Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Perceptual confusions among permissible and impermissible english consonant clusters Newton, Colleen Nora 1972

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_1972_A6_7 N49.pdf [ 6.8MB ]
Metadata
JSON: 831-1.0101707.json
JSON-LD: 831-1.0101707-ld.json
RDF/XML (Pretty): 831-1.0101707-rdf.xml
RDF/JSON: 831-1.0101707-rdf.json
Turtle: 831-1.0101707-turtle.txt
N-Triples: 831-1.0101707-rdf-ntriples.txt
Original Record: 831-1.0101707-source.json
Full Text
831-1.0101707-fulltext.txt
Citation
831-1.0101707.ris

Full Text

PERCEPTUAL CONFUSIONS AMONG PERMISSIBLE AND IMPERMISSIBLE ENGLISH CONSONANT CLUSTERS by COLLEEN NORA NEWTON B . S c , U n i v e r s i t y of B r i t i s h Columbia, 1970 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n the Department of PAEDIATRICS D i v i s i o n of Audiology and Speech Sciences We accept t h i s t h e s i s as conforming t o the re q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA June, 1972 In present ing th i s thes is in pa r t i a l fu l f i lment o f the requirements for an advanced degree at the Un ivers i t y of B r i t i s h Columbia, I agree that the L ibrary sha l l make it f r ee l y ava i l ab le for reference and study. I fu r ther agree that permission for extensive copying of th is thes i s fo r scho la r l y purposes may be granted by the Head of my Department or by his representat ives . It is understood that copying or pub l i c a t i on of th i s thes is fo r f i nanc i a l gain sha l l not be allowed without my wr i t ten permiss ion. Department of A ^ e Q / A T/ZiC $ The Un ivers i ty o f B r i t i s h Columbia Vancouver 8, Canada Date ~Xlt / f 7 3, . ABSTRACT The present study i s an attempt t o gain i n s i g h t i n t o the perceptual mechanism f o r consonant c l u s t e r s and t o d i s c e r n i t s r e l a t i o n s h i p to e x i s t i n g t h e o r i e s of speech p e r c e p t i o n . Nonsense s y l l a b l e s were c o n s t r u c t e d , c o n s i s t i n g of the vowel / I / and 18 two consonant c l u s t e r s composed of one of the f r i c a t i v e s Is/, \t\, and /// and one of the p l o s i v e s /p/, / t / , and /k/ i n e i t h e r order. Some of these c l u s t e r s form p e r m i s s i b l e E n g l i s h c l u s t e r s , the others impermissible c l u s t e r s . These c l u s t e r s appeared i n three p o s i t i o n s ; i n i t i a l , medial (followed by / i ^ / , ) , and f i n a l . They were recorded by a t r a i n e d p h o n e t i c i a n . Pink noise at a +3 dB s i g n a l - t o - n o i s e r a t i o (determined from peak readings) was used t o mask the s i g n a l . Nine experimental tapes (three f o r each c l u s t e r p o s i t i o n ) were constructed; 108 items were presented on each tape. Eighteen subjects (nine males and nine females) each l i s t e n e d t o one tape f o r each p o s i t i o n and responded according t o a fo r c e d choice paradigm t o each item by w r i t i n g the c l u s t e r they p e r c e i v e d . Responses f o r a l l subjects and f o r each p o s i t i o n were t a b u l a t e d i n confusion matrices and analyzed according 1) t o p e r m i s s i b i l i t y as E n g l i s h c l u s t e r s , 2) t o manner of a r t i c u l a t i o n , 3) to place of a r t i c u l a t i o n and 4) to d i f f e r e n t i a t i o n according to the d i s t i n c t i v e features of [ a n t e r i o r ] , [ c o r o n a l ] , and [ d i s t r i b u t e d ] . In s p e c t i o n of these confusion matrices i n d i c a t e d t h a t there may be some d i f f e r e n c e i n the perception of p e r m i s s i b l e and impermissible c l u s t e r s ; however, r e s u l t s are not c o n c l u s i v e . The proposal t h a t c l u s t e r s are perceived as a u n i t i s r e f u t e d by the r e s u l t s obtained. Manner of a r t i c u l -a t i o n was i d e n t i f i e d c o r r e c t l y more f r e q u e n t l y than place of a r t i c u l a t i o n . Place of a r t i c u l a t i o n f o r f r i c a t i v e s was i d e n t i f i e d c o r r e c t l y more f r e q u e n t l y than place of a r t i c u l a t i o n f o r p l o s i v e s . Analyses according t o d i s t i n c t i v e - i i -- i i i -f e a tures provides some support f o r d i s t i n c t i v e f e a t u r e s . The r e l a t i o n s h i p of the r e s u l t s of t h i s study t o four major t h e o r i e s of speech perception i s d i s c u s s e d . Three of these can be used to e x p l a i n some aspect of the responses observed. TABLE OF CONTENTS Page Abstract i i Table of Contents i v L i s t of Tables • v i L i s t of Figures . • i x Acknowledgments x Chapter 1. Introduction 1 1.1 Introduction 1 1.2 Literature Review 3 1.21 Building Blocks of Speech 3 1.22 Sound Sequences 7 1.23 F r i c a t i v e s 12 1.24 Plosives 14 1.25 Clusters 17 1.26 Effect of Noise on Signal 20 1.27 Models of Speech Perception 22 Chapter 2. Statement of the Problem 37 Chapter 3. Method 39 3.1 Experimental Design 39 3.2 Subjects 41 3.3 P i l o t Study . 41 3.31 Determination of Noise Level . . . 41 3.32 Preparation of P i l o t Tapes . 42 3.33 P i l o t Experiment Procedure 43 3.34 Results of P i l o t Study 45 3.4 Preparation of Experimental Tapes 47 3.41 C a l i b r a t i o n Signals 47 3.42 Recording of Numbers 47 3.43 Determination of 5ignal-to-Noise Ratio . . . . 47 3.44 Equipment Set-up 48 3.45 Recording Stimuli on Experimental Tapes . . . . 48 3.46 Evaluating Completed Tape 48 3.5 Experimental Procedure 49 - i v -- v -Page Chapter 4. Results 51 4.1 Confusion Matr ices for Ind i v idua l C lus te r s 51 4.2 Responses to Each C lus te r 52 4.3 Permiss ib le versus Impermissible C lus te r s i n Eng l i sh 53 4.31 Types of Responses 54 4.4 Manner of A r t i c u l a t i o n 56 4.5 P lace of A r t i c u l a t i o n for P los i ves and F r i c a t i v e s . . 56 4.6 Analyses According to c e r t a i n Jakobson, Fant and Ha l l e D i s t i n c t i v e Features 56 Chapter 5. D i s cuss ion 86 5.1 L im i ta t ions of the Experiment 86 5.2 D i scuss ion of Results 87 • 5.21 To ta l Confusion Matr ices 88 5.22 Responses to Each C lus t e r 88 5.23 Permiss ib le versus Impermissible C lus te rs . . . 89 5.24 Manner of A r t i c u l a t i o n 91 5.25 Place of A r t i c u l a t i o n 92 5.26 Analyses Accord ing to c e r t a i n Jakobson, Fant and Ha l le D i s t i n c t i v e Features 93 5.3 Re la t ionsh ip to Models of Speech Percept ion 95 Chapter 6. Conclus ion and Impl icat ions for Further Research . . . 98 B ib l iography 100 Appendix A 104 Appendix B 105 Appendix C 106 Appendix D 115 Appendix E 116 LIST OF TABLES Table Page I. Total Errors for each Signal-to-Noise Level in Pilot Study 46 II. Confusion Matrix for Clusters in Initial Position 59 III. Confusion Matrix for Clusters in Medial Position. ..... IV. Confusion Matrix for Clusters in Final Position &1 V. Number of Correct Responses of Each Cluster in Each Position 62 VI. Significant Differences observed between Means from Table V 63 VII. Total Responses of Each Cluster in Each Position 64 VIII. Confusion Matrix: Responses to Initial Clusters Combined According to Permissibility for Initial Position 65 IX. Confusion Matrix: Responses to Medial Clusters Combined According to Permissibility for Medial Position 65 X. Confusion Matrix: Responses to Final Clusters Combined According to Permissibility for Final Position 65 XI. Frequency of Types of Errors Produced in Response to Initial Clusters Combined According to their Permissibility as Initial English Clusters . 66 XII. Frequency of Types of Errors Produced in Response to Medial Clusters Combined According to their Permissibility as Medial English Clusters 67 XIII. Frequency of Types of Errors Produced in Response to Final Clusters Combined According to their Permissibility as Final English Clusters 68 XIV. Confusion Matrix: Responses for Clusters in Initial Position Combined According to Manner of Articulation. . 69 XV. Confusion Matrix: Responses for Clusters in Medial Position Combined According to Manner of Articulation. . 69 - v i -v i i -Table Page XVI. Confusion Matrix: Responses for Clusters in Final Position Combined According to Manner of Articulation. . 69 XVII. Confusion Matrix: Responses for Clusters in Initial Position Combined According to Place for Plosives 70 XVIII. Confusion Matrix: Responses for Clusters in Medial Position Combined According to Place for Plosives 71 XIX. Confusion Matrix: Responses for Clusters in Final Position Combined According to Place for Plosives 72 XX. Confusion Matrix: Responses for Clusters in Initial Position Combined According to Place for Fricatives. ... 73 XXI. Confusion Matrix: Responses for Clusters in Medial Position Combined According^ to Place for Fricatives. ... 74 XXII. Confusion Matrix: Responses for Clusters in Final Position Combined According to Place for Fricatives. ... 75 XXIII. Confusion Matrix: Responses for Clusters in Initial Position Combined According to the Distinctive Feature Anterior for Fricatives 76 XXIV. Confusion Matrix: Responses for Clusters in Medial Position Combined According to the Distinctive Feature Anterior for Fricatives 76 XXV. Confusion Matrix: Responses for Clusters in Final Position Combined According to the Distinctive Feature Anterior for Fricatives 76 XXVI. Confusion Matrix: Responses for Clusters in Initial Position Combined According to the Distinctive Feature Anterior for Plosives 77 XXVII. Confusion Matrix: Responses for Clusters in Medial Position Combined According to the Distinctive Feature Anterior for Plosives. 77 XXVIII. Confusion Matrix: Responses for Clusters in Final Position Combined According to the Distinctive Feature Anterior for Plosives. 77 XXLX. Confusion Matrix: Responses for Clusters in Initial Position Combined According to the Distinctive Feature Anterior for both Plosive and Fricative 78 - v i i i -Table Page XXX. Confusion M a t r i x : Responses f o r C l u s t e r s i n M e d i a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature A n t e r i o r f o r both P l o s i v e and F r i c a t i v e 79 XXXI. Confusion M a t r i x : Responses f o r C l u s t e r s i n F i n a l P o s i t i o n Combined According to the D i s t i n c t i v e Feature A n t e r i o r f o r both P l o s i v e and F r i c a t i v e 80 XXXII. Confusion M a t r i x : Responses f o r C l u s t e r s i n I n i t i a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature Coronal f o r P l o s i v e s 81 X X X I I I . Confusion M a t r i x : Responses f o r C l u s t e r s i n Medial P o s i t i o n Combined According to the D i s t i n c t i v e Feature Coronal f o r P l o s i v e s 81 XXXIV. Confusion M a t r i x : Responses f o r C l u s t e r s i n F i n a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature Coronal f o r P l o s i v e s 81 XXXV. Confusion M a t r i x : Responses f o r C l u s t e r s i n I n i t i a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature Coronal f o r both P l o s i v e and F r i c a t i v e 82 XXXVI. Confusion M a t r i x : Responses f o r C l u s t e r s i n Medi a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature Coronal f o r both P l o s i v e and F r i c a t i v e 83 XXXVII. Confusion M a t r i x : Responses f o r C l u s t e r s i n F i n a l P o s i t i o n Combined According t o the D i s t i n c t i v e Feature Coronal f o r both P l o s i v e and F r i c a t i v e 84 XXXVIII. Confusion M a t r i x : Responses f o r C l u s t e r s i n I n i t i a l P o s i t i o n Combined According to the D i s t i n c t i v e Features: Coronal and D i s t r i b u t e d f o r F r i c a t i v e s 85 XXXIX. Confusion M a t r i x : Responses f o r C l u s t e r s i n M e d i a l P o s i t i o n Combined According t o the D i s t i n c t i v e Features: Coronal and D i s t r i b u t e d f o r F r i c a t i v e s 85 XL. Confusion M a t r i x : Responses f o r C l u s t e r s i n F i n a l P o s i t i o n Combined According t o the D i s t i n c t i v e Features: Coronal and D i s t r i b u t e d f o r F r i c a t i v e s 85 LIST OF FIGURES Figure Page 1. Mcxiel f o r the speech-generating and speech-perception p r o c e s s . 25 2. An o u t l i n e of production of the nine experimental tapes. 40 3. Block diagram of equipment used i n monitoring s t i m u l i t o o b t a i n the d e s i r e d S i g n a l - t o - N o i s e l e v e l s used i n the P i l o t Study 44 - i x -ACKNOWLEDGMENTS The completion of t h i s study brings to a c lose a number of r e l a t i onsh ips which have been e s sen t i a l to the formulat ion of t h i s thes is and a l so instrumental i n my personal development. To John G i l b e r t , Head of the D i v i s i o n of Audiology and Speech Sc i ences , I say thanks, fo r h is u n f a i l i n g concern fo r the wel fare of the Master 's program and i t s s tudents , and fo r h is s e n s i t i v i t y and understanding i n t r ea t ing me as a person as we l l as a s tudent ; and may I always be a " G i l b e r t g i r l . " To my c l i n i c a l superv i sor , Joyce Edwards, thanks fo r being there ; and may I measure up to the standard she has set for us . To Andre-Pierre Benguerel , thanks fo r amazing pat ience i n teaching me how to use the equipment and wr i t e c l e a r l y . Through his unwavering ins i s t ence on p r e c i s i o n , I have f i n a l l y learned the importance of accuracy i n a l l t h ings . I have a l so apprec iated the time and i n s t r u c t i o n g iven me by the other professors and i ns t ruc to r s i n the D i v i s i o n . I wish to thank Meredith Kimbal l for serv ing on my thes is committee and p a t i e n t l y reading my t h e s i s . Without the moral support and confidence of my f r i ends and fami ly , t h i s venture would have been more d i f f i c u l t . I e s p e c i a l l y appreciated those who e n t h u s i a s t i c a l l y p a r t i c i p a t e d as subjects i n my experiment. And f i n a l l y a s p e c i a l thanks to Margaret and Barbara for t h e i r f r i endsh ip through a l l the t r i a l s and triumps we have shared i n the l a s t two years . CHAPTER 1 INTRODUCTION 1.1 Introduction Every person develops motivations to share a fragment of his emotions, experiences and knowledge with another being. It i s believed that from such abstractions of thought, (through mechanisms for c l a s s i f y i n g and sequencing the elements of speech), come neurophysiological s p e c i f i c a t i o n s for precise movements of the anatomical structures involved i n phonation and a r t i c u l a t i o n : g l o t t i s , tongue, l i p s and velum. The a r t i c u l a t o r s act upon the a i r flow produced by the lungs, creating a sound wave characterized by the acoustic parameters of frequency, i n t e n s i t y and duration. The sound wave impinges on the auditory mechanism of a l i s t e n e r where the acoustic parameters are convert-ed to neurophysiological c o r r e l a t e s . At a perceptual l e v e l higher than the auditory system these neurophysiological c o r r e l a t e s are decoded and organized i n a l i n g u i s t i c a l l y meaningful manner. Thus, s t r u c t u r i n g of the abstract thought of a speaker is recreated by a l i s t e n e r and i t i s i n f e r r e d that communication has-occurred. Verbal communication i s the most v i t a l aspect of human s o c i a l i z a t i o n . Human communication has the p o t e n t i a l to create an i n f i n i t e number of messages from a f i n i t e number of elements (consonants and vowels) continuously rearrang-ed i n t o meaningful sequences. The placement of such elements i s l i n g u i s t i c a l l y predetermined, being based on the order of the two types of elements. In spoken En g l i s h , t y p i c a l sequences of consonants and vowels are: /VC/, /CV/, /VCV/, /CVC/, /CCV/ and /CCCV/. A l l arrangements for the thousands of e x i s t i n g words are based on combinations and permutations of these sequences or s i m i l a r ones. Each consonant i s produced i n a d i f f e r e n t manner when combinations in v o l v i n g two or three consecutive consonants are placed together. In E n g l i s h in a two consonant sequence, f o r example, a f r i c a t i v e can be followed by a stop, a nasal or a sonorant, e.g. /st/,/sn/,or / s i / . In a three consonant -1--2-sequence a f r i c a t i v e can be followed by a stop and then a sonorant, e.g. / s t r / or /spr/. Elements i n a sequence are perceptually d i f f e r e n t i a t e d . However, i n the movements of the a r t i c u l a t o r s , and in the acoustic wave produced for speech there are no i n d i c a t i o n s where one element (C or V) i s completed and another begins. Therefore, i t is necessary to extract the d i s c r e t e elements from the continuous pattern and i d e n t i f y each i n order to reconstruct an o r i g i n a l message. The process of segmenting speech s i g n a l s presents a basic problem for the development of a model of speech perception. I d e n t i f i c a t i o n of one element from a s i n g l e set of acoustic c h a r a c t e r i s t i c s cannot explain the e f f i c i e n c y and complexity of the segmenting process. A number of theor-ies have been proposed to account for the added complexity of speech percept-ion. The theories may be c l a s s i f i e d i n t o two types; 1) Ac t i v e and 2) Passive. 1) Ac t i v e theories of speech perception include those which propose that neural e x c i t a t i o n s created i n t e r n a l l y are compared with e x c i t a t i o n s created by an incoming acoustic s i g n a l , for example, the Motor Theory (Liberman, et a l . , 1966; 1967a; 1967b; and Studdert-Kennedy et a l . , 1970) and Analysis-by-Synthesis (Stevens, 1960; Halle and Stevens, 1962). 2) Passive theories propose mechanism u t i l i z i n g only the c h a r a c t e r i s t i c s of the incoming s i g n a l f o r a n a l y s i s , for example, D i s t i n c t i v e Features (Jakobson, Fant and Ha l l e , 1963 and Chomsky and H a l l e , 1968) and Context-S e n s i t i v e Coding (Wickelgren, 1966; 1969a; and 1969b). Models of speech perception are dependent on an understanding of r e l a t i o n s h i p s between acoustic parameters of speech elements and the psycho-l o g i c a l r e a l i t y of perceptual u n i t s . Such an understanding can be expanded by conducting experiments in the mechanical synthesis and recogn i t i o n of speech, and by the i d e n t i f i c a t i o n of speech sounds under various conditions of d i s t o r t i o n . U n t i l r e c e n t l y , a phoneme was assumed to be the basic unit of speech perception. A phoneme was characterized by a s i n g l e set of acoustic -3-s p e c i f i c a t i o n s • r e s u l t i n g from a r t i c u l a t o r y processes involved i n producing a s i n g l e element. The assumption that a phoneme i s the minimal unit of speech perception i s now being reconsidered. Speech perception may possibly occur through a v a r i e t y of perceptual processes u t i l i z i n g the phoneme for one l e v e l of perception, the s y l l a b l e f o r another l e v e l of perception and p o s s i b l y phrases at a higher l e v e l . No theories have yet been proposed to explain speech perception according to such mechanisms because of the high degree of complexity needed. In speech production, occurrence >of two or three s p e c i f i c consonants i n a c l u s t e r i s frequent. I t has therefore been proposed that these consonant c l u s t e r s may be perceived as one u n i t , rather than as a sequence of two or three d i f f e r e n t elements. Empirical evidence i n d i c a t i n g perception of the consonant c l u s t e r as a s i n g l e unit would demand re-examination of e x i s t i n g theories of speech perception. The perception of i n d i v i d u a l consonants has been the topic of a number of studies (Cooper, D e l a t t r e , Liberman, Borst and Gerstman; 1962; Hughes and H a l l e , 1956; H a l l e , Hughes and Radley, 1957; and H a r r i s , 1958), whereas the perception of consonants i n a c l u s t e r has only r e c e n t l y become a topic f or extensive i n v e s t i g a t i o n , ( U l d a l l , 1964; Bond, 1971). Thus, the hypothesis that consonant c l u s t e r s are perceived as one perceptual unit i s of considerable i n t e r e s t . Such a hypothesis was tested in the present i n v e s t i g a t i o n . 1.2 L i t e r a t u r e Review 1.21 Building Blocks of Speech As early as 1953, C y r i l Harris tested the hypothesis that speech i s •composed of blocks or units of consonants and vowels combined in various sequences. He presumed that, i f the hypothesis was v a l i d , i t would then have been possible to rearrange the units of a sequence to produce a second, -4-e q u a l l y i n t e l l i g i b l e sequence. H a r r i s c o n s t r u c t e d speech by s p l i c i n g s e c t i o n s of tape c o n t a i n i n g segments having a s u f f i c i e n t number of c h a r a c t e r -i s t i c s e s s e n t i a l t o the perception of the vowel or consonant. He found t h a t such speech sounded unnatural and, f o r the most p a r t , u n i n t e l l i g i b l e . Subsequently, he attempted t o determine how many b u i l d i n g blocks were r e q u i r e d to achieve adequate speech i n t e l l i g i b i l i t y . Sequences were produced c o n s i s t -ing of consonants preceding d i f f e r e n t vowels. The consonant i n the sequence was replaced by another s i m i l a r consonant which had preceded another d i f f e r e n t vowel, f o r example, the /k/ i n / k l k / was replaced by the it j i n /tok./. Upon p r e s e n t a t i o n of the newly constructed sequences t o l i s t e n e r s , H a r r i s observed that the m a j o r i t y of e r r o r s were type e r r o r s , i e . one p l o s i v e was r e p l a c e d by another p l o s i v e , r a t h e r than a p l o s i v e being replaced by a f r i c a t i v e , e.g. / t / was replaced by /k/ r a t h e r than / s / . C e r t a i n consonants were recognized i r r e s p e c t i v e of the f o l l o w i n g vowel, s p e c i f i c a l l y : It//, / s / , / z / , and / t / . In order t o recognize these sounds, H a r r i s assumed that only one b u i l d i n g block was s u f f i c i e n t . The r e s u l t s of H a r r i s ' s experiments were not s u r p r i s i n g i n the l i g h t of a c o u s t i c knowledge that p l o s i v e i d e n t i f i c a t i o n i s dependent on the second formant t r a n s i t i o n t o the f o l l o w i n g vowel (Cooper et a l . , 1952) w h i l e /s/ and /// i d e n t i f i c a t i o n i s not a l t e r e d by varying the f o l l o w i n g t r a n s -i t i o n , ( H a r r i s , 1958). H a r r i s (1953) concluded t h a t , except f o r /tff, / s / , /z/ and It/, speech i s composed of allophones r a t h e r than phonemes. In c o n t r a s t to H a r r i s ' s c o n c l u s i o n . Fry (1964) s t a t e d t h a t : The phonemic type of s t r u c t u r e appears to be one of the brute f a c t s about speech and language which cannot be conjured away by any amount of t h e o r i z i n g . ( F r y , 1964, p. 59) He i n d i c a t e d that segmentation i n t o phonemic u n i t s by the normal speaker and l i s t e n e r i s supported by the widespread e x i s t e n c e of a l p h a b e t i c w r i t i n g . According to F r y , t h i s phonemic u n i t must be i n v o l v e d i n the s t r u c t u r i n g of a c o u s t i c i n f o r m a t i o n i n the b r a i n . Thus, two u t t e r a n c e s , s u p p l y i n g the same - 5 -information to the brain of each l i s t e n e r , may be i n d i s t i n g u i s h a b l e to a speaker-listener of one language and indisputably d i f f e r e n t to a speaker-l i s t e n e r of another language. A number of studies conducted at Haskins Laboratories (Liberman, H a r r i s , Hoffman and G r i f f i t h , 1957; Liberman, Harris, Kinney and Lane, 1961; and Liberman, H a r r i s , Eimas, Lisker and Bastian, 1961) support the theory that perception occurs at a phonemic l e v e l . * These studies indicate that l i s t e n e r s segment a continuum of small acoustic changes into d e f i n i t e c l a s s e s . Specif-i c a l l y , native E n g l i s h speakers categorized consonant s t i m u l i as /b/, /d/ and /g/ when a continuum of minimal changes were effected i n synthesized second formant t r a n s i t i o n s before the vowel /a/. For an experimental set of t r a n s i t i o n s produced on The Pattern Playback* there was great consistency i n l a b e l l i n g some s t i m u l i as /b/, other s t i m u l i as /d/, and others as /g/. Between these three s e t s , the l i s t e n e r s would sometimes i d e n t i f y a given s t i m u l i as /b/, sometimes as /d/ or at other times /d/ and /g/. The area of uncertainty of i d e n t i f i c a t i o n was termed, "the phoneme boundary." If phonemic grouping operates i n the perception of speech, i t i s as a r e s u l t of language learning. Thus, a c h i l d must attend to relevant acoustic differences between s t i m u l i and ignore i n s i g n i f i c a n t v a r i a t i o n s . Following t h i s reasoning, i t might be predicted that a l i s t e n e r w i l l not be equally s e n s i t i v e to sound d i f f e r e n c e s over the whole range of p o s s i b l e v a r i a t i o n . Such an hypothesis was tested (Liberman, H a r r i s , Hoffman and G r i f f i t h , 1957) by presenting to l i s t e n e r s , s t i m u l i from the experiment described above embedded i n an ABX d i f f e r e n t i a t i o n t e s t . This task involved acoustic d i s c r i m i n a t i o n , rather than l i n g u i s t i c l a b e l l i n g . Two s t i m u l i d i f f e r e n t i a t e d by one minimal change i n second formant t r a n s i t i o n were said to be one 'step' "The Pattern Playback i s an instrument which works l i k e a speech spectro-graph i n reverse....The Pattern Playback, on the other hand, accepts a spectrogram-like pattern, scans i t with a l i g h t beam and produces the corresponding sound wave." ( D e n e s a n d p i n s o n > 1 9 6 3 , p . 