THE ACQUISITION OF ENGLISH VERB MORPHOLOGY BY NONNATIVE SPEAKERS Dissertation for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY EDWARD EUGENE HECKLER 197-5 ABSTRACT / THE ACQUISITION OF ENGLISH VERB MORPHOLOGY BY NONNATIVE SPEAKERS By Edward Eugene Heckler This study of the acquisition of English grammar by non- native speakers was based on both transformational and structural viewpoints. The transformational viewpoint provided the theory of language acquisition and the theory of English grammar while the structural viewpoint provided the testing and the analytical methodologies. The model for the present study was the psycho- linguistic research study on English morphology done by Jean Berko in I958 on native English-speaking children. A modification and expansion of Berko's study and of other similar studies, the present study consisted of three tests, each with seventy items: an oral productive test, a written productive test, and a written receptive test. For each verb structure being tested, a pair of questions was included, one with a nonsense word and one with a real word. Thirty-six foreign students studying English at Michigan State University served as the subjects. /, Twelve of them spoke Arabic as their native language; twelve SPOke Japanese; twelve spoke Spanish. Each of these language groups was Edward Eugene Heckler composed of four beginners, four intermediates, and four advanced students. Several patterns emerged from the study. The level of a student's proficiency in English gave a good indication of his per- formance; thus advanced students gave the most correct responses; the beginners, the fewest. The native language also gave an indi- cation of performance as the data showed an ascending Spanish-Arab- Japanese order in the number of correct responses. More correct responses were found on the written tests than on the oral test. Nonnative speakers tended to learn the structures of English verb morphology in a nonrandomized order: the infinitive (§9_nv) was acquired earlier than the gerund (ing_fly); tense (past), before tense (present s); the gn_of passive, before the be; the gn_of perfect, before the have; and the bg_of progressive, before the jgg, 0f the structures that can follow a modal, fly_and bg_were acquired before have, For present s, /s/ and /z/ allomorphs were acquired before /Iz/; for past, and for §n_of perfect /t/ and /d/ allomorphs were acquired before /Id/. THE ACQUISITION OF ENGLISH VERB MORPHOLOGY BY NONNATIVE SPEAKERS By Edward Eugene Heckler A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of English 1975 Copyright by EDWARD EUGENE HECKLER 1975 ACKNOWLEDGMENTS I should like to express my deep and grateful appreciation to Dr. Robert J. Geist, the chairman of my guidance committee, for his judicious advice and unstinting help throughout the course of this research. Thanks also are due to Dr. William W. Heist for his valuable suggestions. I would also like to thank Dr. Stephen Judy and Dr. John Yunck for their fine support and help. In addition, I would like to thank the staff of the English Language Center, in particular Dr. Ralph Barrett, Dr. Paul Munsell, Mrs. Wu Yi So, and Dr. James Stalker, for their excellent coopera- tion. A great deal of credit must go to all the students who so willingly participated in the testing; without their voluntary cooperation this research would not have been possible. Finally, I wish to thank my mother and my friends for their encouragement and continued moral support during the many years of my doctoral program. 11' lilll'l UUIHIUUUI lll'll TABLE OF CONTENTS LIST OF TABLES . INTRODUCTION The Transformational Theory of Native or First Language Acquisition . The Acquisition of a Second Language Concluding Remarks . . . Chapter I. THE RELATED LITERATURE The Berko Study . . Other Berko- -Type Studies Longitudinal Studies II. THE EXPERIMENTAL DESIGN Limitations of Berko's Study . The Experimental Design The Experimental Test . . Limitations of the Experimental Design. III. THE EXPERIMENTAL RESULTS . The Oral Test . . The Fill- In Test and the Multiple- -Choice Test The Control Group . . Statistics . . . Analysis by Level of Proficiency in English . Analysis by Native Language . . IV. SUMMARY AND CONCLUSION APPENDICES A. THE VERB MORPHOLOGY TEST. . B. STATISTICAL RESULTS BY LEVEL OF PROFICIENCY C. STATISTICAL RESULTS BY NATIVE LANGUAGE LIST OF REFERENCES iii Page iv KDU‘IN IIIII‘Ulll'll'lIllll‘lllll‘ lull! [II Table osooowasm-bw NN—Jc—l—d—l—J—l—l—ldu—l #ommwmmth—I LIST OF TABLES All Tense (Oral Test) All Modal (Oral Test) All Perfect (Oral Test) All Progressive (Oral Test) All Passive (Oral Test) All Infinitive (Oral Test) All Tense (Written Tests) . All Modal (Written Tests) . All Perfect (Written Tests) All Progressive (Written Tests) . All Passive (Written Tests) All Infinitive (Written Tests) Tense by Level Modal by Level Perfect by Level . Progressive by Level Passive by Level . Infinitive by Level . Gerund by Level Tense by Language Modal by Language iv Page 52 56 59 63 66 69 73 77 80 86 90 93 lOl lO4 lO7 llO ll2 114 ll5 ll9 123 [[llu'l‘ll‘ [[[lll‘l‘l‘ll‘ll‘ Table 22. 23. 24. 25. 26. Perfect by Language . Progressive by Language Passive by Language . Infinitive by Language . Gerund by Language Page l25 l28 l3l I33 l34 [HUN U[[[I[[[‘|‘[ll[.llfl INTRODUCTION Studies concerning the acquisition of language comprise two main areas of investigation: those on a child's acquisition of his native language and those on second or foreign language acquisition. In general, research in foreign language learning is based on behavior- ist, structural linguistic theories, while most of the work in child language acquisition stems from concepts and hypotheses drawn from the rationalist, transformational framework. (Falk 1974: 5) Child language studies usually provide a descriptive analysis of the acquisition process with little or no consideration being given to any pedagogical implication. Second language studies, on the other hand, are generally undertaken in order to improve classroom instruction. Because of these differing theoretical bases and interests, research studies remain relatively dichotomized; consequently, insights gained in one area remain unnoticed by scholars working in the other. While differences do exist between the learning of one's native language and the learning of a foreign language, information gained by investigating one field may be "gf_relevance" to the other (Natalicio 1969: 9). A comparison of relevant data in both areas may result in a clearer understanding of the language learning process. Only a few studies (e.g., Natalicio 1969, Martinez-Bernal 1972) have tried to bridge this gap. By using 1 [[[Il‘ll‘Il-Il‘f‘l'll [[.[[[l[lllllllll [[(lll’llll Ulll‘l'l child language methodology, the present study of second language acquisition will be another attempt to bridge it. The Transformational Theory of Native or First Language Acquisition Of all the phases of child development, the learning of lan- guage has traditionally attracted most attention because of the complexity of language and the apparent ease and swift- ness of learning. (Carroll 1961: 331) Such "apparent ease and swiftness" in learning "the com- plexity of language" has interested psycholinguists greatly. In recent years they have conducted longitudinal studies on child language acquisition, some of which have lasted for several years. The children generally are not over seven years old because after this age they have mastered for the most part the phonology and the grammar of the native language. After tape-recording and ana- lyzing the data, the psycholinguists construct a descriptive grammar of each child's language. One of the psycholinguists' goals has been the formulation of an accurate, detailed model of language acquisition. One cur- rent theory of language acquisition, the transformationalist one, will serve as the model for this study. The following ideas com- prise the basic transformational model: 1. language acquisition is species specific; 2. language acquisition is biologically determined and, hence, innate; and 3. a universal, maturational pattern of acquisition occurs. The capacity for language acquisition remains a species- specific characteristic of homo sapiens; only man can acquire language. Most animals have developed systems of communication, some of them extremely complex. But none of these sytems repre- sent even "primitive stages of human communication" (Lenneberg 1964: 71). Recent experiments with Washoe and other chimpanzees may require new definitions of species-specificity (see Linden 1974). Being species-specific, the capacity for language acquisi- tion is biologically determined. Lenneberg (1967a: 394) states that the rocesses by which the realized, outer structure of a natural language comes about are deeply-rooted, spec1es- specific, innate properties of man's biological nature. Language acquisition, an "innate" property of man, consequently, cannot be culturally learned behavior; the environment plays no critical role in language learning except in the selection of the specific language being learned. If the language had to be "learned" from the environment, the physically handicapped and the mentally regarded would likely be unable to acquire more than the most "primitive stages of human communication." But this is not true. The congenitally blind acquire language nearly as well as the average person, and the con- genitally deaf, despite extreme physical disability, can master reading, writing, and even sign language. Intelligence has little bearing on language development since even the mentally retarded acquire language: Children whose IQ is 50 at age 12 and about 30 at age 20 are completely in possession of language though their articulation may be poor and an occasional grammatical mis- take may occur. (Lenneberg 1964: 80) Finally, according to the transformationalist theory, a universal, maturational pattern occurs in language development in every culture. All the evidence suggests that the capacities for speech production and related aspects of language acquisition develop according to built-in biological schedules. They appear when the time is ripe and not until then, when a state of . . . "resonance" exists. The child somehow becomes "excited," in phase with the environment, so that the sounds he hears and has been hearing all along sud- denly acquire a peculiar prominence. (Lenneberg 1967b: 8) Children in every culture throughout the world appear to be endowed with this state of "resonance," which is so necessary for the acquisition of language. This