i n g C sentences, i t would be possible for the subjects to make a choice, based on pr o b a b i l i t y , without reference to the following sentence. Second possible r e l a t i o n Had the s l i d e s not been presented u n t i l a f t e r auditory presentation of the sentences had been completed, then aadl-toKy mzmosiy would have s i g n i f i c a n t l y affected response time. Presenting s l i d e s after auditory presentation would, however, have eliminated the problem associated with where response time should be measured from (when dealing with sentences of unequal length); the response time could simply have been measured from the instant at which the s l i d e s were projected. (This problem w i l l be dealt with i n the following chapter). Upon further consideration, however, t h i s argument proved specious - one would merely be trading a more obvious problem (the length of time i t takes for each stimulus sentence to be uttered) for a more subtle problem (the length of time i t takes for each s l i d e stimulus to be v i s u a l l y processed, which, i n order for the previous argument to be v a l i d , must be assumed to be the same for each s l i d e - a doubtful assumption) . 55 Third possible r e l a t i o n Associated with the presentation of both st i m u l i simultaneously, was the disadvantage of determining a suitable point i n the sentence from which to measure response time. However, t h i s solution held the a t t r a c t i o n that i t would not contaminate the task with either v i s u a l or auditory memory. This i s also the temporal r e l a t i o n which most nearly mimicks natural language; just as i n spontaneous language sentences are generally uttered with reference to a context, i n t h i s investigation the s l i d e s provide a reference for the accompanying sentences at the time of utterance. On the basis of the foregoing, simultaneous presentation of both s l i d e and sentence st i m u l i appeared to be the best experimental option. Subjects Sixteen subjects, 8 men and 8 women, were recruited to pa r t i c i p a t e i n t h i s investigation. A l l were young, healthy adults, i n t h e i r twenties, with no known history of v i s u a l or auditory perception d e f i c i t s . These subjects a l l reported that English was t h e i r native language; only one subject con-sidered himself to be reasonably fluent i n another language (German). Every e f f o r t was made to draw these subjects from as wide a variety of educational and occupational backgrounds as possible. No two subjects were involved i n pr a c t i c i n g the same occupation or i n studying the same d i s c i p l i n e . The 56 mean amount of post-secondary education for these subjects was 2.59 years, with a range of from one year less than high school completion to 6 years of post-secondary schooling. For "the male subjects, the mean educational l e v e l was 3.44 years, with a range of from zero to six years. For the female subjects, the mean educational l e v e l was 2.2 years of post-secondary education, with a range from -1 to f i v e years. Mean age of the subjects at the time of testing was 25 years, 5.3 months, with a range from 23 years, 1 month to 29 years, 1 month. Broken down with respect to sex, the mean ages of the subjects were 25 years'11.6 months for men and 24 years, 10.9 months for women, ranging from 24 years, 2 months to 28 years 11 months, and 23 years 1 month to 29 years, 11 months, respectively. Balancing Subjects, Tasks, and Subsets Each of the 16 subjects participated i n both Task 1 and Task 2 of the experiment. Since there were four t e s t subsets of sentence s t i m u l i , t h i s required balancing these four sets, a, b, c, and d), and the two Tasks (1 and 2) with two groups of eight subjects (male and female). Half of the subjects participated i n Task 1 f i r s t , and half i n Task 2 f i r s t . Each te s t subset was given with equal frequency for Task 1 and Task 2, and with equal frequency as f i r s t or second task i n which the subject pa r t i c i p a t e d . Using these c r i t e r i a for tes t construction, eight groups of subsets were formulated. Each of these p a i r s determined the t e s t i n g content and order f o r one male and one female s u b j e c t . These e i g h t p a i r s are given i n the f o l l o w i n g t a b l e : Table 4 a Counterbalancing of Tasks, Subsets and Subjects Subset Task Subset Task Number of Subjects Subset Task Subset Task Male Female a 1 d 2 1 1 a 2 c 1 1 1 b 2 a 1 1 1 b 1 a 2 1 1 c 2 d 1 1 1 d 2 b 1 1 1 c 1 b 2 1 1 d 1 c 1 1 1 T o t a l 16 Presentation order i s read l e f t to r i g h t , i . e . , The f i r s t row indicates that subset "a" was presented as Task 1, followed by presentation of subset "d" as Task 2. 58 Procedure The experiment was conducted i n a sound-proofed lab-oratory (Nc -22) with minimal ambient noise and a comfortable climate. Physical arrangement of the equipment was held con-stant for each subject. Experiments were run on subjects i n d i v i d u a l l y , with only the examiner and subject present. Prior to tes t i n g , the subjects were required to sign an i n -formed consent form and to supply some biographical informa-t i o n . Anonyminity of the subjects was guaranteed i n a l l cases. The consent form was composed such that the informa-ti o n given to the subject at that time, although accurate, was s u f f i c i e n t l y vague that the subjects would be unaware both of which s p e c i f i c parameters of the sentences were of i n t e r e s t to the examiner and which subject behaviours were being examined. Each subject was informed that the purpose of the experiment would be explained to his/her s a t i s f a c t i o n upon completion of the tasks. Once these preliminaries had been accomplished, the subject was seated within easy reach of the response box. (The pushbuttons were l a b e l l e d " l e f t / y e s " and "right/no" to serve as a reminder to the subject during the experiment.) Pre-recorded instructions were presented at t h i s point (re-produced i n the following section.) These in s t r u c t i o n s included a sample of speech i n which the subject was asked i f the volume was at a comfortable l i s t e n i n g l e v e l ; i f not, i t was adjusted to the subject's s p e c i f i c a t i o n s . Two sets 59 of i n s t r u c t i o n s were prepared; the set which the subject heard f i r s t depended on whether s/he p a r t i c i p a t e d i n Task 1 or Task 2 f i r s t . F o l l o w i n g p r e s e n t a t i o n of i n s t r u c t i o n s , the subject was presented w i t h three t r a i n i n g sentences. (These sentences were s y n t a c t i c a l l y d i f f e r e n t from the t e s t sentences and were designed to ensure t h a t the subjects had understood the i n s t r u c t i o n s . ) A l l subjects c o r r e c t l y responded to a l l three t r a i n i n g sentences. F o l l o w i n g p r e s e n t a t i o n of t r a i n i n g sentences, more i n f o r m a t i o n , regarding the nature of the t e s t -in g sentences, was given. I n s t r u c t i o n s were then b r i e f l y r e i t e r a t e d , and t e s t i n g commenced. During t e s t i n g , the experimenter and the c o n t r o l apparatus were placed behind the subject and out of h i s / h e r view. The o n l y experimental a p p a r a t i v i s i b l e to the subject were the p r o j e c t o r s c r e e n ( s ) , and the pushbutton response box. Each ta s k , w i t h i n s t r u c t i o n s , took about 20 minutes to ad-m i n i s t e r . The subject then had a break. During t h i s break, the examiner prepared f o r the f o l l o w i n g task. (The s l i d e s t i m u l i had been organized to correspond w i t h the order of the f i r s t t e s t subset; i t was, t h e r e f o r e , necessary to r e -arrange the s l i d e s i n the c o r r e c t order f o r the second task. A l l s l i d e s were coded f o r t h i s purpose.) The subject then returned to the l a b o r a t o r y , a t which time s/he was r e - i n -s t r u c t e d and r e t a i n e d f o r the second task. The e n t i r e experiment, i n c l u d i n g the break, took approximately one hour f o r each s u b j e c t . I n s t r u c t i o n s Verbatim I n s t r u c t i o n s O b t a i n i n g the most comfortable l i s t e n i n g l e v e l The f o l l o w i n g was read to each s u b j e c t once, a t the be g i n n i n g of the experiment: Before we b e g i n , I'd l i k e t o make sure the volume i s a t a comfortable l e v e l f o r you to l i s t e n t o . L e t me know i f you would l i k e i t turned up or down? Is i t a t a comfortable l e v e l f o r you now? S p e c i f i c i n s t r u c t i o n s : Task 1 The f o l l o w i n g was read t o each s u b j e c t , p r i o r to Task 1: Now you are going to hear some sentences come over the loudspeaker, one a t a time. When you hear each sentence, you w i l l a l s o see two s l i d e s . One s l i d e w i l l go w i t h the sentence, and one s l i d e w i l l not. I want you to d e c i d e which s l i d e goes wit h the sentence. In f r o n t of you there are two bu t t o n s . I f the c o r r e c t s l i d e i s on the l e f t , I want you to pr e s s the bu t t o n on the l e f t . I f the c o r r e c t s l i d e i s on the r i g h t , p r e s s the bu t t o n on the r i g h t . You w i l l hear each sentence o n l y once, so l i s t e n c a r e f u l l y . When you've d e c i d e d which i s the c o r r e c t s l i d e , push the a p p r o p r i a t e b u t t o n . A few seconds l a t e r , y o u ' l l hear the next sentence. For each sentence, i n d i c a t e your c h o i c e by pushing the bu t t o n on the same s i d e as the c o r r e c t s l i d e appears. Do you have any qu e s t i o n s ? (Pause.) We are going t o t r y some p r a c t i c e sentences now... S p e c i f i c i n s t r u c t i o n s : Task 2 The f o l l o w i n g was read to.each s u b j e c t , p r i o r to Task 2: 61 Now you are going to hear some sentences come over the loudspeaker, one at a time. When you hear each sentence, you w i l l also see a s l i d e . The s l i d e may or may not go with the sentence you hear. Some sl i d e s go with the sentences and some s l i d e s do not. I want you to decide whether or not the s l i d e c o r r e c t l y depicts the sentence. In front of you there are two buttons. If the s l i d e goes