1 2 9 ) -6-apart; the number of 'steps' between two s t i m u l i were the number of minimal changes i n t r a n s i t i o n s . The B s t i m u l i used i n the ABX d i f f e r e n t i a t i o n t e s t were one, two or three 'steps' removed from the A s t i m u l i . Fewer 'steps' were required for d i f f e r e n t i a t i o n at the phoneme boundary than in the area between phoneme boundaries (the area of consistent phoneme i d e n t i f i c a t i o n ) . An experiment of the same type was conducted using non-speech-like s t i m u l i (Liberman, Ha r r i s , Kinney and Lane, 1961). The purpose of t h i s study was to investigate the p o s s i b i l i t y that v a r i a t i o n i n d i f f e r e n t i a l s e n s i t i v i t y observed with second formant t r a n s i t i o n s was c h a r a c t e r i s t i c of a l l acoustic perception. V a r i a t i o n i n d i f f e r e n t i a l thresholds were not observed when the subjects responded to s t i m u l i throughout the continuum. According to Fry the d i f f e r -ence between responses to speech and non-speech s t i m u l i suggested that: ...the features revealed by the previous l a b e l l i n g and d i s c r i m i n a t i o n experiments r e a l l y are associated with language functioning and are a genuine expression of phonemic s t r u c t u r i n g . (Fry, 1964, p. 65) Investigations at Haskins Laboratories suggested that observed d i s t i n c t i v e n e s s of phoneme boundaries, i n perception, might depend p a r t l y on a r t i c u l a t o r y d i f f e r e n c e s between two phonemes. Vowels are produced along a continuum of a r t i c u l a t o r y p o s i t i o n s ; on the other hand consonants are produced i n d i s c r e t e , a r t i c u l a t o r y p o s i t i o n s , (Fry, 1964). A d i s c r i m i n a t i o n experiment (Fry, 1964) using F^ and F^ patterns f or the synthetic vowels / I / , /e/ and /ae / as s t i m u l i on the Pattern Playback, revealed vowel phoneme boundaries to be less sharply defined than phoneme boundaries observed for consonants. The i d e n t i f i c a t i o n of a vowel was influenced by s t i m u l i which had preceded i t . Although grouping was more d i f f i c u l t with vowels, subjects could d i s t i n g u i s h many more sub-phonemic d i f f e r e n c e s , and only one 'step' d i f f e r e n c e s were found to be perceptually d i f f i c u l t i n the ABX tasks. The degree of d i f f e r e n t i a l s e n s i t i v -i t y was consistent throughout the range of formant values tested. In summary, the r e s u l t s of the foregoing experiments i n d i c a t e that the perception of -7-vowels i s more l i k e the pe r c e p t i o n of non-speech sounds, than l i k e the per c e p t i o n of consonants. In order t o i n v e s t i g a t e phonemic l a b e l l i n g by speakers of four d i f f e r e n t languages ( E n g l i s h , Puerto-Rican S p a n i s h , Hungarian and T h a i ) , Lotz et a l . (1950) presented spoken s y l l a b l e s c o n s i s t i n g of v o i c e l e s s a s p i r a t e d , v o i c e l e s s u n a s p i r a t e d , and voiced unaspirated s t o p s . Each group e x h i b i t e d d i f f e r e n t p o s i t i o n s f o r phoneme boundaries, i . e . the same s t i m u l i were l a b e l l e d d i f f e r e n t l y by the four d i f f e r e n t language groups. Lotz's r e s u l t s i n d i c a t e d t h a t , i n d e a l i n g w i t h s i m i l a r s e t s of s t i m u l i , d i f f e r e n t language users apply c r i t e r i a e s t a b l i s h e d through t h e i r n a t i v e phonemic system. Although evidence f o r phonemic c l a s s i f i c a t i o n of stop consonants now seems adequate, c o n s i d e r a t i o n of c o - a r t i c u l a t i o n e f f e c t s , together w i t h p e r c e p t i o n of subphonemic fe a t u r e s and suprasegraentals, i n d i c a t e s that speech p e r c e p t i o n cannot be completely e x p l a i n e d by phonemic c a t e g o r i z a t i o n . M i l l e r (1962) proposed t h a t p e r c e p t u a l u n i t s of connected speech are i n f a c t l a r g e r than the phoneme, and a f f e c t e d by the morphological and grammatical r u l e s i n v o l v e d i n an u t t e r a n c e . In h i s stud y , such u n i t s were considered t o be p o s s i b l y as long as sh o r t phrases. 1.22 Sound Sequences I n s i g h t i n t o the per c e p t i o n of consonant sequences i s gained from a d i s c u s s i o n of the per c e p t i o n of non-speech sequences. This l a t t e r phenomenon has been i n v e s t i g a t e d i n terms o f , (a) temporal o r d e r i n g , i . e . the a b i l i t y t o pe r c e i v e the order of two or more consecutive sound s t i m u l i , (b) tone d i f f e r e n t i a t i o n , i . e . the a b i l i t y t o p e r c e i v e as two separate sounds two s t i m u l i separated by a short i n t e r v a l ; and (c) sound i n t e r a c t i o n , i . e . the i n f l u e n c e of a sound s t i m u l u s on the p e r c e p t i o n of the adjacent sound s t i m u l u s . A summary of these s t u d i e s i s presented i n the f o l l o w i n g s e c t i o n . -8-H i r s h (1959) and Broadbent and Ladefoged (1959) s t u d i e d the p e r c e p t i o n of temporal order using a v a r i e t y of sound s t i m u l i . In H i r s h ' s study, 500 msec, tones, 500 msec, noise b u r s t s , and c l i c k s which had v a r i a b l e i n t e r s t i m u l u s i n t e r v a l s , were presented to s u b j e c t s . A f t e r c o n s i d e r a b l e p r a c t i c e , the order of two random sounds could s t i l l be p e r c e i v e d when the i n t e r v a l between the onset of each stimulus was reduced to 15 - 20 msec. The author concluded th a t the minimal i n t e r v a l necessary f o r the p e r c e p t i o n of order was not i n f l u e n c e d by the d u r a t i o n and q u a l i t y of the sounds i n the sequence. The sounds used by Broadbent and Ladefoged (1959) were a 120 msec, high frequency n o i s e , a 30 msec.-800 Ha-tone, and a 30 msec.-171 Ha-square wave. On f i r s t p r e s e n t a t i o n of the s t i m u l i an onset i n t e r v a l of 150 msec, was needed before order could be p e r c e i v e d . As p r e v i o u s l y observed, the minimal onset i n t e r v a l was reduced to 30 msec, by s u b j e c t s , w i t h t r a i n i n g . According to H i r s h (1959), the minimal i n t e r v a l necessary f o r the p e r c e p t i o n of order was independent of the q u a l i t y of the sounds i n the sequence. Broadbent and Ladefoged (1959), however, suggested that the q u a l i t y of the p a i r as a whole might have i n f l u e n c e d the p e r c e p t i o n of order i n t h e i r experiment. In a set of four experiments, C h i s t o v i c h (1960) i n v e s t i g a t e d tone d i f f e r e n t i a t i o n , p e r c e p t i o n of temporal o r d e r , and sound i n t e r a c t i o n . C h i s t o v i c h presupposed that the perception of a sound depended on a c e r t a i n i n t e r v a l of exposure and was a f f e c t e d by the d u r a t i o n of the sound. The f i r s t experiment was designed to determine the i n t e r v a l between two tones necessary f o r the d i s c r i m i n a t i o n of two separate tones. C h i s t o v i c h presented 20 msec, of 500 Ha and 4000 Ha-tones, w i t h v a r i a b l e i n t e r s t i m u l u s i n t e r v a l s . V/ith an i n t e r v a l s h o r t e r than 80 m s e c , two tones were seldom p e r c e i v e d as two; w i t h an i n t e r v a l of 110-115 msec, two tones were d i f f e r e n t i a t e d f i f t y percent of the t i m e . In a second experiment, C h i s t o v i c h r e p l i c a t e d Hirsh's and Broadbent and Ladefoged's i n v e s t i g a t i o n s of p e r c e p t i o n of temporal o r d e r . The s u b j e c t ' s -9-task i n v o l v e d i d e n t i f i c a t i o n of a sequence of 500 Ha-and 4000 Ha-tones separated by v a r i a b l e i n t e r v a l s . Responses to the s t i m u l i i n d i c a t e d two types of pe r c e p t u a l d i s t u r b a n c e s . The f i r s t d i s t u r b a n c e , i n which one tone was heard and subjects attempted t o guess whether i t began high or low, was observed up t o 100 msec. The second d i s t u r b a n c e , i n which s u b j e c t s heard two sounds but cou l d not i d e n t i f y them, operated f o r average i n t e r v a l s . One hundred percent r e c o g n i t i o n of a.sequence r e q u i r e d an i n t e r s t i m u l u s i n t e r v a l up to 300 msec. The discrepancy between t h i s value and that obtained by Broadbent and Ladefoged might be explai n e d by the type of s t i m u l i used i n each st u d y . In C h i s t o v i c h ' s i n v e s t i g a t i o n , the s t i m u l i were two pure tones;, whereas i n Broadbent and Ladefoged's study the s t i m u l i were d i f f e r e n t types of s i g n a l s . In her t h i r d and f o u r t h experiments, C h i s t o v i c h examined sound i n t e r -a c t i o n . The e f f e c t of a f o l l o w i n g complex tone on the d e t e r m i n a t i o n of frequency of a s i g n a l was i n v e s t i g a t e d i n the t h i r d of the s e r i e s . Two pure tones, A^ and A^, were each f o l l o w e d by a complex sound B. A^ was held c o n s t a n t , w h i l e A^ was v a r i e d by the subjec t u n t i l the two sequences, A^B and A^B, were perceived as uniform. The value of the frequency d i f f e r e n c e over the abs o l u t e frequency, (A^-AgWA. d i m i n i s h e d as the i n t e r v a l between the A and B tone increased to about 120 msec. For i n t e r v a l s longer than 120 m s e c , the value i n c r e a s e d s l i g h t l y . Subjects reported that t h e i r method of comparison a l t e r e d w i t h v a r i a t i o n i n the l e n g t h of i n t e r v a l . When short i n t e r v a l s (up to 90 msec.) were presented, s u b j e c t s compared the beginning of the combined sounds; f o r average i n t e r v a l s (90 t o 160 msec.) s u b j e c t s compared the whole sequences; f o r the longest i n t e r v a l s (160 t o 300 msec.) su b j e c t s perceived separate tones but apparently the f i r s t complex sound hindered the r e c a l l of the f i r s t pure tone. -10-Th e f o u r t h experiment i n C h i s t o v i c h ' s s e r i e s i n v o l v e d comparison of the frequency of an i s o l a t e d pure tone w i t h the frequency of a pure tone d e l i v e r e d i n a sound sequence. The complex sound which f o l l o w e d the tone d e l i v e r e d i n the sound sequence i n f l u e n c e d d i s c r i m i n a t i o n of the frequency of the tone over a range of i n t e r v a l s up t o 300 msec. With short i n t e r v a l s , the d i f f e r e n t i a l t h r e s h o l d (frequency d i f f e r e n c e between the two tones d i v i d e d by the absolute frequency) p r o g r e s s i v e l y decreased w i t h an i n c r e a s e i n i n t e r s t i m u l u s i n t e r v a l . With average i n t e r v a l s , the i n f l u e n c e of the second sound was c o n s i d e r a b l e but d i d not depend on the length of the i n t e r v a l . C h i s t o v i c h compared the r e s u l t s of experiments three and f o u r and noted t h e i r s i m i l a r i t y t o speech p e r c e p t i o n . R e s u l t s of the t h i r d experiment were b e t t e r f o r short i n t e r v a l s , s i n c e i d e n t i c a l types of s t i m u l i were being compared. Comparison of two sound sequences, as presented i n the t h i r d experiment, was considered to be analogous to a comparison of two s y l l a b l e s or consonant c l u s t e r s i n speech. I t i s noted that two processes were invo l v e d i n e v a l u a t i o n of a sound sequence, both comparing the c h a r a c t e r -i s t i c s of the two f i r s t sounds, and d i f f e r e n t i a t i n g the f i r s t and second sound. An analogy to speech, t h e r e f o r e , i m p l i e s that a comparison of two c l u s t e r s , f o r example, /sp/ compared w i t h /sk/ or /ks/, u t i l i z e d both the absolute c h a r a c t e r i s t i c s of the f i r s t element and the d i f f e r e n c e between the f i r s t and second element. R e s u l t s of the f o u r t h experiment were b e t t e r f o r longer i n t e r v a l s (160 to 300 msec.) s i n c e the two tones to be compared were adjace n t . I t was suggested that the absolute c h a r a c t e r i s t i c s of the tones were the b a s i s f o r e v a l u a t i o n of an i s o l a t e d tone w i t h a tone d e l i v e r e d i n a sound sequence. A comparison of such s t i m u l i was considered to be analogous to that of a s i n g l e consonant w i t h a consonant i n a s y l l a b l e or c l u s t e r , f o r -11-examp-le Is I compared w i t h / s t / . The r e s u l t s appear to imply that p e r c e p t i o n . i n the l a t t e r case would be based on the absolute c h a r a c t e r i s t i c s of the s i n g l e speech sound which i n the example c i t e d i s / s / . V.'arren and Warren (1S70) i n v e s t i g a t e d the pe r c e p t i o n of temporal order without the cues of onset and t e r m i n a t i o n . They hypothesized t h a t : "sequence p e r c e p t i o n w i t h p a i r s of sounds represents a s p e c i a l case and i s r e a l l y the perception of onset and t e r m i n a t i o n . " (Warren and Warren, 1970, P. 3 4 ) . A v a r i e t y of sequences, c o n s i s t i n g of a 1000 Ha-tone, a 40 Ha-square wave, a' 796 Ha-tone and an octave band of noise 'centered at 2000 Hz, were recorded on tape loops w i t h no i n t e r s t i m u l u s i n t e r v a l . When s t i m u l i were 200 msec, i n d u r a t i o n , c o r r e c t i d e n t i f i c a t i o n of the sequence occurred o n l y at chance l e v e l . In order to o b t a i n 100 percent c o r r e c t i d e n t i f i c a t i o n of the sequence from h a l f the subjects ( s a m p l e ) , ' i t was necessary to in c r e a s e the d u r a t i o n of each stimulus t o between 300 and 700 msec. When four spoken d i g i t s , each 200 m s e c , were presented the subjects p e r c e i v e d the order a t once and gave c o n s i s t e n t responses. This r e s u l t i n d i c a t e d t h a t : ...except w i t h great d i f f i c u l t y , accurate p e r c e p t i o n of temporal order may be p o s s i b l e only f o r sequences that resemble those encountered i n speech and i n music -- s p e c i a l sequences i n which the component sounds are l i n k e d together f o l l o w i n g s p e c i f i c r u l e s , i n t o coherent passages. (Warren and Warren, 1970, p. 36) A comparison w i t h the p e r c e p t i o n of non-speech s i g n a l s adds i n s i g h t i n t o the perception of speech sound sequences. Temporal ordering of speech elements i s e s s e n t i a l t o the understanding of a l i n g u i s t i c message. Warren and Warren's r e s u l t s have shown that i t i s not p o s s i b l e to order non-speech s t i m u l i at the same rat e as speech s t i m u l i . I t has been shown that an i n t e r -stimulus i n t e r v a l of 150 m s e c , ( C h i s t o v i c h , 1960) i s needed between sound s t i m u l i f o r accurate sequence i d e n t i f i c a t i o n . These i n t e r v a l s are not found between speech sounds, and yet accurate sequencing occurs w i t h no d i f f i c u l t y . According to C h i s t o v i c h , an i n t e r v a l of 110 - 115 msec i s needed between - 1 2 -two tones i f they are t o be d i f f e r e n t i a t e d f i f t y percent of the time; consonant c l u s t e r s have s h o r t e r i n t e r s t i m u l u s i n t e r v a l s , yet both consonants are per c e i v e d at a l l t i m e s . For consonants, the subsequent a c o u s t i c s i g n a l a i d s i n p e r c e p t i o n , whereas C h i s t o v i c h ' s r e s u l t s i n d i c a t e d that sound p e r c e p t i o n of a pure tone was hindered by a f o l l o w i n g complex sound. The r e s u l t s of experiments d e s c r i b e d i n the foregoing paragraphs appear to i n d i c a t e that p e r c e p t i o n of consonant sequences i s b e t t e r than pe r c e p t i o n of sound sequences i n three r e s p e c t s ; i ) f o r temporal ordering, i,e. the a b i l i t y t o perceive the order of two consecutive sounds, i i ) f o r tone d i f f e r e n t i a t i o n , i . e . the a b i l i t y t o p e r c e i v e as two separate sounds, two adjacent s t i m u l i , and i i i ) f o r sound i n t e r a c t i o n , i . e . the i n f l u e n c e of a sound stimulus on the p e r c e p t i o n of the adjacent sound s t i m u l u s . Some of the g r e a t e r e f f i c i e n c y of the speech system may be due t o the redundancy i n speech such as the t r a n s i t i o n s between consonants and vowels. 1.23 F r i c a t i v e s The consonant c l u s t e r s s t u d i e d were composed of f r i c a t i v e s and p l o s i v e s . Each of these types of sound w i l l be d i s c u s s e d i n d i v i d u a l l y f o l l o w e d by a d i s c u s s i o n of these sounds combined i n consonant c l u s t e r s . The f r i c a t i v e i n i s o l a t i o n has been found t o be c h a r a c t e r i z e d by a p e r i o d of noise c a l l e d f r i c t i o n ; i n a CV s y l l a b l e , t h i s f r i c t i o n i s f o l l o w e d by a segment w i t h well-marked formant s t r u c t u r e s termed by K. H a r r i s (1958) the v o c a l i c p o r t i o n . Hughes and H a l l e , (1956), considered the f r i c t i o n p o r t i o n of the i s o l a t e d f r i c a t i v e s / f / , /v/, / s / , / z / , ///, /0/, and / § / . C o n s i s t e n t d i f f e r e n c e s i n s p e c t r a l a n a l y s i s were i d e n t i f i e d among the va r i o u s a r t i c u l a t o r y p o s i t i o n s f o r these sounds. A c o u s t i c s i g n a l s w i t h s p e c t r a l energy peaking between 2000 Ha and 4000 Ha were i d e n t i f i e d as ///; s i g n a l s w i t h s p e c t r a l energy peaking below 2000 Ha were i d e n t i f i e d as / f / ; - 1 3 -and s i g n a l s w i t h s p e c t r a l ene rgy p e a k i n g above 4000 Ha were i d e n t i f i e d as / s / . In 1960, S t r e v e n s s t u d i e d n i n e v o i c e l e s s f r i c a t i v e s and found t h a t t h e p o s i t i o n o f c o n s t r i c t i o n r e s u l t e d i n i d e n t i f i a b l e d i f f e r e n c e s o f sound q u a l i t y . He a l s o showed t h a t s l i g h t changes i n mouth o p e n i n g c o u l d c o n t r i b u t e t o s p e c t r a l d i f f e r e n c e s among the f r i c a t i v e s . In S t r e v e n ' s s t u d y , i t was found t h a t a i r p r e s s u r e was a l t e r e d by b o t h a f f e c t v a r i a t i o n o f s p e e c h l e v e l and i n t r i n s i c i n t e n s i t y o f a phoneme. T h i s i n t r i n s i c i n t e n s i t y was n o t e d t o be dependen t on the e x t e n t and d u r a t i o n o f t he o b s t r u c t i o n . S t r e v e n s o b s e r v e d t h a t , f o r a g i v e n s p e a k e r , a f r i c a t i v e c o u l d be h a b i t u a l l y p r o d u c e d a t a h i g h e r o r lower i n t e n s i t y l e v e l t han the same f r i c a t i v e p r o d u c e d by a n o t h e r s p e a k e r . The o r d e r o f t h e f o l l o w i n g a c o u s t i c d e s c r i p -t i o n s i s ba sed on S t r e v e n ' s a s c e n d i n g rank o r d e r o f f r i c a t i v e i n t e n s i t y . The a c o u s t i c d e s c r i p t i o n s f o r each sound were d e r i v e d f rom b road band s p e c t r o g r a m s and a m p l i t u d e s e c t i o n s . It I'. Lowest f r e q u e n c y i s a r o u n d 1500 - 1700 c p s . Low peaks o f e n e r g y t end t o o c c u r a round 1900 c p s . , 4000 c p s . , and o c c a s i o n a l l y 5000 c p s . Uppe r l i m i t o f f r e q u e n c y i s r a r e l y be low 70C0 c p s . , u s u a l l y a round 7500 cps Is/: Lowest f r e q u e n c y a lmos t a lways above 3500 c p s . Peaks of e n e r g y t e n d t o o c c u r w i t h no a p p a r e n t p a t t e r n , e x c e p t t h a t t hey do no t l i e c l o s e r t o one a n o t h e r t h a n 1000 c y c l e s . Upper l i m i t o f f r e q u e n c y exceeds 8000 c p s . i n most c a s e s . . . . / / / : Lowest f r e q u e n c y v a r i e s between 1600 and -2500 c p s . Peaks o f e n e r g y t end t o o c c u r not l e s s t h a n 1000 c y c l e s a p a r t and the a s p e c t o f a m p l i t u d e c r o s s - s e c t i o n s shows a w e i g h t i n g towards the b o t t o m o f t he p a t t e r n . Upper l i m i t o f f r e q u e n c y shows a s h a r p c u t - o f f a round 7000 c p s . . . . ( S t r e v e n s , 1960, p. 41) The measurements , r e p o r t e d by S t r e v e n s , were o b s e r v e d b o t h f o r f r i c -a t i v e s i n spon taneous s p e e c h u t t e r a n c e s , and f r i c a t i v e s p r o l o n g e d f o r s p e c t r o g r a p h i c a n a l y s i s . The p r o d u c t i o n o f s y n t h e t i c f r i c a t i v e s , u s i n g s i m u l a t e d s p e c t r a emp loy ing t he measurement d e s c r i b e d above , l e d t o improved i n t e l l i g i b i l i t y . -14-F r i c a t i v e s / s / , ///, If/, /v/, and /9/ i n the context of a s y l l a b l e , and the r e l a t i v e c o n t r i b u t i o n of cues i n the f r i c t i o n and v o c a l i c p o r t i o n of a sound, were studied by H a r r i s (1958). The f r i c t i o n p o r t i o n of a f r i c a t i v e was separated from i t s v o c a l i c segment, and combined w i t h the v o c a l i c p o r t i o n of a d i f f e r e n t f r i c a t i v e . This procedure was performed on both v o i c e l e s s and v o i c e d f r i c a t i v e s , although f o r v o i c e d f r i c a t i v e s the boundary between f r i c t i o n and v o c a l i c p o r t i o n was more d i f f i c u l t t o d e f i n e . I t was observed t h a t f r i c t i o n , i n the consonants Is I and ///, when p a i r e d w i t h the v o c a l i c p o r t i o n of any f r i c a t i v e , produced a judgment of Is I and ///. Correct judgment of / f / and /©/, however, depended on the v o c a l i c p o r t i o n of the s y l l a b l e , f o r example, when the v o c a l i c p o r t i o n of If I was p a i r e d w i t h If I or /0/ f r i c t i o n , the r e s u l t a n t sound was judged as If I; when any other v o c a l i c p o r t i o n was p a i r e d w i t h If I or /0/ f r i c t i o n , the r e s u l t a n t sound was judged as /9/. S i m i l a r , although l e s s e x p l i c i t , r e s u l t s were obtained w i t h voiced f r i c a t i v e s . 1.24 P l o s i v e s Parameters which d i f f e r e n t i a t e p l o s i v e s are frequency of the s p e c t r a l peak of the b u r s t , d u r a t i o n of c l o s u r e and shaping of the formants. The v o i c i n g d i s t i n c t i o n was not considered i n the present i n v e s t i g a t i o n s i n c e i n most cases E n g l i s h consonant c l u s t e r s c o n s i s t of a v o i c e l e s s f r i c -a t i v e w i t h a v o i c e l e s s p l o s i v e . The length of i n t e r v a l between the b u r s t and the beginning of the v o c a l i c p o r t i o n has been shown to be the d i f f e r e n t i a t -ing f a c t o r between voiced and v o i c e l e s s p l o s i v e s ; length of i n t e r v a l : f o r /p/ mean i s 120 msec, w i t h a range of 90-140 m s e c ; f o r /b/ mean i s 75 msec, w i t h a range of 65-90 msec. ( L i s k e r and Abramson, 1964). The d i s c r i m i n a t i o n of place f o r d i f f e r e n t p l o s i v e s has been shown t o depend both on the frequency of the b u r s t and the formant t r a n s i t i o n f o l l o w i n g -15-c l o s u r e . With s y n t h e t i c p l o s i v e s , / t / was i d e n t i f i e d when high frequency b u r s t s were followed by any vowel, whereas /p/ and /k/ i d e n t i f i c a t i o n depended on the p o s i t i o n of the burst i n r e l a t i o n to the formants of the f o l l o w i n g vowel; /k/ was i d e n t i f i e d when the burst was l e v e l or s l i g h t l y above the second formant and /p/ was i d e n t i f i e d when the bur s t was below the second formant, (Cooper et a l . 1952). Researchers at Kaskins Laboratory have e x t e n s i v e l y s t u d i e d the e f f e c t of formant t r a n s i t i o n s on p l o s i v e i d e n t i f i c a t i o n . In an experiment conducted i n 1952, Cooper et a l . v a r i e d the extent and d i r e c t i o n of second formant t r a n s i t i o n s on P a t t e r n Playback recordings i n an attempt to determine e f f e c t s on voiced and v o i c e l e s s p l o s i v e s , and nasal i d e n t i f i c -a t i o n . The f i r s t and only part of the experiment r e l e v a n t to t h i s t h e s i s d e a l t w i t h the p a i r i n g of second formant t r a n s i t i o n s and s y n t h e t i c two-formant vowels /i,e fG,a,e,o,u/. Eleven second formant t r a n s i t i o n s were used v a r y i n g i n 120 Ha steps from f o u r harmonics (480 Ha) below to s i x harmonics (720 Ha) above the steady s t a t e p o r t i o n . From the 77 separate s t i m u l i (11 t r a n s i t i o n s w i t h seven v o w e l s ) , two experimental sets were created; one set using a s m a l l negative f i r s t formant and a second set using a l a r g e r f i r s t formant.. These sets were i d e n t i f i e d f i r s t i n the/p-t-k/ frame work and secondly i n the/b-d-g/ framework. The p r o b a b i l i t y of s t i m u l i being i d e n t i f i e d as v o i c e l e s s d i d not increase w i t h the small negative f i r s t formant t r a n s i t i o n . Data f o r second formant t r a n s i t i o n s which Cooper et a l . analyzed according to place i d e n t i f i c a t i o n gave the f o l l o w i n g r e s u l t s . / b/-/p/responses - when vowels / i / , /e/, /el, /a/, / Q/ were preceded by minus t r a n s i t i o n s . - few w i t h /o/ and /Q/; most of these were a l s o w i t h i n the minus range. -16-/d/-,/t/responses - i d e n t i f i c a t i o n depended on the vowel. - w i t h IH no t r a n s i t i o n gave /d/ as dominant response. - w i t h lei, 1^1 when t r a n s i t i o n s were i n the v i c i n i t y of 0. - w i t h /a/, Iol, /ol, /u/ w i t h p r o g r e s s i v e l y l a r g e r p o s i t i v e t r a n s i t i o n s . Ig/-•/-/ responses - w i t h p o s i t i v e t r a n s i t i o n s which are not d/t judgments. - w i t h extreme p o s i t i v e t r a n s i t i o n s f o r HI, lei, lei, /a/. - w i t h small p o s i t i v e t r a n s i t i o n f o r /o/. - w i t h t r a n s i t i o n s between +4 and 0 f o r /o/, /u/. Liberman and h i s colleagues concluded that d i r e c t i o n and degree of change" of second formant t r a n s i t i o n s serve as cues f o r place d i s t i n c t i o n . They a l s o observed that a g r e a t e r number of c o r r e c t i d e n t i f i c a t i o n s were made when the t h i r d formant of each vowel was i n c l u d e d . In a f u r t h e r a n a l y s i s of st o p consonants, D e l a t t r e et a l . (1955) proposed the ex i s t e n c e of second formant l o c i , independent of p o s i t i o n of the formants i n the f o l l o w i n g vowel. The quest i o n of whether the l o c i was a s t a r t i n g p o i n t f o r t r a n s i t i o n s or a poin t f o r t r a n s i t i o n s t o poin t toward was i n v e s t i g a t e d . In the f i r s t p a r t of t h i s experiment, s t i m u l i w i t h v a r i a b l e s t r a i g h t second formants and r i s i n g f i r s t formant t r a n s i t i o n s were presented. Data from t h i s experiment i n d i c a t e d t h a t : 1. best Igl was produced w i t h F at 3000 Ha; best Id/ was produced w i t h F at 1800 Ha; best /b/ was produced w i t h F 2 at 720 Ha. 2. except when the s t r a i g h t second formant was between Ig/ and /d/ l o c i the second formant l o c i were independent of vowel c o l o u r . 