state generally lasts from the sec- ond to the twelfth year of life.1 After that time "progress in language development usually ceases" (Lenneberg 1967b: 9). Child language data suggest that language learning occurs in certain stages. Many linguists have agreed on the general order in which the sounds are acquired; for example, /p/ and /m/ are nearly always acquired before /d/ (see Carroll, 1961, Smith and Miller 1966, McNeill 1970). Linguists have begun establishing the stages of grammatical development. Two such stages are the holophrastic and the IArnold Gesell (1941) in Wolf Child and Human Child docu- mented the language difficulties that confronted Kamala, the "wolf" girl who missed the early years of this critical state of "reso- nance." telegraphic stages. Before his first birthday a child usually utters single, meaningful words like /mama/ and /dada/. Before this time, the child's utterances have been random, meaningless cries and sounds. Since these utterances are now meaningful, language acquisition may be said to have begun. During the holo- phrastic stage, a one-word utterance indicates a thought; "dog," for instance, may mean "Do you see the dog?" Before his second birthday, a child will form two-word sentences such as "see dog" to express the same thought. This stage in which a few words express a thought is called the telegraphic stage. The Acquisition of a Second Language_ The present study concerns the acquisition of English by second-language learners, not bilingual learners. By "second language acquisition" we mean the acquisition of another language after having acquired the basics of the first, whereas "bilingual acquisition" is the acqui- sition of two languages simultaneously. (Dulay and Burt 1972: 235) The second language may be called the target language to differenti- ate it from the native or primary language, the "mother tongue“ that has been learned first. Children often learn a target language with relative ease. Most parents who have lived abroad have marveled at how easily their children pick up a foreign language, and per- haps have wondered about their child's unusual talent. Many children, without the benefit of formal classroom instruction, learn the language of a new country in the first year they are there. (Dulay and Burt 1972: 235) III I $(ll’l‘ll‘. It...“ [[[I‘ llll-‘l Although children may learn a foreign language with remarkable ease, their parents do not. Lenneberg (1967a: 176) states that most individuals of average intelligence are able to learn a second language after the beginning of their second dec- ade, although the incidence of "language-learning-blocks" rapidly increases after puberty. Also automatic acquisi- tion from mere exposure to a given language seems to disappear after this age, and foreign languages have to be taught and learned through a conscious and labored effort. Foreign accents cannot be overcome easily after puberty. However, a person can_learn to communicate in a foreign language at the age of forty. ' / At this point a weakness in the transformationalist model r, f, LIZ. arises: if the state of "reSonance" is indeed critical to the //__., language-learning process, no adult would theoretically be able to learn a second language but, as Lenneberg says, an adult gan_with "conscious" effort. Lenneberg (1967a: 176) tries to explain away this seeming contradiction of fact and theory by saying: This does not trouble our basic hypothesis on age limita- tions because we may assume that the cerebral organization for language learning as such has taken place during childhood, and since natural languages tend to resemble one another in many fundamental aspects . . . the matrix for language skills is present. Although plausible, this does not seem, a priori, any more con- vincing than the behaviorists' explanation. Carroll (1971: 109), while agreeing "in some measure" with the transformationalist model, offers a behaviorist explanation: The evidence for a "critical period” and a decline in language acquisition ability during the middle school years is not strong, however, and even if there is some decline, I am not persuaded that one must appeal to biology to explain it. An alternative hypothesis about this decline is that it is due to the consolidation of the habits estab- lished in primary language acquisition and their interfer- ence with the acquisition of new habits. Further, it may be that the large individual differences in foreign language aptitude that can be observed reflect individual differences in the rate of this decline. Carroll stresses the difficulty of acquiring new language habits after the establishment of the first ones. Other explanations can be given. Possibly the state of "resonance" remains critical only for first language learning. This puzzling problem of adult lan- guage learning clearly indicates the need for additional research. Besides possible age limitations, other differences exist in the learning of a target language. The second language learner is exposed to a smaller amount of carefully controlled linguistic data while the child is exposed to a larger amount of generally uncontrolled data, some of which may include nonsense words and "nonsentences" (e.g., Tabakowska 1969: 47-8). In addition, in second language learning (i) the individual's cognitive development is at a later and more advanced stage; (ii) he is already in possession of the grammatical structures of a language . . . (iii) he already possesses concepts and meanings. (Jakobovits 1968: 269) Other factors playing a role in second language learning include the learner's motivation, his general language aptitude, his previous language learning experience, the status of the native and the target languages, the usefulness of the target language in communication, and the method of instruction (see Weinreich 1970: 3-4, 72-80; Lambert 1972: 160-96). The theory of contrastive analysis states that the second language learner will tend to use his native language structures in his second lan uage speech, and where structures in his first language (L1 and his second language (L2) differ, he will goof. For example, in Spanish, subjects are often dropped, so Spanish children learning English should tend to say Wants Miss Jones for He wants Miss Jones. (Dulay and Burt 1972: 236) In short, there is interference, "the use of elements from one lan- guage while speaking or writing another" (Mackey 1965: 239). This usually occurs during the acquisition of a target language. Three kinds of interference exist: (1) phonological, (2) syntactic, and (3) lexical. Phonological interference often consists of under-differentiation of phonemes. A Japanese speaker, for example, may pronounce both the /r/ and the /1/ in English the same; he under-differentiates these two English phonemes. In Japanese these two sounds are nondistinctive (i.e., allophonic) but in English they are distinctive (i.e., phonemic). Syntactic interference often results in incorrect word order in the target language as this example shows: A German speaker says in English this woman loves the man on the model of German diese Frau liebt der Mann, intend- ing to communicate the message “the man loves this woman," but producing the opposite effect. (Weinreich 1970: 37) Errors of verb tense, subject—verb agreement, and idiomatic usages of prepositions may also occur. “ Lexical interference usually involves the inappropriate transferral of a word from the native to the target language; for example, a German speaker, substituting Fleisch for meat, may say, "My Fleisch (or flesh) is cold." Native words may be mistakenly identified; for instance, a nonnative German speaker, seeing some German goods labeled "DIE DEUTSCHEN WAREN FUR INDIA," translated it as "The Germans were for India" instead of the more appropriate "German goods for India" (Kelly 1959: 225).2 To minimize such possible interference, drills based on the distinctive contrasts in phonology and syntax of the native and the target languages are often implemented in the classroom. But native language interference may not create as many problems for the sedond language learner as has often been assumed. Jakobovits (1969: 65) notes that similarities in the native and the target languages may result in "positive transfer" and a speeding up of the language learning process; this does not rule out, a priori, interference from dissimilar elements. Recently some researchers have questioned the usefulness of instruction based on the theory of contrastive analysis. They believe this theory “inadequate, theoretically and practically, to predict the interference problems of a language learner" in the area of syntax (Whitman and Jackson 1972: 40). They do not rule out the possibility that contrastive analysis may prove useful for phonology. Additional research is needed to establish more accurately the extent of interference in second language learning. Concluding Remarks This present study of the learning of English verb mor- phology by 36 nonnative speakers is based on both transformational 2See also Weinreich (1970: 7-71) for a detailed analysis of interference. 