with the sentence, push the button on the l e f t . I f the s l i d e does not go with the sentence, press the button on the r i g h t . You w i l l hear each sentence only once, so l i s t e n care-f u l l y . When you've decided i f the s l i d e goes with the sentence or not, press the appropriate button - l e f t for correct, r i g h t for incorrect. A few seconds l a t e r y o u ' l l hear the next sentence. For each sentence, indicate your decision by pressing the l e f t button i f the s l i d e goes with the sentence, or the button on the r i g h t i f i t does not. Do you have any questions? (Pause.) We are going to t r y some practice sentences now... General instructions Task 1 and Task 2 The following was read to each subject after presentation of the practice sentences: The sentences and s l i d e s which follow are somewhat d i f f e r e n t . The sentences describe a sequence of two events, which are joined by either hzfaofiz or afate.fi. As i n a cartoon, the l e f t picture represents the events which happened f i r s t , and the ri g h t picture represents the following event. Therefore, your task i s to determine i f the order of events depicted i n each s l i d e corresponds with the order of events as described by the sentence. Reiteration of i n s t r u c t i o n s : Task 1 The following was read to each subject, immediately p r i o r to presentation of Task 1 s t i m u l i : So each time, decide which of the two sl i d e s shows the.two events i n the same order as In the sentence, and press the button on the same side as the correct s l i d e appears. 62 Reiteration of instructions : Task 2 The following was read to each subject, immediately p r i o r to presentation of Task 2 s t i m u l i : So, each time, decide whether or not the s l i d e shows the two events i n the same order as i n the sentence, and press the button marked yes i f i t does, or the button marked no i f i t does'nt. A Comment on these Instructions One point concerning these instructions merits discussion. The wording of instructions can strongly bias a subject's res-ponses by the amount of information they contain regarding the task under experimental scrutiny. In the present case, i t was necessary to decide whether a subject should be informed that the dimension of the response being measured was speed, not accuracy. In order not to bias the subjects, instructions to "respond as quickly as possible" were omitted. F a i l u r e to include such a statement may have resulted in responses being more accurate but less rapid than they would have been had the subjects been instructed to act quickly. The rationale for t h i s decision was to recreate as n a t u r a l i s t i c a language environment as can be accomplished i n the context of a con-t r o l l e d laboratory experiment. In a normal conversation, a l i s t e n e r generally places more emphasis on his/her under-standing of the content of what i s being said, than on the speed at which s/he understands the message. Thus, suggesting that subjects answer "as quickly as possible" could have caused subjects to engage a strategy for comprehension which 63 might be co n s i d e r a b l y d i f f e r e n t than t h a t used i n n a t u r a l i s t i c language environments. Subjects were, t h e r e f o r e , not informed t h a t the speed of t h e i r responses was the v a r i a b l e being i n v e s t i g a t e d . In the event t h a t subjects asked how much time they had i n which to respond (as some did) they were t o l d non-commitally t h a t no s t i m u l i would be presented u n t i l the previous one had re c e i v e d a response. 64 CHAPTER THREE TREATMENT OF DATA AND RESULTS Overview The data from both Task 1 and Task 2 were analyzed by means of a computer-assisted analysis of variance. The raw data were transformed to square roots of the reaction times for t h i s purpose. Two major factors were of int e r e s t ; Szmanttc and Syntactic. I t should be noted that for the pur-pose of analysis, the terms Syntactic and Semantic' are used loosely. I t i s acknowledged that a l l of those factors described as Syntactic could also be defined semantically. The purpose of such labels i n the analysis was organizational only. For t h i s purpose, a Syntactic factor was here defined as one which has the a f f e c t of changing the surface s t r i n g of the elements within a sentence while retaining the same operational meaning; a Semantic factor i s one which functions to a l t e r the operational meaning of a sentence. In a l l analyses, the Semantic factor was analyzed at two le v e l s ; Con&tKdlnzd and UnconltfiOilYizd. The Syntactic factor 65 was a n a l y z e d by t h r e e s e p a r a t e a n a l y s e s , a t two l e v e l s o f one o f the f o l l o w i n g : 1) C o n j u n c t i o n c h o i c e 2) Order o f ment ion 3) C l a u s e p l a c e m e n t . F o r Task 2 a n a l y s e s , an a d d i t i o n a l f a c t o r , Ttiuth \Jdtixe., was i n v e s t i g a t e d ; t h i s r e f e r r e d to whether the r e q u i r e d response to a p a r t i c u l a r s t i m u l u s had been yz& o r no. In a l l c a s e s , the g e n e r a l n u l l hypotheses were t h a t t h e r e e x i s t e d no s i g n i f i c a n t d i f f e r e n c e between the l e v e l s o f the v a r i o u s f a c t o r s . F a i l u r e to r e j e c t the n u l l h y p o t h e -s i s was e v i d e n t f o r a l l n u l l hypotheses except those c o n c e r n i n g C l a u s e Placement and T r u t h V a l u e . S i g n i f i c a n t d i f f e r e n c e s between the mean v a l u e s were e v i d e n t f o r these f a c t o r s , w i t h the mean h a v i n g s m a l l e r v a l u e s f o r m a i n - c l a u s e -i n i t i a l s entences and f o r yzi r e s p o n s e s , r e s p e c t i v e l y . A t no p o i n t i n the a n a l y s e s were any s i g n i f i c a n t i n t e r a c t i o n e f f e c t s e v i d e n t . S i n c e the d a t a from Task 1 and Task 2 were t r e a t e d i d e n t i c a l l y , i n most r e s p e c t s , s e p a r a t e d i s c u s s i o n o f the a n a l y s i s f o r these exper iments would prove r e d u n d a n t . T h i s c h a p t e r , t h e r e f o r e , i s d i v i d e d i n t o s e c t i o n s d e a l i n g w i t h t rea tment o f a l l d a t a and not i n t o s e p a r a t e s e c t i o n s d e a l i n g w i t h Tasks 1 and 2, e x c e p t as a p p r o p r i a t e . 66 Treatment of Data Prior to Analysis S u i t a b i l i t y of Data Raw data was excluded from the analysis on the basis of either subject or experimenter error. Several kinds of error or mishap occurred during t e s t i n g . For example, i f the sub-je c t responded with an incorrect answer, the reaction time necessary to a t t a i n t h i s incorrect response was not analyzed. These subject errors f e l l into two categories; those which appeared to be true miscomprehensions of the sentences and those which resulted from the subject accidentally depressing the wrong response button. (The l a t t e r were usually imme-d i a t e l y i d e n t i f i e d by the subject.) Occasionally, the sub-j e c t would attempt to engage the investigator i n conversation just as the taped stimuli began, thus obscuring the f i r s t few words of the sentences. S i m i l a r i l y , occasional investigator errors occurred, e.g. : 1) i n s e r t i n g a s l i d e i n c o r r e c t l y ; either i n incorrect order or i n an incorrect orientation, 2) a s l i d e appearing our of focus. In any of the above instances, or i n any other case where the taped and/or v i s u a l stimuli were not consistent with each other, the p a r t i c u l a r stimulus item i n question was excluded from the analysis. I t should be noted that the t o t a l number of rejected s t i m u l i was very small. Of a possible 640 responses for each 67 task, 604 and 614 were recorded and analyzed for Task 1 and 2, respectively. Measurement Procedure D e f i n i t i o n and measurement of reaction time posed a s i g n i f i c a n t problem. As described i n Chapter Two, an e l e c t r o -nic stopwatch was used to record elapsed time between onset of each stimulus sentence and depression of the button i n d i -cating a subject's choice. Since sentences varied somewhat i n length, i t was inappropriate to use t h i s recorded time as an i n d i c a t i o n of the time required for the subject to respond to each stimulus. (Obviously, a raw response time of four seconds, as measured from the onset of each sentence, would not equivalently measure response to two d i f f e r e n t sentences, one of which was two seconds i n duration and the other three seconds.) I t was necessary, therefore, to choose some point other than the onset of the sentence as an a r b i t r a r y zero on the time scale. The f i r s t p o s s i b i l i t y considered was to measure response time from the end of each sentence. This seemed to be a l o g i c a l point to define as "zero" time since i t could be con-s i s t e n t l y applied across a l l sentences and since,, at t h i s point, a l l information necessary for sentence processing would have occurred. In the course of running the experiment, how-ever, i t was noted that a large number of responses were made pr i o r to the completion of the sentence s t i m u l i . This indicated that, once the task was known, i t was not necessary for a subject to hear each sentence i n i t s e n t i r e t y i n order 68 to perform the t a s k . C l e a r l y , a l a r g e p o r t i o n of sentence p r o c e s s i n g had o c c u r r e d p r i o r t o completion of the a u d i t o r y s i g n a l . Given t h i s o b s e r v a t i o n , i t seemed meaningless to d e f i n e as "zero" a p o i n t i n the sentence where much of the p r o c e s s i n g had a l r e a d y been accomplished; i t was, t h e r e f o r e , necessary to d e f i n e a d i f f e r e n t p o i n t as zero time. Measurement of response time was, t h e r e f o r e , begun f o l l o w i n g the f i r s t c l a u s e of the sentence (plus the con-j u n c t i o n bzfaoA.z or afatzn. i n the case of m a i n - c l a u s e - f i r s t s e n t e n c e s ) . In the f o l l o w i n g examples, the c u t - o f f p o i n t i s i n d i c a t e d : She packed the s u i t c a s e before/she got on the plane. Before she got on the plane/she packed the s u i t c a s e . She got on the plane a f