3. the extent of the f i r s t formant t r a n s i t i o n has no e f f e c t on the i d e n t i t y of stop consonants. Using a /d/ l o c i of 1800 Ha as o r i g i n a t i n g p o i n t , p a t t e r n s were drawn to a v a r i e t y of steady s t a t e vowel formant p o s i t i o n s . These s t i m u l i d i d not -17-e l i c i t 100 percent of /d/ judgments. However, a l l s t i m u l i were pe r c e i v e d as /d/ when the f i r s t 50 msec, of the t r a n s i t i o n s were era s e d . These r e s u l t s suggested that i n a r t i c u l a t o r y movements from consonant to vowel, some part of the movement must be completed before sound i s produced. When the le n g t h of s i l e n t i n t e r v a l was v a r i e d , the most c o n s i s t e n t responses were obtained when the s i l e n t p o r t i o n was made to equal h a l f of the t o t a l t r a n s i t i o n . L o c i f o r the f i r s t formant were a l s o t e s t e d i n t h i s s e r i e s of experiments. Most c o n s i s t e n t c o r r e c t responses were obtained when the f i r s t formant l o c i were set at 240 Ha and p r o p o r t i o n of c o r r e c t responses decreased as the frequency of the formant l o c i were r a i s e d . I t i s p o s s i b l e that the locus was between 240 Ha and 0 Ha, however, l i m i t a t i o n s i n in s t r u m e n t a t i o n prevented lowering of locus frequency. The frequency of the stop burst and the vowel t r a n s i t i o n f o l l o w i n g the s i l e n t i n t e r v a l have been shown to be a c o u s t i c cues f o r the per c e p t i o n of stop consonants. 1.25 C l u s t e r s The perception of f r i c a t i v e and p l o s i v e combinations have been s t u d i e d using s y n t h e t i c speech c l u s t e r s , and n a t u r a l speech c l u s t e r s i n n o i s e . P r i o r to U l d a l l ' s experiment i n 1964, no sy n t h e s i s of stops adjacent t o f r i c a t i v e s had been attempted. U l d a l l s y n t h e s i z e d the consonant c l u s t e r s /sps/, / s t s / , and /sks/ i n the words " c l a s p s " l i s p s " l a s t s " , ' l i s t s ? 'asks" " r i s k s , " using a number of a c o u s t i c s p e c i f i c a t i o n s . S u c c e s s f u l s y n t h e s i s of these c l u s t e r s depended on the use of ' t r a n s i t i o n ' type frequency changes i n the f r i c a t i v e n o i s e . The f o l l o w i n g a c o u s t i c d e s c r i p t i o n s by U l d a l l gave the most s u c c e s s f u l s y n t h e s i s : Words w i t h the c l u s t e r / s t s / , l a s t s and l i s t s , are s a t i s f a c t o r i l y made by using a segment of h i s s w i t h a steady peak frequency of 4000 c p s . fo l l o w e d by s i l e n c e f o r the It I, f o l l o w e d by another segment of the same steady frequency h i s s w i t h no s o r t of t r a n s i t i o n . -18-The most s a t i s f a c t o r y /sks/ c l u s t e r s were produced by adding to the steady /s/ h i s s at 4000 c p s . short b u r s t s - - 3 to 5 centiseconds--of h i s s i n the region of 1400-1800 cps. before and a f t e r the s i l e n c e f o r the /k/ c l o s u r e . . . The a d d i t i o n of burst s of h i s s of changing frequency i n the region 1000-2000 c p s . a l s o produced acceptable /sks/ c l u s t e r s . . . We have only been able t o produce acceptable /sps/ c l u s t e r s as i n 'clasps" by s y n t h e s i z i n g very marked t r a n s i t i o n s i n the f r i c a t i v e n o i s e . In the f i r s t method, the peak frequency of the h i s s which i s being used a t a steady value f o r most of the /s/ i s made t o move r a p i d l y ( i n a few centiseconds) down from 4000-1000 cps. immediately before the s i l e n c e f o r the s t o p , and up again r a p i d l y a f t e r the s i l e n c e . In the second method, the same r e s u l t , a r a p i d t r a n s i t i o n i n the f r i c a t i v e n o i s e , i s obtained by u s i n g the other h i s s , which c o n s i s t s i n p u t t i n g h i s s i n s t e a d of v o i c e through Formant 2 and Formant 3 and i s most commonly used f o r /h/'s and a s p i r a t i o n . In the second method of s y n t h e s i z i n g /sps/, bursts of t h i s h i s s , of r a p i d l y changing frequency, are added to the steady h i s s at 4000 cps., before and a f t e r the s i l e n c e f o r the s t o p . The range covered i n 3 to 5 centiseconds by t h i s h i s s of changing frequency was 3400 t o 1600 cps. and back a g a i n . ( U l d a l l , 1964, pp. 13-14) The data i n d i c a t e d that the c l u s t e r s /sks/ and/sps/ r e q u i r e d a c o u s t i c c h a r a c t e r i s t i c s from both adjacent consonants f o r the most e f f e c t i v e s y n t h e s i s . The t r a n s i t i o n e f f e c t may a l s o have been operating w i t h / s t s / , but the s p e c t r a l energy of the h i s s f o r Is I may have masked the high frequency b u r s t of It I. U l d a l l ' s data f o r /sps/ and /sks/ and the p o s s i b l e i n t e r p r e t a t i o n of data f o r / s t s / , support the hypothesis that consonant c l u s t e r s are perceived as a u n i t , r a t h e r than a sequence of d i s t i n c t elements. The a c o u s t i c c h a r a c t e r i s t i c s noted f o r the three consonant c l u s t e r s s t u d i e d by U l d a l l were inv o l v e d i n the experimental c l u s t e r s , s i n c e the s t i m u l i presented included some having the f r i c a t i v e f i r s t and others having the p l o s i v e f i r s t . The mechanism employed i n the perception of consonant c l u s t e r s was i n v e s t i g a t e d by Bond (1971). She suggested t h a t : By observing the p a t t e r n of confusion of obstruent c l u s t e r s i n the presence of n o i s e , i t i s p o s s i b l e to make some inferences about the pe r c e p t u a l mechanism u n d e r l y i n g the p e r c e p t i o n of these c l u s t e r s . (Bond, 1971, p. 48) -19-Bond presented f i f t e e n p a i r s of meaningful words, each d i f f e r i n g only i n the order of elements i n the c l u s t e r s , f o r example, "caspian" and "c a p s i a n " . F i v e p a i r s of words were made for each c l u s t e r combination used: /ps/-/sp/, Its I - / s t I , and /ks/-/sk/, which appeared i n f i n a l and medial p o s i t i o n . Three randomized l i s t s of s i x t y words each were co n s t r u c t e d from the t h i r t y s t i m u l u s words presented t w i c e . White noise at a s i g n a l - t o - n o i s e r a t i o of e i t h e r +12, OdB or -6dB was added t o each l i s t . A l l of the prepared s t i m u l i were presented to nineteen s u b j e c t s from the L i n g u i s t i c s Department of the Ohio S t a t e U n i v e r s i t y . Fourteen s u b j e c t s gave w r i t t e n responses and f i v e subjects gave v e r b a l responses. Confusion matrices were const r u c t e d from the t o t a l l e d r e s u l t s . As a n t i c i p a t e d , scores d e t e r i o r a t e d as noise l e v e l i n c r e a s e d . Reversal e r r o r s , ( e r r o r s i n which the perceived c l u s t e r has the same elements as the stimulus c l u s t e r i n reverse order) were the most common e r r o r s f o r a l l consonant c l u s t e r s . S t o p - f r i c a t i v e c l u s t e r s were perceiv e d c o r r e c t l y more f r e q u e n t l y than corresponding f r i c a t i v e - s t o p c l u s t e r s . A n a l y s i s of w r i t t e n versus spoken responses, one versus two s y l l a b l e words and mono-morphemic versus bi-morphemic words i n d i c a t e d no d i f f e r e n c e i n confusion p a t t e r n s . Reaction time was c o n s i s t e n t l y f a s t e r f o r c o r r e c t responses than i n c o r r e c t responses, although t h i s d i f f e r e n c e was not s i g n i f i c a n t . A decrease f o r words ending i n /ps/ at +12 dB s i g n a l - t o - n o i s e r a t i o was the only s i g n i f i c a n t d i f f e r e n c e observed i n r e a c t i o n time. Methodological inadequacies became apparent on a n a l y s i s of t h i s s tudy. The s t i m u l i presented were meaningful words or were shown to the subject p r i o r to the l i s t e n i n g t e s t . D i f f e r e n t words were used f o r each consonant c l u s t e r ; f o r example, "caspian" and "capsian" were used but not " c a t s i a n " , " c a s k i a n " or " c a k s i a n " . The only exception to t h i s procedure was that of the stimulus p a i r s "apse, "asp" and "ax", "ask". Since only responses which f e l l i n t o -20-the d e s i r e d category could have been produced, responses which were considered due to the p e r c e p t i o n of c l u s t e r s were i n a c t u a l i t y due to the perception of whole words. The e f f e c t of the described procedure was t o bias responses so that the d e s i r e d r e s u l t s , of more frequent r e v e r s a l e r r o r s , were obtained without involvement of the perceptual mechanism proposed by Bond. Other inadequacies i n the study i n c l u d e , i ) the s e l e c t i o n of s u b j e c t s , i i ) the de t e r m i n a t i o n of the s i g n a l - t o - n o i s e r a t i o and i i i ) the t a b u l a t i o n of r e s u l t s . i ) I t i s h i g h l y probable that the use of p h o n e t i c a l l y knowledgable s u b j e c t s would have a f f e c t e d the number of c o r r e c t responses o b t a i n e d . Greater accuracy of response t o the experimental tapes was observed w i t h two p h o n e t i -c a l l y t r a i n e d l i s t e n e r s than w i t h p h o n e t i c a l l y naive l i s t e n e r s . i i ) The speaker of the stimulus words monitored the i n t e n s i t y of h i s speech s i g n a l d u r ing p r o d u c t i o n . A s i g n a l - t o - n o i s e r a t i o determined by t h i s method has been shown t o have c o n s i d e r a b l e v a r i a b i l i t y due to the speaker's imprecise monitoring of h i s own i n t e n s i t y l e v e l . (Brady, 1971). i i i ) The t o t a l number of responses i n the confusion matrices d i d not correspond t o the number of responses expected. The confusion matrices which were compared did not have the same t o t a l number of responses. In the t e x t no e x p l a n a t i o n was made of these d i s c r e p a n c i e s . A need f o r r e l i a b l e data on the perception of consonant c l u s t e r s was e v i d e n t . 1.26 E f f e c t of Noise on S i g n a l P e r c e p t i o n of consonants d i s t o r t e d by the a d d i t i o n of noise or by f i l t e r i n g was analyzed by M i l l e r and N i c e l y (1955). Responses f o r consonants presented at vario u s s i g n a l - t o - n o i s e r a t i o s were t a b u l a t e d i n co n f u s i o n m a t r i c e s . The r e s u l t s f o r these s i g n a l - t b - n o i s e l e v e l s were analyzed i n -21-terms of the f e a t u r e s : v o i c i n g , n a s a l i t y , a f f r i c a t i o n , d u r a t i o n and p l a c e . The production of turbulence or f r i c a t i o n i n the sounds / f / , /v/, / s / , / z / , //// 1^1i w a s termed a f f r i c a t i o n ; the long intense high-frequency noise c h a r a c t e r i s t i c f o r the sounds / s / , ///, / z / , /^/ c h a r a c t e r i z e d the f e a t u r e d u r a t i o n . M i l l e r and N i c e l y noted t h a t : V o i c i n g and n a s a l i t y are much l e s s a f f e c t e d by a random masking noise than are the other f e a t u r e s . A f f r i c a t i o n and d u r a t i o n which are so s i m i l a r that a s i n g l e f u n c t i o n c o u l d represent them both, are somewhat su p e r i o r to place but f a r i n f e r i o r to v o i c i n g and n a s a l i t y . V o i c i n g and n a s a l i t y are d i s c r i m i n a b l e at s i g n a l - t o - n o i s e r a t i o s as poor as -12 dB whereas the place of a r t i c u l a t i o n i s hard t o d i s t i n g u i s h at r a t i o s l e s s than 6 dB, a d i f f e r e n c e of some 18 dB i n e f f i c i e n c y . ( M i l l e r and N i c e l y , 1955, p. 349) House et a l . (1965) presented eighteen l i s t e n e r s w i t h masked CVC words t o i evaluate an a r t i c u l a t o r y t e s t i n g method. Twenty-two phonemes occurred i n the CVC words; s i x t e e n occurred i n both i n i t i a l and f i n a l p o s i t i o n , three (///, /hi, and /w/) occurred only i n i n i t i a l p o s i t i o n and three (/z/, / t / / , and /n/) occurred only i n f i n a l p o s i t i o n . S t i m u l i were masked by s p e c t r a l envelope shaped white n o i s e . "The noise spectrum envelope was uniform up to about 500 cps. and decreased i n l e v e l at 9 dB/oct.above that frequency." (House et a l . , 1965, p. 161). S t i m u l i were presented at s i x s i g n a l - t o -noise r a t i o s ranging from +4 to -16 dB. Percent c o r r e c t responses f o r each phoneme were averaged over the s i x c o n d i t i o n s . For p l o s i v e s considered i n the present study the descending order of the percent c o r r e c t was / t / , /k/, and /p/. For f r i c a t i v e s considered i n the present s t u d y , the descending order of the percent c o r r e c t f o r i n i t i a l p o s i t i o n was / s / , If/, and ///; and f o r f i n a l p o s i t i o n was / s / , and / f / . House et a l . noted that s u b j e c t s more e a s i l y recognized i n i t i a l consonants than f i n a l consonants and there was "a tendency f o r the v o i c e l e s s forms of cognate p a i r s to be heard more s u c c e s s f u l l y than the v o i c e d form, v i z . , /p/ /b/..." (House et a l . 1965, p. 163). -22-1.27 Models of Speech Perception To t h i s p o i n t , the s t u d i e s presented have each d e a l t w i t h i n d i v i d u a l and l i m i t e d aspects of speech and sound p e r c e p t i o n . Models of speech per c e p t i o n attempt to e x p l a i n mechanisms f o r the whole process of speech p e r c e p t i o n , from the d i s c r i m i n a t i o n of elements to the comprehension of. meaning. These models of speech p e r c e p t i o n f a l l i n t o two c a t e g o r i e s ; a c t i v e and p a s s i v e . A c t i v e models, as p r e v i o u s l y noted, are those i n which p e r c e p t i o n i s based on a comparison of an i n t e r n a l l y generated s i g n a l w i t h an incoming a c o u s t i c s i g n a l . Such t h e o r i e s which w i l l be discussed are C r o s s - c o r r e l a t i o n , A n a l y s i s - b y - S y n t h e s i s and the Motor Theory. Passive models, as p r e v i o u s l y mentioned, are those i n which p e r c e p t i o n i s based on d i r e c t a n a l y s i s of an incoming a c o u s t i c s i g n a l . D i s t i n c t i v e Features and Context S e n s i t i v e Coding are such t h e o r i e s which w i l l be d i s c u s s e d . The model t o be presented f i r s t d i scusses both a c t i v e and pas s i v e mechanisms. L i c l c l i d e r (1952) proposed three stages i n the p e r c e p t i o n of speech. The f i r s t stage i n v o l v e d the t r a n s l a t i o n of the speech wave i n t o a form s u i t a b l e f o r the nervous system. Mechanisms f o r such a process operated mainly i n the c o c h l e a . Through t h i s process the necessary a c o u s t i c i n f o r m a t i o n was e x t r a c t e d from the speech spectrum. The second s t a g e , the i d e n t i f i c a t i o n of elements from a c o u s t i c parameters, was considered to be accomplished by e i t h e r i ) an a c t i v e mechanism u s i n g c r o s s - c o r r e l a t i o n or i i ) a passive mechanism usi n g matched f i l t e r i n g . i ) The operations f o r c r o s s - c o r r e l a t i o n which were o u t l i n e d f o r L i c k l i d e r ' s h y p o t h e t i c a l neuronal c o r r e l a t o r can be s u p p l i e d by the nervous system, according to the author. In h i s proposed c o r r e l a t o r , the input s i g n a l went d i r e c t l y to f i f t y channels each having a l o c a l generator of a p a t t e r n f o r one sound, a m u l t i p l i e r and an i n t e g r a t o r . As the s i g n a l went -23-through the generator, the m u l t i p l i e r m u l t i p l i e d the incoming and generated s i g n a l s and the i n t e g r a t o r produced an i n t e g r a l f o r that v a l u e . This i n t e g r a l was sent to a comparator which i n d i c a t e d the value of g r e a t e s t magnitude. The system i m p l i e d the a c t i v e mechanism of withdrawing the s t o r e d standard patterns and comparing the i n t e g r a l s which were produced. For the operation of such a process, a l l of the f i f t y channels would have s t a r t e d g e n e r a t i n g t h e i r s p e c i f i e d s t o r e d s i g n a l as soon as the f i r s t sound a r r i v e d . i i ) Matched f i l t e r i n g was the passive mechanism proposed as an a l t e r n a t i v e to cross c o r r e l a t i o n . In such a system,i the patterns f o r f i l t e r i n g would have been b u i l t i n t o the s t r u c t u r e through which the input was d i r e c t e d . L i c k l i d e r f e l t that i d e n t i f i c a t i o n of speech elements bore more r e l a t i o n to f i l t e r i n g than to c o r r e l a t i o n . L i c k l i d e r proposed that the t h i r d stage i n speech p e r c e p t i o n was the comprehension of meaning. The author f e l t that t h i s process was more a matter of c o r r e l a t i o n than f i l t e r i n g . His j u s t i f i c a t i o n s were that under-standing seemed t o be an a c t i v e process and that concepts c o u l d have been sto r e d elsewhere and d i r e c t e d to the c o r r e l a t i n g mechanism f o r comparison w i t h the incoming message. The most obvious disadvantage of t h i s system i s the i m p l i e d n e c e s s i t y of immense storage and channel f a c i l i t i e s . The problem of segmentation of the continuous speech wave has not been s o l v e d i n L i c k l i d e r ' s model. The second model to be presented was e n t i r e l y based on a c t i v e mechanisms. ' A n a l y s i s - b y - S y n t h e s i s ' was o u t l i n e d by Stevens, (1960); H a l l e and Stevens, (1962); and Stevens and H a l l e , (1967). This model po s t u l a t e d t h a t the p e r c e p t i o n of speech i n v o l v e s the i n t e r n a l s y n t h e s i s of pa t t e r n s according to c e r t a i n r u l e s and a matching of these i n t e r n a l l y generated patterns against the patterns under a n a l y s i s . , JT ,, (Stevens and H a l l e , 1967, p. 88) -24-The model as o u t l i n e d i s presented i n Figure 1. The process begins w i t h a p r e l i m i n a r y a n a l y s i s of an a c o u s t i c speech s i g n a l at the e a r . From the output of the p r e l i m i n a r y a n a l y s i s and s t o r e d c o n t e x t u a l i n f o r m a t i o n an hypothesis i s made ( i n a c o n t r o l component) concerning the a b s t r a c t r e p r e s e n t a t i o n of the u t t e r a n c e . The p h o n o l o g i c a l r u l e s used i n speech prod-u c t i o n generate an hypothesized sequence t o produce an a u d i t o r y p a t t e r n which i s compared w i t h the p a t t e r n under a n a l y s i s . The sequence proceeds to higher l e v e l s of a n a l y s i s i f there i s agreement. I f a d i f f e r e n c e i s noted between the two p a t t e r n s , the c o n t r o l component notes the e r r o r . A new t r i a l sequence i s then assembled by the c o n t r o l component. The p r o c e s s i n g and comparison i s repeated u n t i l the c o r r e c t output i s e s t a b -l i s h e d . A requirement of the model i s that a l i s t e n e r must know the p h o n o l o g i c a l r u l e s that the speaker uses i n speech p r o d u c t i o n . Repeated comparisons r e q u i r e a great amount of time f o r p r o c e s s i n g . The 'Motor Theory of Speech P e r c e p t i o n ' i s the t h i r d a c t i v e theory to be proposed. This model was developed by Likerman, Cooper, and t h e i r co-workers at Haskins L a b o r a t o r i e s i n an attempt to e x p l a i n data from t h e i r r e s e a r c h on speech p e r c e p t i o n . The authors hypothesize that the p e r c e p t i o n of speech sounds i s l i n k e d to the production of speech sounds. Three types of evidence i n d i c a t i n g that p e r c e p t i o n was more c l o s e l y l i n k e d to a r t i c u l a t o r y production than to a c o u s t i c i n v a r i a n c e support t h e i r hypothes i s . The f i r s t type i n d i c a t e s that s i m i l a r a c o u s t i c s t i m u l i could produce d i v e r s e p e r c e p t u a l responses. A number of s t u d i e s support such an a s s e r t -a t i o n . One such study i n v o l v e d the perception of s y n t h e t i c unvoiced stops produced on the P a t t e r n Playback. In the s y n t h e s i s of p l o s i v e - v o w e l DATA FROM PRECEDING ANALYSES TRIAL RULES CONTROL V. TRIAL —1 'TRIAL ERROR COMPARATOR STORE PRELIMINARY ANALYSIS ARTICULATORY MECHANISM AUDITORY MECHANISM Figure 1. Model for the speech-generating and speech-perception process. (The dashed line encloses components of a hypothetical analysis-by-synthesis scheme for speech perception. The shaded component indicates the capability of the model for effecting transformations between articulatory iristructioiis and auditory patterns.) (Figure and Figure Text from Stevens and Halle, 1967, p.100) r e -sequences, segments producing the consonant were i d e n t i c a l w h i l e f o l l o w i n g vowel formants were v a r i e d i n s m a l l s t e p s . S h i f t s i n p l o s i v e i d e n t i f i c a t i o n occurred w i t h minor changes i n vowel formant t r a n s i t i o n s , (Cooper et a l . 1952) The same type of evidence was obtained from s t u d i e s of v a r i a t i o n s of onset of f i r s t formants r e l a t i v e t o second and t h i r d formants, (Liberman, H a r r i s , Kinney and Lane, 1961), v a r i a t i o n s of i n t e r v o c a l i c s i l e n t i n t e r v a l s , (Liberman, et a l . 