10 and structural viewpoints. The transformational viewpoint provides a theoretical model of language acquisition and a theoretical model of English grammar while the structural viewpoint provides testing and analytical methodologies. Serving as the model for the present investigation, the psycholinguistic research study on English morphology done by Jean Berko in 1958 is deeply rooted in transformational theory, but it utilizes a basically structural testing methodology. By this rather unorthodox joining of two com- peting theories, the present investigation of second language learning will, hopefully, reveal new insights about the language- learning process. CHAPTER I THE RELATED LITERATURE The Berko Study According to Noam Chomsky (1970: 43) and the other trans- formationalists, a grammar is not a description of the performance of the speaker, but rather of his linguistic competence . . . a description of competence and a description of perform- ance are different things. By competence Chomsky means the unconscious, internalized knowledge knowledge of the phonological, morphological, syntactic, and lexi- cal rules of a language. Performance denotes the use of this internalized knowledge. Measuring this competence is difficult since "a direct description" of a child's “actual verbal output“ reveals only performance. If anything far-reaching and real is to be discovered about the actual grammar of the child, then rather devious kinds of observations of his performance . . . in many different kinds of circumstance will have to be obtained . . . to determine . . . his underlying linguistic competence at each stage of development. (Chomsky 1970: 44) One such "devious" observation of performance was con- ducted by Jean Berko. Investigating the acquisition of English M‘ morphology by native English-speaking children, Berko (1958: 150) stated the rationale underlying her research: 11 12 . we undertake to discover the psychological status of a certain kind of linguistic description. It is evident that the acquisition of language is more than the storing up of rehearsed utterances, since we are all able to say what we have never before heard. In bringing descriptive linguistics to the study of language acquisition, we hope to gain knowledge of the systems and patterns used by the speaker. Berko, in other words, wanted to ascertain the children's uncon- scious knowledge of English morphology. On her test Berko asked her subjects to inflect correctly several nonsense words. Berko (1958: 150) defended her use of these nonsense words by arguing: . if the subject can supply the correct plural ending, for instance, to a noun we have made up, he has inter- nalized a working system of the plural allomorphs in English, and is able to generalize to new cases and select the right form. If a child knows that the plural of witch is witches, he may simply have memorized the plural form. If, however, he tells us that the plural of *gutch is *gutches, we have evidence that he actually knows, albeit unconsciously, one of those rules which the descriptive linguist, too, would set forth in his grammar. 0f the morphological inflections that occur in English, Berko tested for the plural and the possessive of the noun, the comparative and the superlative of the adjective, the third person singular of the present tense, the past tense, the progressive, and the derivational suffixes -gr_(as in *zibber) and -1gt_(as in *wuglet). The allomorphs of the noun plural morpheme, the possessive morpheme, and the third-person-singular present-tense morpheme are /Izmzms/. They occur in these environments: /12/ occurs after stems ending in the sibi- lants /E j s z E 5/ as in buzzes and wishes; 13 /2/ occurs after stems ending in vowels or the voiced consonants /b d g d m n g 1 r/ as in :99s_and $1.55 /s/ occurs after stems ending with the voice- less consonants /p t k 0 f/ as in gyps_and £132: These allomorphs are phonologically conditioned; that is, the allo- morph is chosen on the basis of the phonological environment (i.e., the final sound of the stem). The rules are regular and predictable. Some of the allomorphs of the past tense morpheme are phonologically conditioned. The phonologically conditioned allo- morphs of the past tense /Id~d~t/ occur in these environments: /Id/ occurs after stems ending in It d/ as in wsgssd_and ggjgsg; /d/ occurs after stems ending in vowels or the voiced consonants /b g j d v z i m n g l r/ as in Lines and 2110129; /t/ occurs after stems ending in the voiceless consonants /p k c 0 f s g/ as in gssgsg_and fishsg, These allomorphs form the so-called "weak" or regular verbs in English. Allomorphs of the past tense morpheme may be morphologi- cally conditioned; that is, the allomorph is not chosen on the basis of the phonological environment of the stem. These so-called "strong" or irregular verbs like gsgg_and swsm_must be memorized individually since the final sound of the uninflected stem pro- vides no clue to the past tense allomorph. While adult native 14 speakers have unconsciously internalized the rules for phonologi— cally conditioned allomorphs, they have only acquired the morphologically conditioned ones by a conscious memorization of individual allomorphs. Berko's psycholinguistic testing procedure was simple. The examiner showed the child a picture of a cartoonlike character per- forming a certain act and then asked for a one-word oral response to statements about the picture. The examiner, for example, showed the picture of a man dangling an object on a string and told the child: This is a man who knows how to bod. He is bodding. He did the same thing yesterday. What did he do yesterday? Yesterday he . (Berko 1958: 156) A response like bodded indicated the child had internalized the grammar rules but a response like ssg_revealed the lack of such internalization. All responses were written down in phonemic nota- tion and recorded on tape. Later the tape was replayed to verify the accuracy of the notation. Berko's preschoolers and first graders showed "consistency, regularity, and simplicity" in their acquisition of English mor- phology: Where they provided inflexional endings, their best per- formance was with those forms that are the most regular and have the fewest variants. With the morphemes that have several allomorphs, they could handle forms calling for the most common of those allomorphs long before they could deal with allomorphs that appear in a limited distribution range. (Berko 1958: 176-7) 15 The progressive morpheme igg_was the most consistently learned 1 Testing for the /12/ present tense allomorph, verb morpheme. Berko discovered that only half of the children used it. The other half gave uninflected forms. Of the three past tense allo- morphs, the children responded more accurately with /t/ and /d/ than with /Id/. If they gave any of these allomorphs as the answer, they usually used it appropriately. They occasionally gave progressive forms (e.g., was ricking) for the simple past (e.g., ijsksg). They gave uninflected forms (e.g., risk), too. For the irregular past morpheme of [igg_no preschooler and only a few first graders gave {Egg} most gave the phonologically conditioned [igggg, Among nouns and possessives /s/ and /z/ allomorphs were better learned than /Iz/. The children gave many uninflected forms, too. Three rare, inappropriate inflections occurred: /hifiz/, /ga€§/ and /kaeis/ (Berko 1958: 162-3). Other Berko-Type Studies Berko's psycholinguistic study of child language acquisi- tion pioneered the way for later studies of competence. In a Berko-type study of the acquisition of English morphology Bellamy and Bellamy (1970) tested 160 school-age children from kindergar- ten through the fourth grade. Their productive phase of testing 1Berko (1958: 157) supplied bs_in her one progressive item: This is a man who knows how to zib. What is he doing? He is In subsequent discussions of the data for progressive in other studies, bg_was included for each test item unless it is noted otherwise. 16 consisted of two tasks: a forward formation task like Berko's in which the child was asked to generate an inflected form of an uninflected word and a backward formation task (not in Berko) in which the child was asked to either generate the uninflected form of an inflected word or to provide a different inflected form. In the comprehension phase the child was asked which pic- ture best fitted a supplied inflected nonsense form. Second, shown one picture, he was asked which of two inflected nonsense words best described that picture. (Bellamy and Bellamy 1970: 202) The Bellamys' data indicated that the /s/ and the /z/ allo- morphs of the plural, the possessive, and the present tense morphemes were mastered before the /Iz/ allomorphs. The /s/ and the /z/ allomorphs were learned at the same rate except in the one backward formation task involving present tense in which the kindergarteners and first graders experienced more difficulty with /s/ than /z/. In backward formation tasks most children formed the progressive more easily for past tense verbs inflected with /Id/ than for those inflected with /t/ and /d/. No overall pattern resulted for either the progressive or the past tense in forward formation tasks. In the comprehension phase the data indi- cated a late mastery of many tasks: . morphological inflection comprehension does not seem to outstrip production by as far as one might expect. At about age eight children seem to be sensitized to morpho- logical inflections and then seem to make rapid progress in mastering both comprehensional and productional tasks at about the same time. This concurrent mastery does not fit the pattern usually found in language development. (Bellamy and Bellamy 1970: 210-1) This late mastery of comprehension tasks appeared most evident in the past tense tasks which proved to be too difficult for even 17 the fourth graders who had, nevertheless, performed well on all other comprehension tasks. In a Berko-type study of twenty preschool children Foley and Locke (1971) noted the late acquisition of the past tense. They found statistically "nonsignificant differences" in the rela- tionship between the child's knowledge of phonological rules and his knowledge of morphological rules. They found "nonsignificant differences" in "the relationship between phonological competence and performance" (Foley and Locke 1971: 261). A Berko-type morphophonemic study by Baird (1973: 229) suggested that ”the principles governing use of the inflectional endings are re—learned within each morpheme." He investigated the children's acquisition of the phonological rule covering the plural, the possessive, and the present tense morphemes (i.e., the rule for the /s/, /z/, and /Iz/ allomorphs).2 Baird tested the validity of the hypothesis that a child's internalization and mastery of a phonological rule for one morpheme meant the simultaneous internalization and mastery of this rule in all cases, including those involving other morphemes. According to this hypothesis, if a child knew the phonological rule for the plural allomorphs, he then knew this same rule for both the pos- sessive and the present tense allomorphs. Baird's findings did not support this view. His data showed that the child's mastery 2Baird found the following kinds of errors: uninflected forms, sound substitutions (e.g., /tovz/ as the plural of /tof/), and rare inappropriate inflections (e.g., /gefIz/ and /vazIz/) (personal communication, 1974). 