t e r / s h e packed the s u i t c a s e . A f t e r she packed the s u i t c a s e / s h e got on the p l a n e . I t was f e l t t h a t , a t t h i s p o i n t , a s u b j e c t would have heard a l l necessary and s u f f i c i e n t i n f o r m a t i o n r e q u i r e d to respond a p p r o p r i a t e l y . Since the stopwatch had p r e v i o u s l y measured time from the b e g i n n i n g of each sentence, recorded time f o r each r e s -ponse now i n c l u d e d t h a t time which has above been d e f i n e d as the r e a c t i o n time, p l u s the time f o r the f i r s t c l a u s e (with bo.faon.2. or afatz/i) to be u t t e r e d . For each recorded sentence, i t was necessary, t h e r e f o r e , to measure the time occupied on the tape up to the c u t - o f f p o i n t , i n order t h a t t h i s c o u l d be s u b t r a c t e d from the recorded response time to o b t a i n r e a c t i o n time. For each sentence, t h i s time was measured using the same stopwatch used for the experiment. Each sentence seg-ment was measured a t o t a l of f i v e times. Some small v a r i a -b i l i t y i n these measures was obtained; therefore, the longest and shorted measures obtained for each sentence were d i s -carded, and the remaining three were averaged. These averaged measurements yielded a master l i s t which contain the time, i n hundredths of a second, from the onset of each sentence to i t s cut-off point. This l i s t was compared to subjects' recorded response times, and, for each response, the pre-measured times were subtracted i n order to obtain the subjects' reaction times. These reaction time data formed the basis for analyses. Transformational of Data The data were transformed, i n order to a t t a i n maximum homogeneity of variance, by taking the square roots of the RTs, while maintaining t h e i r signs. Analysis The transformed data were analyzed by a t o t a l of six computer-assisted Analyses of Variance (ANOVAs); three dealt with data from Task 1 and three dealt with data from Task 2. It was necessary to perform three separate analyses for each task i n order to separate out the three syntactic variables of i n t e r e s t i n the sentence. Table V should make clear the T a b l e V F a c t o r s D i s t i n g u i s h i n g t h e E i g h t S e n t e n c e T y p e s FACTORS Sentence Type Semantic Constraints Order o f mention Conjunction choice Clause placement Examples Constrained 0ME=O0E Before I n i t i a l Main Clause 1 . She packed the suitcase - + + + BEFORE she washed the dishes. 2 . BEFORE she washed the - - + dishes, she packed the suitcase. .3. She washed the dishes - - + AFTER she packed the suitcase. 4. AFTER she packed the - + -suitcase, she washed the dishes. 5 . She ate dinner BEFORE + + + + she washed the dishes. 6. BEFORE she washed the + - + dishes, she ate dinner. 7. She washed the dishes + - - + AFTER she ate dinner. 8. AFTER she ate dinner + + she washed the dishes. 71 various variables pertinent to each sentence type. In order to analyze the data for the three syntactic variables, the data from the d i f f e r e n t sentence types had to be conflated i n various ways. That i s : 1) For the variable bzfioiz versus afite.1, data from sentence types 1, 2, 5 and 6 were grouped together versus data from sentence types 3, 4, 7 and 8. 2) For the syntactic variable OOE=OME versus OOE=OME, data from sentence types 1, 4, 5 and 8 were grouped versus data from sentence types 2, 3, 6 and 7. 3) For the syntactic variable Main-clause-first versus Subordinate-clause-first, data from sentence types 1, 3, 5 and 7 were grouped versus data from sentence types 2, 4, 6 and 8. 4) For the semantic v a r i a b l e , Constrained versus Un-constrained, data from sentences 1-4 were grouped versus data from sentences 5-8. In each of the analyses, one of the three groupings, described previously, was defined as the "syntactic variable under consideration" for that p a r t i c u l a r computer run. This syntactic variable always had two values, plus or minus, as shown i n Table V. The semantic variable was constant for a l l runs and had two le v e l s also; these corresponded to "Con-strained" versus "Unconstrained". For data from Task 2 only, a t h i r d v a r i a b l e , "Truth Value", was analyzed. "Truth Value" r e f e r s to whether a p a r t i c u l a r stimulus item had required a 72 pos i t i v e or negative response. In a l l cases the 16 subjects served as blocks. Prior to the main analyses, several preliminary analyses were conducted i n order to determine whether two other v a r i -ables could r e l i a b l y be treated as random variab l e s . These were Subject Sex and Subject Type. (Subject Type refe r s to those subjects who had participated i n Task 1 f i r s t versus those subjects who had par t i c i p a t e d i n Task 2 f i r s t . ) Both Subject Sex and Type were found to be non-significant and were, therefore, treated as random variables for the major analyses. The computer program by which the six following ANOVAs were accomplished also gave three frequencies, means and standard deviations for each l e v e l of each factor. No s i g n i -f i c a n t i n t e r a c t i o n e f f e c t s were obtained. For a l l analyses, p = .05. Task 1 Analysis IA ; Syntactic factor s i g n i f i e s conjunction  choice A 2 x 2 ANOVA was performed, with the Syntactic factor, i n t h i s case, r e f e r r i n g to conjunction choice. The Syntactic factor had two l e v e l s , Bcfaoic or kfatch.; the Semantic factor also had two l e v e l s , Constrained and Unconstrained; the s i x -teen subjects served as blocks. The following nutt hypotkzie.6 were tested; H Q ^ : There i s no s i g n i f i c a n t difference between the mean values for the sixteen l e v e l s of the Subject factor. HQ2 : There i s no s i g n i f i c a n t difference between the mean values for the two le v e l s (Constrained and Unconstrained) of the Semantic factor. HQ2Z There i s no s i g n i f i c a n t difference between the mean values for the two leve l s (Before and After) of the Syntactic factor. Re.Au.ltA are indicated i n Table VI: Table VI Results of Analysis IA : Analysis of Variance Table Sum of Mean Proba-Source squares DF square F-ratio b i l i t y Test Term SUBJ 26.001 15. 1.7334 22.009 0.00000 RESIDUAL SEM 0.73136E-02 1. 0.73136E-02 0.92862-01 0.76068 RESIDUAL SYN 0.34676E-01 1. 0.34676E-01 0.44029 0.50725 RESIDUAL SEM*SYN 0.83713E-01 1. 0.83713E-01 1.0629 0.30298 RESIDUAL Residual 46.073 585. 0.78758E-01 Total 72.166 603. The tntznpfio,t<xti.on of these s t a t i s t i c s i s as follows: 1. HQ^ i s rejected; that i s , a s i g n i f i c a n t difference was found between the 16 l e v e l s of the Subject factor. Thus, some of the v a r i a b i l i t y i n the data i s accounted for by i n t e r -subject differences. (This finding was evident i n each analysis.) 2. f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values for Constrained and Unconstrained sentences. (This finding was evident i n 74 analysis. 3. f a i l s to be rejected: that i s , no s i g n i f i c a n t difference was found between the mean values for the two lev e l s of the Syntactic factor, Be.fiofie. and kfitz.fi. Analysis IB : Syntactic factor s i g n i f i e s Order-of-mention A 2 x 2 ANOVA was performed, with the Syntactic factor r e f e r r i n g to Order-of-mention. The Syntactic factor had two le v e l s corresponding to 0ME=00E and OME^OOE; the Semantic factor had two l e v e l s , corresponding to Constrained and Un-constrained; the 16 subjects served as blocks. The following null hypothe.i>zi> were tested: HQ^ : There i s no s i g n i f i c a n t difference between the mean values for the 16 le v e l s of the Subject factor. HQ2: There i s no s i g n i f i c a n t difference between the mean values for the two le v e l s (Constrained and Unconstrained) of the Semantic factor. HQ^ : There i s no s i g n i f i c a n t difference between the mean values of the two le v e l s (0ME=00E and OME^OOE) of the Syntactic factor. Re.iu.ltA are indicated i n Table 7. 75 Table VII Results of /Analysis IB : Analysis of Variance Table Sum of squares Source DF square F-ratio b i l i t y Test term SEM*SYN Residual Total SUBJ SEM SYN 25.863 0.69783E-02 0.16881 0.15530E-01 46.007 72.166 15. 1.7242 21.924 0.00000 RESIDUAL 1. 0.69783E-02 0.88732E-01 0.76590 RESIDUAL 1. 0.16881 2.1465 0.14343 RESIDUAL 1. 0.15530E-01 0.19747 0.65694 RESIDUAL 585. 0.78645E-01 603. Interpretation of these s t a t i s t i c s i s as follows: 1. HQ^ i s rejected; that i s , a s i g n i f i c a n t difference was found between the 16 le v e l s of the Subject factor. Thus, some of the v a r i a b i l i t y of the data i s accounted for by i n t e r -subject differences. 2. f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values for Constrained and Unconstrained sentences. 3. HQ^ f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values for the two l e v e l s , (OOE=OME and OOE^OME) of the Syntactic factor. Analysis IC : Syntactic factor s i g n i f i e s Clause placement A 2 x 2 ANOVA was performed. The Syntactic factor had two l e v e l s corresponding to Clause placement (Main-clause-f i r s t and Subordinate-clause-first); the Semantic factor had two l e v e l s corresponding to Constrained and Unconstrained; the 16 subjects served as blocks. The following hypo the.* were tested: HQ^: There i s no s i g n i f i c a n t difference between the mean values of the 16 l e v e l s of the Subject factor. HQ2^ There i s no s i g n i f i c a n t difference between the mean values of the two l e v e l s (Constrained and Unconstrained of the Semantic factor. HQ^: There i s no s i g n i f i c a n t difference between the mean values for the two l e v e l s (Main-clause-first and Sub-ordinate-clause-first) of the Syntactic factor. RzAultA are indicated i n Table VIII: Table VIII Results of Analysis IC : Analysis of Variance Table Sum of Mean Proba-Source squares DF squares F-ratio b i l i t y Test term SUBJ 26.078 15. 1.7385 22.698 0.00000 RESIDUAL SEM 0.10845E-01 1. 0.10845E-01 0.14159 0.70684 RESIDUAL SYN 1.2668 1. 1.2668 16.540 0.00005 RESIDUAL SEM*SYN 0.11746 1. 0.11746 1.5335 0.21609 RESIDUAL Residual 44.807 585. 0.76594E-01 Total 72.166 603. Intzfipfiztatton of these s t a t i s t i c s i s as follows: 1. H Q 1 i s rejected; that i s , a s i g n i f i c a n t difference was found between the 16 le v e l s of the Subjects factor. Thus, some of the v a r i a b i l i t y i n the data i s accounted for by inter-subject differences. 