1961) and d i s c r i m i n a t i o n of inter-phonemic and i n t r a -phonemic d i f f e r e n c e s (Liberman et a l . , 1957). Secondly, the Motor Theory has a l s o been supported by evidence showing th a t d i f f e r e n t a c o u s t i c s t i m u l i produce s i m i l a r p e r c e p t u a l responses. A wide v a r i e t y of s t i m u l i presented to s u b j e c t s were a l l p e r c e i v e d as /k/, (Cooper et a l . , 1952). In a d d i t i o n , the pe r c e p t i o n of i n i t i a l /g/ remained i n v a r i a n t between CV p a i r s , having l a r g e d i f f e r e n c e s i n formant t r a n s i t i o n s , ( D e l a t t r e et a l , 1955). A study by Peterson and Barney (1952) revealed f u r t h e r evidence of t h i s type w i t h response t o vowels. Although f o r a s i n g l e vowel, d i f f e r e n t speakers produced a v a r i e t y of vowel-formant f r e q u e n c i e s , vowel i d e n t i f i c a t i o n remained c o n s t a n t . The t h i r d type of evidence supporting the Motor Theory i s data which s p e c i f i e s that changing p e r c e p t i o n c o r r e l a t e s more w i t h p a t t e r n s of a r t i c u l -a t i o n than a c o u s t i c p a t t e r n s . Evidence of t h i s type has been demonstrated i n a study of the pe r c e p t i o n of loudness. Ladefoged and McKinney (1963) found that i n judging loudness subj e c t ' s responses were more r e l a t e d t o muscular e f f o r t of the speaker than the absolute i n t e n s i t y l e v e l of speech. From the evidence o b t a i n e d , Liberman et a l . conclude t h a t p e r c e p t i o n of speech i s more h i g h l y c o r r e l a t e d w i t h a r t i c u l a t o r y parameters r a t h e r * A motor theory of vowel pe r c e p t i o n was suggested by Joos (1948). He proposed t h a t a l i s t e n e r may normalize the vowel system of each speaker i n order to map i t onto a f i x e d set of vowel i d e n t i f i c a t i o n s . -27-than w i t h a c o u s t i c parameters. The authors r a t i o n a l i z e that a p e r c e i v e r of speech i s a l s o a speaker and that i t would be unreasonable t o assume t h a t there are two e n t i r e l y d i f f e r e n t processes f o r the encoding and decoding of language. One process, l i n k i n g both sensory and motor components, i s thought to be more probable. The authors have s t a t e d i n support of t h e i r theory: ...the l i s t e n e r uses the inconstant sound as a b a s i s f o r f i n d i n g h i s way back to the a r t i c u l a t o r y gestures t h a t produced i t and thence, as i t were, to the speaker's i n t e n t . ., , , , Q C - Acn\ ' * (Liberman et a l . , 1967, p. 453) One basic assumption of the motor theory i s t h a t subphonemic f e a t u r e s e x i s t f o r production and perception of speech. In reference to production these features are assumed to e x i s t i n the C e n t r a l Nervous System as a s e t of n e u r a l i n s t r u c t i o n s . Each unique i n s t r u c t i o n i s sent t o a separate p a r t of the motor mechanism which i t moves independently. Thus, a correspondence between neural s i g n a l s and the dimensions of the subphonemic s t r u c t u r e i s i m p l i e d . The c o n t r a c t i o n of the muscles of the v o c a l t r a c t predetermine the r e s u l t i n g shape of the o r a l c a v i t y . The t o t a l neural a c t i v i t y f o r each phoneme i s thought t o be composed of the neural s p e c i f i c a t i o n s f o r each f e a t u r e . The components of neural a c t i v i t y f o r each phoneme are r e f e r r e d t o as motor commands and are assumed to be i n v a r i a n t . Every new set of v o c a l t r a c t movements was assumed to s t a r t from whatever c o n f i g u r a t i o n e x i s t e d and to t y p i c a l l y occur before the previous set has ended. This i s f e l t t o produce the ef-fect t h a t the shape of the t r a c t at any time represented the merged e f f e c t s of past and present i n s t r u c t i o n s . The p e r c e p t u a l mechanism i s hypothesized as a r e v e r s a l of the production mechanism which works i n c o n j u n c t i o n w i t h i t . The phonemes are broken down i n t o t h e i r c o n s t i t u e n t f e a t u r e s . These features are p o s s i b l y processed i n d i v i d u a l l y i n a p a r a l l e l arrangement. Thus, the processing - 2 8 -of consecutive phonemes could overlap i n time. This i s observed w i t h the second formant t r a n s i t i o n s which simultaneously c a r r y i n f o r m a t i o n about both the preceding consonant and the f o l l o w i n g vowel. The proposed percept u a l mechanism uses t h i s process to recover a r t i c u l a t o r y events from t h e i r encoded t r a c e s i n the sound stream. The motor theory has not been immune to c r i t i c i s m or opposing evidence i n s p i t e of i t s experimental evidence and r a t i o n a l b a s i s . Lane (1965) c r i t i c i z e d the experimental methodology used by Liberman and h i s c o l l e a g u e s ; suggesting there had been p r e - s e l e c t i o n of su b j e c t s f o r the perceptual experiments. To v e r i f y the exi s t e n c e of a s p e c i a l p e r c e p t u a l mechanism f o r speech Lane compared the perceptual f u n c t i o n s of consonants w i t h those f o r vowels and f o r s t i m u l i i n c l u d i n g sound, c o l o u r s and c i r c l e s e c t i o n s . Comparable g r a d i e n t s f o r i d e n t i f i c a t i o n and d i s c r i m i n a t i o n were obtained f o r a l l three p e r c e p t u a l f u n c t i o n s . He concluded, t h e r e f o r e , t h a t : ...the p o s t u l a t i o n of a s p e c i a l perceptual mechanism f o r the d i s c r i m i n a t i o n of speech i s unwarranted. (Lane, 1965, p. 307) Liberman and h i s co-workers (1970) responded to the c r i t i c i s m presented by repeating Lane's experiments i n v o l v i n g non-speech s t i m u l i . T h e i r s t u d i e s revealed a s u b s t a n t i a l d i f f e r e n c e between the d i s c r i m i n a t i o n of speech and non-speech s t i m u l i . This appeared to i n d i c a t e that Lane's c r i t i c i s m s were unfounded. Denes, (1967) i n an attempt to o b t a i n experimental v e r i f i c a t i o n of the motor the o r y , i n v e s t i g a t e d the i n f l u e n c e of production on p e r c e p t i o n of speech processed through a vocoder. S t i m u l i were presented to two groups, one which could use the microphone to r e l a t e a r t i c u l a t o r y movements t o processed sound. Learning curves were d e r i v e d f o r d i s c r i m i n a t i o n t e s t e d on f i f t y randomized words. No d i f f e r e n c e s were observed between the two -29-groups i n d i c a t i n g t h a t there was no c o r r e l a t i o n of production and p e r c e p t i o n . Denes suggested that there may be more than one method of speech p e r c e p t i o n i n v o l v e d i n the t o t a l p rocess. In c o n c l u s i o n , he remarked: The i n c o n c l u s i v e r e s u l t s may t h e r e f o r e j u s t i n d i c a t e the fundamental v a r i e t y of ways i n which speech pe r c e p t i o n can be achieved r a t h e r than that a r t i c u l a t o r y experiences do not p l a y an important part i n the p r o c e s s . (Denes, 1967, p. 314) Denes d i d not, t h e r e f o r e , b e l i e v e the motor theory had been c o n c l u s i v e l y r e f u t e d . Morton and Broadbent (1967) have been c r i t i c a l of a c t i v e models, three of which have been discussed above. They have s t a t e d that i n order to compare two s i g n a l s meaningfully i t i s e s s e n t i a l that the generated and incoming s i g n a l s or neural patterns be at the same l e v e l . A ccording to Morton and Broadbent: In the Halle-Stevens model the comparison i s at the neuroacoustic l e v e l . A generator produces a t r i a l phoneme sequence which i s converted i n t o a comparison spectrum presumably v i a the correspond-in g motor commands. In the Haskins model, on the other hand, i t would appear that the comparison takes place on the n e u r o a r t i c u l a t o r y l e v e l . Presumably then, the input spectrum would have to be converted from i t s neuroequivalent t o a n e u r o a r t i c u l a t o r y form f o r the comparison. • (Morton and Broadbent, 1967, p. 103) No d i s c r e t e d i f f e r e n t i a t i o n of n e u r a l l e v e l s such as proposed by Morton and Broadbent has been observed i n speech p e r c e p t i o n r e s e a r c h . The a u t h o r s ' argument i s weakened by the vagueness of the terminology which they p r e s e n t . Another argument which i s presented i s t h a t i f decoding has proceeded to a p o i n t of comparison i t could e q u a l l y w e l l proceed f u r t h e r without the need of any a r t i c u l a t o r y r e f e r e n c e . Morton and Broadbent f e l t that evidence f o r a c t i v e models was not r e a l l y i n c o n s i s t e n t w i t h a passive e x p l a n a t i o n , and concluded that w i t h respect to a c t i v e t h e o r i e s , the homunculus i s being given more work than i s s t r i c t l y necessary. No -30-adequate e x p l a n a t i o n or experimental v e r i f i c a t i o n i s gi v e n to support t h i s statement. Even though no a c t i v e theory which gives a complete explanation of speech p e r c e p t i o n has yet been proposed, the a c t i v e p r i n c i p l e ( i . e . , of ge n e r a t i n g an i n t e r n a l s i g n a l f or comparison w i t h an incoming s i g n a l ) has d e f i n i t e c r e d i b i l i t y . In a d d i t i o n to a c t i v e models, s o - c a l l e d passive models have a l s o been proposed to e x p l a i n the process of speech p e r c e p t i o n . The f i r s t p a s s i v e model, that of " D i s t i n c t i v e Features" was presented by Jakobson, Fant and H a l l e (1963). The authors recognized that two words, f o r example " b i l l " and " p i l l " may have d i f f e r e n t meaning d e s p i t e minimal s t r u c t u r a l d i f f e r e n c e s . A d i s t i n c t i o n between sounds i s termed 'minimal' i f i t cannot be r e s o l v e d i n t o f u r t h e r d i s t i n c t i o n s capable of d i f f e r e n t i a t i n g words. Minimal d i s t i n c t i o n s between sounds confront the l i s t e n e r w i t h b i n a r y d e c i s i o n s . The choice between the two opposites may be termed d i s t i n c t i v e f e a t u r e s . The d i s t i n c t i v e features are the u l t i m a t e d i s t i n c t i v e e n t i t i e s of language s i n c e no one of them can be broken down i n t o s m a l l e r l i n g u i s t i c u n i t s . The d i s t i n c t i v e features combined i n t o one simultaneous concurrent bundle form a phoneme. (Jakobson, Fant and H a l l e , 1963, p. 3) D i s t i n c t i v e features i n an utterance are both superimposed and s e q u e n t i a l . Decisions made on superimposed features d i s t i n g u i s h one word from a l l p o s s i b l e words o c c u r r i n g w i t h a s i m i l a r s t r u c t u r e . For example, d e c i s i o n s on features f o r the i n i t i a l phoneme would d i s t i n g u i s h " b i l l " • from " p i l l " , " f i l l " , " m i l l " , " t i l l " , " d i l l " , " s i l l " , and " k i l l " . D e c i s i o n s made on consecutive f e a t u r e s , segment one phoneme from another adjacent phoneme i n the sequence. For example, d e c i s i o n s on the features [ s t r i d e n t ] and [continuant] segment Is/ from /tl i n " s t i l l " . D i s t i n c t i v e features f o r every l i n g u i s t i c environment are not crea t e d f o r every p o s s i b l e minimal d i f f e r e n c e between sounds. Features r e s u l t i n g i n -31-semantic di f f e r e n c e s are c l a s s i f i e d as d i s t i n c t i v e for that s p e c i f i c language. In English, for example, velar /k/ as in " c o o l " and p a l a t a l /k/ as i n "k e e l " can be interchanged; i n Romanian, however, t h i s d i f f e r e n c e d i s t i n g u i s h e s words. The combination of d i s t i n c t i v e features forming each phoneme i s not random. In English, for example, the feature [nasal ] implies the feature [ v o i c e d ] . Features may in d i c a t e the a r t i c u l a t o r to be moved and the target p o s i t i o n for the movement; for example, the feature [high] designates the tongue p o s i t i o n i n the o r a l c a v i t y . Features may al s o i n d i c a t e gesture to be made by the a r t i c u l a t o r ; f o r example, the feature p a i r [ i n t e r r u p t e d -continuant] d i f f e r e n t i a t e abrupt onset as i n stops from gradual onset as i n f r i c a t i v e s . For some features there i s a one-to-one r e l a t i o n with an acoustic a t t r i b u t e , whereas for others the acoustic a t t r i b u t e s depend on the context. In decoding an utterance i n t o the underlying feature representations, therefore, a l i s t e n e r (or device that attempts to simulate aspects of the performance of a l i s t e n e r ) can extract a few features through d i r e c t d e t ection of appropriate acoustic a t t r i b u t e s , but the decoding of some features can proceed only a f t e r other features i n the same or i n adjacent segments have been i d e n t i f i e d , since the acoustic a t t r i b u t e s are context-dependent. (Halle and Stevens, 1972, p. 194) One of the d i f f i c u l t i e s associated with acoustic analysis of features i s that acoustic cues are often absent or d i s t o r t e d i n normally-spoken material. D i r e c t analysis of properties of the a r t i c u l a t o r s and auditory mechanism and also information from l i n g u i s t i c studies provide evidence f o r the inventory of features.- Analysis of phonetic and morphological r e g u l a r i t i e s of languages suggests that: ...the s t r u c t u r a l organization of a language i s best captured by means of the same feature set as that derived from observation of man's c a p a b i l i t y f o r the production and auditory processing of speech sounds. (Halle and Stevens, 1972, p. 195) -32-Since the theory was proposed the i n v e n t o r y of f e a t u r e s has undergone refinement and r e s p e c i f i c a t i o n . M i l l e r and N i c e l y (1955) s p e c i f i e d E n g l i s h d i s t i n c t i o n by the f e a t u r e s : v o i c i n g , n a s a l i t y , a f f r i c a t i o n , d u r a t i o n and p l a c e . The f e a t u r e system o u t l i n e d by Jakobson, Fant and H a l l e i n 1963 was based on a r t i c u l a t o r y and o f t e n a c o u s t i c d i f f e r e n c e s and i n c l u d e d the f e a t u r e s : V o c a l i c v s , Non-vocalic Consonantal v s . Non-consonantal In t e r r u p t e d v s . Continuant Checked v s . Unchecked S t r i d e n t v s . Mellow Compact v s . D i f f u s e Grave v s . Acute F l a t v s . P l a i n Sharp v s . P l a i n Tense v s . Lax Nasal v s . O r a l The system o u t l i n e d by Wickelgren (1965) i n c l u d e d the f e a t u r e s : v o i c i n g , n a s a l i t y , openness (three dimensions) and place ( f i v e dimensions). An extensive c l a s s i f i c a t i o n of f e a t u r e s was o u t l i n e d by Chomsky and H a l l e i n 1968. Features were designated according to four c a t e g o r i e s : i ) major c l a s s f e a t u r e s , i i ) c a v i t y f e a t u r e s , i i i ) manner of a r t i c u l a t i o n  f e a t u r e s , and i v ) source f e a t u r e s . i ) The major c l a s s features focus on d i f f e r e n t aspects of the open phase (when a i r flows out f r e e l y through the v o c a l t r a c t ) and the closed phase (when a i r flow i s e i t h e r impeded or stopped). The features i n t h i s category i n c l u d e : sonorant, v o c a l i c and c o n s o n a n t a l . -33-i i ) C a v i t y features i n d i c a t e p o s i t i o n s or target p o s i t i o n s of the a r t i c u l -a t o r s . The features i n c l u d e : c o r o n a l , a n t e r i o r , tongue features ( h i g h , low, back), round, d i s t r i b u t e d , covered, g l o t t a l c o n s t r i c t i o n s and secondary apertures ( n a s a l and l a t e r a l ) . i i i ) Manner of a r t i c u l a t i o n f eatures i n d i c a t e the type of movement or gesture made by the a r t i c u l a t o r s . Features i n t h i s category i n c l u d e : c o n t i n u a n t , r e l e a s e features (primary r e l e a s e and secondary r e l e a s e ) , supplementary movements ( s u c t i o n , and pressure) and tense. i v ) Source features focus on d i r e c t e f f e c t s on the a i r stream. These fe a t u r e s i n c l u d e : heightened s u b g l o t t a l p r e s s u r e , v o i c e , and s t r i d e n t . Although fe a t u r e systems have been e x t e n s i v e l y d e s c r i b e d , no completely adequate and e m p i r i c a l l y j u s t i f i e d l i s t of d i s t i n c t i v e features has been o u t l i n e d s i n c e the theory was proposed. The adequacy of a f e a t u r e system can be assessed by i t s a b i l i t y t o p r e d i c t responses and e r r o r s of s u b j e c t s . Wickelgren (1966) analyzed e r r o r s i n r e c a l l to determine the p r e d i c t i v e a b i l i t y of the f e a t u r e systems of M i l l e r and N i c e l y , of Wickelgren and of an e a r l i e r system by Chomsky and H a l l e . None of these systems was found to be adequate i n p r e d i c t i n g e r r o r s of r e c a l l . W i ckelgren, t h e r e f o r e , concluded t h a t one type of code, i . e . the f e a t u r e system, c o u l d not perform the t o t a l i n t e r p r e t a t i o n of speech and language. Feature systems were analyzed mathematically by K l a t t (1968). He was attempting t o devise a system of optimal f e a t u r e s . His o b s e r v a t i o n was t h a t : ...no absolute statement, e.g. 'binary f e a t u r e s A,B,C, are the features d e s c r i b i n g the d a t a , ' i s p o s s i b l e . This i s t r u e because s l i g h t changes i n f e a t u r e d e f i n i t i o n s do not always produce a s i g n i f i c a n t change i n the s i m i l a r i t y m a t r i c e . Therefore the optimal features (mathematically derived) are not very meaningful except f o r t h e i r a b i l i t y t o serve as a standard against which to compare systems of l i n g u i s t i c a l l y motivated f e a t u r e s . ( K l a t t , 1968, p. 407) -34-K l a t t , t h e r e f o r e , concluded t h a t : We cannot say t o the l i n g u i s t that t h i s feature system, w i t h a p p r o p r i a t e m o d i f i c a t i o n s , i s the only p e r c e p t u a l l y v a l i d model of language processing i n v o l v i n g short-term memory. ( K l a t t , 1968, p. 407) Although the p r i n c i p l e of breaking the task of phonemic c a t e g o r i z a t i o n i n t o a number of b i n a r y d e c i s i o n s i s a t t r a c t i v e , the theory of d i s t i n c t i v e f e atures has not yet proposed p o s s i b l e mechanisms f o r t h i s p r o c e s s . According t o such a t h e o r y , mechanisms f o r making the b i n a r y d e c i s i o n s and f o r the c a t e g o r i z a t i o n of phonemes from the s p e c i f i e d features are necessary f o r speech p e r c e p t i o n . Wickelgren (1969a, 1969b) has proposed an a l t e r n a t i v e passive t h e o r y . Most t h e o r i e s propose t h a t a phoneme i s the primary u n i t f o r speech p e r c e p t i o n . Wickelgren suggested, however, that a u n i t more c l o s e l y r e l a t e d t o the allophone would be more a p p r o p r i a t e . Thus, w i t h the use of "Context-S e n s i t i v e Coding" a code f o r words c o n s i s t s of: ...an unordered set of symbols f o r every word, where each symbol r e s t r i c t s the choice of i t s l e f t and r i g h t neighbours s u f f i c i e n t l y t o determine them uniquely out, of the unordered set f o r any g i v e n word. In t h i s case, the unordered s e t , i n c o n j u n c t i o n w i t h the dependency r u l e s , c o n t a i n s a l l the i n f o r m a t i o n necessary t o r e c o n s t r u c t a unique o r d e r i n g of the symbols f o r each word. (Wickelgren, 1969b, p. 86) The f o l l o w i n g example i l l u s t r a t e s t h i s p r i n c i p l e . The unordered set of c o n t e x t - s e n s i t i v e symbols coded f o r the v/ord " s t o p " i s {I^s^l•> /gfp/* /\T> /, / P „ / ) . These symbols are c o n s i s t e n t w i t h one, and only one ordering t p D tf of the a s s o c i a t e d c o n t e x t - f r e e symbols / s t r i p / . Each c o n t e x t - s e n s i t i v e allophone i s assumed t o have an i n t e r n a l r e p r e s e n t a t i o n i n the b r a i n . A c o n j u n c t i o n of unique a c o u s t i c features over a maximum p e r i o d of t i m e , i . e . 200-300 m s e c , s p e c i f i c a l l y a c t i v a t e s each of these i n t e r n a l r e p r e s e n t a t i o n s . The a c o u s t i c input i s analyzed by a l l a l l o p h o n i c r e p r e s e n t a t i v e s , s i m u l t a n e o u s l y . A unique subset of a l l o p h o n i c - 3 5 -r e p r e s e n t a t i v e s w i l l have been a c t i v a t e d above a s p e c i f i e d t h r e s h o l d by the time the input word i s completed. The word having the c l o s e s t c o r r e l a t i o n to the set of determined allophones i s then s e l e c t e d . According t o Wickelgren, segmentation i s not problematic f o r t h i s mechanism, s i n c e , because of p a r a l l e l p r o c e s s i n g , i t i s unnecessary to d i v i d e the a c o u s t i c waveform i n t o segments. However, s i n c e d i s c r e t e u n i t s are proposed, segmentation must occur at some point d u r i n g the p e r c e p t u a l process. The advantageous c h a r a c t e r i s t i c of t h i s process i s t h a t the order of r e c o g n i t i o n of allophones i s l e s s c r i t i c a l . The major problem i n the p e r c e p t i o n of speech, i . e . v a r i a b i l i t y of a c o u s t i c cues f o r a phoneme across d i f f e r e n t l e f t and r i g h t c o n t e x t s , i s e l i m i n a t e d by the proposed process of context-s e n s i t i v e c o d i n g . A c t u a l l y c o n t e x t - s e n s i t i v e coding i s dependent on v a r i a b i l i t y of a c o u s t i c cues f o r phonemes i n d i f f e r e n t c o n t e x t s . I f there i s any i n v a r i a n c e i t i s more l i k e l y t o be i n the allophone than i n the phoneme. Wickelgren s t a t e s t h a t i f the number of perc e p t u a l u n i t s i s expanded to i n c l u d e consonant c l u s t e r s and vowels having d i f f e r e n t segmental s t r e s s , then the e f f e c t of remote phonetic context on a c o u s t i c features can be e x p l a i n e d . The simultaneous p e r c e p t i o n of segmentals and suprasegmentals, i e . i n t o n a t i o n , s t r e s s and d u r a t i o n , i s a l s o d i s c u s s e d by Wickelgren. He s t a t e s that i f cues f o r suprasegmental s t r e s s do not i n t e r a c t s t r o n g l y w i t h s t r e s s cues f o r allophones, then there would be no problem i n h a n d l i n g such cues i n a c o n t e x t - s e n s i t i v e t h e o r y . V a r i a b l e d u r a t i o n of segment le n g t h and r a t e of speech would be a s s o c i a t e d w i t h v a r i a b l e time windows f o r a l l a l l o p h o n i c r e p r e s e n t a t i v e s . The author suggests that a separate mechanism to recognize 'intonation patterns might be operating p a r a l l e l t o a l l o p h o n i c r e c o g n i t i o n . This second mechanism would r e q u i r e the time window of the e n t i r e utterance f o r i n t o n a t i o n r e c o g n i t i o n . - 3 6 -Wickelgren concludes that h i s model solves the problems of segmentation, temporal ordering and simultaneous p e r c e p t i o n of segmentals and suprasegmentals. However, the 'c o s t ' which must be paid by t h i s model i s the immense amount of processing needed f o r the a c o u s t i c input to be channeled through r e p r e s e n t a t i v e s f o r every a l l o p h o n i c c o n t e x t . The per c e p t i o n of speech i s such a complex phenomenon that i t i s no s u r p r i s e i f none of the t h e o r i e s proposed to date has been able to f u l l y e x p l a i n the mechanisms i n v o l v e d . CHAPTER 2 STATEMENT OF THE PROBLEM From the preceding discussion of speech perception, i t appears that no s p e c i f i c theory can be c o n c l u s i v e l y supported at t h i s time. If there i s a d i r e c t r e l a t i o n s h i p between perception and production, i t i s possible that perception of units may be influenced by the frequency of t h e i r production. Certain v o i c e l e s s f r i c a t i v e and pl o s i v e consonant combinations v.'hich are s p e c i f i e d c l u s t e r s , e.g. /sp/ as i n ' s p i l l ' , or r e s u l t from p l u r a l i t y r u l e s , e.g. /ps/ as i n 'taps', are frequently produced in E nglish. From the above p r i n c i p l e , i t could be i n f e r r e d that for English speakers, the perception of these combinations would be f a c i l i t a t e d by the frequency of t h e i r production, whereas the perception of infrequently produced f r i c a t i v e - p l o s i v e combinations such as /fp/ or //k/ would not be influenced i n such a manner. If a s p e c i a l perceptual mechanism i s postulated f o r the perception of consonant c l u s t e r s , then a d i f f e r e n c e i n response to permissible and impermissible c l u s t e r s should be evident. Bond (1971) observed that i n response to permissible E n g l i s h c l u s t e r s there was a greater p r o b a b i l i t y for absolute r e v e r s a l e r r o r s , i . e . consonants produced are c o r r e c t l y perceived but in reverse order. A s i g n i f i c a n t d i f f e r e n c e between the frequency of absolute reversals and other types of errors f o r permissible and impermissible c l u s t e r s would imply a d i f f e r e n c e i n the mechanism of perception f o r these two kinds of consonant combinations. The present i n v e s t i g a t i o n of speech perception sought answers to the following s p e c i f i c questions: When s t i m u l i of c e r t a i n p l o s i v e - f r i c a t i v e combinations are presented with noise: 1) Is a l i s t e n e r ' s response a l t e r e d by the occurrence of these consonant combinations as permissible English c l u s t e r s ? -37--38-2) To what extent i s the l i s t e n e r ' s response a f f e c t e d by the Place of of a r t i c u l a t i o n of each member of t h i s c l u s t e r ? 3) To what extent i s the l i s t e n e r ' s response a f f e c t e d by the Manner of a r t i c u l a t i o n of each member of t h i s c l u s t e r ? 4) Can the errors made by the l i s t e n e r s be predicted by a D i s t i n c t i v e Feature Theory and how are these predictions l i k e l y to be affec t e d by a p a r t i c u l a r type of noise masker? CHAPTER 3 METHOD 3.1 Experimental Design Experimental c l u s t e r s included a l l combinations of one of the three v o i c e l e s s f r i c a t i v e s : / f / , /s/, and ///, with one of the three v o i c e l e s s p l o s i v e s : /p/, / t / , and /k/. (See Appendix A for the 18 possible combinations.) The c l u s t e r s were presented: 1) in f i n a l p o s i t i o n preceded by the vowel / I / e.g. /Isp/; 2) i n i n i t i a l p o s i t i o n followed by the vowel III e.g. / s p l / ; 3) i n medial p o s i t i o n preceded by the vowel / I / and followed by /In/ e.g. /Ispln/ with equal stress on each s y l l a b l e . One set of 18 s t i m u l i was created for each of the three c l u s t e r p o s i t i o n s ; a t o t a l of 54 s t i m u l i were created. These s t i m u l i are given i n Appendix B. One hundred and eight items (18 s t i m u l i repeated s i x times) were randomized into l i s t s f o r each set. Three l i s t s were compiled for each of the three sets producing a t o t a l of nine l i s t s . These nine l i s t s are given i n Appendix C. A diagrammatic o u t l i n e of the production of the nine l i s t s i s presented i n Figure 2. Nine experimental tapes were made with items ordered according to the nine prepared l i s t s . Pink noise as a s i g n a l masker was added to the tapes at a si g n a l - t o - n o i s e l e v e l determined by a method to be discussed. Insight into perception i s obtained by analyses of perceptual confusions. Noise was added to the speech s i g n a l to produce confusions i n perception which could be analyzed. Each subject responded to three tapes, i . e . one for each p o s i t i o n for a t o t a l of 324 items. Ordering of the tapes was randomized across subjects and responses were written on prepared response sheets. An example of the response sheets used i s presented i n Appendix D. -39--40-Figure 2. An outline of production of the nine experiments tapes. 18 Clusters Final 1 Set=18 Stimuli Repeated 6X 1 List=108 Items A\. Randomized I n i t i a l 18 Stimuli 3X Repeated 6X 108 Items A. Randomized 3X Medial 18 Stimuli Total 54 Stimuli Repeated 6X 108 Items Randomized 3X List List List 1 2 3 Tape Tape Tape 1 2 3 List List List 4 5 6 Tape Tape Tape 4 5 6 List List List 7 8 9 Tape Tape Tape 7 8 9 -41-Data were analyzed i n t o confusion matrices which were tabulated i n terms of: 1) place of a r t i c u l a t i o n for p l o s i v e s , 2) place of a r t i c u l a t i o n for f r i c a t i v e s , 3) manner of a r t i c u l a t i o n , 4) d i s t i n c t i v e features: i ) [ c o r o n a l ] i i ) [ d i s t r i b u t e d l a n d i i i ) [ a n t e r i o r ] and 5) permissible versus impermissible confusions. The data were a l s o analyzed for types of errors and frequency of responses since i t was f e l t that t h i s information would add i n s i g h t i n t o the mechanism for the perception of consonant c l u s t e r s . 