18 of the phonological rule for one morpheme (e.g., the plural) did not mean the simultaneous mastery of this rule in the case of other morphemes (e.g., the possessive and the present tense); instead, the child had to relearn the rule for each morpheme. Children do not generalize phonological rules across morpheme boundaries. In addition to being given to middle class children, Berko- type tests have been given to the socially and the economically disadvantaged. Some educators such as Bereiter and Engelmann (1966: 41) contend that cultural deprivation, "a lack of those particular kinds of learning that are important for success in school," can be found in lower class children. Language depriva- tion represents one form of social deprivation: Many disadvantaged children of preschool age come very close to the total lack of ability to use language as a device for acquiring and processing information. (Bereiter and Engelmann 1966: 39) Linguists refute this idea of language deprivation; they say the real issue is dialectal differences (see Williams 1971). Shriner and Miner (1968) administered to 25 culturally advantaged and 25 culturally disadvantaged preschool children a test consisting of a productive part like Berko's and a receptive part on noun plurals. Mental age was controlled. Although only some allomorphs of the plural, the present tense, and the past tense morphemes were included as test items, Shriner and Miner's data revealed no statistically significant difference in the use of English morphological rules between the two groups. In a study of 145 kindergarteners from rural Florida Vogel (1970) investigated the morphological competence of white Y I l l Y I..II .I| II U I 19 and black children. Using a slightly modified Berko test, Vogel (1970: 52) concluded: Significant differences were found that were attributed to the variables of intelligence and of race. No signifi- cant difference was found that could be attributed to the variable of sex alone. In analyzing interaction effects, significant differences were found between and among all variables--intelligence, race, and sex. Among verb allomorphs the kindergarteners learned /z/ before /Iz/ and /d/ before /Id/. They gave nonstandard responses more often for the verbs than for the nouns. Their most common nonstandard response was the uninflected form of an inflected word. In 35 out of 38 group tabulations the black children answered with the unin- flected form more often than the whites did. Both male and female white children gave a standard response more often than did their male and female black counterparts. These differences in responses among the blacks and the whites resulted because of dialectal dif- ferences. Whereas standard English regularly adds inflectional endings to denote the present and the past morphemes, black English often does not. Except for a brief conment, Vogel failed to explore this possibility of dialectal differences. Several shortcomings in the Vogel study sharply reduce the number of generalizations that can be made from it. The most noticeable weakness concerns her data charts. Having classified all responses as either standard or nonstandard, Vogel recorded them by sex and race in charts in the appendix. All responses were recorded in standard orthography, but Vogel did not indicate how she had transcribed the variant allomorphs of a single mor- pheme. Apparently she listed those responses with the correct 20 allomorph as standard responses and all others as nonstandard but the reader cannot be certain. Her notation should have been made clearer by use of phonemic transcription or the inclusion of com- mentary concerning her answers in standard orthography. In her discussion of the data, Vogel made an erroneous statement about the past tense allormorph of one nonsense verb and, thereby, added to the confusion concerning her data transcription. Vogel (1970: 57) stated: Two items (8, 10) are in the language category requiring, in linguistic symbols, the allomorph /ad/, and four are in the category requiring /-d/, after stems ending in voiced sounds. Careful analysis of her test items and charts reveals that she included risk./r1k/ as one of the four nonsense verbs taking /d/. But [isk_should take /t/, not /d/, since its stem ends in a voice- less stop. Vogel, though, never stated that any verb on the test required /t/. Risk_appeared in the Berko (1958: 164-5) study in which a high percentage of children, 73% in fact, inflected it with /t/. Berko reported no instance of /d/ with risk, Vogel never directly stated which allomorph the children used with gigs but her earlier comment referred to /d/. Clearly, /rIkt/, not /rIkd/, is the standard response. One nonstandard response, E12525; points to a possible variant past tense allomorph, either /t/ or more likely /It/, which would be a variant of /Id/ (Vogel 1970: 89). If this is a variant pronunciation of the past tense allomorph, Vogel should have noted that the word was inflected for past tense but an 21 inappropriate allomorph was used. Vogel never indicated such a possibility. Vogel, at times, failed to discuss her results as thor- oughly as she might have; e.g., she failed to comment about [isisi_ as a possible past tense variant form. Even more clearly, bing-ed, given as the nonstandard response by 2.9% of the white females for the past tense of siggg represents the use of /Id/ (Vogel 1970: 96). Or is it /d/? Another possible past tense response, bsgg, was given by 4.7% of the black males and 6.2% of the black females (Vogel 1970: 97). Bsgg_could be patterned after the ring-rang-rung paradigm. Berko (1958: 165) indicated an awareness of this possi- bility: Adults clearly felt the pull of the irregular pattern, and 50% of them said *bang or *bung for the past tense of *bing, while 75% made *gling into *glang or *glung in the past. Only one child of the 86 interviewed on these items said *bang. One also said *glang. At one point Vogel (1970: 56) was vague when she commented on "the nuances of verb tense formation" concerning spow. A 1973 study by Ramer and Rees explored dialect differences of blacks and whites. They administered the Berko test, with modifications, to black children ranging in age from five to fif- teen years in several preschool, kindergarten, first grade, fifth grade, and eighth grade classes in the New York City schools. For each morphology item on the test, the dialects of standard American English and black English use different morphemes. Ramer and Rees (1973: 575-6) learned that 22 for the morphemes examined, the black children of low socioeconomic background in the population of this study know and use the rules of morphological construction of both black English and standard American English. As the children grow in age and maturity, they demonstrate increased use of standard American English forms, but in no case in this study did even the oldest group, the eighth-graders, use any of these forms to the exclusion of the alternate black English forms. These children knew /s/ and /2/ present allomorphs better than /12/ and they knew the /d/ past allomorph better than either the /t/ or the /Id/ forms. They used black English forms more frequently in the present tense than in the past. First graders who showed a slightly greater use of black English forms in the past tense (68% to 61.5% in the present) proved to be the only exception to this trend (Ramer and Rees 1973: 572). Berko-type tests have been given to the physically handi- capped and the mentally retarded. One study by Goodglass and Berko (1960) of the language acquisition by agrammatic and nonagrammatic aphasics revealed an order of learning differing from the normal one. As previous studies have indicated, nonaphasic subjects experienced the most difficulty with /Iz/ and /Id/. The aphasics, in contrast, were more influenced by "grammatical function" than phonology (Goodglass and Berko 1960: 266). For the possessive and the present tense inflections, they tended to respond more accu- rately with /s/ or /2/ than /Iz/ but for these same three allomorphs of the plural they showed essentially no difference in learning. On past tense items the aphasics, in contrast to nonaphasic chil- dren, performed slightly better on inflected verbs with /Id/ than those with /t/ or /d/. 23 Cooper (1967), in a comparison of deaf and hearing females, found the progressive less well learned than the 1958 Berko study had indicated. But, like Berko, he found past tense inflections better learned than present tense ones. On the receptive part both deaf and hearing females knew the /5/ present allomorph better than the /Iz/; however, the hearing females correctly produced /5/ 1% of the time more often than /Iz/, whereas the deaf correctly produced /Iz/ 7% of the time more often than /s/ (Cooper 1967: 83). The language acquisition of the mentally retarded has been studied by several researchers (Newfield and Schlanger 1968, Dever and Gardner 1970, Dever 1972, and Lovell and Bradbury 1971). In a study containing both real and nonsense words Newfield and Schlanger (1968) concluded that normal subjects showed marked superiority over the mentally retarded in the rate of morphological acquisi- tion, but all subjects tended to acquire the inflections in the same order. Both groups responded better to real words than to nonsense syllables. Items of verb morphology revealed the greatest significant differences in the performance of normal and retarded subjects. Both tended to learn /t/ and /d/ before /Id/. The retarded, however, scored 23% correct for /Id/ items involving real words but only 17% on /d/ ones (Newfield and Schlanger 1968: 698). The normal subjects excelled most noticeably over the retarded in the areas of the morphologically conditioned past tense (rang, sang) and the /d/ allomorph. The mentally retarded scored low on items requiring /Iz/, the only present tense allomorph on 24 the test. Overall, the Newfield and Schlanger results paralleled those of Berko (1958). Dever and Gardner (1970) studied two groups of educable mentally retarded boys. One group, the MRMA, was paired with the normal boys on the basis of mental age while the other group, the MRCA, was paired with the normal boys on the basis of chronological age. The normal boys consistently surpassed both the MRMA's and the MRCA's in the number of correct responses. The MRMA's outdid the MRCA's on many items, especially those for /Id/ and [359, Research conducted by Lovell and Bradbury (1971) on educa— tionally subnormal special school children generally supported the findings of Berko, Newfield and Schlanger, and Dever and Gardner. The progressive was learned early; /t/ and /d/ were acquired before /Id/; and /Iz/ was learned very late. Correct responses to real words occurred more frequently than those to nonsense words. The subnormal and retarded did not score as well as normal subjects. Other Berko-type studies have been carried out. Investiga- ting the acquisition of plurals by children, Anisfeld and Tucker (1967), Koziol (1971), and Graves and Koziol (1971) learned that /s/ and /z/ plural allomorphs were acquired before /Iz/. Investi- gating plural stem endings, Solomon (1972: 50) found that the phonetic characteristics of the stem endings at least partially determine whether young children will inflect them correctly. For example, children tend to omit inflections altogether on stems ending in fricatives or affricates, regardless of which allomorph would be appro- priate for the stem. 25 Children experienced difficulty adding either the /s/ to /O/ and /f/ stems or the /z/ to /d/ and [0/ stems. No conclusions could be reached concerning /Iz/. Kernan and Blount (1966) used Berko's methodology to study the acquisition of Spanish noun and verb inflections by Mexican children. Because of vast differences in inflections in the English and Spanish languages, their data do not necessarily illuminate the acquisition of English morphology. The Kernan and Blount study, however, can be useful in comparative studies of the acquisition of the first and the second languages by Spanish speakers. Of all researches using the Berko methodology, only three "“‘ M xH--_~w' ‘ '""‘ "_L\_ W“ concern the acquisition of English morphology—by'non-English speak- ‘ Y -M._m ’ MW” ——.