77 2. HQ2 f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values for Constrained and Unconstrained sentences. 3. HQ^ i s rejected. A s i g n i f i c a n t difference was found between the mean values of the two le v e l s (Main-clause-first and Subordinate-clause f i r s t ) of the Syntactic factor. As i s shown i n Table IX, the mean response value for Main-clause-f i r s t sentences i s smaller than the mean value for Subordinate-c l a u s e - f i r s t sentences. Table IX Frequencies, Means and Standard Deviations for Analysis IC : Syntactic Factor Main clause Subordinate clause f i r s t f i r s t 302 302 0 MEAN 1.2780 1.3678 P MEAN 1.2771 1.3688 0 STDV 0.30147 0.38056 S ERR M 0.15930E-01 0.15930E-01 Task 2 Due to task differences between Tasks 1 and 2, i t was necessary to add an extra factor to analysis of data from Task 2. This factor, c a l l e d Truth Value, r e f l e c t e d the fac t that, for t h i s task, the subject was required to decide i f a pa r t i c u l a r s l i d e appropriately depicted the events described 78 by the sentence. Thus, a po s i t i v e or negative response {yz.1 or no) was demanded of the subject. Since reaction time may have been influenced by whether the correct response was yz& or no, i t was necessary to add t h i s factor to the analysis. Truth Value had two l e v e l s i n each of the following analyses, corresponding to yii> or no as the appropriate response to each s p e c i f i c s t i m u l i . In a l l other respects, the analyses for Task 2 are i d e n t i c a l to the analyses for Task 1. Analysis 2A : Syntactic factor s i g n i f i e s Conjunction  choice A 2 x 2 ANOVA was performed. They Syntactic factor had two l e v e l s corresponding to Conjunction choice (Be^o-te or kfit&n); the Semantic factor had two le v e l s corresponding to Constrained and Unconstrained; Truth Value had two le v e l s corresponding to Vzi or Wo; and the 16 subjects served as blocks. The following null hypothz&ZA were tested: HQ-^: There i s no s i g n i f i c a n t difference between the mean values for the 16 l e v e l s of the Subject factor. H Q 2 : There i s no s i g n i f i c a n t difference between the mean values of the two l e v e l s (Constrained and Unconstrained) of the Semantic factor. HQ.J : There i s no s i g n i f i c a n t difference between the mean values for the two le v e l s CBifioni and AfitZA.) of the Syntactic factor. Hgg: There i s no s i g n i f i c a n t difference between the mean values for the two l e v e l s (ye^orNo) of the Truth Value factor. 79 Re.6u.ltA a r e i n d i c a t e d i n T a b l e X: T a b l e X R e s u l t s o f A n a l y s i s 2A : A n a l y s i s o f V a r i a n c e T a b l e Sum of Mean Pro- Test Source squares DF square F - r a t i o b a b i l i t y term SUBJ 31.638 15. 2.1092 29.320 0.00000 RESIDUAL SEM 0.93467E-01 1. 0.93467E--01 1.2993 0.25481 RESIDUAL SYN 0.25473 1. 0.25473 3.5410 0.06036 RESIDUAL TRUTH 2.7551 1. 2.7551 38.299 0.00000 RESIDUAL SEM*SYN 0.48704E-01 1. 0.48704E--01 0.67703 0.41094 RESIDUAL Sm*TRUTH 0.97837E-02 1. 0.97837E--02 0.13600 0.71242 RESIDUAL SYN*TRUTH 0.10709 1. 0.1070 9 1.4886 0.22291 RESIDUAL Sm*SY^*TRUTH 0.64210E-01 1. 0.64210E--01 0.89257 0.34517 RESIDUAL Residual 42.515 591. 0.71938E--01 T o t a l 77.414 613. lnt£A.pn.e.ta£ton o f t h e s e s t a t i s t i c s i s a s f o l l o w s : 1. HQ^ i s r e j e c t e d ; t h a t i s , a s i g n i f i c a n t d i f f e r e n c e was f o u n d between t h e 16 l e v e l s o f the. S u b j e c t f a c t o r . T h u s , some o f t h e v a r i a b i l i t y i n t h e d a t a i s a c c o u n t e d f o r b y i n t e r - s u b j e c t d i f f e r e n c e s . 2. H Q 2 f a i l s t o be r e j e c t e d ; t h a t i s , no s i g n i f i c a n t d i f f e r e n c e was f o u n d between t h e mean v a l u e s f o r C o n s t r a i n e d and U n c o n s t r a i n e d s e n t e n c e s . 3. HQ.J f a i l s t o be r e j e c t e d ; t h a t i s , no s i g n i f i c a n t d i f f e r e n c e was f o u n d between t h e mean v a l u e s o f t h e two l e v e l s o f t h e S y n t a c t i c f a c t o r , Be^oAe and kfate,n.. 80 Hgg i s rejected; that i s , a s i g n i f i c a n t difference i s evident between the mean values of the two le v e l s (Ye.* and Wo) of the Truth value factor. As i s shown i n Table XI, the mean values for responses i s smaller than the mean values for Wo responses. Table XI Frequencies, Means and Standard Deviations for Analysis 2A : Truth Value Factor Correct response value "Yes" "No" 0 MEAN P MEAN 0 STDV S ERR M 309 1.2889 1.2882 0.37619 0.15304E-01 305 1.4222 1.4230 0.31985 0.15405E-01 Analysis 2B : Syntactic factor s i g n i f i e s Order-of-mention A 2 x 2 ANOVA was performed. The Syntactic factor had two le v e l s , corresponding to Order-of-mention (OME=OOE and OME^OOE); the Semantic factor had two le v e l s corresponding to Constrained and Unconstrained, the Truth Value factor had two l e v e l s corresponding to YtA and Wo; the 16 subjects served as blocks. The following Yia.lt hypothec a were tested: HQ^: There i s no s i g n i f i c a n t difference between the mean values for the 16 le v e l s of the Subject factor. 81 H „: T h e r e i s no s i g n i f i c a n t d i f f e r e n c e b e t w e e n t h e mean v a l u e s f o r t h e two l e v e l s ( C o n s t r a i n e d a n d U n c o n s t r a i n e d ) o f t h e S e m a n t i c f a c t o r . HQ^: T h e r e i s no s i g n i f i c a n t d i f f e r e n c e b etween t h e mean v a l u e s o f t h e two l e v e l s (OME=OOE and OME^OOE) o f t h e S y n t a c t i c f a c t o r . Enc: T h e r e i s no s i g n i f i c a n t d i f f e r e n c e b etween t h e mean 06 v a l u e s o f t h e two l e v e l s {Vzi and Mo) o f t h e T r u t h V a l u e f a c t o r . RzAultA a r e i n d i c a t e d i n T a b l e X I I . Resu T a b l e X I I I t s o f A n a l y s i s 2B : A n a l y s i s o f V a r i a n c e T a b l e F - r a t i o Pro- Test b a b i l i t y term SUBJ SEM SYN TRUTH SEM*SYN SEM*TRUTH SYN*TRUTH SM*SYU*TRUTH Residual T o t a l 31.645 0.89640E-0.19132E-2.6835 0.42408E-0.10574E-0.57570E-0.70152E-42.855 77.414 1 5 . 2 .1097 01 1 . 0 . 9 8 6 4 0 E - 0 1 •02 1 . 0 . 1 9 1 3 2 E - 0 2 1 . 2 .6835 •01 1 . 0 . 4 2 4 0 8 E - 0 1 -01 1 . 0 . 1 0 5 7 4 E - 0 1 •02 1 . 0 . 5 7 5 7 0 E - 0 2 -01 1 . 0 . 7 0 1 5 2 E - 0 1 5 9 1 . 0 . 7 2 5 1 3 E - 0 1 613. 29.094 1.2362 0.26384E-01 37.008 0.58483 0.14582 0.79392E-01 0.96744 0.00000 0.26666 0.87102 0.00000 0.44473 0.70270 0.77822 0.32572 RESIDUAL RESIDUAL RESIDUAL RESIDUAL RESIDUAL RESIDUAL RESIDUAL RESIDUAL I n t z t i p f l o t a t i o n , i s a s f o l l o w s : 1.. H Q 1 i s r e j e c t e d ; t h a t i s , a s i g n i f i c a n t d i f f e r e n c e was f o u n d between t h e 16 l e v e l s o f t h e S u b j e c t f a c t o r . T hus, some o f t h e v a r i a b i l i t y i n t h e d a t a i s a c c o u n t e d f o r by i n t e r - s u b j e c t d i f f e r e n c e s . 2. f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values of Constrained and Unconstrained sentences. 3. HQ^ f a i l s to be rejected; that i s , no s i g n i f i c a n t difference was found between the mean values for the two le v e l s , (OME=OOE and OME^OOE) of the Syntactic factor. 4. HQ£ i s rejected; that i s , a s i g n i f i c a n t difference was found between the mean values of the two level s (yo.A and No) of the Truth Values factor. As i s shown i n Table XIII, the mean values for Vzi responses was smaller than the mean values for No responses. Table XIII Frequencies, Means and Standard Deviations for Analysis 2B : Truth Value Factor Correct response value Yes No 0 MEAN P MEAN 0 STDV S ERR M 309 1.2889 1.2890 0.37619 0.15372E-01 305 1.4222 1.4222 0.31985 0.15473E-01 Analysis 2C : Syntactic factor s i g n i f i e s Clause-place-ment A 2 x 2 ANOVA was performed. The Syntactic factor had two l e v e l s corresponding to Clause-placement (Main-clause-f i r s t and Subordinate-clause-first); the Semantic factor had two l e v e l s , corresponding to Constrained and Unconstrained; the Truth Values factor had two le v e l s corresponding to YZA 83 and No; the 16 subjects served as blocks. The following nail hypothz&t& were tested: HQ^: There i s no s i g n i f i c a n t difference between the mean values of the 16 le v e l s of the Subject factor. H Q 2 : There i s no s i g n i f i c a n t difference between the mean values fo the two le v e l s (Constrained and Unconstrained) of the Semantic factor. H Q 5 : There i s no s i g n i f i c a n t difference between the mean values for the two le v e l s (Main-clause-first and Sub-ordinate-clause-first) of the Syntactic factor. HQ^: There i s no s i g n i f i c a n t difference between the mean values of the two le v e l s {Yzi and No) of the Truth Value factor. Rz6ult-& are indicated i n Table XIV. Table XIV Results of Analysis 2C : Analysis of Variance Table Source Sum of squares DF Mean square F-ratio Proba-b i l i t y Test tern SUBJ 31.784 15. 2.1189 30.897 0.00000 RESIDUAL SEM 0.87420E--01 1. 0.87420E--01 1.2747 0.25935 RESIDUAL SYN 1.9637 1. 1.9637 28.633 0.00000 RESIDUAL TRUTH 2.8452 1. 2.8452 41.486 0.00000 RESIDUAL SEM*SYN 0.11628 1. 0.11628 1.6955 0.19338 RESIDUAL SEM*TRUTH 0.11871E--01 1. 0.11871E--01 0.17309 0.67753 RESIDUAL SYN*TRUTH 0.15002 1. 0.15002 2.1874 0.13968 RESIDUAL SEM*SYN*TRUTH 0.23882 1. 0.23882 3.4823 0.06252 RESIDUAL Residual 40.532 591. 0.68581E-01 Total 77.414 613. Intzn.pKiLta.tton of the s t a t i s t i c s i s as follows: 1. HQ^ i s rejected; that i s , a s i g n i f i c a n t difference was found between the 16 le v e l s of the Subject factor. Thus, some of the v a r i a b i l i t y i n the data i s accounted for by intersubject differences. 2. H Q 2 f a i l s to be rejected; that i s no s i g n i f i c a n t difference was found between the mean values for Constrained and Unconstrained sentences. 