3.2 S e l e c t i o n of Subjects Ten males and eleven females, unpaid, chronological age range 19-28 years, (median - 22 years) served as subjects for both p i l o t and e x p e r i -mental sessions. Masked Bekesy audiograms, speech reception thresholds and speech d i s c r i m i n a t i o n scores were obtained for each subject; a l l subjects demonstrated hearing threshold l e v e l s w i t h i n normal l i m i t s f o r the tests performed. Two females and one male were subjects i n the p i l o t -study; nine females and nine males served as subjects in the f i n a l experiment. 3.3 P i l o t Study The purposes of the p i l o t study were: f i r s t , to determine a s i n g l e s i g n a l - t o - n o i s e r a t i o to be used in the experiment proper; second, to t e s t the f e a s i b i l i t y of using a u d i t o r i l y determined s i g n a l - t o - n o i s e r a t i o s as described below; and f i n a l l y , to test the adequacy of the prepared i n s t r u c t i o n s . 3.31 Determination of Noise Level White noise has equal i n t e n s i t y for a l l frequencies to 10,000 Ha, whereas the i n t e n s i t y l e v e l of pink noise, slopes downward 3 dB per octave from 20 Ka to 20.000 Ha. Pink noise was selected to mask the s i g n a l s i n the present study s i n c e , as demonstrated by Strevens, f r i c a t i v e s and -42-p l o s i v e s c o n t a i n energy i n the higher frequencies of the speech spectrum. The vowel p o r t i o n s of the s t i m u l i were masked more than the f r i c a t i v e s and p l o s i v e s by t h i s type of n o i s e . I t has been observed that t r a i n e d l i s t e n e r s could repeatedly o b t a i n a s i g n a l - t o - n o i s e r a t i o w i t h i n +2 dB by using only a u d i t o r y cues. A procedure was designed t o determine whether the present i n v e s t i g a t o r could o b t a i n s i g n a l - t o - n o i s e r a t i o s repeatedly w i t h i n +2 dB l i m i t s . The experimenter's r e p e a t a b i l i t y of s i g n a l - t o - n o i s e l e v e l s on measurements f o r ten t r i a l s was w i t h i n +2 dB. S t i m u l i using s i g n a l - t o - n o i s e r a t i o s obtained by employing a u d i t o r y c r i t e r i o n were presented i n the p i l o t study t o examine the f e a s i b i l i t y of using t h i s procedure to produce the experimental tapes. 3.32 P r e p a r a t i o n of P i l o t Tapes S t i m u l i presented i n the p i l o t study had been recorded by a t r a i n e d p h o n e t i c i a n using a Nagra IVD tape-recorder w i t h a s s o c i a t e d AKG-202 m i c r o -phone on Ampex 605 Low Noise Tape i n an IAC-1204 sound-proofed room. One ut t e r a n c e f o r each stimulus was s p l i c e d from a dubbing of the o r i g i n a l tape and inc o r p o r a t e d i n t o a t a p e - l o o p . Steady pink noise from a General Radio Co. Noise Generator, Type 1382 was then recorded on t r a c k 2 of the t a p e - l o o p s . Both t r a c k s of these tape-loops were played on a S c u l l y 2 80 tape-recorder and "added" (by p l a y i n g the two tr a c k tape monaurally) on a Revox A77 t a p e - r e c o r d e r . Record and playback l e v e l s f o r each channel of the Revox A77 were kept unchanged throughout the monitoring and r e c o r d i n g . For tape-loops used f o r the f i r s t p i l o t tape, the s i g n a l i n t e n s i t i e s were monitored by peak readings on a B r u e l and Kjaer Voltmeter #2409. For tape-loops used f o r the second tape, the s i g n a l i n t e n s i t i e s were monitored a u d i t o r i l y through Grason-Stadler headphones #TDH 39. The d e s i r e d s i g n a l - t o - n o i s e r a t i o was obtained by a d j u s t i n g the playback l e v e l of the two channels on the -43-S c u l l y tape-recorder. The adjustment on each channel necessary t o o b t a i n the s i g n a l - t o - n o i s e r a t i o was noted f o r each tape-loop. This procedure i s o u t l i n e d on F i g u r e 3. In p reparing p i l o t t apes, item numbers were v e r b a l l y recorded on Channel 1 of the Revox A77 t a p e - r e c o r d e r . Playback l e v e l on the S c u l l y 280 tape-recorder was adjusted to the pre-determined l e v e l f o r each s t i m u l u s . A f t e r the item number, each stimulus was recorded on the Revox. The 108 items f o r each tape were copied f o l l o w i n g t h i s procedure. Tapes f o r the p i l o t study contained only s t i m u l i w i t h the consonant c l u s t e r i n f i n a l p o s i t i o n . Two tapes of 108 items each were made f o r the p i l o t s t u d y . The f i r s t tape used three d i f f e r e n t s i g n a l - t o - n o i s e r a t i o s monitored d i r e c t l y by peak readings on the V o l t m e t e r . The three s i g n a l - t o - n o i s e r a t i o s used were: 1) 0 dB f o r /s/ and / f / s t i m u l i and -6 d3 f o r /// s t i m u l i ; 2) +2 dB f o r /s./ and If I s t i m u l i and -4 dB f o r /// s t i m u l i ; 3) -2 dB f o r /s/ and / f / items and -8 dB f o r /// items. Because of the g r e a t e r i n t e n s i t y a s s o c i a t e d w i t h Iff a d i f f e r e n t s i g n a l - t o - n o i s e r a t i o was used f o r /s/ and /// and f o r If I and /// items. S i g n a l - t o - n o i s e r a t i o was adjusted a u d i t o r i l y f o r Tape 2 at a l e v e l at which a p l o s i v e i n a c l u s t e r was a u d i b l e but at which i t c o u l d a l s o be confused. 3.33 P i l o t Experimental Procedure The three s u b j e c t s were g i v e n the two p i l o t tapes i n d i f f e r e n t orders: two l i s t e n e d to tape 2 f i r s t and one l i s t e n e d to tape 1 f i r s t . S ubjects were given a r e s t between the two tapes. V e r b a l i n s t r u c t i o n s g i v e n to the s u b j e c t s by the experimenter were as f o l l o w s : The f o l l o w i n g sounds w i l l i n c l u d e the vowel ' I ' as i n 'DID' together w i t h one of ' f , ' s ' , or 'sh' and a l s o one of 'p', ' t ' , or 'k'. Write the sound you hear i n the c o r r e c t response space. Scully 280 Tape-recorder Revox A77 Tape-recorder Playback level "Adding" of adjusted. speech and noise.*" Channel 1 Channel 1, Speech Signal Speech Signal Channel 2 Channel 2 Pink Noise Pink Noise 7 Bruel and Kjaer Voltmeter #2409 Monitoring of Intensity Level for specified Signal-to-Noise Ratio /H \—Grason-Stadler headphones #TDH39 V J \__Jfov monitoring signals according to auditory criterion. Figure 3. Block diagram of equipment used in monitoring stimuli to obtain the desired Signal-to-Noise levels used in the Pilot study. * Adjustment necessary to obtain desired Signal-to-Noise levels were noted. Record and Playback adjustments were unchanged. -45-When the three subjects had completed the task t h e i r answers were tabulated and the results analyzed. 3.34 Results of P i l o t Study The t o t a l number of errors ( a l l responses which were not i d e n t i c a l to the presented s t i m u l i ) for items presented at each s i g n a l - t o - n o i s e l e v e l are shown in Table 1. These r e s u l t s indicated that when a signa l - t o - n o i s e l e v e l determined by the auditory c r i t e r i o n was used, the number of errors produced was greater than when a 0 dB or a +2 dB si g n a l - t o - n o i s e r a t i o was used. The number of errors for items s p e c i f i e d according to the auditory c r i t e r i o n was more than d e s i r e d ( f i f t y percent c o r r e c t ) . Therefore, i t was decided that, using auditory c r i t e r i o n to determine s i g n a l - t o - n o i s e r a t i o was not an -adequate method. The r e s u l t s indicated that errors for place of a r t i c u l a t i o n f o r f r i c a t i v e s were few in comparison with errors f o r place of a r t i c u l a t i o n for p l o s i v e s . Thus i t was decided that the noise l e v e l used for masking should be determined i n r e l a t i o n to the l e v e l causing place errors for p l o s i v e s rather than the l e v e l causing place errors f o r f r i c a t i v e s . Therefore, the increased noise l e v e l used f o r /// items was unnecessary since such a noise l e v e l increased p l o s i v e errors rather than f r i c a t i v e e r r o r s . The t o t a l number of errors for each stimulus was determined. Items with s p e c i f i e d s i g n a l - t o - n o i s e r a t i o s showed a s i m i l a r trend with items whose si g n a l - t o - n o i s e r a t i o s were determined by the auditory c r i t e r i o n . The average number of errors per stimulus was greater with impermissible c l u s t e r s than with permissible c l u s t e r s . This trend w i l l be discussed when data from the experimental study are reported i n Section 4.3. The large percentage of errors obtained (74 percent for the f i r s t t e s t and 69 percent for the second t e s t ) was f e l t to be due both to the noise -46-Table.I. Total Errors for each Signal-to-Moise Ratio in Pilot Study. S/N Ratio (dB) 0 +2 -2 Auditory Criterion Total Errors 69/108 64/108 87/108 74/108 -47-lev e l s used and to the inadequacy of the i n s t r u c t i o n s given each subject. In the experiment proper, i n s t r u c t i o n s were modified to minimize misunder-standing. 3.4 Preparation of Experimental Tapes The preparation of the experimental tapes involved the recording of s t i m u l i to be used, preparing tape-loops of selected utterances, and recording c a l i b r a t i o n s i g n a l s , item numbers and s t i m u l i on the experimental tapes. The method of recording the utterances, s e l e c t i n g those to be used as s t i m u l i and incorporating them into tape loops was the same for p i l o t and experimental tapes. 3.41 C a l i b r a t i o n signals The experimental tapes were made on a Revox Model #A77 tape-recorder. A c a l i b r a t i o n tone and c a l i b r a t i o n noise were recorded at the beginning of each tape. The 1000 Ha c a l i b r a t i o n tone was generated by a Bruel and Kjaer Eeat Frequency O s c i l l a t o r #1022 and taped on track 1; the c a l i b r a t i o n pink noise was generated by the noise generator and taped on track 2. Both si g n a l s were recorded at 0 V.U. 3.4 2 Recording of numbers A f t e r the c a l i b r a t i o n signals had been recorded on each tape the 108 item numbers were recorded using an A l t e c Microphone #681 ALO. Item numbers were separated by approximately 10 seconds with s l i g h t deviations due to v a r i a t i o n i n feet meter readings occurring with r o t a t i o n s of the spools. 3.43 Determination of Signal-to-Noise Ratio In view of the r e s u l t s of the P i l o t Study, i t was f e l t that a s i g n a l -to-noise r a t i o of +3 dB would be appropriate for the purposes of t h i s study. This d e c i s i o n was based on the observation that a large number of errors occurred i n the P i l o t Study when a +2 d3 si g n a l - t o - n o i s e r a t i o was used. -48-3.44 Equipment Set-up. The organization of equipment used for taping of s t i m u l i was the same, as that used in the p i l o t study. In obtaining d e s i r e d s i g n a l - t o - n o i s e l e v e l s , adjustments were :r.ade only on the playback l e v e l of the S c u l l y 280 tape-recorder; record and playback le v e l s on the Revox A77 tape-recorder were maintained at a constant l e v e l . Intensity of the speech and noise s i g n a l s was ind i c a t e d by d e f l e c t i o n of peak reading on a Bruel and Kjaer Voltmeter #2409. 3.45 Recording S t i m u l i on Experimental Tapes. In the p i l o t study, each tape-loop had a s p e c i f i e d playback adjustment for each channel of the S c u l l y 280 tape-recorder. It was observed that items which had been recorded by the Revox A77 tape-recorder onto the p i l o t tapes from the same tape-loop using i d e n t i c a l playback l e v e l s on the S c u l l y were not consistent i n i n t e n s i t y output l e v e l s . Due to the observed imprecision between items, t h i s method for obtaining the desired s i g n a l - t o - n o i s e r a t i o was modified for the experimental tapes. The a l t e r n a t i v e procedure used was to vary the S c u l l y playback l e v e l s f or each item u n t i l the recorded out-put i n t e n s i t y l e v e l s f or each stimulus and masking noise were monitored (by peak reading) at a +3 dB s i g n a l - t o - n o i s e r a t i o . The s t i m u l i were ordered on the experimental tapes according to the nine randomized l i s t s mentioned i n Section 3.1. Using the appropriate tape-loop and following the procedure just described each item was recorded by the Revox A77 tape-recorder onto the experimental tapes immediately succeed-ing the spoken item number. 3.46 Evaluating Completed Tape. When the 108 items on each tape had been recorded the experimenter l i s t e n e d to Channel 1 of the tape to v e r i f y the order of stimulus presentation and then to both channels "added" to evaluate the t o t a l e f f e c t . -49-This procedure was performed on a l l of the nine tapes as each was completed. 3.5 Experimental Frocedure The equipment to be used i n the t e s t i n g s e s s i o n s , that i s a Revox A77 tape- r e c o r d e r , a set of TDH-39 Grason S t a d l e r head-phones w i t h MX-41/AR cushions, and nine experimental tapes were t r a n s p o r t e d to two o f f i c e s which were a v a i l a b l e f o r use. Subjects were, given l i s t e n i n g t e s t s i n one o f f i c e and r e s t s e s s i o n s i n the o t h e r . L i s t e n i n g t e s t s were scheduled f o r evenings or weekends when no other persons were i n the b u i l d i n g and the only extraneous n o i s e was produced by out s i d e t r a f f i c . Nine t e s t i n g sessions were scheduled, two subjects t o each s e s s i o n . This arrangement enabled each subject t o a l t e r n a t e between l i s t e n i n g task and r e s t s e s s i o n . Experimental tapes were randomized across s u b j e c t s and s e s s i o n s . The f o l l o w i n g i n s t r u c t i o n s were read t o and g i v e n each s u b j e c t t o read, before t h e i r f i r s t l i s t e n i n g t e s t : You w i l l now be p a r t i c i p a t i n g i n a l i s t e n i n g t a s k . You w i l l be l i s t e n i n g to three tapes of 108 items each. Each tape takes approximately f i f t e e n minutes t o run through. You w i l l be given a r e s t between each tape w h i l e your partner p a r t i c i p a t e s i n the t a s k . You w i l l be hearing speech utterances i n n o i s e . Sometimes you w i l l understand the utterance and sometimes you may f i n d i t d i f f i c u l t to understand. Each utterance c o n s i s t s of the vowel ' I ' as i n DID and two consonants i n a c l u s t e r : one of F, S, or SH and one of P, T, or K. The f i r s t tape you w i l l hear w i l l have the c l u s t e r s : Every time you hear an item, w r i t e the utterance you hear i n the ap p r o p r i a t e response space on the answer sh e e t . At the beginning of the tape you may f i n d the time between the items i s not s u f f i c i e n t to make a d e c i s i o n about the consonants. You may stop the playback on the tape-recorder u n t i l you have made your d e c i s i o n . However, you KAY NOT go over an item once you have heard i t . When you become -SO-f a m i l i a r w i t h the task t h i s w i l l probably be unnecessary . Do you understand? Are you ready t o begin? The i n s t r u c t i o n s g i v e n i n the m i s s i n g s e c t i o n depended on which c l u s t e r p o s i t i o n , i . e . i n i t i a l , m e d i a l and f i n a l , the s u b j e c t was be ing presented w i t h . When necessary , the experimenter v e r b a l l y r e - i t e r a t e d the i n s t r u c t i o n s and wrote p o s s i b l e combinat ions of f r i c a t i v e and p l o s i v e , f o r example: ' S T / ' T F ' o r ' K S H . ' The s u b j e c t s u s u a l l y wrote the t h r e e f r i c a t i v e s and the t h r e e p l o s i v e s at the t o p of t h e i r response sheet f o r r e f e r e n c e . For the f i r s t l i s t e n i n g t e s t , the s u b j e c t s were i n s t r u c t e d t o s t o p the tape a f t e r ten i t e m s , so t h a t comprehension of the task c o u l d be v e r i f i e d . Examples of the i n s t r u c t i o n s f o r the second and t h i r d t e s t were : Second "The second tape you w i l l hear w i l l have the c l u s t e r s i n the MIDDLE of the u t t e r a n c e f o l l o w e d by ' I N G ' as i n 'DOING' f o r example: I3FING, ITFING or IK3HING." and T h i r d "The f i n a l tape you w i l l hear w i l l have the c l u s t e r s at the BEGINNING of the u t t e r a n c e , f o r example: S F I , F T I or K S H I . " The order of i n s t r u c t i o n s g i v e n c o r r e l a t e d w i t h the order of c l u s t e r p o s i t i o n s presented t o the s u b j e c t . To i n s u r e a l l s u b j e c t s were g i v e n the t e s t s a t the same i n t e n s i t y , p r i o r to each t e s t , the playback l e v e l of the t a p e - r e c o r d e r was a d j u s t e d u n t i l the output i n t e n s i t i e s of the c a l i b r a t i o n s i g n a l s recorded on the tape were moni tored at 0 VU. When t h i s adjustment had been made, the s u b j e c t s proceeded w i t h the t a s k . The d e s c r i b e d procedure was f o l l o w e d f o r each s u b j e c t u n t i l a l l of the n ine males and nine females had been t e s t e d . CHAPTER 4 RESULTS . . . t r u t h can be revealed by e r r o r s and order can r e s u l t from confus i o n s . ( M i l l e r , 1956, p. 360) The p r i n c i p l e espoused by M i l l e r i s r e l e v a n t to the present a n a l y s e s : c o r r e c t and i n c o r r e c t responses t o consonant combinations were analyzed to r e v e a l some ' t r u t h ' about pe r c e p t i o n of consonant c l u s t e r s ; responses were t a b u l a t e d i n confusion matrices to r e v e a l some 'order' i n perceptual e r r o r s . Each of the eighteen s u b j e c t s responded t o 108 items f o r each c l u s t e r p o s i t i o n ( i n i t i a l , medial and f i n a l ) . T o t a l number of responses f o r e i g h t -een s u b j e c t s f o r each p o s i t i o n was 1944. T o t a l number of responses f o r a l l s u b j e c t s i n a l l three p o s i t i o n s was 5832. These responses were t a b u l a t e d i n three (one f o r each c l u s t e r p o s i t i o n ) confusion m a t r i c e s . The con f u s i o n matrices were f u r t h e r t a b u l a t e d f o r responses combined acc o r d i n g t o 1) p e r m i s s i b i l i t y v s . i m p e r m i s s i b i l i t y as an E n g l i s h c l u s t e r , 2) manner of a r t i c u l a t i o n , 3) place of a r t i c u l a t i o n f o r p l o s i v e s and f r i c a t i v e s , and 4) c e r t a i n Jakobson, Fant and H a l l e - d i s t i n c t i v e f e a t u r e s . 4.1 Confusion Matrices f o r I n d i v i d u a l C l u s t e r s In the experiment some responses occurred which were not designated c h o i c e s , e.g. / t i / , / s i t / or / t p l / . The t o t a l of such responses was 94. A randomized response choice was made f o r these items i n order t h a t the number of responses i n d i c a t e d would be i d e n t i c a l to the number of s t i m u l i presented. A l l c o n f u s i o n matrices are presented i n a c o n s i s t e n t format; the s t i m u l i are given on the l e f t s i de ana responses at the t o p . Each c e l l of the matrix i n d i c a t e s the number of times each response ( i n d i c a t e d by the column) occurred f o r each stimulus ( i n d i c a t e d by the row). -51--52-The frequency of responses t o s i x p r e s e n t a t i o n of a c l u s t e r v a r i e d among s u b j e c t s ; f o r example, f o r s t i m u l u s /sk/ i n i n i t i a l p o s i t i o n , the frequency of /sk/ responses ranged from 0 to 6. Due to the range among subjects i t was f e l t more reasonable to analyze t o t a l responses f o r a l l s u b j e c t s rather than i n d i v i d u a l responses. Responses f o r a l l s u b j e c t s , i n each p o s i t i o n (1944 responses), were t a b u l a t e d i n confusion matrices i n d i c a t i n g the t o t a l responses to each consonant combination. C o r r e c t i d e n t i f i c a t i o n s of s t i m u l i are i n d i c a t e d on the diagonal of each m a t r i x . A t o t a l of 108 responses to each consonant combination were g i v e n , s i n c e each c l u s t e r was presented s i x times to eighteen s u b j e c t s . S e v e r a l rows t o t a l 114 or 102 responses s i n c e , i n s p i t e of c a r e f u l checking, some s t i m u l i occurred on c e r t a i n experimental tapes f i v e or seven times ( i n s t e a d of s i x ) and the tape was presented t o s i x l i s t e n e r s . These rows are i n d i c a t e d by an a s t e r i s k at the r i g h t s i d e , A computed t o t a l of s i x responses c o u l d have occurred i n each c e l l of the matrix by chance. Confusion Matrices f o r c l u s t e r s i n each p o s i t i o n are presented i n Table I I , I I I and IV. 4.2 Responses t o Each C l u s t e r The number of c o r r e c t responses f o r consonant combinations i n each and every p o s i t i o n was obtained from the diagonals of the confusion m a t r i c e s . These values are t a b u l a t e d i n Table V. Student - t values were c a l c u l a t e d on Mean C o r r e c t Response per c l u s t e r c a l c u l a t e d from the r e s u l t s presented i n Table V, the r e s u l t s of which are presented i n Table VI. An example of such a c a l c u l a t i o n i s presented i n Appendix E. The number of times each consonant combination was used- as a response to a l l s t i m u l i i n each t e s t i s sh^wn i n Table VII. C a l c u l a t i o n s of Student-t values showed no s i g n i f i c a n t d i f f e r e n c e between the mean number of responses per c l u s t e r to /s/ c l u s t e r s , / f / c l u s t e r s and /// c l u s t e r s . -53-4.3 P e r m i s s i b l e versus Impermissible C l u s t e r s i n E n g l i s h I f a consonant combination a c t u a l l y occurs w i t h i n E n g l i s h morphemes f o r a s p e c i f i c p o s i t i o n i t w i l l be termed a " p e r m i s s i b l e " c l u s t e r f o r th a t p o s i t i o n . I f a consonant combination does not occur w i t h i n E n g l i s h morphemes f o r a s p e c i f i c p o s i t i o n i t w i l l be termed " i m p e r m i s s i b l e " f o r that p o s i t i o n . The source used t o determine i f a consonant combination occurred w i t h i n E n g l i s h morphemes f o r a s p e c i f i c p o s i t i o n was Trnka (1968). Of the consonant combinations presented, those which are impermissible f o r a l l p o s i t i o n s a r e: / f p / , / f k / , / p f / , / t f / , / k f / , / / t / , //p/, and Ifk/. Of the consonant combinations presented, those which are p e r m i s s i b l e i n i n i t i a l p o s i t i o n i n c l u d e : /sp/ as i n ' s p i l l ' , / s t / , as i n ' s t i l l ' , /sk/ • as i n ' s k i l l ' and / t / / as i n 'church'. Besides the c l u s t e r s impermissible i n a l l p o s i t i o n s , the c l u s t e r s impermissible i n i n i t i a l p o s i t i o n i n c l u d e : /ps/, / t s / , /ks/, / f t / , /pf/, and Ikfl. Except f o r the c l u s t e r s i mpermissible i n a l l p o s i t i o n s , a l l of the other c l u s t e r s were p e r m i s s i b l e i n medial p o s i t i o n . These i n c l u d e : /sp/ as i n ' c l a s p i n g ' , / s t / as i n ' b l a s t i n g ' , /sk/ as i n 'masking', /ps/ as i n ' c a p s i z e ' , / t s / as i n ' s c h i z o p h r e n i a ' , /ks/ as i n 'waxed', / f t / as i n ' a f t e r ' , /pf/- as i n ' o p t i o n ' , /tf/ as i n ' i t c h i n g ' , and /kf/ as i n -'action'. Tne consonant combinations presented which are p e r m i s s i b l e c l u s t e r s i n f i n a l p o s i t i o n i n c l u d e : /sp/ as i n 'wasp', / s t / as i n 'guest', /sk/ as i n 'ask', /ps/ as i n 'taps', / t s / as i n ' r a t s ' , /ks/ as i n ' s i x ' , / f t / as i n ' l e f t ' and/t// as i n 'church'. Besides the consonant combinations which were impermissible i n a l l p o s i t i o n s , /pf/ and /kf/ were impermissible i n f i n a l p o s i t i o n . * Trnka c l a s s i f i e d /tf/ as a phoneme but i n t h i s study i t i s assumed to be a c l u s t e r made up of / t / followed by /f(. -54-For each c l u s t e r p o s i t i o n , a confusion matrix was constructed i n which s t i m u l i and responses were d i f f e r e n t i a t e d according to t h e i r p e r m i s s i b i l i t y f o r that p o s i t i o n . The number of responses which co u l d have occurred by chance (assuming s i x responses by chance per c e l l i n Tables I I , I I I , and IV), i s i n d i c a t e d i n the lower r i g h t corner of each block i n the m a t r i x . The c o n f u s i o n matrices f o r p e r m i s s i b i l i t y are presented i n Tables V I I I , IX and X. 4.31 Types of Responses When a consonant c l u s t e r was presented there were at l e a s t e i g h t types of i n t e r p r e t a t i o n s f o r the response which could have occurred: 1) The c l u s t e r could be perceived c o r r e c t l y a Correct Response, e.g. a response /sp/ f o r a stimulus /sp/. 2) A S i n g l e P l o s i v e E r r o r i s a response i n which the only p e r c e p t u a l e r r o r i s a confusion f o r place of a r t i c u l a t i o n of the p l o s i v e , e.g. a response / s t / f o r a stimulus /sp/. 3) A S i n g l e F r i c a t i v e E r r o r i s a response i n which the only perceptual e r r o r i s a confusion f o r place of a r t i c u l a t i o n of the f r i c a t i v e , e.g. a response /fp/ f o r a stimulus /sp/. 4) A Double S u b s t i t u t i o n i s a response i n which there i s confusion f o r place of a r t i c u l a t i o n of the p l o s i v e and the f r i c a t i v e but the order of manner of a r t i c u l a t i o n i s c o r r e c t , e.g. a response / f t / f o r a stimulus /sp/. 