—” _ers: ”Natalicio (1969), Johnson (1974), and MaPETHEEZBernal (1972) studied the acquisition of English morphology by Spanish speaking bilinguals. Natalicio tested 144 students in the first, second, third, and tenth grades in San Antonio (Texas) schools; one-half of her sample were native English speakers and the other half were native Spanish speakers. Aptitude and socioeconomic status were controlled. Modifying Natalicio's test and testing procedure, Johnson tested 147 Spanish-English bilinguals in the third, sixth, eighth, and tenth grades in San Antonio schools. Martinez-Bernal tested her bilingual students ranging in age from five to eight years old in the kindergarten, the first, and the second grade classes of schools in Tucson, Arizona. All Martinez-Bernal's sample were enrolled in a bilingual-bicultural program, but no 26 monolingual English speakers learning Spanish were included in the testing. Both Natalicio and Martinez-Bernal concluded that Spanish speakers tended to learn English plural allomorphs in the same order that native speakers did. Johnson, however, noted several problems (such as devoicing and null endings) that make any analy- sis of plural allomorphs extremely difficult. Natalicio and Johnson restricted their studies to nouns, but Martinez-Bernal tested for both nouns and verbs. Martinez- Bernal (1972: 127) found the order of the acquisition of English verb morphology among bilinguals to be the following: /s/ and /z/ present allomorphs are acquired first at a 40% accuracy-in- response rate; then /t/ and /d/ past allomorphs at 30%; the pro- gressive at 23%; the /Id/ past allomorph at 18%; and the /Iz/ present allomorph at 9%. Both /s/ and /z/ plural and possessive allomorphs received higher accuracy scores than those for any verb inflection. Martinez—Bernal's evidence for present progressive con- trasted sharply with that of the 1958 Berko study. Attempting to clarify this discrepancy, Martinez-Bernal admitted her data on the progressive resulted from a flaw in the test question. To such a question as: The boy knows how to dap. What is he doing right now? the children sometimes gave the response daping without any form of the verb gs, Since it directly answered the question, daping did not necessarily indicate that the children omitted ss_in their normal conversation. This one-word answer was ambiguous. 27 After several children gave this response, Martinez-Bernal (1972: 120, 127) changed her question by adding a pronoun: The boy knows how to dap. What is he doing right now? He With the revised question no ambiguity exists; the pronoun in the test item requires a form of bs_in the response. Since Martinez- Bernal counted all daping responses as incorrect, her figures on the progressive most likely reflect too low a score of those chil- dren who knew this morpheme. The role of interference in target language learning was discussed in all three studies. Finding only a little interference, Natalicio (1969: 157) stated: Since our data failed to provide any support for predic- tions derived from the contrastive analysis of Spanish and English in terms of such notions as "interference" and different types of "transfer," we find little justi- fication for their continued use, at least as regards plural formation in English. Johnson (1974: 119-20), however, disagreed: interference does play an important role in native Spanish speakers' performance in plural formation . . . and in any case, study l's [i.e., Natalicio's] discussion of inter- ference is misleading in not offering a distinction between phonological and syntactic interference. Martinez-Bernal (1972: ii) noted a minimal amount of interference: Interlingual interference was found to be present in 2.7 percent of all the responses. Of this, 2.5 percent was interference of Spanish in English, and 2.9 percent was interference of English in Spanish. The matter of interference, thus, remains unresolved. Both Natalicio and Martinez-Bernal support the transforma- tionalist theory of a universal, maturational pattern of language 28 acquisition. As supportive of Lenneberg's theory of "specific periods of language development," Natalicio (1969: 159-60) cited the fact that her tenth graders failed to show "a markedly better performance" than her third graders. Likewise, Martinez-Bernal (1972: 186) pointed to the striking similarity of order of ease (and therefore probable order of acquisition) of comparable items in Spanish and English morphology by bilingual children as additional evidence strengthening "develop- mental theories of language acquisition." Although suggesting a pattern of language development, these studies involve carefully controlled test statements in an artificial situation. In a comparison of the results of a revised version of Berko's test with the free speech of mentally retarded children, Dever (1972: 173) concluded: . . the test responses did not do a good job of pre- dicting errors or the lack of errors in the free speech of the subjects participating in this experiment. In free speech the children rarely used an incorrect inflection. While useful in indicating which inflections the children knew, the Berko test failed to predict the occurrence or nonoccurrence of morphological errors in free speech. The greatest correlation between test scores and free speech occurred for the progressive, the irregular past, and the /2/ present allomorph. Real lexical items predicted errors slightly better than nonsense ones. Longitudinal Studies‘ Dever's experiment raises questions about the extent to which Berko-type data reflect the actual order of acquisition of 29 morphological inflections in the spontaneous, natural speech of young children. Several longitudinal studies help provide an answer. Weir (1962), analyzing the pre-sleep monologues of a two- and-a-half-year old child, found evidence supporting Berko's con- clusions. For plural and possessive inflections, Weir's young subject used /s/ and /z/ more often than /Iz/. He used the pro- gressive morpheme the most frequently of all the verb inflections, although sometimes without a form of 9s, Past tense came next, followed by third person singular present tense. Studying the language development of several monolingual children, Ervin (1964) reached the same conclusion regarding noun plurals. Her subjects, moreover, inflected irregular nouns according to the rules of phonological conditioning (e.g., iggis_ for feet; mans for msn) (1964: 175). They also failed to produce the syllabic /Id/ past tense allomorph with any regularity. Ervin's subjects produced the past tense of some irregular verbs (e.g., ssms) before they produced any regular past tense (e.g., wsikgg) (1964: 178). This early learning of irregular past tense morphemes was due most likely to their extremely frequent usage in adult speech. Once an awareness of the regular past tense inflec- tions developed, the children overregularized the irregular verbs by adding regular endings for past tense; their free speech revealed such linguistic creations as breaked, buyed, comed, and doed (1964: 178-9). Discussing this phenomenon, Ervin (1964: 178) wrote: 30 The odd, and to me, astonishing thing is that these exten— sions occurred in some cases before the child had produced sgy_other regular past tense forms. Cazden (1968), reporting on the acquisition of inflections by three children over a five-year span, noted the learning of the plural allomorphs before the possessive ones by two of her subjects; the third child acquired both simultaneously. Among verbs, the criterion of 90% accurate usage was reached first for present pro- gressive. Cazden found no clear trend for the acquisition of the present and past tenses since her subjects acquired them in a dif- ferent order. Menyuk (1969) grouped the inflectional errors of nursery school children, kindergarteners, and first graders into three categories: (1) omission errors, (2) substitution errors, and (3) redundancy errors. Errors of omission involve the use of no inflection (zero allomorph) as in he talk for he talks. Substi- tution errors refer to the use of phonologically conditioned allomorphs for morphologically conditioned allomorphs, e.g., msgs_ (plural), catched (past), and hsigs_(present) (1969: 60-1). The use of more than one inflection is termed a redundancy; words like /laIktId/3 for liked, /kemId/ for came, and /gozIz/ for goes 3Berko (1958: 165) noted one redundant answer on her test, the response /spozd/ as the past of /spo/. In this case the present /z/ allomorph occurs, too. Baird (personal communication) noted two redundant responses on his test: /wagIgz/ with the present /2/ allomorph added to the progressive /Ig/ and /vazIz/ with both /2/ and /Iz/. I'll Hill Y I III 31 illustrate the error of redundancy (1969: 60-1).4 Menyuk's sample usually committed errors of omission. In summary, both the Berko-type studies and the longi- tudinal ones indicate a sequential learning of the English inflections by children. Morphological inflections seem to be acquired in a non-randomized, non-simultaneous order. While not unanimous in their conclusions, these investigations strongly indicate a mastery of the regular noun plural rules before those involving the possessives. The present progressive of verb inflec- tions is learned first, followed by regular past tense inflections, and then the present tense ones. The /s/ and /z/ allomorphs are acquired before the /12/ and the /t/ and /d/ allomorphs are learned before the /Id/. Such a developmental pattern of mor- phology learning is not incompatible with the transformational theory of language acquisition. 