3. HQJ. i s rejected. A s i g n i f i c a n t difference was found between the mean values of the two level s (Main-clause-first and Subordinate-clause-first) of the Syntactic factor. As i s shown i n Table XV, the mean value for Main-clause-first sentences i s smaller than the mean response value for Sub-or d i n a t e - c l a u s e - f i r s t sentences. Table XV Frequencies, Means and Standrad Deviations for Analysis 2C : Syntactic Factor Main-clause-first Subordinate-clause^-f i r s t 0 MEAN P MEAN 0 STDV S ERR M 309 1.3042 1.2989 0.33366 0.14912E-01 305 1.4066 1.4122 0.36953 0.15009E-01 4. HQ£ i s rejected; that i s , a s i g n i f i c a n t difference was found between the mean values of the two lev e l s (Ye.6 and No) of the Truth Value factor. As i s shown i n Table XVI the mean values for 9ZA responses was smaller than the mean value for 85 No responses. Table XVI Frequencies, Means and Standard Deviations for Analysis 2C : Truth Value Factor Yes No 309 305 O MEAN 1.2889 1.4222 P MEAN 1.2871 1.4241 0 STDV 0.37619 0.31985 S ERR M 0.14941E-01 0.15039E-01 Summary of Results 1. In every analysis, a s i g n i f i c a n t difference was found between the mean response values for the 16 subjects. Thus, some of the variance in the data i s accounted for by inter-subject differences. 2. At no point i n the analysis was a s i g n i f i c a n t difference found between the mean values of the two le v e l s of the Semantic factor (Constrained and Unconstrained). 3. At no point i n the analysis was a s i g n i f i c a n t difference found between the mean response values of the two lev e l s of the Syntactic factor, when t h i s factor indicated Conjunction choice, (Be.faon.e. or Afate.fi) . 86 4. At no point i n the analysis was a s i g n i f i c a n t difference found between the two level s of the Syntactic factor, when t h i s factor indicated Order-of-mention (0ME=00E and OME^OOE). 5. In both Task 1 and Task 2, a s i g n i f i c a n t difference was found between the two le v e l s of the Syntactic factor when th i s factor indicated Clause-placement. For both Tasks, t h i s difference was i n the d i r e c t i o n of smaller mean response values for Main-clause-first sentences than for Subordinate-c l a u s e - f i r s t sentences. 6. Whenever Truth Value was a factor ( i . e . , i n a l l Task 2 analyses), a s i g n i f i c a n t difference was found between the two le v e l s of t h i s factor. In a l l cases, t h i s difference was i n the d i r e c t i o n of Vzi responses having smaller mean values than No responses. Q 7. At no point i n the analysis was any s i g n i f i c a n t i n t e r a c t i o n e f f e c t noted. 87 CHAPTER FOUR DISCUSSION The r e s u l t s o f t h i s i n v e s t i g a t i o n i n t o t h e e f f e c t s o f v a r i o u s f a c t o r s on t h e a u d i t o r y comprehension o f complex se n t e n c e s c o n j o i n e d w i t h bdfaonz. and <x{t<LH a r e documented i n th e p r e v i o u s c h a p t e r . I n t h e p r e s e n t c h a p t e r , t h e s e r e s u l t s a r e d i s c u s s e d w i t h r e f e r e n c e t o e x p e r i m e n t a l d e s i g n , p r e v i o u s r e s e a r c h and t h e o r i e s o f s e n t e n c e s comprehension. D i s c u s s i o n o f R e s u l t s I n t e r - s u b j e c t D i f f e r e n c e s As n o t e d i n Cha p t e r Three, s t a t i s t i c a l l y s i g n i f i c a n t d i f f e r e n c e s were found f o r t h e mean v a l u e s o f re s p o n s e s f o r d i f f e r e n t s u b j e c t s . F o r t h e purpose o f t h i s i n v e s t i g a t i o n , i n d i v i d u a l r e s ponse p a t t e r n s were n o t a n a l y z e d . S u b j e c t s d i d , however, appear t o adopt i d i o s y n c r a t i c r e s ponse s t r a t e g i e s . For example, some s u b j e c t s tended t o r a c e t h r o u g h t h e t a s k , as i f speed o f response r e p r e s e n t e d a c h a l l e n g e . Other s u b j e c t s appeared more co n c e r n e d w i t h making an a c c u r a t e r e s p o n s e , 88 r e s u l t i n g in a slowing of response time. I t i s of i n t e r e s t to note, however, that no noticeable difference i n response accuracy distinguished these two response strategies. It i s suggested that, had instructions e x p l i c i t l y stated that speed of response was the parameter of i n t e r e s t , then inter-subject response differences would have been reduced. That i s , the way i n which the instructions were worded allowed some freedom of i n t e r p r e t a t i o n with respect to speed and/or accuracy of response. I t i s speculated that those subjects who responded more cautiously may not have done so had the instructions stressed speed. No s t a t i s t i c a l l y s i g n i f i c a n t difference i n mean response values was found for subjects' sex. However, there was evidence of a nonsignificant trend towards smaller res-ponse values for female subjects. Although nonsignificant, t h i s trend i s of i n t e r e s t since i t i s i n accord with a r e s u l t obtained by Slobin (1966). In a study investigating picture-v e r i f i c a t i o n for r e v e r s i b l e and non-reversible active and passive sentences, Slobin found that, at a l l ages tested, female subjects gave shorter reaction times than male sub-jects . True/False Differences For Task 2, i n which subjects were required to determine whether or not a single s l i d e matched the given sentence, mean response values were smaller for s t i m u l i requiring a "yes" response. This indicates that subjects found i t easier to v e r i f y that a s l i d e and sentence matched, than that they did 89 not match. Similar finding have been previously documented. Slobin (1966), found that reaction times were shorter for true active and true passive sentences, than for f a l s e active and f a l s e passive sentences ( i . e . , sentences which either matched or did not match an illuminated picture, r e s p e c t i v e l y ) . Slobin suggested that "there may be a tendency to c a l l a f f i r -mative sentences true" (p. 224) . Results obtained i n the present study appear to lend support to t h i s notion. Sentences used i n the present study ( a l l affirmative) appeared easier to process when they were true with respect to the given context, i n t h i s case an accompanying s l i d e . Constrained/Unconstrained Differences An important finding i s that no differences were found between responses to sentences which were either constrained or unconstrained by general knowledge. In l i g h t of the re-sults of French and Brown (1977) and Kavanaugh (1979), obtained with children, t h i s point merits some discussion. As pre-viously mentioned, re s u l t s of these studies demonstrated childrens' performance was superior for sentences i n which two events were meaningfully ordered. Two explanations for t h i s contradiction between the present and previous re s u l t s are offered. F i r s t , as detailed i n Chapter Two, unconstrained sen-tences i n the present study were constructed i n such a way as to avoid implausible sentences a r i s i n g as a r e s u l t of cross-matching. This caution was considered necessary since i t was found that, i f cross-matching were performed without t h i s constraint, highly questionable sentences resulted. French and Brown, however, constructed t h e i r unconstrained sentences by randomly cross-matching clauses from t h e i r constrained sentences. This procedure must have produced some uncon-strained sentences which were of questionable p l a u s i b i l i t y . I f t h i s was indeed the case, then French and Brown's compre-hension of plausible and implausible sentences, rather than differences i n comprehension between constrained and uncon-strained sentences, as they assumed. I t i s also possible that the subject populations used may account for t h i s difference i n r e s u l t s . Perhaps the f a c i l i t a t i o n e f f e c t of l o g i c a l context on the comprehension of these sentences i s remarkable only for children. It i s speculated that, while children are acquiring the meanings of the terms bz.faoh.il and afatzn, supportive context i s helpful i n inte r p r e t i n g sentences containing these words. However, as the meanings of the words become more f u l l y understood, less support i s sought from the semantic constraints of the sentence. Support for t h i s notion comes from Slobin (1966) who noted that the influence of n o n r e v e r s i b i l i t y on the com-prehension of passive sentences decreases with age. One f i n a l point regarding t h i s r e s u l t requires comment. S u p e r f i c i a l l y , the lack of a constrained/unconstrained difference may be considered to support the "constancy hypo-thes i s " , as outlined by Forster and Olbrei (1973). This hypo-thesis (discussed i n Chapter One) contends that semantic 91 differences across sentences do not a f f e c t aspects of sentence comprehension which are att r i b u t a b l e to syntactic processing. Although the r e s u l t s of the present study do not c o n f l i c t with t h i s hypothesis, neither do they lend i t any d i r e c t support. No attempt was made, i n the present study, to i s o l a t e the syntactic processing components of sentence processing. Rather, the procedures used were designed i n the hope of attaining some more h o l i s t i c i n d i c a t i o n of the sub-jects' understanding of the sentences. Comprehension, as such, i s not a consideration i n the "constancy hypothesis". Order-of-mention Differences No s i g n i f i c a n t e f f e c t of order of mention was found i n the present study; sentences i n which the order-of-mention and order-of-occurrence of events correspond were responded to no d i f f e r e n t l y than sentences without such a correspondence. This r e s u l t i s contrary to some previously reported r e s u l t s , e s p e c i a l l y those of Clark (1971), of children who seemed to r e l y strongly on an order-of-mention strategy i n interpreting such sentences. I t i s suggested that one reason no such strategy was evident i n the present study may be that an order-of-metnion strategy i s one primarily used by children acquiring the meaning of bzfiofie. and afito.fi, and which becomes less dominant with age. Support for t h i s notion comes from Clark's own study, i n which i t was noted that younger c h i l d -ren r e l i e d more heavily on t h i s strategy than older c h i l d r e n . That i s , apparent use of an order-of-mention strategy de-creased with increasing age and understanding of the terms bifiofie. and afitzn. One could speculate that i f t h i s trend i n -crease with age u n t i l adulthood, then the ro l e of t h i s strategy f o r adults, i f