5) An Absolute R e v e r s a l E r r o r 'is a response i n which the c o r r e c t consonants are per c e i v e d i n reverse o r d e r , e.g. a response /ps/ f o r a stimulus /sp/. 6) A Reversal w i t h a P l o s i v e E r r o r i s a response i n which there i s a confusion f o r place of a r t i c u l a t i o n of the p l o s i v e and a r e v e r s a l i n order of manner of a r t i c u l a t i o n , e.g. a response / t s / f o r a stimulus /sp/. 7) A Re v e r s a l w i t h a F r i c a t i v e E r r o r i s a response i n which there i s a -55-confusion f o r place of a r t i c u l a t i o n of the f r i c a t i v e and a r e v e r s a l of the order of manner of a r t i c u l a t i o n , e.g. a response /pf/ f o r a stimulus /sp/. 8) A Rev e r s a l w i t h a P l o s i v e and a F r i c a t i v e E r r o r i s a response i n which there i s a confusion f o r place of a r t i c u l a t i o n of the p l o s i v e and the f r i c a t i v e and a r e v e r s a l of the order of manner of a r t i c u l a t i o n , e.g. a response / t f / f o r a st i m u l u s /sp/. The frequency of each of the s p e c i f i e d types of e r r o r s are ta b u l a t e d f o r the group of p e r m i s s i b l e and impermissible c l u s t e r s f o r each p o s i t i o n . The r e s u l t s f o r the three p o s i t i o n s are presented i n Tables XI, XII and XIII. To the lower r i g h t of the observed number of responses i s i n d i c a t e d the number of responses which could have occurred i n that c e l l by chance. For each type of e r r o r , there i s a d i f f e r e n c e i n number of responses by chance between p e r m i s s i b l e and impermissible c l u s t e r s due to the d i f f e r e n c e i n number of c l u s t e r s i n each group, i . e . f o r i n i t i a l p o s i t i o n there are four p e r m i s s i b l e c l u s t e r s and fourteen impermissible c l u s t e r s , f o r medial p o s i t i o n there are ten p e r m i s s i b l e and e i g h t impermissible c l u s t e r s , and f o r f i n a l p o s i t i o n there are e i g h t p e r m i s s i b l e and ten impermissible c l u s t e r s . The d i f f e r e n c e s i n chance l e v e l s among the types of e r r o r s are due to the v a r i a b l e number of responses which can produce the d i f f e r e n t t y p e s , f o r example, i n response to /sp/ a s i n g l e p l o s i v e e r r o r could be made through the responses / s t / or /sk/, w h i l e an absolute r e v e r s a l e r r o r c o u l d be made only through the response /ps/. Student-t values were c a l c u l a t e d between the mean number of responses per c l u s t e r f o r p e r m i s s i b l e and impermissible c l u s t e r s f o r each type of e r r o r . An example of such a c a l c u l a t i o n i s presented i n Appendix E. The 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 obtained are i n d i c a t e d . -56-4.4 Manner of A r t i c u l a t i o n Responses i n each c l u s t e r p o s i t i o n were analyzed according t o manner of a r t i c u l a t i o n , i . e . those c l u s t e r s w i t h a f r i c a t i v e f i r s t and those c l u s t e r s w i t h a p l o s i v e f i r s t . The group of c l u s t e r s which had the f r i c a t i v e f i r s t c o n s i s t e d of: /sp/, / s t / , /sk/, / f p / , / f t / , / f k / , Ifpl, /ft/, and //k/. The group of c l u s t e r s which had the p l o s i v e f i r s t c o n s i s t e d of /ps/, / t s / , /ks/, / p f / , / t f / , / k f / , /pj7, / t / / , and /k//. The number of responses which could occur by chance was i d e n t i c a l f o r each block of the c o n f u s i o n m a t r i x . The r e s u l t s of these anaylses are presented i n Tables XIV, XV ana XVI. 4.5 Place of A r t i c u l a t i o n f o r P l o s i v e s and F r i c a t i v e s R e s u l t s were analyzed f o r place of a r t i c u l a t i o n f o r both p l o s i v e s and f r i c a t i v e s . C l u s t e r s were grouped f o r analyses of place f o r p l o s i v e s by combining a l l the s t i m u l i which ended i n one s p e c i f i c p l o s i v e ; f o r example, the c l u s t e r s w i t h /sp/, /fp/ and / f p / were combined as Fr.-p; the c l u s t e r s /ps/, /pf/, and /pf/ were combined as p-Fr. The r e s u l t s of these analyses are presented i n Tables X V I I , X V I I I and XIX. Number of responses which cou l d have occurred by chance i s i d e n t i c a l i n each c e l l i n the m a t r i x . Analyses of p l a c e of a r t i c u l a t i o n f o r f r i c a t i v e s : c l u s t e r s were grouped by combining a l l the s t i m u l i which began w i t h one s p e c i f i c f r i c a t i v e or a l l the s t i m u l i which ended w i t h one s p e c i f i c f r i c a t i v e ; f o r example, the s t i m u l i /sp/, / s t / , and /sk/ were combined as s - P l . , and the s t i m u l i /ps/, / t s / , and /ks/ were combined as P l . - s . The confusion matrices r e s u l t i n g from these analyses are presented i n Tables XX, XXI and X X I I . Responses at chance l e v e l are i d e n t i c a l f o r each c e l l of the m a t r i x . 4.6 Analyses According to C e r t a i n Jakobson, Fant and K a l l e D i s t i n c t i v e Features Responses were analyzed according to the f e a t u r e s i ) [ a n t e r i o r ] , -57-i i ) [ c o r o n a l ] and i i i ) [ d i s t r i b u t e d ] . i ) According to Chomsky and K a l l e : A n t e r i o r sounds are produced v/ith an o b s t r u c t i o n that i s l o c a t e d i n f r o n t of the p a l a t o - a l v e o l a r region of the mouth; non a n t e r i o r sounds are produced without such an o b s t r u c t i o n . (Chomsky and H a l l e , 1968, p. 304) The f e a t u r e [ a n t e r i o r ] d i f f e r e n t i a t e s the f r i c a t i v e s / f / and /s/ which are [+ a n t e r i o r ] from /// which i s [ - a n t e r i o r ] . Confusion matrices f o r c l u s t e r s combined according t o [ a n t e r i o r ] f o r f r i c a t i v e s are presented i n Tables X X I I I , XXIV and XXV. The p l o s i v e s /p/ and / t / are [ t a n t e r i o r ] w h i l e /k/ i s [ - a n t e r i o r ] . Confusion matrices f o r c l u s t e r s combined according to [anterior] f o r p l o s i v e s are presented i n Tables XXVI, XXVII and X X V I I I . When the c l u s t e r s are d i f f e r e n t i a t e d according to the f e a t u r e [ a n t e r i o r ] (value f o r f r i c a t i v e w r i t t e n f i r s t ) c l u s t e r s which are [++anterior ] i n c l u d e : /sp/, / s t / . / f p / , / f t / , /ps/, / t s / , / p f / , / t f / ; c l u s t e r s which are [+- a n t e r i o r ] i n c l u d e : / s k / , / f k / , /ks/, and / k f / ; c l u s t e r s which are [-+ a n t e r i o r ] i n c l u d e : / / p / , I f t / , /pf/, and / t f / ; c l u s t e r s which are [-- a n t e r i o r ] i n c l u d e : //k/, and /kf I. Confusion matrices f o r c l u s t e r s combined according to [ a n t e r i o r ] f o r f r i c a t i v e s and p l o s i v e s are presented i n Tables XXIX, XXX, and XXXI. i i ) According to Chomsky and H a l l e : Coronal sounds are produced w i t h the blade of the tongue r a i s e d from i t s n e u t r a l p o s i t i o n ; noncoronal sounds are produced w i t h the blade of the tongue i n the n e u t r a l p o s i t i o n . (Chomsky and H a l l e , 1968, p. 304) The feature [ c o r o n a l ] d i f f e r e n t i a t e s the f r i c a t i v e s Is I and /// which are [•+ c o r o n a l ] from the f r i c a t i v e / f / which i s [- c o r o n a l ] . This feature d i f f e r e n t i a t e s the p l o s i v e / t / which i s [+ c o r o n a l ] from the p l o s i v e s /p/ and /k/ which are [- c o r o n a l ] . -58-P l o s i v e s were f i r s t analyzed according t o the f e a t u r e [ c o r o n a l ] . Consonant combinations i n c l u d i n g the p l o s i v e It/ were compared w i t h consonant combinations i n c l u d i n g /p/ and fk/. The r e s u l t s are shown i n the confusion matrices of Tables XXXII, XXXIII and XXXIV. Responses were then analyzed according t o the f e a t u r e [coronal"], f o r f r i c a t i v e s and p l o s i v e s . The f e a t u r e value f o r the f r i c a t i v e i s i n d i c a t e d f i r s t . C l u s t e r s which are [++ c o r o n a l ] i n c l u d e d : / s t / , /ft}, / t s / and /tf/. C l u s t e r s which are [+- c o r o n a l ] i n c l u d e d : /sp/, /sk/, I f p l , Ifk/, /ps/, /ks/, /pf/, and /k//. C l u s t e r s which are [-+ c o r o n a l ] i n c l u d e d : / f t / and / t f / . C l u s t e r s which are [-- c o r o n a l ] i n c l u d e d : / f p / , / f k / , / p f / , and / k f / . Results of analyses according to t h i s more thorough c a t e g o r i z a t i o n are presented i n Tables XXXV, XXXVI, and XXXVII. c i i i ) D i s t r i b u t e d sounds are produced w i t h a c o n s t r i c t i o n t h a t extends f o r a c o n s i d e r a b l e d i s t a n c e along the d i r e c t i o n of a i r flow; n o n d i s t r i b u t e d sounds are produced w i t h a c o n s t r i c t i o n that extends o n l y f o r a sho r t d i s t a n c e i n t h a t d i r e c t i o n . ,_, , , „ ,, , _ c o (Chomsky and H a l l e , 1968, p. 312) The f e a t u r e , [ d i s t r i b u t e d ] d i f f e r e n t i a t e s the f r i c a t i v e s /s/ and /// which are [+ d i s t r i b u t e d ] from the f r i c a t i v e If I which i s [- d i s t r i b u t e d ] . The same d i f f e r e n t i a t i o n of f r i c a t i v e s occurs w i t h the f e a t u r e [ c o r o n a l ] . V.'hen the f r i c a t i v e s are analyzed f o r the two features [ c o r o n a l ] and L d i s t r i b u t e d J the consonant combinations are e i t h e r -H- (/sp/, / s t / , / s k / , I f p l , Iftl, /fk/, /ps/, / t s / , / k s/, I p f l , Itfl, and Ikfl) or - - ( f p / , / f t / , / f k / , / p f / , Itfl, and / k f / ) . A comparison of the two groups i s presented i n Tables XXXVIII, XXXIX, and XL. Table II. Confusion Matrix for Clusters in I n i t i a l Position /.sp/ /st/ /sk/ /ps/ /ts/ ' /ks/ /ip/ / f t / /fk/ /pf/ /tf/. /kf/ /ft/ /fk/ /pf/ Ml Ml. /sp/ 100 1 1 1 5 /st/ 53 18 22 1 1 1 2 1 2 3 3 l /sk/ 39 6 48 2 3 1 1 3 4 1 /ps/ 14 22 11 22 30 2 2 3 1 1 /ts/ 11 11 18 16 37 3 2 . 2 1 1 1 1 4 /ks/ 11 14 27 20 18 9 2 1 1 3 1 1 /fp/ 18 4 4 2 1 24 15 8 8 4 2 3 4 9 2 / f t / 7 1 3 2 24 20 4 29 8 4 1 1 2 2 /fk/ 13 2 22 2 2 1 12 3 24 11 2 6 1 6 1 /pf/ 8 3 2 4 21 16 12 28 3 4 3 1 3 /tf/ 13 6 8 2 1 13 6 13 27 2 1 7 3 1 1 4 /kf/ 18 2 1 3 2 33 10 4 19 5 2 5 ' 1 1 1 1 //p/ 6 1 1 2 1 1 67 9 16 4 A r t / 8 5 1 2 1 7 2 1 12 55 10 1 2 1 //k/ 1 1 5 1 42 4 50 1 3 /p// 4 . 3 1 1 12 19 18 25 22 3 / t / / 1 3 15 7 7 75 /k// 3 1 1 1 1 1 11 15 23 21 20 10 Table III. Confusion Matrix for Clusters in Medial Position /sp/ /st/ /sk/ /ps/ /ts/ /ks/ /fp/ / f t / /fk/ /pf/ /tf / /kf/ //p/ /ft/ //k/ /pf/ /tf/ /kf/ /sp/ 79 7 16 1 1 1 1 1 1 /st/ 6 90 3 1 1 1 2 1 2 1 /sk/ 63 7 34 1 3 /ps/ 3 1 55 14 34 1 /ts/ 2 3 1 33 37 18 2 3 1 3 2 •1 2 /ks/ 1 1 10 4 70 1 1 1 12 2 1 2 2 /fp/ 2 1 . 1- 5 2 31 15 17 8 4 14 2 1 2 1 2 / f t / 5 1 1 2 1 1 6 63 18 3 2 2 1 1 1 /fk/ 1 7 1 3 12 4 44 2 1 15 2 3 2 4 1 /pf/ 4 1 1 14 2 1 16 4 5 40 4 11 3 2 /tf / 3 1 13 3 5 11 5 22 26 4 1 1 1 7 3 2 /kf/ 4 1 1 13 5 7 6 4 4 18 12 31 2 3 2 1 //p/ 3 1 1 • 1 29 11 61 1 / / t / 1 2 6 68 29 2 //k/ 1 1 • 4 1 1 2 1 12 4 78 3 /p// 1 4 1 2 81 16 3 / t / / 1 1 1 1 1 1 23 64 15 /k// 4 1 2 2 13 26 60 Table IV. Confusion Matrix for Clusters in Final Position V /st/ /sk/ /ps/ /ts/ /ks/ /±p/ /it/ /fk/ /pi/ /tf/ /kf/ /fp/ /ft/ /fk/ /p// Ml /kf/ /sp/ 11 4 11 2 1 1 14 12 8 11 5 3 6 8 1 2 2 /st/ 9 4 9 5 2 4 20 13 10 8 6 11 3 2 8 3 3 /sk/ 16 7 42 2 5 3 3 2 4 2 1 1 4 3 9 3 3 4 /ps/ 2 7 7 16 3 4 9 10 12 8 12 2 2 4 5 1 /ts/ 1 5 4 14 8 3 8 10 9 13 11 2 1 3 1 12 3 /ks/ 3 1 5 8 17 18 2 3 7 1 3 9 3 5 2 12 3 /fp/ 4 3 5 2 1 2 10 12 16 17 4 9 4 4 4 4 5 2 /ft/ 4 4 10 4 4 1 24 17 11 6 3 6 5 2 6 3 3 1 /fk/ 1 5 1 1 5 21 15 17 10 2 3 4 7 5 4 5 2 /pf/ 2 2 6 5 8 8 11 13 7 11 3 1 4 18 3 /tf/ 4 3 1 3 2 7 2 7 8 13 16 1 2 6 4 27 2 /kf/ 4 3 3 6 3 1 15 7 7 42 1 8 3 5 //p/ 1 2 2 1 38 15 36 6 6 1 //t/ 2 2 2 1 39 23 24 9 4 2 //k/ 2 2 2 1 1 29 21 27 13 6 4 /p// 1 1 1 1 4 4 1 12 77 6 /tf/ 1 1 1 4 2 2 2 6 1 7 68 7 Ml 1 1 1 2 6 3 4 80 10 -62-Table V, Number of Correct Responses of Each Cluster in Each Position. Stimulus I n i t i a l Medial Final Total /sp/ 100 79 11 190 /st/ 18 90 4 112 /sk7 48 34 42 124 /ps/ 22 55 7 84 /ts/ 37 37 14 88 /ks/ 9 70 18 97 /fp/ 24 31 10 65 / f t / 20 63 17 100 /fk/ 24 44 17 85 /pf/ 28 40 13 81 /t f / 2 26 13 41 /kf/ 2 31 42 75 //p/ 67 29 38 134 / / t / 55 68 23 146 //k/ 50 78 27 155 /p// 25 81 12 118 Ml 75 64 68 207 lyji 10 60 10 80 Table VI. Significant Differences observed between Means from Table V. Comparison of Means t-value Mean than number correct for /s/ clusters greater mean number correct for If/ clusters. 2. 28** Mean than number correct for /// clusters greater mean number correct for Ifl clusters. 3. 140*** Mean than number correct for /// clusters greater mean number correct for Isl clusters. 1. 02 Mean than number correct for Medial Position greater mean number correct for I n i t i a l Position. 2. 50** Mean than number correct for Medial Position greater mean number correct for Final Position. 5. 30**** Mean than number correct for I n i t i a l Position greater mean number correct for Final Position. 1. 75* * Significantly different at .1 level. * Significantly different at .05 level. * Significantly different at .01 level. * Significantly different at .001 level - 6 4 -Table VII. Total Responses of Each Cluster in Each Position. Stimulus I n i t i a l Medial Final Total /sp/ 316 175 64 555 /st/ 107 120 29 256 /sk/ 182 73 114 369 /ps/ 75 154 45 274 /ts/ 100 75 72 247 /ks/ 17 134 55 206 /fp/ 135 79 122 336 / f t / 80 107 105 292 /fk/ 75 103 133 311 /pf/ 124 98 105 327 /t f / 26 63 75 164 /kf/ 22 79 139 240 //p/ 174 60 150 384 /ft/ 137 91 9.9 327 /fk/ 145 180 156 481 /pf/ 74 137 84 295 /tf/ 137 128 339 604 /kf/ 18 88 58 164 -65-TableVIII.Confusion Matrix: Responses to I n i t i a l clusters combined according to Perrnissibility for I n i t i a l Position. Perrnissible Impermissible Permissible 368 ^ 6 5 336 InperriLLSsible 375 ^35 1136 ^ Table IX. Confusion Matrix: Responses to Medial clusters combined according to Permissibility for Medial Position. Perrnissible Jjriperrnis s ible Permissible 902 1 4 5 480 Inpermissible 192 7 0 5 384 Table X. Confusion Matrix: Responses to Final clusters combined according to Permissibility for Final Position. Permissible Inperrnissible Perrnissible 4 6 3 384 4 0 7 480 Irripermissible 3 6 0 480 714 600 -66-Table XI. Frequency of Types of Errors produced in response to I n i t i a l clusters combined according to their Permissibility as I n i t i a l English clusters. Error Permissible Impermissible t-value Correct Response 241 4 5 0 84 2 . 5 4 * * Single Plosive 1 2 9 48 3 7 3 168 0.60 Single Fricative 1 5 48 97 3 168 0.68 Double Substitution 1 1 96 6 8 336 0.71 Absolute Reversal q 24 1 5 7 84 3 . 1 8 * * * Reversal with Plosive Error 1 3 48 3 0 6 168 2.88** Reversal with Fricative Error 4 48 5 1 168 1.46 Reversal with Plosive and Fricative 2 z 96 93 3 0 336 1.30 Error Significant at .05 Level. Significant at .01 Level. - 6 7 -Table XII. Frequency of Types of Errors produced in response to Medial clusters combined according to their Permissibility as Medial English clusters. Error Permissible Irnpermissible t-value Correct Response 6 3 3 60 3 4 7 48 2.26** Single Plosive 335 1 2 Q 2 4 2 96 0.55 Single Fricative 2 0 120 5 4 96 2.98*** Double Substitution 3 9 240 6 3 192 0.86 Absolute Reversal 5 60 5 6 48 2.93*** Reversal with Plosive Error 2 0 120 5 9 96 2.78** Reversal with Fricative Error 8 120 1 8 96 1.19 Reversal with Plosive and Fricative Error 17 x 240-. 2 8 192 1.61 ** Significant at .05 Level. *** Significant at .01 Level. -68-Table XIII. Frequency of Types of Errors produced in response to Final clusters combined according to their Permissibility as Final English clusters. Error Permissible Impermissible t-value Correct Response 1 8 1 48 2 0 5 60 0.25 Single Plosive 1 6 1 96 4 4 1 120 2.51** Single Fricative 1 0 9 96 77 120 1.54 Double Substitution 1 7 7 192 8 7 240 2.88** Absolute Reversal 2 5 48 6 6 60 2.02* Reversal with Plosive Error 5 2 96 1 1 2 120 1.54 Reversal with Fricative Error 5 9 96 3 3 120 2.09* Reversal with Plosive and Fricative 106 i g 2 4 8 240 2.85** Error Significant at .1 Level. Significant at .05 Level. Table XIV. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to Manner of Articulation. Fricative-Plosive Plosive-Fricative Fricative-Plosive 846 486 126 486 Plosive-Fricative 504 486 468 486 Table XV. Confusion matrix: Responses for clusters combined according to Manner of Articulation. in Medial Position Fricative-Plosive Plosive-Fricative Fricative-Plosive 872 486 94 486 Plos ive-Fricative 116 486 862 486 Table XVI. Confusion matrix: combined according to Manner Responses for clusters of Articulation. in Final Position Fricative-Plosive Plosive-Fricative Fricative-Plosive 733 486 257 486 Plosive-Fricative 239 486 715 486 - 7 0 -Table XVII. Confusion Matrix: Responses for clusters in I n i t i a l Position combined according to Place for Plosives. Fr.-p Fr.-t Fr.-k p-Fr. t-Fr. k-Fr. Fricative-p 225 34 32 19 10 4 Fricative-t 100 115 . 52 34 16 7 Fricative-k 107 20 161 17 10 9 p-Fricative 63 65 47 83 56 9 t-Fricative 46 T,5 48 56 124 6 k-Fricative 84 44 62 64 48 22 The number of responses which could have occurred in each block by chance was 54. -71-Table XVIII. Confusion Matrix: Responses for clusters in Medial Position combined according to Place for Plosives. Fr.-p Fr.-t Fr.-k p-Fr. t-Fr. k-Fr. Fricative-p 146 37 98 17 10 16 Fricative-t 24 227 56 8 6 3 Fricative-k 94 18 172 7 8 19 p-Fricative 21 8 9 197 38 51 t-Fricative 13 19 14 100 139 39 k-Fricative 16 9 7 60 65 173 The number of responses which could have occurred in each block by chance was 54. Table XIX. Confusion Matrix: combined according to Place Responses to for Plosives. clusters in Final Position Fr.-p Fr.-t Fr.-k p-Fr. t-Fr. k-Fr. Fricative-p 88 52 90 43 24 21 Fricative-t 107 .67 80 38 25 25 Fricative-k 82 56 111 35 24 22 ,p-Fricative 25 27 34 59 136 37 t-Fricative 20 20 40 34 152 52 k-Fricative 14 11 48 25 125 95 The number of responses which could have occurred in each block by chance was 54. - 7 3 -Table XX. Confusion Matrix: Responses to clusters in I n i t i a l Position combined according to Place for Fricatives. s-Pl. f-Pl. /-PI. . Pl.-s Pl.-f PI.-/ s-Plosive 288 6 20 8 0 2 f-Plosive 74 134 18 10 74 14 /-Plosive 27 14 265 5 1 12 Plosive-s 139 5 11 157 4 8 Plosive-f 61 128 22 12 91 10 Plosive-/ 16 3 120 0 2 183 The number of responses which could have occurred in each block by chance was 54. -74-Table XXI. Confusion Matrix: Responses for clusters in Medial Position combined according to Place for Fricatives. s-Pl. f-Pl. /-PI. Pl.-s Pl.-f PI.-/ s-Plosive 305 4 7 4 2 2 f-Plosive 19 210 11 15 49 14 /-Plosive 14 4 298 1 1 6 Plosive-s 12 7 6 275 17 7 Plosive-f 16 60 5 58 168 23 Plosive-/ 2 4 4 10 3 301 The number of responses which could have occurred in each block by chance was 54. -75-Table XXII. Confusion matrix: Responses for clusters in Final Position -combined according to Place for Fricatives. s-Pl. fVPl. /-PI. Pl.-s Pl.-f Pl.-f s-Plosive 113 86 43 25 48 21 f-Plosive 36 143 41 21 60 29 /-Plosive 13 6 252 0 2 51 Plosive-s 28 56 18 95 78 43 Plosive-f 15 62 22 29 124 66 Plosive-/ 2 7 29 2 7 271 The number of responses which, could have occurred in each block by chance was- 54. -76-Table XXIIl. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distincitive Feature Anterior for Fricatives. 1191 105 864 432 68 580 432 216 Table XXIV. Confusion matrix: Responses for clusters in Medial Position combined according to- the Distinctive Feature Anterior for Fricatives. 1221 75 864 432 39 609 432 216 Table XXV. Confusion Matrix: Responses for clusters in Final Position combined according to the Distinctive Feature Anterior for Fricatives. 1019 283 864 432 39 603 432 216 -77-["able XXVI. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distinctive Feature <Anterior for Plosives. 1091 205 864 432 394 254 432 216 Table XXVII. Confusion matrix: Responses for clusters in Medial Position combined according to the Distinctive Feature Anterior for Plosives. 1010 286 864 432 277 371 432 216 Table XXVIII. Confusion matrix: Responses for clusters in Final Position combined according to the'Distinctive Feature Anterior for Plosives. 917 379 864 432 372 276 432 216 -78-Table XXIX. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distinctive Feature Anterior for both Plosive and Fricative. (Value for Fricative written f i r s t ) . ++ +_ -+ — ++ 663 384 127 192 65 192 9 96 +- 254 192 147 96 15 96 16 48 37 192 15 96 325 96 55 48 8 96 8 48 117 48 83 24 -79-Table XXX. Confusion Matrix: Responses for clusters in Medial Position combined according to the Distinctive Feature Anterior for both Plosive and Fricative. (Value for Fricative written f i r s t ) . ++ +- -+ — ++ 659 384 160 192 33 192 12 96 +- 186 192 216 96 23 96 7 48 -+ 16 192 3 96 302 96 111 48 10 96 10 48 58 138 48 24 -80-Table XXXI. Confusion matrix: Responses for clusters in Final Position combined according to the Distinctive Feature Anterior for both Plosive and Fricative. (Value for Fricative written f i r s t ) . ++ +- -+ — ++ 41+7 384 235 192 137 192 51 96 +- 151 192 186 96 54 96 41 48 -+ 13 192 15 96 320 96 78 48 — 6 95 5 48 161 48 44 24 -81-Table XXXII. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distinctive Feature Coronal for Plosives. 299 349 216 432 288 1008 432 864 Table XXXIII. Confusion matrix: Responses for clusters in Medial Position combined according to the Distinctive Feature Coronal for Plosives. 391 257 216 432 193 1103 432 864 Table XXXIV. Confusion matrix: Responses for clusters in Final Position combined according to the Distinctive Featrure Coronal for Plosives. 264 396 216 432 455 829 432 864 -82-Table XXXV. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distinctive Feature Coronal for both Plosive and Fricative. (Value for Fricative written f i r s t ) . ++ +- _+ — ++ 236 96 178 192 9 48 9 96 +_ 203 192 644 384 3 96 14 192 -+ 18 48 47 96 36 24 115 48 24 96 132 192 58 48 218 96 -83-Table XXXVI. Confusion matrix: Responses for clusters in Medial Position combined according to the Distinctive Feature Coronal for both Plosive and Fricative. (Value for Fricative vrritten f i r s t ) . ++ +- _+ — ++ 273 96 145 192 6 48 8 96 +- 107 192 729 384 16 96 12 192 -+ 12 48 39 96 100 24 65 48 22 96 88 192 48 48 274 96 -84-Table XXXVII. Confusion matrix: Responses for clusters in Final Position combined according to the Distinctive Feature Coronal for both Plosive and Fricative. (Value for Fricative written f i r s t ) . ++ +- -+ — ++ 142 96 165 192 42 48 89 96 +- 297 192 403 384 46 96 112 192 -+ 46 48 55 96 31 24 90 48 56 96 101 192 56 48 213 96 -85-Table XXXVIII. Confusion matrix: Responses for clusters in I n i t i a l Position combined according to the Distinctive Features Coronal and Distributed for Fricatives. ++ ++ 1261 35 864 432 — 221 432 427 216 Table XXXIX. Confusion matrix: Responses for clusters in Medial Position combined according to : the Distinctive Features Coronal and Distributed for Fricatives. ++ ++ 1254 42 864 432 161 487 432 216 Table XL. Confusion Matrix: Responses for clusters in Final Position combined according to the Distinctive Features: Coronal and Distributed for Fricatives. ++ ++ 1006 290 864 432 259 432 389 216 CHAPTER 5 DISCUSSION The o b j e c t i v e o f t h i s i n v e s t i g a t i o n was t o g a i n i n s i g h t i n t o the mechan ism i n v o l v e d i n p e r c e p t i o n o f c o n s o n a n t c l u s t e r s . To what e x t e n t t h i s o b j e c t i v e was a c h i e v e d , as d e m o n s t r a t e d by r e s u l t s o b t a i n e d f r om the e x p e r i -ment , w i l l be d i s c u s s e d i n t h i s c h a p t e r . A f t e r c o n s i d e r i n g l i m i t a t i o n s o f e x p e r i m e n t a l p r o c e d u r e , t he t r e n d s r e v e a l e d and o b s e r v a t i o n s n o t e d w i l l be p r e s e n t e d . 5.1 L i m i t a t i o n s o f t he E x p e r i m e n t C e r t a i n p r o c e d u r a l l i m i t s were n e c e s s a r y due t o t he l e n g t h o f t h i s i n v e s t i g a t i o n . C l u s t e r s were p r e s e n t e d i n c o m b i n a t i o n w i t h o n l y one vowe l / I / w h i c h a l l o w e d the c r e a t i o n o f a l a r g e number o f nonsense w o r d s . I f s t i m u l i had been c r e a t e d u s i n g o t h e r vowels i n c o m b i n a t i o n w i t h the c o n s o n a n t c l u s t e r s more g e n e r a l i z e d c o n c l u s i o n s c o u l d have been r e a c h e d . However , t h i s wou ld have r e q u i r e d much more e x t e n s i v e d a t a c o l l e c t i o n and a n a l y s e s than was p o s s i b l e i n t he t ime a v a i l a b l e f o r t h i s s t u d y . S i n c e u t t e r a n c e s were p r o d u c e d by o n l y one s p e a k e r , i t i s p o s s i b l e t h a t u n i q u e c h a r a c t e r i s t i c s o f t h i s s p e a k e r ' s s p e e c h m i g h t have i n f l u e n c e d s u b j e c t s ' p e r c e p t i o n . A r e p e t i t i o n o f the p r o c e d u r e w i t h a number o f s p e a k e r s m igh t r e d u c e t h i s e f f e c t . S t i m u l i were p r e s e n t e d a t o n l y one s i g n a l - t o - n o i s e r a t i o (+3 d B ) , t hus a l l o w i n g g r e a t e r f l e x i b i l i t y i n v a r y i n g c l u s t e r p o s i t i o n . A more d e t a i l e d i n v e s t i g a t i o n wou ld a l l o w e x a m i n a t i o n o f s t i m u l i p r e s e n t e d a t s e v e r a l s i g n a l - t o - n o i s e r a t i o s , wh i ch wou ld a l l o w o b s e r v a t i o n o f d e t e r i o r -i o n o f p e r c e p t i o n w i t h i n c r e a s e s i n n o i s e . Fo r an e x p e r i m e n t i n wh i ch b o t h s i g n a l - t o - n o i s e r a t i o and c l u s t e r p o s i t i o n s were v a r i e d , t h e t ime r e q u i r e d f o r r e c o r d i n g the e x p e r i m e n t a l t a p e s and the a c t u a l l e n g t h o f t he -86--87-l i s t e n i n g task would have been u n r e a l i s t i c w i t h i n the time r e s t r a i n t s f o r t h i s s t u d y . F r i c a t i v e s show v a r i a b i l i t y i n i n t r i n s i c i n t e n s i t y , ( S t r e v e n s , 1960). In t h i s s t u d y , the phoneme /// has g r e a t e r i n t r i n s i c i n t e n s i t y than the phonemes /s/ and / f / . Thus, any masking noise w i l l mask the f r i c a t i v e s /s/ and If I t o a g r e a t e r extent than i t w i l l mask ///. This d i f f e r e n c e i n mask-ing c o u l d not be overcome usi n g standard types of noise (pink n o i s e , white noise and USASI noise) at one s i g n a l - t o - n o i s e r a t i o . Using more than one s i g n a l - t o - n o i s e l e v e l f o r the d i f f e r e n t items was not f e a s i b l e . ( S e c t i o n 3.3). The only method of compensating f o r t h i s problem would be to use speech shaped noise i n which amplitude envelope of the noise f o l l o w s the spectrum of the speech s i g n a l ( H o r i i et a l . , 1971). Due t o t e c h n i c a l d i f f i c u l t i e s i t was not p o s s i b l e to use t h i s type of n o i s e i n the present s t u d y . When the task was g i v e n t o s u b j e c t s , no p r a c t i c e items were presented. Thus, at the beginning of t e s t i n g each subject had to i d e n t i f y s t i m u l u s c l u s t e r s as he was " l e a r n i n g " the t a s k . This may have r e s u l t e d i n e r r o r s not due t o i n a c c u r a t e p e r c e p t i o n . The subjects a l s o seemed a f f e c t e d by the length of the t a s k . The three experimental tapes f o r each s u b j e c t were presented i n one s e s s i o n . I t was apparent t h a t s u b j e c t s became f a t i g u e d and d i s i n t e r e s t e d towards the end of t e s t i n g ; had s u b j e c t s been given one t e s t only per session, t h i s f a t i g u i n g e f f e c t might have been reduced. 