4The redundancy /kemId/ is especially unusual since it involves the use of the wrong past allomorph after /m/. CHAPTER II THE EXPERIMENTAL DESIGN Limitations of Berko's Study Certain limitations and omissions in procedure and test con- struction can be found in Berko's pioneering psycholinguistic research methodology. Aware of these limitations, several later researchers (Shriner and Miner, 1968, Natalicio and Natalicio 1969, Bellamy and Bellamy 1970, among others) made changes in Berko's work to avoid its shortcomings. Only those limitations of rele- vance to the present study will be enumerated. Anisfeld and Tucker (1967: 1203) noted that Berko tested for productive control of the morphological rules but not for receptive control. While the children were required to inflect nonsense words with appropriate inflections, they were never given an opportunity to choose the answer. Anisfeld and Tucker (1967: 1204) also indicated Berko's test items contained "an unequal dis- tribution of the three allomorphs" which "precludes making all possible error comparisons." Of the three present tense allomorphs, for instance, Berko included in her study two nonsense words, isgggs_and figs, for /Iz/ but none for either /s/ or /z/. Natalicio and Natalicio (1969: 207) expressed dissatisfac- tion with Berko's selection of subjects "based solely on the availability of S5 . . . in the particular schools used for the 32 33 research," the only criterion being "(1) preschooler and (2) first- grader"; "the sample," thus, "was one of convenience." Such variables as sex, 10, and socioeconomic status were not controlled. The Natalicios noted that Berko's subjects may not have accurately perceived all stimuli, thus precluding an accurate evaluation of all responses. Because of "the occurrence of unfa- miliar phonemes in the stimulus" or because of the similarity of distinctive features in some phonemes (e.g., /9/ and /f/), the subjects may have "confused" the sounds they heard: Assuming the possibility of such perceptual confusions, let us suppose that a S is presented with a stimulus item *heaf (Berko 158), and is asked to provide the plural. According to Berko's discussion, if the S replied with either *heafs or *heaves . . . the plural response was considered correct. If, on the other hand, the S responded with *heaf (no ending) or *heafes, the plural response was considered incorrect (158). The question arises when a S provides a plural response of, for example, *heases. According to the adult sample cri- terion, this last would be judged incorrect as a plural for *heaf. But, was the singular stimulus actually *heaf? (Natalicio and Natalicio 1969: 210-1) Besides possible perceptual confusion resulting from an inaccurate hearing of the nonsense words, some of Berko's children may have been unable to produce some phonemes correctly even if they had accurately perceived them. This inability may have resulted from a physiological disability or an incompletely developed phono- logical system. It would be possible to learn whether this was the case by requesting the subjects to repeat the nonsense stimulus (see Anisfeld and Tucker 1967). None of the above limitations necessarily invalidate the results of Berko's investigation; they do, however, point out the 34 difficulty involved in data interpretation and sharply delimit any generalization concerning the results. Overall, though, the Berko methodology remains a useful tool in measuring linguistic compe- tence. The Experimental Design A modification and expansion of Berko's work, the present experimental study was restricted to the acquisition of verb mor- phology. To be more precise, this study mainly concerned the acquisition of the English verb auxiliary by second language learners. It consisted of three tests, each with seventy items. The first test, which measured oral productive ability, was pat- terned very closely after Berko's methodology. While the student looked at the picture of a brightly colored cartoon figure perform- ing a certain task, the examiner read a statement about the cartoon illustration and requested the student to fill in a missing blank with a one-word response. The examiner, for example, showed a picture of a m n hitting a drum with sticks. The examiner pro- nounced the nonsense 0rd as it first appeared in the statement and asked the student to rep t it. If the student pronounced it incorrectly the first time, t s procedure was repeated, several times if necessary, until e'ther t student gave the desired word or further effort seemed/unproductive. onunciation difficulties and deviant responses/Here written down. Afte it had been deter- mined whether or not/the student could repeat the nonsense word, the whole statement was read as in: 35 This man zicks very hard. He has to do it every week. What must he do next week? Next week he must very hard. Each statement was usually repeated twice. On some occasions, at the insistence of the student, it was repeated more often. Such a procedure, however, was not followed for questions like: The boy can klant the house. In other words, the house can klanted by the boy. Here the student was not asked to make a verbal repetition since no form of the main verb was the correct answer. A form of ss_could have been given to the student but this would have provided a clue to the answer; giving the student this clue would make it more difficult to determine what he actually knew. Whenever the answer was is_or a form of bs_or Eggs, no repetition to determine pro- nunciation was requested of the student. If a form of the main verb was the answer, the earlier procedure involving the word repetition was carried out. Since test one concerned oral productive ability, all cues were given verbally. None of the words appeared in writing on the pictures themselves. Berko, in contrast, apparently printed her test statements on the cards. In the present study all responses to the oral test were written down in phonemic notation and the entire oral test was taped on a Panasonic tape recorder. Later each tape was replayed to determine the accuracy of the phonemic notation. During this oral test the examiner tested each student individually. The oral productive test was given so that each 36 student's pronunciation might be studied, especially as it con- cerned the allomorphs of the present and the past tenses. No written test could provide such direct data. This oral test pro- vided an Opportunity to test the student's listening comprehension, too. While the Berko investigation studied only oral productive ability, the present one also included a written productive test. After the oral test had been administered, the written productive test was given. The seventy items on this second test matched those on the oral test in grammatical structure: different verbs were used on the second test to prevent memorization of first-test items and to reduce monotony in the wording of questions. This second time the student wrote a one-word response in the appropri- ate blank in each test item but he saw no pictures illustrating the statements. After this written productive test was completed, a written receptive test was administered. Once again key words were changed but grammatical structures remained the same. On this test the student was asked to choose the correct answer out of four possi- bilities, as this item shows: The woman is tooging the bridge. The girl also wants to the bridge. a. tooge b. tooges c. tooged d. tooging1 While nearly all of the previous Berko-type studies focused attention on the acquisition of noun mOrphology, especially that of 1See Appendix A, which lists all test items on all three tests. 37 the plural allomorphs, the present study focuses its attention exclusively and more extensively on the acquisition of verb mor- phology. This apparent lack of interest in the English verb system reflected in previous research studies likely results from the complexity of the verb system; the English noun morphological system, in contrast, is relatively simple. A study by Martinez-Bernal (1972: 96) showed that chil- dren found no problem in using nonsense words "except for some puzzlement at first by the concept of an imaginary verb." Part of this puzzlement certainly was caused by the difficulty of assigning meaning to a nonsense verb. This difficulty became apparent during the construction of our test. While nonsense words are mere creations, they do have "varying degrees of meaning- fulness" (Shriner and Miner 1968: 609). All morphological inflec- tions occurring in the sentence and the syntactic pattern of the sentence itself help provide some of this meaningfulness. Berko (1958: 151) chose not to create any nonsense pronouns because of the difficulty involved in making up a nonsense pronoun, and because the pronouns are so few in number and so irregular that we would hardly expect even adults to have any generalized rules for the handling of new pro- nouns. Moreover, we do not encounter new pronouns, whereas new verbs, adjectives, and nouns constantly appear in our vocabularies. In our study, likewise, no nonsense verbs were created to parallel the modals, the forms of 9e, and the forms of have. In the 38 terminology of C. C. Fries, these are "function words," words that one must know as items.2 Substitutes will not work. 0f the 27 items on Berko's test, all except three were non- sense words. The noun glass and the verbs ring and melt were the { only real words on her test. In our study one real word was i:; I included for each nonsense word. For each allomorph and each structure being investigated, our test includedagpajrse:;eees;’ E1o6§,Co6§:nithao_oooo§ooe word and one with a real word. The 2 Aummmv mmcmh Accov New 2mmmv RON Acmev Rom __< 2mmv an“ Amsv New Aemv Rue >_< ANMNV “mm AeMNV “mm “s_wv “Pa “_32 AONV 2mm Ampv um“ ANPV Rom “>3: AGNNV umm ANQFV as“ 2_m>2 “on _Fwd 222W go“ 20_v “no Amps gem >F_a 222.2 22% 2m~_v use Amopv “an _ato 20_ new 2¢_v 2mm 2mg 22m _ago Aummm vcm.m pcmmmcav wmcmp >>< >2 cm+mn am ucmmmgav omcm> AOGNV New ANQNV am“ AQmNV umm >_< 2mmv an“ Ammv Rem Aomv Rom __< Amopv “mm Anopv umw Aemv am“ o_=z ANNV N>¢ 2_~v 22m AFNV Num “,3: Anopv umm AFmV um“ AMNV Rom >_Pe Am_v 2m“ AMFv Rem Ampv 2N0 _me Aemv Nos Amov 23m Amov umm Page 2422 2mm Amy New flow as _ato pmmm __< >2 cm+mn cm+m>m2 mm pcmmmgnv mmcmp 2ommv Rpm Asmmv awe Aommv 22m __< 2mmv 22m Amev Rom Amev “we __< 2¢m_v amm AmNPV 2mm Aompv “mm “>32 ANNV RFm