present at a l l , would be too reduced to be evident i n many tasks. The preceding explanation does not, however, account for the r e s u l t s obtained by Clark and Clark (196 8) i n t h e i r study of adults' memory for sentences of t h i s kind. They noted that verbatim r e c a l l i s better for sentences i n which the order-of-mention preserves the order-of-occurrence of events. On the contrary, order-of-mention was not found to be a s i g n i -f i c a n t factor either i n the present study or i n the study by Smith and McMahon (1970). Smith and McMahon suggested that order-of-mention plays a d i f f e r e n t r o l e i n memory tasks than i n comprehension tasks. The re s u l t s reported here support t h e i r view, although i t i s suggested that t h i s notion re-quires further experimental elaboration. Coker (1978) suggested a more s p e c i f i c explanation for the difference i n r e s u l t s regarding order-of-mention, among the d i f f e r e n t studies using children as subjects. As ex-plained more f u l l y i n Chapter One, Coker postulated that when a subject i s required to attend to both clauses (as they c e r t a i n l y are i n the Clarks' verbatim r e c a l l task) an order-of-mention strategy predominates. When, however, the sub-jec t s ' attention i s not cued to both clauses, a strategy of p r e f e r e n t i a l l y attending to the main clause i s dominant. It i s suggested that the tasks used i n the present investigation did not require subjects to attend f u l l y to both clauses. (Indeed, the f a c t that many responses were made p r i o r to the completion of the sentence would appear to indicate that such an explanation i s l i k e l y . ) The r e s u l t i n g dominance of a main-clause strategy over an order-of-mention strategy could, therefore, be seen as support for Coker's postulation. Clause Placement Differences Sentences which begin with a main clause were found to r e s u l t i n smaller mean response values than sentences with an i n i t i a l subordinate clause. As shown i n Chapter Three, t h i s r e s u l t was highly s i g n i f i c a n t i n each analysis. I t would appear that, for the tasks used i n the present study, the main clause plays a primary r o l e i n sentence comprehension. This i s i n accord with previous evidence of p r e f e r e n t i a l attention to the main clause, as reported by Smith and McMahon (1970), Amidon and Carey (197 2), and Coker (197 8). Coker suggested that one (of two) main clause strategies used by children i n her study could be described as follows: Subjects act out only the events i n the main clause and ignore the events i n the subordinate clause. A similar strategy appears to have been followed by subjects i n the present study. This strategy can be para-phrased as follows: Jfi thz matn cZauiz ti hzaxd fitn.it avid. luppZtzi znough tnfioimatto n to compZztz thz talk, izipond tmmzdtatzZy; tfi, hovizvzn, thz iubon.dtn.atz cZauiz ti hzaKd fith.it, zvzn tfi tt luppZtzi znough tnfioh.matton to compZztz thz talk, watt unttZ hzafitng thz matn cZauiz bzfion.z Kzipondtng. Why should i t be that i n i t i a l main clauses are interpre-ted more r e a d i l y than i n i t i a l subordinate clauses? Townsend and Bever (1977) discussed t h i s question i n some d e t a i l . They claimed that main and subordinate clauses can be com-pared to v i s u a l figures and grounds, respectively. Main clauses, l i k e v i s u a l figures, contain information that i s new; t h i s information constitutes the assertion made by the speaker. Subordinate clauses, l i k e v i s u a l grounds, contain older i n f o r -mation; t h i s information i s presupposed by the speaker to be known to the l i s t e n e r , and provides a context for the asser-tion of the main clause. Townsend and Bever claimed that these c h a r a c t e r i s t i c s of main and subordinate clauses imply that the information contained i n the main clause i s con-considered to be more important. The r e s u l t s of the present study support t h i s view; i f the l i s t e n e r expects the more important information to be found i n the main clause, i t i s not surprising that s/he attends to t h i s clause preferen-t i a l l y . One f i n a l point regarding t h i s r e s u l t should be mentioned. I t might be suspected that the s p e c i f i c measurement procedure used i n t h i s i n v e s t i g a t i o n may have biased the analyses of clause-placement differences. As described i n Chapter Two, the point chosen for the onset of response time measurement was within the sentence, following the f i r s t clause and the conjunction. I t i s acknowledged that t h i s d e f i n i t i o n of "zero time", although the most agreeable of several alternatives, i s an imperfect compromise. The problem arises because t h i s p a r t i t i o n i n g procedure does not divide the sentences i n an equivalent manner. Although the same number of words precede the p a r t i t i o n i n both types of sentence, the conjunction i s heard at a r e l a t i v e l y d i f f e r e n t position with respect to "zero time" for sentences beginning with a main clause, as opposed to sentences beginning with a subordinate clause. Consider the sentences, She packed the &vcttca&e befaoh.e the wa&hed the dtihe* and Befaofie *he washed the dti>hei>, 6he packed the 6utt-cai,e. In the f i r s t case, the subject must wait u n t i l "zero time" (following the word befaofie) u n t i l s/he has a l l the necessary and s u f f i c i e n t information required to make the res-ponse. For the second sentence, however, i t i s possible that the subject may have a l l the necessary information p r i o r to t h i s point. That i s , upon hearing only Begone &he washed, the subject may have decided that the s l i d e depicting wathtng must be the one describing the f i n a l event. I f t h i s i s the case, the subject could possibly begin responding at a point p r i o r to the onset of response time measurement. It is clear, however, that i f such a response pattern occurred, i t s e f f e c t would be to re-duce the response time of subo r d i n a t e - c l a u s e - i n i t i a l sentences, r e l a t i v e to those beginning with a main clause. Any suspected bias due to the measurement procedure i s , therefore, i n the d i r e c t i o n opposite to that of the observed r e s u l t . If such a bias was inherent i n the measurement, i t would serve only to strengthen the obtained r e s u l t of smaller response values for m a i n - c l a u s e - i n i t i a l sentences. Before/After Differences No s i g n i f i c a n t difference was found between responses to sentences containing bzfaonz and afatzn. Although c h i l d -ren's superior performance for bzfaon.z sentences was a r e s u l t stressed by Clark (1971), both Smith and McMahon (1970) and Coker (1978) have pointed out that the reason underlying t h i s r e s u l t may not be that bzfaofiz i s the more basic of the two terms, as Clark suggested, but that bzfaotiz and afatzh. are confounded with other factors i n these sentences. In bzfaoKZ sentences, the f i r s t event i s also asserted i n the main clause; thus, better performance on bz&oAz sentences can be viewed as further evidence of p r e f e r e n t i a l attention to the main clause. Coker further suggested that the response pattern of superior performance on bzfaonz sentences i s the way i n which preference for the main clause i s manifest when an order-of-mention strategy i s dominant; otherwise, the primacy of the main clause w i l l be evident i n a more obvious manner, e.g., a subject e n t i r e l y ignoring the subordinate clause. The r e s u l t s of t h i s study are consistent with Coker's speculation. P r e f e r e n t i a l attention to the main clause has been evidenced as the dominant strategy, i . e . , by smaller 97 subject response values to sentences beginning with a main clause. The f a c t that i t i s not also evident i n smaller response values to bzfaoiz sentences would be predicted by Coker, and i s supported by the r e s u l t of Smith and McMahon's study; i n the presence of a dominant e f f e c t of response latencies being smaller for m a i n - c l a u s e - i n i t i a l sentences, no sim i l a r e f f e c t was noted for bzfaoA.z sentences. A Comment on the Relative Importance of  Semantic and Syntactic Factors I t i s tempting to view these r e s u l t s as evidence of the primacy of syntactic factors over semantic factors i n the auditory comprehension of complex sentences conjoined with bzfaoAz and afatzfi. Such a conclusion, however, i s not only pre-mature, but also, i n a l l p r o b a b i l i t y , i n v a l i d . Labelling a factor (and subsequently a strategy) as either semantic or syntactic i s by no means a t o t a l l y objective decision. Such a judgement i s necessarily coloured by one's t h e o r e t i c a l bent, and ultimately, depends on l i n g u i s t i c analyses which may or may not be v a l i d . I t i s the author's contention that such labels are useful only i f they serve to communicate d e t a i l e f f i c i e n t l y , and should be used cautiously since a lab e l may inadvertently serve to obscure d e t a i l rather than communicate i t . A case i n point i s the major finding of t h i s paper, i . e . , evidence for a strategy by which subjects attend p r e f e r e n t i a l l y 98 to the main clause i n a sentence. Although t h i s strategy has been described i n syntactic terms, i t would be highly mislead-ing and inappropriate to l a b e l i t as a "syntactic" strategy (as Coker has l a b e l l e d a s i m i l a r strategy). I t i s not at a l l cle a r that the reason subjects attend d i f f e r e n t l y to main clause i n i t i a l sentences i s due to these sentences being s y n t a c t i c a l l y simpler, as such a l a b e l implies. It i s at lea s t equally probable that semantic reasons, concerning the nature of main clause content, underly t h i s r e s u l t (Bever, 1970; Townsend and Bever, 1977). By l a b e l l i n g such a strategy as either semantic or syntactic, one i s not only assuming more than the evidence supports, but one also runs the dangerous r i s k of presupposing the nature of the process. As i s indicated above, the reason underlying the r e s u l t of smaller response values to m a i n - c l a u s e - i n i t i a l sentences i s