5.2 D i s c u s s i o n of R e s u l t s The r e s u l t s obtained are now c o n s i d e r e d , bearing i n mind the e x p e r i -mental l i m i t a t i o n s d e s c r i b e d above. -88-5.21 T o t a l Confusion M a t r i c e s (Tables I I , I I I , and I V ) . A ge n e r a l tendency f o r responses to f a l l along the di a g o n a l i s evi d e n t i n the confusion matrices f o r i n i t i a l c l u s t e r s and i s more apparent f o r medial c l u s t e r s . The g r e a t e r degree of s c a t t e r i n g observed f o r f i n a l c l u s t e r r e s u l t s i n i ) a g r e a t e r number of each type of e r r o r (Table X I I D a n d i i ) a 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 from medial c l u s t e r s f o r mean c o r r e c t responses. (Table V I ) . I t i s of i n t e r e s t to note t h a t f o r t h i s t e s t there are fewer s c a t t e r e d responses to /// than to /s/ or / f / c l u s t e r s , c o n t r i b u t -ing t o a s i g n i f i c a n t l y g r e a t e r mean c o r r e c t response f o r a l l /// c l u s t e r s (Table V I ) . 5.22 Responses to Each C l u s t e r (Tables V, V I , and V I I ) . Mean number of c o r r e c t responses f o r c l u s t e r s i n medial p o s i t i o n was s i g n i f i c a n t l y g r e a t e r than the mean number of c o r r e c t responses to c l u s t e r s i n f i n a l and i n i t i a l p o s i t i o n s . Subjects f r e q u e n t l y s t a t e d t h a t they f e l t the s t i m u l i i n the medial t e s t to be more l i k e r e a l words than s t i m u l i i n the other two p o s i t i o n s . The s t i m u l i w i t h medial c l u s t e r s had three t r a n s -i t i o n a l cues; those i n the vowel p r i o r to the c l u s t e r , those between the two consonants i n the c l u s t e r , and those i n the vowel f o l l o w i n g the c l u s t e r and beginning /In,/. This would have g i v e n a c o u s t i c i n f o r m a t i o n about both consonants i n c l u s t e r , w h i l e s t i m u l i f o r c l u s t e r s i n i n i t i a l and f i n a l p o s i t i o n had two t r a n s i t i o n a l cues: those between the two consonants and those between the consonant and the adjacent vowel. The mean number of c o r r e c t responses f o r c l u s t e r s i n i n i t i a l p o s i t i o n was g r e a t e r than f o r c l u s t e r s i n f i n a l p o s i t i o n , s i g n i f i c a n t a t the .1 l e v e l . This o b s e r v a t i on i s c o n s i s t e n t w i t h that made by House et a l . (1965) i . e . that i n n o i s e , i n i t i a l consonants are more e a s i l y i d e n t i f i e d than f i n a l consonants. -89-In the same study by House and h i s co-workers, the descending order of c o r r e c t i d e n t i f i c a t i o n f o r f r i c a t i v e s , presented i n n o i s e was' Is/, If I, and ///. This order f o r Is I and If I was confirmed by the order of mean number c o r r e c t observed f o r /s/ and If I c l u s t e r s (Table V I ) . However, /// c l u s t e r s were c o r r e c t l y i d e n t i f i e d more f r e q u e n t l y than Is I or It I c l u s t e r s , whereas House et a l . observed the reverse t r e n d . The trend observed i n t h i s study i s c o n s i s t e n t w i t h the g r e a t e r i n t r i n s i c i n t e n s i t y of /// (Strevens, 1960) and inadequacy of the noise to mask /// which was d i s c u s s e d above. I t i s noted that i n a n a l y s i s of the frequency of c o r r e c t and t o t a l responses (Tables V and V I I ) , the g r e a t e s t number of responses occurred w i t h the c l u s t e r s /sp/ and /tf/. These two c l u s t e r s have a high frequency of occurrence i n spoken E n g l i s h . This o b s e r v a t i o n confirms the p r i n c i p l e proposed i n Chapter 2 that frequency of production of a u n i t f a c i l i t a t e s i t s p e r c e p t i o n . 5.23 P e r m i s s i b l e versus Impermissible C l u s t e r s (Tables V I I I - X I I I ) To observe p o s s i b l e d i f f e r e n c e s i n p e r c e p t i o n of p e r m i s s i b l e and impermissible c l u s t e r s , i t was necessary t o present both kinds of c l u s t e r s f o r each p o s i t i o n . In each p o s i t i o n , c l u s t e r s were combined acc o r d i n g t o t h e i r permiss-i b i l i t y f o r that s p e c i f i c p o s i t i o n . This grouping has been o u t l i n e d i n S e c t i o n 4.3. Confusion matrices according to such grouping are presented i n Tables V I I I , IX, and X. For c l u s t e r s i n i n i t i a l and medial p o s i t i o n there i s a trend f o r impermissible c l u s t e r s to be perceived as p e r m i s s i b l e more f r e q u e n t l y than p e r m i s s i b l e c l u s t e r s to be p e r c e i v e d as i m p e r m i s s i b l e , thus i n d i c a t i n g that when a stimulus i s p e r c e i v e d ambiguously a s u b j e c t i s -SO-more l i k e l y t o respond to i t as i f i t were f a m i l i a r . A reverse of t h i s trend i s observed f o r c l u s t e r s i n f i n a l p o s i t i o n ; more p e r m i s s i b l e c l u s t e r s were perceived as impermissible than impermissible c l u s t e r s were perceived as p e r m i s s i b l e . The g r e a t e r degree of s c a t t e r i n g of responses observed i n the confusion m a t r i x f o r f i n a l c l u s t e r s may have c o n t r i b u t e d to the d i s c r e p -ancy between Table X and Tables V I I I and IX. D i f f e r e n c e s i n p e r c e p t i o n of p e r m i s s i b l e and impermissible c l u s t e r s would be manifest i n d i f f e r e n c e s i n the frequency of each type of e r r o r between the two types of c l u s t e r s . For each p o s i t i o n , the frequency of each type of e r r o r was tabulated f o r both p e r m i s s i b l e and impermissible c l u s t e r s . Student-t values were c a l c u l a t e d between means f o r the two types of consonant combinations. The l e v e l s of s i g n i f i c a n c e s p e c i f i e d by those t-values are i n d i c a t e d . For c l u s t e r s presented i n I n i t i a l P o s i t i o n , (Table XI) there are s i g n i f i c a n t d i f f e r e n c e s between the means of p e r m i s s i b l e and impermissible c l u s t e r s f o r three out of e i g h t types of e r r o r s . For clusters.presented i n Medial P o s i t i o n , (Table X I I ) , four out of ei g h t types of e r r o r s have s i g n i f i c a n t d i f f e r e n c e s between the means of p e r m i s s i b l e and impermissible c l u s t e r s . There are f i v e s i g n i f i c a n t d i f f e r e n c e s between p e r m i s s i b l e and impermissible c l u s t e r s f o r consonant combinations presented i n F i n a l P o s i t i o n . Eond's (1S71) proposal that p e r m i s s i b l e c l u s t e r s are perceptual u n i t s was based on her observation of a high frequency of r e v e r s a l e r r o r s . In the present study t h i s high frequency of r e v e r s a l e r r o r s was not observed f o r e i t h e r p e r m i s s i b l e or impe r m i s s i b l e c l u s t e r s i n any p o s i t i o n . S u b s t i t u t i o n e r r o r s , i . e . s i n g l e p l o s i v e and s i n g l e f r i c a t i v e e r r o r s and double -91-s u b s t i t u t i o n e r r o r s were much more frequent than r e v e r s a l e r r o r s . Thus, the r e s u l t s of t h i s study are not i n agreement w i t h those obtained by Bond. A s i g n i f i c a n t d i f f e r e n c e was observed between the frequency of absolute r e v e r s a l e r r o r s f o r p e r m i s s i b l e and impermissible c l u s t e r s i n a l l three p o s i t i o n s ; impermissible c l u s t e r s had a s i g n i f i c a n t l y g r e a t e r number of absolute r e v e r s a l e r r o r s than p e r m i s s i b l e c l u s t e r s . For a l l three c l u s t e r p o s i t i o n s , s i n g l e p l o s i v e e r r o r s were most fre q u e n t . I t has been reported p r e v i o u s l y ( S e c t i o n 1.23), that d i f f e r e n t -i a t i o n of place of a r t i c u l a t i o n among p l o s i v e s depends on frequency of the burst and second formant t r a n s i t i o n s . In the present c o n t e x t , i n t e n s i t y d i f f e r e n c e s which seemed to be used to d i f f e r e n t i a t e p l a c e of a r t i c u l a t i o n among f r i c a t i v e s seemed not to be e f f e c t i v e f o r p l o s i v e s . For the s t i m u l i p r e sented, when the p l o s i v e preceded the f r i c a t i v e i n i n i t i a l c l u s t e r s as i n / p s l / or followed the f r i c a t i v e i n f i n a l c l u s t e r s as i n /tsp/ the vowel formant cues were reduced or absent. When the p l o s i v e was d i r e c t l y f o l l o w e d by a vowel as i n / s p l / or /TspTr)/ the p l o s i v e was s h o r t e r than a f i n a l p l o s i v e so l e s s a c o u s t i c i n f o r m a t i o n was a v a i l a b l e from the b u r s t . R e s u l t s presented i n Tables X I , X I I and X I I I show s i g n i f i c a n t d i f f e r e n c e s between the frequency o f • p e r m i s s i b l e and impermissible c l u s t e r s f o r twelve e r r o r s (three i n i t i a l , four medial and f i v e f i n a l ) , which would suggest some d i f f e r e n c e i n the perceptual mechanism of p e r m i s s i b l e and impermissible c l u s t e r s although the r e s u l t s cannot be considered c o n c l u s i v e . 5.24 Manner of A r t i c u l a t i o n . (Tables X I I - X I V ) . The order f o r manner of a r t i c u l a t i o n ( f r i c a t i v e - p l o s i v e , or p l o s i v e -f r i c a t i v e ) was i d e n t i f i e d c o r r e c t l y f o r 67 percent of the s t i m u l i i n i n i t i a l p o s i t i o n , f o r 89 percent of s t i m u l i i n medial p o s i t i o n and f o r 75 percent of s t i m u l i i n f i n a l p o s i t i o n . For c l u s t e r s i n i n i t i a l p o s i t i o n (Table X I I ) , -92-when p l o s i v e - f r i c a t i v e c l u s t e r s were presented there were a g r e a t e r number of f r i c a t i v e - p l o s i v e responses than of p l o s i v e - f r i c a t i v e responses. The g r e a t e r accuracy of d i f f e r e n t i a t i o n f o r manner of a r t i c u l a t i o n as compared t o pla c e of a r t i c u l a t i o n i s c o n s i s t e n t w i t h the r e s u l t s of M i l l e r and N i c e l y (1955). These authors observed that the feat u r e [ a f f r i c a t i o n ] , which d i f f e r e n t i a t e s f r i c a t i v e s and p l o s i v e s , was per c e i v e d c o r r e c t l y a t a much lower s i g n a l - t o - n o i s e r a t i o than was p l a c e of a r t i c u l a t i o n . 5.25 Place of A r t i c u l a t i o n (Tables XV - X X ) . In the confusion m a t r i x f o r c l u s t e r s combined according t o pla c e of a r t i c u l a t i o n , s u b s t i t u t i o n e r r o r s were more frequent than r e v e r s a l e r r o r s . . For c l u s t e r s i n i n i t i a l and medial p o s i t i o n ( / s t / , / f t / , and / / t / ) and (/sk/, / f k / , and /fk/) were perceived more o f t e n as (/sp/, / f p / , and Ifpl) than the opposite type of c o n f u s i o n . The high frequency of /sp/ responses to a l l c l u s t e r s (Table V I I ) might e x p l a i n t h i s phenomenon. D i f f i c u l t y w i t h place d i f f e r e n t i a t i o n f o r p l o s i v e s was d i s c u s s e d i n re f e r e n c e t o s i n g l e p l o s i v e e r r o r s . In the a n a l y s i s f o r p l a c e of a r t i c u l a t i o n f o r f r i c a t i v e s , there was a g e n e r a l trend f o r more r e v e r s a l s i n responses: f o r example. If I-plosive response f o r p l o s i v e -IfI s t i m u l i were more frequent than s u b s t i t u t i o n of p l a c e , e.g. /sp/ response f o r / f p / s t i m u l u s . This e f f e c t might have been due t o the s i m i l a r i t y of f r i c a t i o n and masking n o i s e . When the f r i c a t i v e preceded the p l o s i v e , the s u b j e c t s might p e r c e i v e the noise i n the c l o s u r e f o l l o w i n g the p l o s i v e as a f r i c a t i v e . When the f r i c a t i v e f o l lowed the p l o s i v e , the s u b j e c t might p e r c e i v e the no i s e preceding the c l o s u r e , as the f r i c a t i v e , and the f r i c a t i v e f o l l o w i n g the c l o s u r e , as masking. -93-5.26 Analyses According t o C e r t a i n Jakobson, Fant and K a l l e D i s t i n c t i v e F e a t u r e s . (Tables X X I I I - XL) Responses were analyzed t o determine the v a l i d i t y , from the point of view of p e r c e p t i o n of some of the d i s t i n c t i v e f eatures proposed by Jakobson, Fant and K a l l e (1963) which were assumed to e x i s t at the production l e v e l . The f e a t u r e s used f o r the present analyses were: i ) [ a n t e r i o r ] , i i ) [ c o r o n a l ] and i i i ) [ d i s t r i b u t e d ] . These features grouped the c l u s t e r s as presented i n S e c t i o n 4.6. A comparison can be made of responses grouped according t o the f e a t u r e [ a n t e r i o r ] f o r f r i c a t i v e s (Tables X X I I I to XXV), w i t h responses grouped according to the same f e a t u r e f o r p l o s i v e s (Tables XXVI to X X V I I I ) . I t i s noted that f o r a l l c l u s t e r p o s i t i o n s , the frequency of c o r r e c t i d e n t i f i c a t i o n f o r the f e a t u r e [ a n t e r i o r ] was g r e a t e r f o r f r i c a t i v e s than f o r p l o s i v e s . In a l l three p o s i t i o n s , s t i m u l i w i t h [+ a n t e r i o r ] f r i c a t i v e s were more f r e q u e n t l y p e r c e i v e d as having [- a n t e r i o r ] f r i c a t i v e s than the opposite type of confusion of c l u s t e r s w i t h [ - a n t e r i o r ] f r i c a t i v e s p e r c e i v e d as having [+ a n t e r i o r ] f r i c a t i v e s . The high frequency of /p/ responses f o r both c o r r e c t and t o t a l responses (Tables V and V I I ) p r e d i c t s a g r e a t e r number of c l u s t e r s w i t h [- a n t e r i o r ] p l o s i v e s p e r c e i v e d as having [+ a n t e r i o r ] p l o s i v e s than c l u s t e r s w i t h [+ a n t e r i o r ] p l o s i v e s perceived as having [- a n t e r i o r ] p l o s i v e s . However, t h i s s i t u a t i o n i s only observed f o r f i n a l c l u s t e r s (Table XXVI). Further analyses according t o the f e a t u r e [ a n t e r i o r ] f o r p l o s i v e plus f r i c a t i v e e.g. [-H- a n t e r i o r ] f o r /sp/ are presented i n Tables XXIX to XXXI. The grouping of c l u s t e r s f o r t h i s a n a l y s i s i s presented i n S e c t i o n 4.6. For a l l three p o s i t i o n s , the c o r r e c t i d e n t i f i c a t i o n of the f e a t u r e f o r both consonants i n the c l u s t e r exceeds chance i n a l l c a s e s . As p r e d i c t e d by r e s u l t s presented i n Tables X X I I I to XXV, f o r a l l three c l u s t e r p o s i t i o n s there i s g r e a t e r s e p a r a t i o n of responses according t o the value of the f e a t u r e f o r f r i c a t i v e s than according to the value of the f e a t u r e f o r p l o s i v e s . • i -94-Analyses according to the f e a t u r e [ c o r o n a l ] were made f o r p l o s i v e s (Tables XXXII t o XXXIV), f o r p l o s i v e plus f r i c a t i v e (Tables XXXV to XXXVIII) and f o r f r i c a t i v e s (Tables XXXVIII to X L ) . The t o t a l of 1913 confusions among p l o s i v e s f o r the f e a t u r e [ c o r o n a l ] (Tables XXXII t o XXXIV) i s of the same order as the t o t a l of 1938 confusions among p l o s i v e s f o r the f e a t u r e [ a n t e r i o r ] (Tables XXVI t o X X V I I I ) . T h i s i m p l i e s that i f d i s t i n c t i v e f eatures are used f o r pe r c e p t u a l d i f f e r e n t i a t i o n , then d i f f e r e n t i a t i o n of p l o s i v e s according to the fea t u r e s [ c o r o n a l ] and [ a n t e r i o r ] occurs to the same e x t e n t . For c l u s t e r s i n i n i t i a l and medial p o s i t i o n (Tables XXXII and XXXIII) there are a g r e a t e r number of s t i m u l i w i t h [+ c o r o n a l ] p l o s i v e s p e r c e i v e d as having [- c o r o n a l ] p l o s i v e s than the opposite type of c o n f u s i o n . This obser-v a t i o n could be explained by the high frequency of /p/ responses i n Tables V and V I I . Fu r t h e r analyses f o r the feat u r e [ coronal] f o r p l o s i v e plus f r i c a t i v e e.g. [++ c o r o n a l ] f o r / s t / , are presented i n Tables XXV to XXVII. The grouping of c l u s t e r s f o r t h i s a n a l y s i s i s presented i n S e c t i o n 4.6. For a l l three c l u s t e r p o s i t i o n s , the frequency of c o r r e c t i d e n t i f i c a t i o n according t o the value of the f e a t u r e f o r both consonants i n - c l u s t e r exceeds the number which c o u l d have occurred by chance. As observed w i t h the a n a l y s i s of the featu r e [ a n t e r i o r ] f o r f r i c a t i v e plus p l o s i v e (Tables XXIV to XXXI), d i f f e r e n t i a t i o n a c c o r d i n g to the value of the fe a t u r e f o r f r i c a t i v e s i s b e t t e r than according t o the value of the same feature f o r p l o s i v e s . The f e a t u r e s [ c o r o n a l ] and [ d i s t r i b u t e d ] f o r f r i c a t i v e s were analyzed together s i n c e s e p a r a t i o n according to these f e a t u r e s produced i d e n t i c a l groups of c l u s t e r s (Tables XXXVIII t o X L ) . Both f r i c a t i v e s /s/ and /// are [+ c o r o n a l ] and [+ d i s t r i b u t e d ] ; the f r i c a t i v e If I i s [- c o r o n a l ] and [ d i s t r i b u t e d ] . The frequency of confusions among these responses i s substan-t i a l l y l e s s than the frequency of confusions among p l o s i v e s combined -95-according to the f e a t u r e [ c o r o n a l ] . For a l l p o s i t i o n s , the frequency of c o r r e c t i d e n t i f i c a t i o n f o r f r i c a t i v e s d i f f e r e n t i a t e d according to the f e a t u r e s [ c o r o n a l ] and [ d i s t r i b u t e d ] (Tables XXXVIII to XL) i s not as g r e a t as that observed-for f r i c a t i v e s d i f f e r e n t i a t e d according to the feature [ a n t e r i o r ] (Tables X X I I I to XXV). From the preceding analyses f o r responses combined according to the Jakobson, Fant and H a l l e d i s t i n c t i v e f e a t u r e s : [ a n t e r i o r ] , [ c o r o n a l ] and [ d i s t r i b u t e d ] , i t can be seen that these features appear to have some v a l i d i t y i n p e r c e p t i o n w i t h i n the present c o n t e x t . Due to more accurate d i s c r i m i n a t i o n of f r i c a t i v e s d i s c u s s e d i n S e c t i o n s 5.1 and 5.23, the d i f f e r e n t i a t i o n of f e a t u r e s f o r f r i c a t i v e s was much more accurate than the d i f f e r e n t i a t i o n of features f o r p l o s i v e s . The g r e a t e s t amount of d i f f e r e n t -i a t i o n occurred w i t h the f e a t u r e [ a n t e r i o r ] f o r f r i c a t i v e s . D i f f e r e n t i a t i o n f o r p l o s i v e s according t o the f e a t u r e s [ a n t e r i o r ] and [ c o r o n a l ] occurred to a s i m i l a r e x t e n t . 5.3 R e l a t i o n s h i p to Models of Speech P e r c e p t i o n . The present study has attempted to determine the p o s s i b l e u n i t used i n the p e r c e p t i o n of consonant c l u s t e r s and to g a i n some i n s i g h t i n t o the p e r c e p t u a l mechanism i n v o l v e d . Bond (1971) presents an argument which s t a t e s that a l i s t e n e r may l e a r n t o d i s t i n g u i s h and p e r c e i v e c l u s t e r s as whole u n i t s r a t h e r than as a sequence . of d i s c r e t e sound elements. Bond suggests f u r t h e r that c o n t e x t - s e n s i t i v e coding (Wickelgren, 1969a, 1969b) can e x p l a i n p e r c e p t i o n of consonant c l u s t e r s . I f c o n t e x t - s e n s i t i v e coding were used f o r such p e r c e p t i o n , the two elements would be coded as an unordered sequence, each element being i d e n t i f i e d by the preceding element and the f o l l o w i n g element. The elements would then be ordered and the intended sequence determined. -96-Bonci s t a t e d that her o b s e r v a t i o n of a high frequency of r e v e r s a l e r r o r s support.both the proposal of c l u s t e r s as whole u n i t s and context-s e n s i t i v e c o d i n g . However, s i n c e her r e s u l t s were not s u b s t a n t i a t e d i n t h i s present study, her conclusions regarding perception of consonant c l u s t e r must be re-examined. From analyses of the r e s u l t s obtained i t i s noted that the frequency of p l o s i v e confusions f o r both p e r m i s s i b l e and i m p e r m i s s i b l e c l u s t e r s i s g r e a t e r than the frequency of f r i c a t i v e c o n f u s i o n s . The d i f f e r e n c e between the frequency of confusions f o r the two types of consonants c o u l d suggest that the p e r c e p t i o n of each consonant i n such combinations i s independent of p e r c e p t i o n of the a s s o c i a t e d consonant. This i m p l i e s that c l u s t e r s are p e r c e i v e d as a sequence of elements, r a t h e r than a s i n g l e p e r c e p t u a l u n i t s i n c e the elements of which they are composed are a c t u a l l y perceived independently. The r e s u l t s f o r p e r m i s s i b l e and i m p e r m i s s i b l e c l u s t e r s suggest a d i f f e r e n c e i n the p e r c e p t u a l mechanism f o r these two types of c l u s t e r s . I f p e r c e p t i o n i s accomplished by c o n t e x t - s e n s i t i v e c o d i n g , a l l o p h o n i c r e p r e s e n t a t i o n s are present f o r each allophone context used i n speech p r o d u c t i o n . Thus, f o r a c l u s t e r appearing i n each p o s i t i o n t h e re would n e c e s s a r i l y be a l l o p h o n i c r e p r e s e n t a t i o n s f o r that c l u s t e r i n t h a t s p e c i f i c p o s i t i o n ; f o r example, the c l u s t e r /sp/ would have r e p r e s e n t a t i o n s f o r a l l three p o s i t i o n s : ( P ), ( S , P ), and ( S , P „ ) ; however, the c l u s t e r nf p, s v v p s v v p s # jplj would only have r e p r e s e n t a t i o n s f o r medial p o s i t i o n (yPy» p^v^' *n such a case, when impermissible c l u s t e r s are presented to the l i s t e n e r no a l l o p h o n i c r e p r e s e n t a t i o n s e x i s t f o r that combination of consonants. There-f o r e , there i s an attempt to f i t the ambiguous (masked by noise) and impermissible s t i m u l i i n t o the e x i s t i n g a l l o p h o n i c r e p r e s e n t a t i o n s . This -97-phenomenon would e x p l a i n the g r e a t e r frequency of p e r m i s s i b l e responses t o impermissible s t i m u l i than the reverse type of c o n f u s i o n . This i n t e r p r e t -a t i o n of the r e s u l t s i s c o n s i s t e n t w i t h Bond's s u g g e s t i o n , that context-s e n s i t i v e coding i s used i n the p e r c e p t i o n of consonant c l u s t e r s . The g r e a t e r accuracy observed f o r d i f f e r e n t i a t i o n of manner of a r t i c u l a t i o n than f o r place of a r t i c u l a t i o n might be e x p l a i n e d by the motor theory of speech p e r c e p t i o n . The motor commands producing the a r t i c u l a t i o n of a f r i c a t i v e are d i f f e r e n t from the motor commands producing the a r t i c u l a t i o n of a p l o s i v e . I f these motor commands are h i g h l y d i f f e r e n t -i a t e d and i f perception i s d i r e c t l y i n v o l v e d w i t h p r o d u c t i o n , i t f o l l o w s t h a t p e r c e p t i o n f o r manner of a r t i c u l a t i o n would be more h i g h l y d i f f e r e n t i a t e d than p e r c e p t i o n f o r place of a r t i c u l a t i o n . As p r e v i o u s l y d i s c u s s e d , analyses were c a r r i e d out on the responses according t o c e r t a i n Jakobson, Fant and H a l l e d i s t i n c t i v e f e a t u r e s . The r e s u l t s i n d i c a t e d there seemed to be some v a l i d i t y t o the p r o p o s a l features are o p e r a t i v e i n p e r c e p t i o n . For p l o s i v e s , d i f f e r e n t i a t i o n occurred to the same extent f o r both d i s t i n c t i v e features ( [ a n t e r i o r ] a nd[coronal] ) . For f r i c a t i v e s , d i f f e r e n t i a t i o n occurred to a g r e a t e r extent w i t h the feature [ a n t e r i o r ] than w i t h the f e a t u r e [ c o r o n a l ] . I t seems th a t consonant c l u s t e r s are not a p e r c e p t u a l u n i t . I f t h i s were the case, the p e r c e p t i o n of consonant c l u s t e r s should u t i l i z e the same mechanism i n v o l v e d i n the perception of speech. From the preceding d i s c u s s i o n i t i s apparent that one theory of speech pe r c e p t i o n cannot f u l l y e x p l a i n the r e s u l t s . Three of the e x i s t i n g t h e o r i e s of speech perception can be used to e x p l a i n some observations made. However, no d e f i n i t e conclusions can be yet made regarding the mechanism u t i l i z e d f o r the p e r c e p t i o n of consonant c l u s t e r s . CHAPTER 6 CONCLUSIONS AND IMPLICATIONS FOR FURTHER RESEARCH This study of the per c e p t u a l confusions among p e r m i s s i b l e and imp e r m i s s i b l e E n g l i s h c l u s t e r s sought answers to questions regarding the e f f e c t on p e r c e p t i o n of 1) the p e r m i s s i b i l i t y of the consonant combination, of 2) the manner of a r t i c u l a t i o n f o r f r i c a t i v e and p l o s i v e , 3) of the pla c e of a r t i c u l a t i o n f o r f r i c a t i v e and p l o s i v e , and 4) of c e r t a i n d i s t i n c t i v e f e a t u r e s . Analyses of the responses obtained to consonant combinations presented i n noise i n d i c a t e d t h a t : 1) Consonant c l u s t e r s seemed t o be perceived as a sequence of elements r a t h e r than a percept u a l u n i t . 