AMNV 2mm 2_~v 22m “>3: 2m~_v “mm 209.2 “ms Away “Po >__2 Aomv 2mm Ampv Rm“ Awpv gm“ _Pwu 2mmv New 2002 &_¢ Amsv 2mm Page 2222 so“ 222 25m Amy 2mm Page .w pcmmmcm —~< >2 mcw+mo =m+m>m2 aw ucmmmcav mmcmh Aumv Rpm ANMV use 2m_v flow __< 2mm_v 25m Ammpv 2m“ AmMFV 2N8 >_< 25_V 2mm Ao_v New 252 Run “>32 Aouv 22m Away 225 “new spa “_sz 2m_v awe AMPV Rem Amy amp >222 Akov 2mm 2mmv N_m “New 2mm >>_2 Amy 2mm Amv 2N_ Amy am, Page 2852 2mm Aemv 22¢ ANNV 22m Paco >2 cm+ma mcw+mn Aummmv mmcm> >2 mm pcmmummv mmcmp nmucm>u< mumvuwegmucH gmccwmmm umucm>n< mumwumscmch cmc=_mwm baa» “map uumggou gossaz van “cougma pumcgou emnszz can ucmugma .Pm>as >2 amcap--.m_ wsm< Away 2mm 2.52 amm Amov ans P_< ACPNV mum 2_o_v use 2~m_v 2mm 32:: Amsv 2mm Anny no“ Romy “mm 6.52 2>m_v um“ 2om_v “No 2552 “.4 .222 AwNV “mm 2mmv 2mm ANNV xms _Pwu 2oe>v gum Aemv mam Amos gnu Paco ANNV New AQNV 22s 2222 RNN _ago Pecos F_< >2 m=_hmm _auoe 53:5» 2m~_v “mm Away awe Aomv “mm >_< 2mmv NMN Anny fims Ammv mos _P< Amev umm 2_mv new AQMV “Na 32:: AMNV 2mm 2~_V “a“ AmFV am“ 3.32 ANev “hm Apmv use 2>NV “ms >_22 APNV 22w 2m_v “em 2>_v ums __Pu Anny was Ammv am“ Ammv new _ato Amy 22m AMV sw. ADV so _aeo .mm >>< >2 cm+ma cm+m>m2 pmuos mmcmh Aemwv “Pm Ao_~2 RNA 2¢m_v 2mm F_< 2omv 25¢ ANNV 22m 2~_v “mm _>< 2cm «no Aemv “km 2252 “mm ups: AQNV 2mm 2N_v New 222 Nam 52:: 2%“ “an Asks amm Akmv “mm >_22 AONV wa Ampv 25m Amv 2mm P_22 Amos ask 2242 was AOMV “_m Page Ao_v 2,5 2N2 2m Amy Rm _ato .mm >>< >2 mcm+mn cm+m>m2 paves mmcmh 24m.v umm 2252 Ram 2mmv 2cm _>< 2_mv 20m Akmv umm 2m_v um _>< 2222 new 2852 $25 2mmv New “>5: Asmv 202 22.2 2mm 20.2 new 3.5: “mus was Acev ape 2_Nv N.N >222 ANMV sec 25.2 saw Amy as _._a Aemv “mm 2__v N__ AMV Rm .520 2m_v s_m 202 2N2 222 gm Paco m>m2 F>< >2 cm+m>mz Faves mmcmh 2om_v usm 2m__v Nmm 2_mv 2mm F_< 2mm_v 2mm 2mmv Ram Aomv 2mm _>< 2252 $25 2052 2mm Akmv 222 “Pa: Anew Nam 2_mv 250 Aomv Rmo 32:2 Amsv aoo_ Amsv xmm 2mmv RNA >__a Amsv 22m 2_mv 250 APNV fins >_wa 2_¢V -nmm ANNV New 2522 23m Page AmMV so“ Aomv 2m“ AGNV wow Page >2 cm +.mm paves mmcmp .mz panes mmcmh vmocm>u< mumwumELmucm gmccwmmm vmucm>u< mummumEmecH gmccwmmm “awe “use uumegoo gmnszz ucm pcmucma uomgcou consaz can ucmucmm .Pasma >2 Pauoz--.¢_ msmqe 105 The advanced students answered more items correctly than did the intermediates, who, likewise, did better than the begin- ners. Three exceptions occur: 1. modal have+en progressive (ora1: beginner 8%, intermediate 8%) 2. modal have+en passive (multiple-choice: beginner 75%, intermediate 70%). 3. modal nging (fill-in: intermediate 68%, advanced 58% . The percentages for advanced students are always higher than those for the beginners with the least difference (13%) for modal ggfing (fill-in: 58%, 45%) and the greatest difference (62%) for modal have+en MV (fill-in: 66%, 4%). Correctness percentages for modal on the multiple-choice test are higher than those on the fill—in test, which are higher than those on the oral test with these exceptions: l. modal My_(beginner: ora1 60%, fill—in 43%; inter- mediate: ora1 75%, fill-in 64%, multiple-choice 64% 2. modal have+en progressive (beginner: fill-in 33%, multiple-choice 29%; intermediate: 54%, 50%; advanced: 83%, 83%) 3. modal gsfen (advanced: fill-in 100%; multiple- choice 97%). For one-half of the structures involving modals, the null hypothesis is rejected. The difference in performance among the three groups of students is significant at the .05 level for modal hsigfen progressive and modal hsygfen passive; it is highly sig- nificant at the .001 level for modal hgygfen MV, the least cor- rectly answered modal structure on the tests. 106 Perfect (Have+en) The 42% correctness for perfect by beginners is less than the 55% by intermediates and the 78% by advanced students. §g_ (beginner 49%, intermediate 66%, advanced 86%) is more correctly answered than hgig_(beginner 32%, intermediate 42%, advanced 68%). Advanced students correctly answered every gg_item for have+gg_ progressive (fill-in and multiple-choice tests) and modal have+gg_ progressive (fill-in test). The percentages for the advanced students are higher than those for the intermediates, which, in turn, are higher than those for the beginners with these exceptions: l. modal have+en progressive (oral: beginner 8%, intermediate 8%) 2. modal have+en passive (multiple-choice: beginner 75%, intermediate 70%) 3. have+gg progressive (fill-in: beginner 75%; inter- mediate 75%; multiple-choice: intermediate 95%, advanced 91%) 4. have+en passive (fill-in: beginner 62%, intermedi- ate 54%; multiple-choice: beginner 87%, intermedi- ate 87%) 5. have+gg_passive (fill-in: intermediate 95%, advanced 95%) 6. modal have+§g_progressive (multiple-choice: begin- ner 83%, intermediate 79%) 7. modal have+§g_passive (multiple-choice: beginner 79%, intermediate 75%). The percentages on the oral test are always lower than the corresponding percentages on the fill-in test; the percentages on the fill-in test, similarly, are lower than those on the multiple- choice test with these exceptions: l. modal hgigfen progressive (beginner: fill-in 33%, multiple-choice 29%; intermediate: 54%, 50%; advanced: 83%, 83%) 107 AAAA RAA AAAV RAA 2AAA RAA AAA AANA RAAA AANA RAA AAAA RAA AAA: AANA RAAA ANNA RAA AAAA RAA AAAA AAAA RAA 2AA RAA AAA RAA AAAA >2 mcw+mo.mm+m>o2 mmcwp 2AAAA RAA AAAAA RAA AANAV RNA AAA ANANA RAA 2AAAV RAA AAAAA RAA AAA 2AANA RAA AAANA RNA AA_NA RAA RAA: 2AAA RAA 2AAA RAA 2AAA RAA AAA: AAANV RAA ANANA RAA AAAAV RAA AAAA AAAA RNA AAAA RNA 2AAA RAA _AAA 2AANA RAA 2AAAA RAA 2AAA RAA Auto 2AAA RAA ANAA RAA ANAA RAA _Ato pumwgmm ppq >2.mm+m>o2 mmcmw 2AAAA RAA 2AAAA RAA 2AANV RAA AAA 2AAA RAA 2AA2 RAA 2AAA RAA AAA 2AAAA RAA 2AAAA RAA ANAAA RAA AAA: ANNA RAA AANA RAA AANV RAA AAA: AAAAV RAA 2AAAA RAA 2AAA RAA AAAA AARA RAA 2A_A RAA AAAA RNA ___A 2AAAA RAA AAAA RAA AAAA RAN AAAA 2AAA RAA 2AA RAN AAA RA AAAA mm >>< >2 co+mn co+m>u2 mmcmh AAANA RAA AAAAA RNA 2ov_v RNA AAA RAAA RAA AAAA RAA AAAA RNA AAA AANAA RAA AAAA RNA 2AAV RAA RAA: ANNA RAA AANA RAA AANA RAA RAA: ANAAA RAA AAAA RAA 2AA. RAA AAAA AANA RAA 2AAA RAA 2AAA RAA AAAA 2AAV RAA 2ANA RAP 2AA RA _Ato 2AAA RAA AAA RAN 2AA RAN _Ato m>o2 ——< >2 mcw+mn cm+m>m2 mmcm> 2AAV RAA 2AAA RNA 2AAA RAA A_< AAAV RAA 2AAA RAA AANV RAA AAA AANA RAA 2AAA RAA 2A_. RAA RAaz AANV RAA 2AAA RAA AAAA RAA RAA: AAPA RAA 2A_v RAA AAAA RNA AAAA AANA RAA 2AAA RAA AAAV RAA AAAA 2AAV RAA 2AA RAA 2AA RAN .AAA AAA RAA 2AA RNA 202 RA _Ato >2 cm+mn mM+m>A2 Ponce wmcmh >2 cw+mo cm+m>m2 _mvoe mmcmh 2AAA RAA 2AAA RAA AAAA RAA AAA 2AAA RAA AANA RAA 2A_A RAN AAA ANNA RAA AAAA RAA RONA RAA RAA: AANA RAA ANAA RAA 2AA RAN A_Az AANA RAAA 2AAA RAA AAAA RAA _AAA AANV RAA 2AAA RAA AAA RAA AAAA AANA RAA AAA RAN 2A2 RNA AALA AAAA RAA ANA R 2N2 RA AAAA >2 mcw+ma dm+m>mz paces omcm> >2 ch+mn cm+m>m2 —Avoe mmcm> 2AAA RAA AmAv RAA 2AAA RAA AAA 2AAA RAA 2AAA RAN 2A_A RA FAA AANA RAA ANNA RAA AANV RAA AAA: 2AAA RAA AAAA RAA AAAA RAN RAA: AANA RAA AANV RAA ANAA RAA AAAA ANAA RAA ARAA RAN 2N. RA _AAA AANV RAA 2AAA RAA 2AA RAN Paco 2AAV RAA 2AA RNP A>A RN AAAA >2 :m+ma.mm+m>mc mmcwk >2 cm+m>m2 Paces wmcmp vmuca>u< muowvoeLmucu chcwmmm nmocm>u< wumwcwEAmucfi mecwomm AAAA AAAA uumgcou Awassz uco ucwugom pumggou Lassa: ucm acmugma ._A>AA AA RAAALAA--.AR AAAAA 108 have+§g_MV (advanced: fill-in 92%, multiple- choice 89%) have+en progressive (advanced: fill-in 100%, multiBTe-choice 100%) have+en passive (intermediate: fill-in 95%, multiple-choice 91%; advanced: 95%, 95%) modal have+gg_progressive (advanced: fill-in 100%, multiple-choice 91%) modal have+gg_passive (intermediate: fill-in 75%, multiple-choice 75%) §g_§all) (advanced: fill-in 93%, multiple-choice 91% . Nam-hum For regular verbs, 51% (111) of the 216 sg_responses show gg_on the oral test (beginner 34%, 25; intermediate 48%, 35; advanced 70%, 51); 69% (151) show sg_on the fill-in test (beginner 40%, 29; intermediate 72%, 52; advanced 97%, 70); and 84% (183) do so on the multiple-choice test (beginner 76%, 55; intermediate 80%, 58; advanced 97%, 70). For irregular verbs 41% (30) show sg_ for sigg_and migg_on the oral test (beginner 25%, 6; intermediate 33%, 8; advanced 66%, 16); 47% (34) show §g_for gigg_and gigg_on the fill-in test (beginner 29%, 7; intermediate 33%, 8; advanced 79%, 19); and 58% (42) show §g_for sprigg_and igigg_on the multiple-choice test (beginner 45%, 11; intermediate 62%, 15; advanced 66%, 15). Regular verbs are answered correctly more often than irregular ones. The null hypothesis is rejected for eight of the ten struc- tures involving perfect. For modal hsysfen progressive and for modal hgygfen passive the difference among the three groups of students is significant at the .05 level. It is highly significant at the .01 level for: 1. have+en progressive 2. have+§g_MV 109 3. have+gg passive 4. modal have+gg_progressive. It is highly significant at the .001 level for modal have+en MV, the least correctly answered perfect structure, and for have+gg_ progressive. Progressive (Be+ing) For progressive the beginners' 38% correctness is lower than the intermediates' 57% and the advanced students' 71%. Bg_ (beginner 42%, intermediate 61%, advanced 74%) is answered more correctly than igg_(beginner 29%, intermediate 49%, advanced 65%). No intermediate student responded correctly with being for be+igg_ passive on the oral test; one beginner did. The data for language levels show the beginners with the lowest figures, the intermediates next, and the advanced students with the highest figures with these exceptions: l. modal sgfing (fill—in: intermediate 68%, advanced 58% 2. nging passive (ora1: beginner 12%, intermediate 12%; multiple-choice: intermediate 66%, advanced 58% 3. modal perfect ggfing (multiple-choice: beginner 83%, intermediate 79%) 4. be+in MV (fill-in: intermediate 75%, advanced 75%)_9' 5. be+igg_passive (oral: beginner 4%, intermediate 0%; multiple-choice: intermediate 79%, advanced 79%). As elsewhere, the percentage accuracy is highest on the multiple-choice test, lower on the fill-in test, and lowest on the oral test. One instance occurs of an oral response equaling or surpassing a fill-in response (beginner: oral 12%, fill-in 12% 110 AAAA RAA 2AAA RAA ANAA RAA AAA AAAA RAA 2AAA RNA AARA RAA AAAz “AAA RAA AAAA RAA AAAA RAA AAAA AAAA RAA AAAA RAA 2AA RAA AAAA >2.mmw+ma owed» ANAAA RRA ANANA RAA AAAAA RAA AAA AAAA RAA AAAA RAA AAAA RAA AAA AAAAA RAA 2AAAA RAA AAAA RAA AAaz ANNA RAA “AAA RAA “ANA RAA RAA: AANAA RAA ANAAA RAA 2AAA RNA AAAA AANV RAAA AAAA RAA AAAA RAA AAAA AAAA RAA AAAA RAN AANA RAA .AAA AANA RAA 2AA RAN 2AA RNA AAAA o>mmmmgmoga PP< >2 mcmtmm cm+m>cn _muos mmcmh AAAA RAA AAAA RAA 2AAA RAN AAA AAAA RAA 2AAA RAA 2AAA RAA AAA AAAA RAA 2AAA RAA ANNA RAA AAAz AANA RAA. AANA RAA 2AAA RAA AAA: RNAV RAA .ANA RAA 2AAA RAA PARA AANA RAAA ANNA RAA AAAV RAA ARAA AANA RAA AAAA RNN AAA RAA AAAA AAAA RAA AAA RAA 2AA RAA AAAA .mmw F>< >2 mcwmmm cm+m>as mmcoh AAANA RAA AANNA RAA RNAAA RNA AAA AAAA RAA ANAA RAA AAAV RAN AAA 2AAAA RAA AAAV RAA AAAA RAA RAA: 2AAV RAA AAAA RAA AA RAA AAAz AAAA RAA 2AAA RAA 2AAA RAA AAAA 2AAA RNA AA_A RAA 2A RNA AAAA AAAA RAA 2AAA RNA AANA RAA AAAA AAA RAA 2AA RN. 2AA RNA AAAA .Am ARA >2 AA+AA AAAAAA AAAAA AAAA RAA AANA RAA AAAA RAA AAA AAAA RAA AAAA RAA 2AAA RAA ARA 2AAA RAA 2AAA RAA 2AA RAA AAA: AAAV RAA AAAA RAA 2AAA RNA RAAz 2AAV RAA 2AA RAA 2N2 RA AAAA AANA RAA AAAV RAA ANNA RAA AAAA 2AAA RAA 2AA RA 2AA R AAAA AANA RAA AANV RAA AAAA RNN AAAA >2 cm+mn.mmw+ma mmcmp >2 mcwamm Autos mmcm> umucm>c< mpmwumEAmuca Amccwmmm umucm>c< mumwvmeAmucH Amccwmmm ummh nomAAou Amnssz new acmugoa ummh uumggou AwnEAZ new ucwugma ‘ ._m>m4 2n m>RmmwAmogmnl.o_ ~4m<> 111 for nging passive); several occur of a fill-in response equaling or surpassing a multiple-choice response: 1. gsfing passive (advanced: fill-in 62%, multiple- choice 58%) perfect ggfing (advanced: fill-in 100%, multiple- choice 100%) modal perfect ggfin (advanced: fill-in 100%, multiple-choice 91%) 'be+igg_MV (beginner: fill-in 58%, multiple-choice 58%; intermediate: 75%, 62%). 