probably not so simple that one could describe t h i s as either a syntactic or semantic strategy. I t i s suggested that t h i s r e s u l t r e f l e c t s an i n t e r a c t i o n of syntactic and semantic de-vices u t i l i z e d by the subject as an aid for sentence comprehen-sion. More s p e c i f i c a l l y , i t i s suggested that the surface structure phenomenon to which the subject cues has a syntactic base (namely, the absence or presence of a subordinating con-junction at the beginning of a sentence). The purpose of such focus on t h i s syntactic term, however, i s to aid organization of the semantic content of the sentence. In other words, the pre-sence of bzfiotiz or afitzA at the beginning of the sentence cues the subject to attend less to the i n i t i a l clause and to await the f i n a l clause, i n order that s/he w i l l be free to process the more important content of the main clause. Conversely, the absence of such a syntactic cue at the onset of a sentence signals the subject that s/he i s free to process the f i r s t clause and, upon i t s completion, i s free to react to the task, as the more important information of the sentence has been processed at t h i s point. Possible Sources of Experimental Error Physical and Mechanical Sources of Error Several aspects of the physical and mechanical set-up and preparation allowed room for error due to the l i m i t a t i o n s of the equipment. In a l l such instances, i t i s anticipated that any e x i s t i n g inaccuracy would be consistent throughout the experiment, and not a source of random, uncontrolled v a r i a t i o n . One such problem arises i n r e l a t i o n to the s p l i c i n g of stimulus tapes. Although t h i s s p l i c i n g was performed with every caution, i t cannot be stated with certainty that every s p l i c e corresponded to the onset of each sentence to one hundredth of a second accuracy. Four aspects of t h i s pro-cedure may be i d e n t i f i e d as possible sources of error; locating the onset of speech for each sentence, the thickness of the razor blade used to make the cut i n the tape, and the accuracy of the cut i t s e l f . Despite meticulous care to mini-mize these sources of inaccuracy, they s t i l l must be acknowledged. 100 The measurement of each sentence to the point of p a r t i -t ioning, ( i . e . , locating "zero time" for response measure-ment) , also i s problematic. As discussed i n Chapter Two, v a r i a t i o n i n measurement was minimized by measuring each sentence f i v e times, discarding the highest and lowest read-ings, and then averaging the remaining three readings. I t i s anticipated that t h i s procedure rendered consistent any v a r i a -t i o n i n measurement between sentences to an accuracy close to one hundredth of a second. A further source of inaccuracy l i e s i n the co-ordination of the onset of the sentence with the illumination of the s l i d e ( s ) . Although both operations begin at the same instant, a small period of time i s required for the l i g h t of the s l i d e projector to reach maximum il l u m i n a t i o n . Thus, f u l l illumina-t i o n of each s l i d e would lag the onset of the sentence by a f r a c t i o n of a second. As t h i s period of time i s not only very small, but also i s consistent for each stimulus, t h i s i s not considered to be a major problem. Sources of Error i n Data C o l l e c t i o n and Analysis As a large amount of data was co l l e c t e d , and as each datum was subjected to a number of operations p r i o r to com-puter analysis, i t i s possible that error may have occurred i n t h i s area. Possible sources of error include; data re-cording, arithmetical manipulations of data (see Chapter Three) and entering the data into the computer. As each of these steps was, at the very l e a s t , t r i p l e checked, i t i s doubtful that such error did occur. 1 0 1 A Problem with the Experimental Design A more serious problem involves the nature of the ex-perimental design, i n p a r t i c u l a r with reference to Con-strained/Unconstrained differences. I f subjects had demon-strated superior performance for constrained sentences, then no conclusion could have been drawn concerning the role of t h i s factor (general knowledge constraints) i n sentence com-prehension. This i s because one could not r e l i a b l y eliminate the p o s s i b i l i t y that the subjects had bypassed the processing of the constrained sentences and had, instead, responded to context as represented by the s l i d e s . If t h i s had been the case, the predicted r e s u l t would have been superior perform-ance for constrained sentences. I t would be impossible, therefore, to determine the cause of any observed difference between responses to constrained and unconstrained sentences. Since, however, no s i g n i f i c a n t difference was noted between responses to constrained and unconstrained sentences, and since a subject must process at l e a s t part of the sentence i n order to perform the task for constrained sentences, i t i s reasonable to conclude that subjects also processed some part of the constrained sentences and were not merely responding to the pictured context. This problem (namely, the i m p o s s i b i l i t y of determining whether subjects would have been responding to the pictured context of the content of the sentences, i n the event of superior performance for constrained sentences having been evident) i s i n t e r e s t i n g for a number of reasons; F i r s t , i t 102 underscores the necessity for keeping experiments "clean", i n the sence of requiring a l l stimuli to be presented i n one modality i n order to avoid contamination factors. The present study, for example, would have been redesigned so as to eliminate the need for v i s u a l cues, had t h i s f a u l t i n the design been noted i n advance. Secondly, an i n t e r e s t i n g question i s raised regarding what one hopes to discover when attempting to examine the r o l e of e x t r a - l i n g u i s t i c information on sentence comprehen-sion. Ultimately, the goal must be to elaborate those factors operating i n the comprehension of natural language, and how such factors aid or i n h i b i t the understanding of language outside of the laboratory. Unfortunately, i n order to keep experiments methodologically clean, most contextual factors, including many which would be i n t e r e s t i n g to i n -vestigate, must be eliminated. Resolution of t h i s problem i s not immediately apparent; hopefully, further research and discussion w i l l lead to methods of investigation which can r e l i a b l y examine the influence of more than one modality on the active process of sentence comprehension. 103 CHAPTER FIVE CONCLUSIONS The purpose of t h i s investigation has been to determine the e f f e c t s of ce r t a i n factors on adult subjects* comprehen-sion of complex sentences conjoined with bifiofii and afitah.. The following factors were investigated: 1) general knowledge constraints 2) conjunction choice 3) order of mention 4) clauses placement The tasks were designed i n an e f f o r t to e l i c i t varying re-action times. These tasks involved matching and v e r i f y i n g the correspondence between s l i d e and sentence s t i m u l i . Major re-su l t s of t h i s investigation are as follows: 1. Subjects demonstrated superior performance for sentences requiring a "true" response. 2. Subjects demonstrated superior performance for sentences beginning with a main clause. 3. No s i g n i f i c a n t difference was noted between subjects' responses for bifioh.1 and afitin. sentences. 104 4. Subjects did not demonstrate use of an order-of-mention strategy i n i n t e r p r e t i n g these sentences. 5. E x t r a - l i n g u i s t i c information, i n the form of general knowledge constraints on the sentences, did not a f f e c t subjects' responses to these sentences. The r e s u l t s of the present study lend support to a grow-ing body of data which suggests that the main clause enjoys a priveleged p o s i t i o n i n the comprehension of complex sentences (Amidon and Carey, 1972; Bever, 1970; Coker, 1978; Smith and McMahon, 1970; Townsend and Bever, 1977). Preference for the main clause, i n the present study, i s manifest by smaller response values to sentences beginning with a main clause. Whether p r e f e r e n t i a l attention to main clauses stems from some added syntactic d i f f i c u l t y of processing subordinate clauses, (Amidon and Carey, 1972; Smith and McMahon, 1970), or from c h a r a c t e r i s t i c s of the semantic content of main clauses (Bever, 1970; Townsend and Bever, 1977) has not been determined i n the present study, but remains a t h e o r e t i c a l l y i n t e r e s t i n g point. I t i s suggested, however, that the re-su l t s of the present study are compatible with a view that a s o l e l y semantic or syntactic explanation for t h i s phenomenon i s too s i m p l i s t i c . Rather, subjects may focus on a syntactic cue i n the surface structure i n order to help them organize the semantic content of complex sentences. When the r e s u l t s of the present study are considered i n conjunction with those of previous studies, one i s struck by the v a r i a t i o n i n response patterns observed across studies, depending on the task requirements. A comment by Coker (1978) provides some framework i n which to note consistencies i n these divergent response patterns. Coker suggested that, i n the comprehension of these sentences, a preference for main clauses i s demonstrated by one of two response patterns; superior performance on be.fion.z sentences (in which the main clause expresses the f i r s t event) or attendance only to the main clause, i n the sense of ignoring the subordinate. She further outlined situations i n which each of these patterns could be expected. When the subject i s cued to attend to both clauses, main clause performance w i l l be manifest as superior performance for bzfioiz sentences. When the tasks requirements do not demand f u l l attention to both clauses (as in the present study) main clause dominance w i l l be manifest i n a more obvious way. Coker also suggested that, i n the former s i t u a t i o n , an order-of-mention strategy w i l l dominate over the main-clause strategy, whereas i n the l a t