2) There were some d i f f e r e n c e s i n response to p e r m i s s i b l e and impermissible c l u s t e r s . However, the r e s u l t s were not c o n c l u s i v e . 3) Manner of a r t i c u l a t i o n was d i f f e r e n t i a t e d b e t t e r than place of a r t i c u l -a t i o n . Place of a r t i c u l a t i o n f o r f r i c a t i v e s was d i f f e r e n t i a t e d b e t t e r than p l a c e of a r t i c u l a t i o n f o r p l o s i v e s . 4) Analyses of responses according to the Jakobson, Fant and H a l l e d i s t i n c t i v e features [ a n t e r i o r ] , [ c o r o n a l ] and [ d i s t r i b u t e d ] brought support f o r the v a l i d i t y of these f e a t u r e s i n p e r c e p t i o n . P l o s i v e s were d i f f e r e n t i a t e d to a s i m i l a r extent by the fea t u r e s £ a n t e r i o r ] . a n d [ coronal]- F r i c a t i v e s were d i f f e r e n t i a t e d to a g r e a t e r extent by the featu r e [ a n t e r i o r ] than the features [ c o r o n a l ] and [ d i s t r i b u t e d ] . 5 ) The models of c o n t e x t - s e n s i t i v e c o d i n g , motor theory and d i s t i n c t i v e f e atures c o u l d be used to e x p l a i n some aspect of the r e s u l t s o b t a i n e d . Further i n s i g h t i n t h i s area of speech pe r c e p t i o n would be gained by i n v e s t i g a t i o n of the f o l l o w i n g q u e s t i o n s : 1) Would voiced c l u s t e r s such as / b l / undergo the same types of percept u a l confusion when masked by noise as those observed f o r v o i c e l e s s c l u s t e r s ? 2) What types of perceptual confusions would be made i n response to c l u s t e r s of three elements such as /spr/? 3) How would p r e s e n t a t i o n i n continuous speech a l t e r the percept u a l confusions among consonant c l u s t e r s ? -98--99-4) Is i t p o s s i b l e to have p e r c e p t u a l confusions among vowel combinations such as g l i d e s and diphthongs? 5) What i s the e f f e c t of morpheme boundaries on pe r c e p t i o n of the c l u s t e r s presented? W i l l the sequence /f#p/ i n " c h i e f please" produce the same types of percept u a l confusions as the c l u s t e r /fp/ i n / I f p / ? 6) Would the pe r c e p t i o n of consonant c l u s t e r s be a l t e r e d i f the l i s t e n e r was unable t o produce such c l u s t e r s , e.g. as w i t h i n d i v i d u a l s demonstrat-in g neuromuscular d i s o r d e r s ? Answers t o these and s i m i l a r questions supplemented by the r e s u l t s of t h i s present i n v e s t i g a t i o n would increase the i n s i g h t i n t o the pe r c e p t i o n of consonant c l u s t e r s . P o s s i b l y w i t h that knowledge, the mechanism o p e r a t i v e i n the perception of speech might be more thoroughly understood. BIBLIOGRAPHY BOND, Z. (1971). "Units in Speech Perception," Working Papers in Linguistics, No. 9_ Ohio State, Columbus, Ohio. BRADY, P. T. (1971). "Need for the Standardization in the Measurement of Speech Level," J. Acoust. Soc. Amer., 50, 712-714. BROADBENT, D. E. and LADEFOGED, P. (1959). "Auditory Perception of Temporal Order," J. Acoust. Soc. Amer., 31, 1539. CHISTOVICH, L. A. (1960). "Perception of a Sound Sequence," Biofizyka, 5_, 671-676. CHOMSKY, N. and HALLE, M. (1968). The Sound Pattern of English (Harper and Row, Publishers N.Y.). COHEN, A. (1967). "Speech, Percepts and Linguistic Forms," in Models for the Perception of Speech and Visual Form Weiant Wathen-Dunn, Ed. (MIT Press, Cambridge), pp. 320-325. COOPER, F. S., DELATTRE, P. C. , LIBERMAN, A. M. , BORST, J. M. , and GERSTMAN, L. J. (1952). "Some Experiments on the Perception of Synthetic Speech Sounds," J. Acoust. Soc. Amer., 24_, 597-606. COOPER, F. S., LIBERMAN, A. M., HARRIS, K. S., and GRUBB, P. M. (1958). "Some input-output relations observed in experiments on the perception of speech," Second International Congress of Cybernetics, Namu. DELATTRE, P. C. , LIBERMAN, A. M., and COOPER, F. S. (1955). "Acoustic Loci and Transitional Cues for Consonants," J. Acoust. Soc. Amer., 27, 769-773. DENES, P., and PINSON, E. N. (1963). The Speech Chain Bell Telephone Laboratories. (Williams and wilkins. Co., Maryland). DENES, P. (1967). "On the Motor Theory of Speech Perception," in. Models for the Perception of Speech and Visual Form Weiant Wathen-Dunn, Ed. (MIT Press, Cambridge), pp. 309-314. FELLOWS, B. J. (1969). The DiscrdjTLJnation Process and Development (Pergamon Press, London). FLANAGAN, J. L. (1965). Speech Analysis, Synthesis and Perception (Academic Press, N.Y.). FRY, D. B. (1964). "Experimental Evidence for the Phoneme," in. In Honour of Daniel Jones David Abercrombie, D. B. Fry, P. A. D. MacCarthy, N. C. Scott, and J. L. M. Trim, Eds. (Longmans, London), pp. 59-72. -100--101-HAGGARD, M. P. (1967). "Models and Data in Speech Perception," in Models for the Perception of Speech and Visual Form Weiant Wathen-Dunn, Ed. (MIT Press, Cambridge), pp. 331-339. HALLE, M. , HUGHES, G. W. , and RADLEY, J. P. (1957). "Acoustic Properties of Stop Consonants," J. Acoust. Soc. Amer., 29, 107-116. HALLE, M. and STEVENS, K. N. (1962). "Speech Recognition: A Model and a Program for Research," IRE Transactions on Information Theory, V no: 17, 155-159. HALLE, M. and STEVENS, K. N. (1972). "On Phonetic Features," in Conference Record on 1972 Conference on Speech Communication  and Proceeding pp. 194-197. HARRIS, C. (1953). "A Study of the Building Blocks in Speech," J. Acoust. Soc. Amer., 2_5, 962-969. HARRIS, K. S. (1958). "Cues for the Discriirrination of American English Fricatives in Spoken Syllables," Lang, and Speech, 3L, 1-7. HEINZ, J. M. , and STEVENS, K. N. (1961). "On the Properties of Voiceless Fricative. .Consonants," J. Acoust. Soc. Amer., 3_3, 589-596. HIRSH, I. J. (1959). "Auditory Perception of Temporal Order," J. Acoust. Soc. Amer., 31, 759-767. HORII, Y., HOUSE, A. S., and HUGHES, G. (1971). "A Masking Noise with Speech Envelope Characteristics," J. Acoust. Soc. Amer., 49, 1849-1856. HOUSE, A. S., WILLIAMS, C. E., HECKER, M. H., and KRYTER, K. D. (1965). "Articulation-Testing Methods: Consonantal Differentiation with a Closed-Response Set," J. Acoust. Soc. Amer. , 37. 158-166. HUGHES, G.; W., and HALLE, M. (1956). "Spectral Properties of Fricative Consonants," J. Acoust. Soc. Amer., 28_, 303-310. JAKOBSON, R., FANT, C. G. M., and HALLE, M. (1963). Preliminaries  to Speech Analysis (MTT Press, Cambridge). JOOS, M. (1948). "Acoustic Phonetics," Language Monographs 24., p. 23. KLATT, D. (1968). "Structure of Confusions in Short-term Memory between English Consonants," J. Acoust. Soc. Amer., 44, 401-407. LADEFOGED, P., and McKU'E\'EY, N. P. (1963). "Loudness, sound pressure, and subglottal pressure in speech," J. Acoust. Soc. Amer., 35, 454-460. LANE, H. (1965). "The Motor Theory of Speech Perception: A Critical Review," Psychol. Rev., 72, 275-309. -102-IJBEFriAN, A. M. , DEIATTRE, P., and COOPER, F. S. (1952). "The Role 1 of Selected stimulus variables in the Perception of the Unvoiced Stop Consonants," Amer. J. of Psychol., 65_, 497-516. LIBERMAN, S. M., HARRIS, K. S., HOFFMAN, H. S., and GRIFFITH, B. C. (1957). "The discrimination of speech sounds within and across phoneme boundaries," J. Exper. Psychol., 54_, 358-368. LIBERMAN, A. M., KARRIS, K. S., EIMAS, P., LISKER, L., and BASTIAN, J. (1961). "An effect of learning on speech perception: The tilis elimination of durations of silence with and without phonemic significance," Lang, and Speech, 175-195. LIBERMAN, A. M., HARRIS, K. S., KINNEY, JOANN, and LANE, H. L. (1961). "The discrimination of relative onset time of the components of certain speech and non-speech patterns," J. of Exper. Psychol., 61, 379-388. LIBERMAN, A. M., COOPER, F. S., STUDDERT-KENNEDY, M., HARRIS, K. S., and SHANKWEILER, D. P. (1966). "Some Observations of the Efficiency of Speech Sounds," XVIII International Congress of Psycholofy, Moscow U.S.S.R. LIBERMAN, A. M., COOPER, F. S., STUDDERT-KENNEDY, M., and SHANKWEILER, D. P. (1967). "Perception of the Speech Code," Psychol. Rev., 74, 431-461. LIBERMAN, A. M., COOPER, F. S., STUDDERT-KENNEDY, M. , HARRIS, K. S. , and MacNEILAGE, P. F. (1967). "Some Observations on a Model for Speech Perception," in Models for the Perception of  Speech and Visual Form Weiant Wathen-Dunn, Ed." (MIT Press, Cambridge), pp. 68-88. LICKLIDER, J. C. R. (1952). "On the Process of Speech Perception," J. Acoust. Soc. Amer., 24, 590-594. LISKER, L. and ABRAMSON, A. (1964). "Stop Categorization and Voice Onset Time," Proceedings of the Fifth International Congress of Phonetic Sciences. LOTZ, J., ABRAMSON, A., GERSTMAN, L., INGEMANN, F., and NEMSER, W. J. (1960). "The Perception of English Stops by Speakers of English, Spanish, Hungarian and Thai," Lang, and Speech, 3_, 71-77. MILLER, G. A., and NICELY, P. E. (1955). "An Analysis of Perceptual Confusion among some English Consonants," J. Acoust. Soc. Amer., 27, 338-352. MILLER, G. A. (1956). "The Perception of Speech," in For Roman Jakobson Morris Halle, Ed. (Mouton, The Hague), pp. 353-360. -103-MILLER, G. A. (1962). "Decision Units in the Perception of Speech," IRE Transactions on Information Theory, V, no: 17, 81-83. MORTON, J. and BROADBENT, D. E. (1967). "Passive versus Active Recognition Models or Is your Homunculus really necessary?," in Models for the Perception of Speech and Visual Form weiant Wathen-Dunn, Ed. (MIT Press, Cambridge), pp. 103-111. PETERSON, G. E. and BARNEY, H. L. (1952). "Control Methods Used in a •Study of the Vowels," J. Acoust. Soc. Amer., 24, 175-184. STEVENS, K. N. (1960). "Towards a Model for Speech Recognition," J. Acoust. Soc. Amer. 32_, 47-55. STEVENS, K. N. , and HALLE, M. (1967). "Remarks on Analysis by Synthesis and Distinctive Features," in Models for the Perception of Speech  and Visual Form Weiant Wathen-Dunn, Ed. (MIT Press, Cambridge), pp. 88-102. . STREVENS, P. '(I960). "Spectra of Fricative Noise in Human Speech," Lang, and Speech, 3_, 32-49. STUDDERT-KENNEDY, M., LIBERMAN, A. M., HARRIS, K. S., and COOPER, F. S. , (1970). "Motor Theory of Speech Perception: A reply to Lane's Critical Review," Psychol. Rev., 77, 234-239. IJLDALL, E. (1964). "Transitions in Fricative Noise," Lang, and Speech, T_, 13-15. WARREN, R. and WARREN, R. (1970). "Auditory Illusions and Confusion," Scientific American, 223, 6, 30-36. WICKELGREN, W. A. (1966). "Distinctive Features and Errors in Short-term Memory for English Consonants," J. Acoust. Soc. of Smer. , 39_, 388-398. WICKELGREN, W. A. (1969a). "Context-Sensitive Coding, Associative Memory and Serial Order in Speech Behavior," Psychol. Rev., 76, 1-15. WICKELGREN, W. A. (1969b). "Context-Sensitive Coding in Speech Recognition, Articulation and Development," in Information Processing in the Nervous System K. N. Leibovic Ed. (Springer-Verlag, N.Y.). -104-APFENDLX A A l l combinations of the three f r i c a t i v e s w i t h the three p l o s i v e s . /sp/ / s t / /sk/ /ps/ / t s / /ks/ /fp / / f t / /fk/ /pf/ / t f / /kf/ Ifpl Ift/ Ifkl Ipfl Itfl Ikfl -105-APPENDLX B 54 S t i m u l i c r e a t e d from using c l u s t e r s i n three p o s i t i o n s 1) F i n a l /Isp/ / I s t / /Isk/ /Ips/ / I t s / /Iks/ / I f p / / I f t / / I f k / / I p f / / I t f / / I k f / /I/p/ / I / t / /I/k/ /Ip// / I t / / /Ik// 2) I n i t i a l / s p l / / s t l / / s k i / / p s l / / t s l / / k s l / / f p l / / f t l / / f k l / / p f l / / t f l / / k f l / / / p l / / / t i / / / k l / /p/I/ / t / I / / k / I / 3) Medial /I s p i n g / / I s t i n g / / I s k i n g / / I p s i n g / / I t s i n g / / I k s i n g / / I f p i n g / / I f t i n g / / I f k i n g / / I p f i n g / / I t f i n g / / I k f i n g / / I / p i n g / / I / t i n g / / I / k i n g / / I p / i n g / / I t / i n g / / I k / i n g / -106-APPENDIX C Randomized L i s t s used to Order Items on Experimental Tapes L i s t #1 C l u s t e r s i n F i n a l P o s i t i o n 1. I k f 2. I f p 3. I f k 4. I p / 5. I / t 6. I/k 7. Ipf 8. 1st 9. I/P 10. Ik/ 11. I f k 12. Ipf 13. I t f 14. I/P 15. Ik/ 16. Is p 17. I t s 18. I t / 19. I/P 20. Ips 21. I t f 22. I / t 23. I k f 24. 1st 25. Ip/ 26. Iks 27. 1st 28. I k f 29. I/k 30. I t f 31. Ips 32. I t f 33. Isk 34. Ips 35. Iks 36. I/k 37. I f p 38. Isk 39. I/k 40. I / t 41. I f t 42. Ipf 43. I f p 44. Ips 45. Iks 46. Isk 47. I f p 48. I f k 49. Ik/ 50. I t / 51. Ips 52. Isp 53. Isp 54. I / t 55. I t / 56. I f p 57. Ips 58. Isk 59. I t / 60. I t s 61. Ip/ 62. I f k 63. I f t 64. I t s 65. 1st 66. Ipf 67. I t s 68. I/P 69. I/k 70. I k f 71. I f k 72. I f t 73. Isk 74. I/P 75. Ik/ 76. Iks 77. I / t 78. Ik/ 79. Ipf 80. I/k 81. Isp 82. Iks 83. I t f 84. I f t 85. I / t 86. I t s 87. Ipf 88. I f t 89. Isp 90. I t / 91. Iks 92. I/P 93. 1st 94. I f k 95. 1st 96. I k f 97. I t s 98. Ikf 99. I f p 100. Ipf 101. Ik/ 102. Ipf 103. I f t 104. Ip/ 105. Isk 106. Isp 107. I f t 108. I t f -107-i L i s t #2 C l u s t e r s i n F i n a l P o s i t i o n 1. 2. I f k 3. I t s 4. Isp 5. I/P 6. I f p 7. Ips 8. Ips 9. Ips 10. Ip/ 11. I/k 12. I f k 13. Ip/ 14. Ik/ 15. Ipf 16. I t f 17. I/k 18. I t / 19. I/P 20. I / t 21. I t s 22. I f p 23. Ipf 24. I/k 25. I t s 26. Ik/ 27. 1st 28. Ip/ 29. I t / 30. Isk 31. I t f 32. I t / 33. I f t 34. I / t 35. 1st 36. I t s 37. I t s 38. I k f 39. I t / 40. I t / 41. Iks 42. Isk 43. 1st 44. I k f 45. I t f 46. I k f 47. I / t 48. I t s 49. I/k 50. Isp 51. I/p 52. I f p 53. Ipf 54. I k / 55. 1st 56. I/P 57. I k f 58. I f t 59. Ikf 60. Isk 61. I f k 62. I / t 63. I f k 64. Isk 65. I t f 66. Iks 67. I/k 68. I / t 69. Ik/ 70. Isp 71. I k f 72. I f p 73. Isp 74. Ipf 75. I f t 76. I f k 77. Isp 78. Ips 79. I / t 80. I f t 81. Ipf 82. Isp 83. I k / 84. Iks 85. Iks 86. I t / 87. I f t 88. 1st 89. Ips 90. Iks 91. I f t 92. I t f 93. 1st 94. I t f 95. Isk 96. Ips 97. I k f 98. I/P 99. Iks 100. 1st 101. Ipf 102. I f p 103. I f p 104. I/P 105. Isk 106. I f k 107. Ipf 108. I/k -108-L i s t #3 C l u s t e r s i n F i n a l P o s i t i o n 1. I t s 2. I k ^ 3. I f p 4. I p f 5. Ips 6. Ipf 7. Isp 8. Isp 9. I t f 10. I/k 11. I f k 12. I/p 13. 1st 14. I t s 15. I f t 16. Ipf 17. Ips 18. Isp 19. I/k 20. I t f 21. I/p 22. I / t 23. I t f 24. I f p 25. Ipf 26. Isk 27. Ipf 28. I f k 29. Isp 30. ifp 31. I/k 32. 1st 33. I / t 34. Ipf 35. I / t 36. I t f 37. I t s 38. Ipf 39. I k f 40. 1st 41. 1st 42. Iks 43. I t / 44. I f t 45. Ips 46. I t s 47. I / t 48. I t s 49. Isp 50. I k f 51. I k f 52. I t / 53. Iks 54. I k / 55. I t / 56. I f p 57. Ipf 58. I / t 59. Iks 60. I f t 61. Isk 62. I/p 63. I k f 64. Isk 65. I/p 66. I / t 67. Ips 68. I/p 69. 1st 70. Ik/ 71. I k f 72. Isk 73. I k f 74. I/k 75. Ipf 76. Ipf 77. 1st 78. I f t 79. 1st 80. I t f 81. Ipf 82. Isk 83. Ikf 84. I f t 85. Isk 86. I t s 87. Ikf 88. I f k 89. I f p 90. I f k 91. I f t 92. I/k 93. I/k 94. Isk 95. Ipf 96. Iks 97. I t / 98. I f p 99. Ikf 100. I t f 101. Ips 102. I f k 103. I f k 104. I t / 105. Itf 106. Ips 107. Iks 108. I f p -109-L i s t #4 C l u s t e r s 1. p s l 2. s k i 3. s t l 7. t s l 8. t s l 9. f p l 13. k s l 14. f t l 15. s p l 19. fpl 20. / k l 21. s k i 25. / k l 26. ps l 27. k f l 31. k f l ' 32. Pfl 33. s t l 37. s k i 38. f p l 39. p s l 43. fkl 44. k s l 45. t / I 49. f p l 50. sp l 51. f t l 55. P/I 56. kfl 57. ftl 61. k f l 62. k s l 63. f k l 67. kfl 68. f p l 69. s t l 73. s k i 74. s p l 75. k f l 79. k s l 80. kfl 81. P f l 85. / t l . ' 86. fpl 87. s p l 91. f p l 92. t s l 93. t f l 97. p f l 98. f t l 99. t / I 103. fpl 104. fpl 105. / k l I n i t i a l P o s i t i o n 4. ftl 5, fpl 6. t f l 10. f k l 11. s p l 12. t / I 1.6. f t l 17. f k l 18. f k l 22. f k l 23. s t l 24. ps l 28. t f l 29. fkl 30. Pfl 34. f p l 35. k f l 36. kfl 40. t f l 41. P f l 42. P f l 46. f t l 47. s p l 48. p s l 52. t / I 53. t s l 54. k s l 58. kfl 59. fpl 60. t / I 64. kfl 65. s k i 66. t f l 70. k f l 71. f t l 72. p s l 76. f k l 77. k s l 78. p f l 82. pfl 83. ftl 84. t s l 88. pfl 89. s t l 90. t / I 94. s t l 95. t s l 96. fkl 100. ftl 101. ftl 102. pfl 106. s k i 107. p f l 108. t f l -110-L i s t #5 C l u s t e r s 1. s k i 2. k f l 3. t f l 7. k s l 8. f k l 9. f p l 13. ftl 14. t/I 15. t/I 19. s k i 20. k/I 21. s t l 25. /Pi 26. s p l 27. p f l 31. t f l 32. k f l 33. k/I 37. t/I 38. P/I 39. k/I 43. s p l 44. P f l 45. f p l 49. k f l 50. psI 51. k/I 55. P f l 56. t f l 57. f t l 61. t/I 62. t s l 63. /k l 67. / k l 68. k s l 69. fpl 73. s p l 74. f p l 75. ftl 79. s p l 80. t f l 81. k/I 85. /k l 86. P f l 87. k/I 91. psI 92. P f l 93. k s l 97. /Pi 98. t s l 99. k f l 103. /p i 104. psI 105. ftl i n I n i t i a l P o s i t i o n 4. s t l 5. s k i 6. f k l 10. ;t"i u - f k l 12. t s l 16. /k l 17. k s l 18. t s l 22. psI 23. P f l 24. f t l 28. t f l 29. pfl 30. f k l 34. fpl 35. tfl 36. f k l 40. k s l 41. psI 42. f t l 46. J k l 47. s t l 48. s k i 52. fpl 53. psI 54. P/T 58. fkl 59. s k i 60. t s l 64. s p l 65. f p l 66. f t l 70. t f l 71. f p l 72. k f l 76. f k l 77. s p l 78. f t l 82. f p l 83. k s l 84. P/I 88. ftl 89. s t l 90. k f l 94. f t l 95. s t l 96. s k i 100. ftl 101. s t l 102. Pfl 106. tfl 107. t s l 108. pfl - I l l -L i s t #6 C l u s t e r s 1. t / I 2. fkl 3. f p l 7. t / I 8. s k i 9. fpl 13. f p l 14. f p l 15. ski 19. t f l 20. t f l 21. tfl 25. k f l 26. s p l 27. P f l 31. pfl 32. p f l 33. p f l 37. fpl 38. p s l 39. k f l 43. fpl 44. s t l 45. ftl 49. kfl 50. ftl 51. fpl 55. ftl 56. ftl 57. f t l 61. s k i 62. t s l 63. Pfl 67. k f l 68. s k i 69. kfl 73. k f l 74. Pfl 75. f t l 79. ftl 80. t s l 81. s p l 85. k s l 86. ftl 87. kfl 91. Pfl 92. s k i 93. kfl 97. P f l 98. p f l 99. pfl 103. s k i 104. k f l 105. fkl i n I n i t i a l P o s i t i o n 4. k s l 5. f t l 6. k s l 10. f k l 11. t f l 12. f k l 16. fpl 17. f t l 18. tfl 22. fpl 23. t s l 24. s p l 28. f p l 29. s p l 30. fkl 34. fkl 35. s t l 36. spl 40. kfl 41. t f l 42. tsl 46. kfl 47. p s l 48. t f l 52. fkl 53. p f l 54. f k l 58. pfl 59. f p l 60. t / I 64. t f l 65. s t l 66. k s l 70. f k l 71. s t l 72. p s l 76. f t l 77. p s l 78. s p l 82. k s l 83. k f l 84. s t l 88. p s l 89. p s l 90. k s l 94. f t l 95. f k l 96. t / I 100. t s l 101. t s l 102. f k l 106. f p l 107. fkl 108. s t l -112-L i s t #7 C l u s t e r s i n Medial P o s i t i o n 1. I s t l n 2. i s p l n 3. i f p l n 4. I k s l n 5. I f t l n 6. I s p l n 7. Ifkln 8. I p s l n 9. I s k l n 10. I s k l n 11. I p s l n 12. I/pIn 13. I p f l n 14. I k s l n 15. I k f l n . 16. I p f l n 17. I f t l n 18. I s t l n 19. I / t I n 20. I s t l n 21. Ip/In 22. I f t l n 23. I t f l n 24. Ip/In 25. I / k I n 26. I t s l n 27. I k s l n 28. I k s l n 29. I t / I n 30. I f p l n 31. I t s l n 32. I f t l n 33. I/pIn 34. I / t l n 35. Ip/In 36. I f k l n 37. I t / I n 38. I s k l n 39. I s p l n 40. I / k l n 41. I p s l n 42. I k f l n 43. I f p l n 44. I f k l n 45. I s k l n 46. I t / I n 47. I f k l n 48. I/pIn 49. I / k l n 50. I t / I n 51. Ipf In 52. I t f l n 53. I k s l n 54. I k f l n 55. I t f l n 56. I f p l n 57. Ip/In 58. I p s l n 59. I s t l n 60. I / k l n 61. I s p l n 62. I s k l n 63. I t s l n 64. Ik/In 65. I k f l n 66. Ip/In 67. I t f l n 68. I p s l n 69. I t f l n 70. I s k l n 71. I s p l n 72. I f p l n 73. I p f l n 74. Ipf In 75. Ik/ I n 76. I / t l n 77. I / t l n 78. I k f l n 79. I t s l n 80. I k s l n 81. I k / I n 82. I t s l n 83. I p s l n 84. I / k l n 85. I k / I n 86. I f t l n 87. I s p l n 88. 'IT t i n 89. I/pIn 90. I f p l n 91. I / p i n 92. I k / I n 93'. I p f l n 94. I s t l n 95. Ip/In 96. I k f l n 97. I / k l n 98. I t f l n 99. I t / I n 100. I f t l n 101, I f k l r , 102. I/pIn 103. I t / I n 104. I / t l n 105. Ik/In 106. I t s l n 107. I s t l n 108. I k f l n -113-L i s t #8 C l u s t e r s i n Medial P o s i t i o n 1. I s p l n 7. I f p i n 13. I s t l n 19. I f t l n 25. I f t l n 31. I s t l n 37. I / k l r , 43. I f p i n 49. I f t l n 55. I k f I n 61. I p f I n 67. I t f l n 73. I p f I n 79. I / k l n 85. I / t l n 91. I k f l n 97. I f p i n 103. I s p l n 2. I / k l n 8. I s p l n 14. I / k l n 20. I t s In 26. I t / I n 32. I p f l n 38. I f t l n 44. I t f l n 50. I p f l n 56. I p f l n 62. Iks I n 68. I t f l n 74. I p f l n 80. I k f l n 86. I f k l n 92. I f p l n 98. I t f l n 104. I f k l n 3. I s t l n 9. I t f l n 15. Ip/In 21. I f k l n 27. I f k l n 33. I f t l n 39. I t / I n 45. I s t l n 51. I s t l n 57. I s t l n 63. I k / I n 69. I k / I n 75. I / k l n 81. I f t l n 87. I s p l n 93. I f k l n 99. I f p l n 105. I / k l n 4. I s p l n 10. Iks In 16. I k f l n 22. I / k l n 28. IpsIn 34. I k / I n 40. Ip/In 46. I t / I n 52. I k / I n 58. I / t l n 64. Ip/In 70. I p s I n 76. I k / I n 82. I s p l n 88. I k f l n 94. I s k l n 100. Ip/In 106. I s p l n 5. Iks In 11. I f t l n 17. I / p I n 23. I t / I n 29. I k / I n 35. I f p l n 41. I k f l n 47. I t / I n 53. I k s l n 59. I/pin 65. I p s I ^ 71. I t / I n 77. I t s l n 83. I t s l n 89. I t s l n 95. I / p I n 101. I k s l n 107. I s k l n 6. Ips In 12. l / p l n 18. Ip/In 24. I t s l n 30. I / t l n 36. I t s l n 42. I s k l n 48. I k s l n 54. I/pIn 60. I / t l n 66. I / t l n 72. I / P i n 78. Ip/In 84. I s k l n 90. Ips In 96. I s k l n 102. I / t l n 108. I s k l n -114-L i s t #9 C l u s t e r s i n Medial P o s i t i o n 1. I f t l n 2. I k s l n 3. I t / I n 4. I / k l n 5. l / k l n 6. I f k l n 7. I k s l n 8. I t s l n 9. I s k l n 10. I/pIn 11. I s p l n 12. I p f l n 13. I s p l n 14. Ik/In 15. I k / I n 16. I s t l n 17. I t / I n 18. I p s l n 19. I k/In 20. Ip/In 21. I k f l n 22. I s t l n 23. I s k l n 24. I t s l n 25. I s t l n 26. I f t l n 27. I / k l n 28. I s k l n 29. I f p l n 30. I s k l n 31. I s t l n 32. I / k l n 33. I t s l n 34. I / k l n 35. I f k l n 36. I p / I n 37. I k f l n 38. I k f l n 39. I / t l n 40. I k / I n 41. I p f l n 42. I k s l n 43. I t f l n 44. I p s l n 45. I f t l n 46. I k f l n 47. I / t l n 48. I p s I n 49. I t s l n 50. I k / I n 51. I p s l n 52. I t f l n 53. I p s l n 54. I p f l n 55. I f p l n 56. I s p l n 57. I / p i n 58. Ip/In 59. I k f l n 60. I f k l n 61. I k s l n 62. I f t l n 63. I s p l n 64. I / P l n 65. I s k l n 66. I p / I n 67 . I t f l n 68. I s p l n 69. I f p l n 70. I / t l n 71. I s p l n 72. I t f l n 73. I t s l n 74. I s t l n 75. I t / I n 76. I t / I n 77. I t f l n 78. I p f l n 79. I / t In 80. Ip/In 81. I f p l n 82. I p / I n 83. I f t l n 84. I f k l n 85. Ip/In 86. I k / I n 87. I f t l n 88. I/pIn 89. I t s l n 90. I k s l n 91. I k f l n 92. I s t l n 93. I / t l n 94. I f p l n 95. I / k l n 96. I p s l n 97. I s k l n 98. I t f l n 99. I f k l 100. I k s l n 101. I / p i n 102. I t / I n 103. I t / I n 104. I/pIn 105. I f k l 106. I / t l n 107. I f p l n 108. I p f l n •US-APPENDIX D Example of Prepared Response Sheet Subjecrt's Initials No. Response. 19 No. Response 20 21 22 24 25 26 27 28 29 30 31 32 33 34 35 36 37 i 38 | 39 41 -116-APPENDIX E An example of the calculation for Student-t value for differences between mean correct responses. Number of Correct Responses for /// . clusters. Number of Correct Responses for /s/ clusters. Number of Correct Responses Squared Number of Correct Responses Squared X l x2 X 2  X l X 2 x2 -134 146 155 118 207 80 190 112 124 84 88 97 17956 21316 24025 13924 42849 6400 36100 12544 15376 7056 7744 9409 Sum of Correct Responses 840 695 126470 88229 Mean Correct Responses per cluster 140 116 Z x 2 = X 2 - (Sum of X.,) 2 Z x x 2 = 126470 - (840)2 y x , 2 = 88229 -N 1 6 2 Y. = 8870 v- 2 £ x 0 = 7725 = 8900 = 7700 D- r x x +s>2 (Nx + N 2 ) - 2 (1/N.j^  + 1/N 2 ) D-8900 + 7700 (1/6 + 1/6) (6 + 6) - 2 23.5 'est Mean^ - Mean2 °x ^Test 140 - 116 23.5 1.02 t-value of 1.38 indicates no statistically significant difference between the mean correct responses per cluster for /// clusters and /s/ clusters. -117-APPENDIX E An example of the calculation for Student-t value for differences between mean frequency for each type of error for perrnissible and irrpermissible clusters. Calculation of difference between Correct Responses for clusters in I n i t i a l position. Number of Number of Number of Number of Correct Correct Correct Correct Responses Responses Responses Responses for for Squared Squared Perrnissible Inpemissible Clusters Clusters X l X2 V X 2 x 2 100 22 10000 484 18 37 324 1369 48 9 2304 81 75 24 5625 576 20 400 24 576 28 784 2 4 2 4 67 4489 55 3025 50 2500 25 625 10 100 Sum of Correct Responses 241 nean Correct Responses per 60 cluster 375 27 18253 15017 X±2 - (Sum of X 1 ) 2 N, *"x \| "^est J"x 2 = 18253 - (241)2 Vx 2 = 15017 - ( 3 7 5 ) 2 H ' 14 27X.J2 = 3 7 3 3 = 3730 £ x 2 2 = 4 9 7 3 = 4 9 7 0 2 . r- 2 x„ i v +z> mx * N 2) - 2 ( 1 / H1 + 1 / N 2 J = Mean, - Mean„ D x N 3730 + 4970 16 -(1/4 + 1/14) est D-x = 13.21 = 60 - 27 13.21 2.54 t-value of 2.54 indicates statistically signficiant difference between the mean correct responses percluster for permissible and irrperrnissible clusters beyond the .05 level. 

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.831.1-0101707/manifest

Comment

Related Items