t e r s i t u a - : tion, a strategy r e f l e c t i n g main clause preference w i l l dominate. Although Coker s p e c i f i c a l l y addressed the v a r i a -tions i n r e s u l t s obtained for children, these comments seem largel y consistent with the re s u l t s found i n the previously reviewed research involving adults, and with the r e s u l t s of the present study. Whereas Coker's suggestions have accom-plished much i n terms of resolving and organizing the apparent inconsistencies found i n the l i t e r a t u r e , they merely organize the various response patterns while f a i l i n g to supply any adequate explanation as to why d i f f e r e n t tasks re-quirements trigger such divergent response patterns and strategies. An in t e r e s t i n g comment on the v a r i a t i o n i n response patterns between tasks has been made by Smith and McMahon (1970) . I t underscores the importance of t h i s phenomenon to the development of any theory of sentence comprehension: "...we are struck by the impression that we are viewing an object (the process of comprehension) through windows made of d i f f e r e n t types of d i s t o r t -ing glass; each window corresponds to a d i f f e r e n t procedure. The image i s , of course, blurred by the inherent v a r i a b i l i t y of our measurements. The question which i s central to the enterprise of understanding the comprehension process i s whether we are viewing the same process through windows which show f i r s t one component process and then another, or whether we are viewing e n t i r e l y d i f f e r e n t processes through each window." (p. 266) In conclusion, many problems remain which require furthe inves t i g a t i o n before our understanding of comprehension of these complex sentences can approach completion. F i r s t , how does one elaborate a theory of sentence comprehension which w i l l not only account f o r , but also explain, the observed variations i n response patterns and apparent strategies used i n sentence comprehension? Second, what i s the e f f e c t of varying instructions on a subject's performance of these and sim i l a r tasks? Third, how can we best examine comprehension a natural language set t i n g , without compromising experimental rigour? F i n a l l y , i t i s hoped that the methodology developed here can provide a framework within which to develop tools for examining sentence comprehension strategies used by language disordered subjects. 107 B I B L I O G R A P H Y TAmes, L.B. (1964) . "The Development of the Sense of Time i n the Young Ch i l d . " J. Genet. Psychol. 68: 97-125. Cited i n Feagans (1980) . Amidon, A., & Carey, P. (1972). "Why Five-Year-Olds Cannot Understand Before and After." J. Venb. Learn. \Jen.b. Behav. 11: 417:423. A n s e l l , B.J., & Flowers, CR. (1982). "Aphasic Adults' Under-standing of Complex Adverbial Sentences." Bnaln and Languale 75: 82-91. Barrie-Blackley, S. (1973). "Six-Year-Old Children's Under-standing of Sentences Adjoined with Time Adverbs." J. Psychollng. Hes. 2 : 153-165. Bever, T.G. (1970a). "The Comprehension and Memory of Sentences with Temporal Relations." In G.B. Flores d'Arcais & W.J.M. Levelt (Eds.), Advance* In Physlo-l l n g u l s t l c s . Amsterdam: North-Holland. Becer, T.G. (1970b). "The Cognitive Basis for L i n g u i s t i c Structures." In J.R. Hayes (Ed.), Cognition and the Ve.ve.lopme.nt ofa Language.. New York: Wiley. Brown, A.L. (1976). "The Construction of Temporal Succession by Pre-operational Children." In A.D. Pick (Ed.), Minnesota Symposia on Child Psychology, 10. Minneapolis. University of Minnesota. Clark, E.V. (1969). Language A c q u i s i t i o n : The Child's Spontaneous Vesch.lptlon ofa Events In Time. Unpublished doctoral d i s s e r t a t i o n , University of Edinburgh. Cited i n Clark (1971). Clark, E.V. (197 0). "How Young Children Describe Events i n Time." In G.B. Flores d'Arcais & W.J.M. Levelt (Eds.), Advances In ? s y c h o l i n g u i s t i c s . Amsterdam: North-Holland. Cited i n Clark (1971). Clark, E.V. (1971) . ''On the Acquisi t i o n of the Meaning of Befaoie and A fa ten.". J. Veib. Leaxn. Veib. Behav. TO: 266-275. 108 Clark, E.V. (1973). "Non-linguistic Strategies and the Acquisition of Word Meanings." CognA.ti.on 2: 161-82. Cited i n Kavanaugh (1979) . Clark, E.V. (1975) . "Knowledge, Context, and Strategy i n the Acquis i t i o n of Meaning." In D.P. Data (Ed.), Ge.OA.g2.town UntvZAitty Round Table, on Languages and LtngutittcA. Washington, D.C.: Georgetown University Press. Cited i n Kavanaugh (1979). Clark, E.V. (1977). "Strategies and the Mapping Problem i n F i r s t Language Acquisition." In J. Macnamara (Ed.), Language., Lzanntng and Thought. New York: Academic Press. Cited i n Kavanaugh (1979). Clark, E.V., & Garnica, O.K. (1974). "Is He Coming or Going? On the Acqui s i t i o n of D e i c t i c Verbs." J. \Je.Ab. Leann. Ve.nb. Behav. 73: 559-72. Cited i n Kavanaugh (1979). Clark, H.H., & Clark, E.V. (1968). "Semantic Distinctions and Memory for Complex Sentences." Q..J. Expt. Viychol. 20: 129-38. Coker, P.L. (1978). "Syntactic and Semantic Factors i n the Acq u i s i t i o n of BefioAe. and Afite.A". J. Ch. Lang. 5: 261-77. Coker, P.L., & Legum, S. (1975). "An Empirical Test of Semantic Hypotheses Relevant to the Language of Young Children." In Working Papers on the Kindergarten Pro-gram : Quality Assurance, SWRL for Educational Research and Development (Los Alamitos, C a l i f o r n i a ) . Cited i n Coker (1978). Coots, J.H. (1976). Chtldnen'A Knowledge, and Acqut&ttton ofi Volan Spattal AdjecttveA. Unpublished doctoral thesis, Harvard University. Cited i n Coker (1978). Derenzi, E., & F a g l i o n i , P. (1978). "Normatic Data and Screening Power of a Shortened Version of the Token Test." CoAtzx 14: 41-49. Cited i n Ansell & Flowers (1982) . Donaldson, M., & Wales, R.J. (1968). "On the Acqui s i t i o n of Some Relational Terms." In J.R. Hayes (Ed.), Cognttton and the. Ve.ve.lopme.nt ofi Language. New Rork: Wiley. Cited i n Clark (1971). E i l e r s , R.E., O i l e r , D.K., & E l l i n g t o n , J . (1974). "The Acquis i t i o n of Word-meaning for Dimensional Adjectives; the Long and Short of i t . " J. Ch. Lang. 1: 195-204. Cited i n Coker (1978). 109 F e a g a n s , Lynne ( 1 9 8 0 ) . " C h i l d r e n ' s U n d e r s t a n d i n g o f Some T e m p o r a l Terms D e n o t i n g O r d e r , D u r a t i o n and S i m u l t a n e i t y . " J . Piycholtng. Re*. 9: 41-57. F o r s t e r , K . I . , & O l b r e i , I . ( 1 9 7 3 ) . " S e m a n t i c H e u r i s t i c s and S y n t a c t i c A n a l y s i s . " Cognttton 2: 319-347. F o r s t e r , K . I . , & R y d e r , L.A. ( 1 9 7 1 ) . " P e r c e i v i n g t h e S t r u c t u r e and M e a n i n g o f S e n t e n c e s . " J . \Jeh.b. Lcan.n. Vcn.b. Bchav. 10: 285-96. C i t e d i n F o r s t e r & Ryder (1973). F r e n c h , L.A. , & Brown, A. ( 1 9 7 7 ) . " C o m p r e h e n s i o n o f Bcfion.c and Afitcn. i n L o g i c a l and A r b i t r a r y S e q u e n c e s . " J . Ch. Lang. 4: 247-56. G l u s c k b e r g , S., Hay, A., & Danks, J.H. ( 1 9 7 6 ) . "Words i n U t t e r a n c e C o n t e x t s ; Young C h i l d r e n do n o t C o n f u s e t h e M e a n i n g s o f Same and Vtfificn.cnt."Ch. Vcv. 47: 737-41. C i t e d i n C o k e r ( 1 9 7 8 ) . G o o d g l a s s , H., & K a p l a n , E. ( 1 9 7 2 ) . The AlACAAmcnt ofi Aphahta and Related VtAOh.den.6. P h i l a d e l p h i a : L e a and F e b i g e r . C i t e d i n A n s e l l and F l o w e r s ( 1 9 8 2 ) . J a k o b s o n , R. ( 1 9 6 8 ) . Chtld Language, Aphaita, and Phono-logtcal Untven.&ali>. The Hague : C i t e d i n A n s e l l & F l o w e r s ( 1 9 8 2 ) . J o h n s o n , H.L. (19 75) . "The M e a n i n g o f Be^o^e and A fit en. f o r P r e - s c h o o l C h i l d r e n . " J . Exp*, ch. P&ych. 19: 88-99. Kavanaugh, R o b e r t D. ( 1 9 7 9 ) . " O b s e r v a t i o n s on t h e R o l e o f L o g i c a l l y C o n s t r a i n e d S e n t e n c e s i n t h e C o m p r e h e n s i o n o f " B e f o r e " and " A f t e r " . " J . Ch. Lang. 6: 353-357. L e e c h , G.N. ( 1 9 7 0 ) . Towa/idl a Semantic Vc6cn.tpt4.onofiEngli.ih. B l o o m i n g t o n , I n d . : I n d i a n a U n i v e r s i t y P r e s s , C i t e d i n C l a r k ( 1 9 7 1 ) . O l s o n , D.R., & N i c k e r s o n , N. ( 1 9 7 7 ) . "The C o n t e x t s o f Com-p r e h e n s i o n : on C h i l d r e n ' s U n d e r s t a n d i n g o f t h e R e l a -t i o n s Between A c t i v e and P a s s i v e S e n t e n c e s . " J . Exp. Ch. Piych. , i n p r e s s a t t i m e o f c i t a t i o n . C i t e d i n F r e n c h & Brown (19 7 7 ) . P i a g e t , J . ( 1 9 6 6 ) . "Time P e r c e p t i o n i n C h i l d r e n . " I n F r a s e r , J . T . ( E d . ) , The Voted ofi Ttme. New Y o r k : Geo r g e B r a z i l l e r . pp. 202-216. C i t e d i n F e a g a n s ( 1 9 8 0 ) . R u b i n s t e i n , H., G a r f i e l d L . & M i l l i k a n , J.A. ( 1 9 7 0 ) . "Homographic E n t r i e s i n t h e I n t e r n a l L e x i c o n . J . Ven.b. Lean.n. \len.b. Behav. 9: 487-492. C i t e d i n F o r s t e r and O l b r e i ( 1 9 7 3 ) . 110 Sasanuma, S., Kamio, A. (1976). "Aphasics' Comprehension of Sentences Expressing Temporal Order of Events." Btialn and Language. 3: 495-506. Sinclair-de-Zwart, H. (1969). "Developmental Psycho-l i n g u i s t i c s . " In D. Elkind & J.H. F l a v e l l (Eds.). Studies in Cognitive Development. New York : O.U.P. Cited i n French & Brown (1977) . Slobin, D.I. (1966) . "Grammatical Transformations and Sentence Comprehension i n Childhood and Adulthood." J. Vetib. Leann. Veib. Behav. 5: 219-227. Smith, K.H., & McMahon, L.E. (1970). "Understanding Order Information in Sentences : Some Recent Work at B e l l Laboratories." In G.B. Flores d'Arcais & W.J.M. Levelt (Eds.), Advances In Psycholongulstlcs. Amsterdam: North-Holland. Stern, W. (1962) . Psychology ofa Early Childhood up to the Sixth Veafi ofa Age. New York : Henry Hold. Cited i n Geagans (1980). Townsend, D.J., & Bever, T.G. (1977). "Main and Subordinate Clauses : a study i n figure and ground." Indiana University L i n g u i s t i c s Club. van Fraassen, B.C. (1970). An Introduction to the Philosophy ofa Space and Time. New York : Random House. Cited i n Feagans (1980). 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0095546/manifest

Comment

Related Items