4500“) Data for ggfing passive reveal significant differences at the .05 level among the three groups of students. Data for several other structures reveal highly significant differences: at the .01 level for modal perfect bgfing and for be+igg_passive; at the .001 level for perfect ggfing. Passive (Be+gg) For passive the 50% correctness by beginners compares with the 69% by intermediates and the 84% by advanced students. The sg_ (beginner 55%, intermediate 76%, advanced 94%) is more correctly answered than bg_(beginner 49%, intermediate 68%, advanced 82%). The responses for progressive Essen (beginner 15%, intermediate 37%, advanced 61%) are markedly below the average Qg_responses. Intermediates answered the items more correctly than begin- ners; advanced students likewise answered them more correctly than intermediates. Three exceptions occur: 1. perfect bgfen (fill-in: intermediate 95%, advanced 2. 32331 perfect be+en (multiple-choice: beginner 79%, intermediate 75%) 3. progressive ngen (oral: beginner 4%, intermediate 0%;)mu1tip1e-choice: intermediate 79%, advanced 79% . 112 AAAA R>A AANV RAA >>AA RAA RAA >A_A RAA AA_A RAA AAA RAA RAA: AAAV RAA 2AA RAA ANA RA >>AA >A>A RAA RAA RA 2AA R AAAA >2 cmamm m:_+mn mmcmh AANAA RAA ANAAV RAA ANANV RAA A>A AAAA RAA 2AAA RNA 2AAA RAA >>< AAARA RAA RAARA RAA ANAAA RAA AAA: A>NA RAA >A_A RAA AARV RAA ARA: AAAAA RAA ANA—A RAA 2AAA RAA APAA >A_v RAA 2AAA RAA AAA RNA >AAA AANAA RAA 2AAA RAA 2AAA RAN _AAA AA_V RAA 2AA RAA AA RAN AAAA .mmwmmmm >>< >2 cmamm cm+m>mc Faves mmcmh 2AAA RAA 2AAA RAA AAAA RAA FAA AAAA RAA 2AAA RAA AAAA RAA _AA AANA RAAA AANA RAA >A>A RAA ARA: RANA RAA ANNA RRA AANA RAA RAA: AANA RAA ARNA RAA >A_A RAA _PAA AANA RAA AANA RAA ANAA RAA AAAA ARNA RAA AAAA RAA >A_A RRA >AAA >>Nv RAA AARA RAA 2AA RAN RAAA .mm >>< >2 cmamm cm+m>A2 mmcmp 2AAAA RNA AAANA RAA ANANV RAA AAA AAARA RAA 2A>AA RNA 2AAA RAA APR AAAAA RAA RANRA RAA RAAA RAA RAA: 2AAA RAA 2AAV RAA AAAA RAA AAAZ APNAA RAA AA>>A RAA 2AAV RAA _RAA 2AAV RAA> 2AAW RAA AAAA RNA AAAA AAAAA RNA AAAV RAA AAAA RAN AAAA AAAA RAA RAN RAA AARA RAA _AAA .mm RAA >2 AAaAm AAAAA AAAAA AAAA RAA AAAA RAA AAAA RAA AAA 2AAV RAA AAAV RAA AAAA RAA _AA AANV RAA. RANA RAA 2A>A RAA ARA: AANA RAA 2A>A RAA RN_A RAA ARA: AANA RAA AANA RAA AARA RAA _RAA 2AAA RAA AA>A RAA AARV RAA _RAA ARNA RAA AAAA RAA AARV RAA RAAA >A>A RAA AA_A RAA AAA RAA AAAA >2.mm+mn mmcmh >2 cmamm mmcm> umocm>v< muARUAEAmch Amccmmmm Amoco>v< mumwumEAmucm Amccwmmm RumAAou Amnsaz can acouema 9mm» pquAou Amaszz new ucmugmA umwh .>m>m4 AAa m>wmmmauu.m> mgmqh 113 While percentages for the fill-in test are always higher than those for the oral test, the following percentages for the fill-in test equal or surpass those for the multiple-choice test: 1. ggfen MV (beginner: fill-in 54%, multiple-choice 50% modal ngen (intermediate: fill-in 95%, multiple- choice 95%; advanced: 100%, 97%) perfect ggfen (intermediate: fill-in 95%, multiple-choice 91%, advanced 95%, 95%) modal perfect ggfen (intermediate: fill-in 75%, multiple-choice 75%) be+eg MV (intermediate: fill-in 87%, multiple- choice 83%). U'I'IDOON For all structures only two show a highly significant difference at the .01 level among the groups of students: perfect sgfen and progressive gsfen. The Infinitive (To MV) For infinitive the 66% beginner response compares with the 79% intermediate response and the 87% advanced response. gv_ responses (beginner 75%, intermediate 88%, advanced 90%) show higher percentages than is_responses (beginner 58%, intermediate 69%, advanced 84%). On the multiple-choice test for ig_+ MV all inter- mediates supplied ig_correctly. Surprisingly, the beginners' 95% for My_on the fill-in test was better than the intermediates' 87% and the advanced students' 91%. A few unexpected results occur: 1. is_MV (fill-in: beginner 62%, intermediate 62%; multiple-choice: intermediate 100%, advanced 95%) 2. to My_(oral: intermediate 87%, advanced 87%; fill-in: beginner 95%, intermediate 87%, advanced 91%; multiple-choice: intermediate 91%, advanced 91%) 3. all (fill-in: beginner 79%, intermediate 75%; multiple-choice: intermediate 95%, advanced 93%). 114 TABLE 18.--Infinitive by Level. Percent and Number Correct Test Beginner Intermediate Advanced Ig_MV Oral 20% (5) 45% (ll) 70% (17) F111 62% (15) 62% (15) 87% (21) Mult 91% (22) 100% (24) 95% (23) All 58% (42) 69% (50) 84% (61) To Mi_ Oral 41% (10) 87% (21) 87% (21) F111 95% (23) 87% (21) 91% (22) Mult 87% (21) 91% (22) 91% (22) All 75% (54) 88% (64) 90% (65) A11 infinitive Oral 31% (15) 66% (32) 79% (38) F111 79% (38) 75% (36) 89% (43) Mult 89% (43) 95% (46) 93% (45) All 66% (96) 79% (114) 87% (126) For the three tests the percentages are the highest on the multiple-choice test and the lowest on the oral test; those on the fill-in test fall in between. The following do not follow this pattern: to Mi_(beginner: fill-in 95%, multiple-choice 87%; intermediate: oral 87%, fill-in 87%; advanced: fill-in 91%, multiple-choice 91%). Both is_MV and to M!_structures reveal a significant difference at the .05 level in performance among the groups of students. The Gerund (Ing MV) The 22% correctness by beginners is less than the 59% by intermediates and the 70% by advanced students; these averages are 115 TABLE 19.--Gerund by Level. Percent and Number Correct Test Beginner Intermediate Advanced Oral 12% (3) 62% (15) 54% (13) Fill 16% (4) 66% (16) 75% (18) Mult 37% (9) 50% (12) 83% (20) A11 22% (16) 59% (43) 70% (51) all lower than those for the infinitive (see immediately above). Two unexpected results occur: oral: intermediate 62%, advanced 54%; intermediate: ora1 62%, fill-in 66%, multiple-choice 50%. A significant difference at the .05 level occurs in the performance of the three groups. The Allomorphs On the oral test the data for present s_and past-tense allomorphs show that advanced students more often supplied the appropriate allomorphs than did the beginners or the intermediates. For present s_the /s/ is correctly given in 39 of the 72 responses (54%) (beginner 29%, 7; intermediate 54%, 13; advanced 79%, 19), the /z/ in 41 of the 72 responses (56%) (beginner 50%, 12; inter- mediate 50%, 12; advanced 70%, 17) and the /Iz/ in only 28 of the 72 responses (38%) (beginner 37%, 9; intermediate 37%, 9; advanced 41%, 10). Almost no difference exists between beginning and advanced students in /12/ usage but considerable difference exists 116 for /s/. In short, whenever the students supplied the present s_ morpheme as the answer, they always supplied the correct present s allomorph. For past—tense allomorphs the difference between beginning and advanced students is greatest for /t/; the difference is about the same for both /d/ and /Id/. The /t/ is supplied correctly in 120 of the 216 responses (55%) (beginner 27%, 20; intermediate 54%, 39; advanced 84%, 61), the /d/ in 111 of the 216 responses (51%) (beginner 37%, 27; intermediate 45%, 33; advanced 70%, 51), and the /Id/ in 56 of the 144 responses (38%) (beginner 25%, 12; intermediate 41%, 20; advanced 50%, 24). 0f the errors made in allomorph usage, most are found in the beginners' responses; com- paratively few occur in the advanced students' responses. The following errors occur: 1. /d/ for /t/ (e.g., /wokd/) (beginner 5, intermedi- 713/lfor /t/ (e.g., /wokId/) (beginner 6, inter- mediate 6, advanced 2) /t/ for /d/ after nasals or liquids (e.g., /smaIlt/) (beginner 6, intermediate 3, advanced 4) /Id/ for /d/ after nasals (e.g., /plinId/) (begin- ner 3, intermediate 2, advanced 4) change of final stem (e.g., froo for frool) (begin- ner 8, intermediate 3, advanced l). 01th Of these 54 errors, beginners made 28, intermediates 15, and advanced students 11. In addition, 7 substitutions of /d/ for final /t/ also occur (e.g., zind for the uninflected stimulus zint) (beginner 4, intermediate 1, advanced 2). 117 Summary Two patterns emerge from the data. For each of the struc- tures. the percentages are normally the lowest on the oral test and the highest on the multiple-choice test with those on the fill-in test in the middle. Occasionally the percentages on the oral test equal or slightly surpass those on the fill-in test. Twice they equal or surpass those on the multiple-choice test: modal gm; (intermediate: oral 75%, multiple-choice 64%) and irregular past igg_forms (beginner: oral 70%, multiple-choice 66%; advanced: 91%, 70%). The latter results probably occur because of the test words used: sigg_on the oral test is more common than sggigg_on the multiple-choice test. Similarly, the level of a student's proficiency in English is a good indicator of how well a student answered the items. Correct responses by beginners surpass those by advanced students once by a single response: to M!_(fill-in: beginner 95%, 23; advanced 91%, 22). In contrast, responses by advanced students often surpass those by beginners by large margins, once, in fact, by 62%: modal hgygfen MV (fill-in: beginner 4%, 2; advanced 66%, 32). Statistically significant differences in the results by the three groups of students can be found in 13 of the 23 struc- tures. At the .05 level they occur in: 1. tense nging passive MV 2. tense modal have+en progressive MV 3. tense modal have+en passive MV 4. is_MV 5. to MV 6. E9!!!- 118 They occur at the .01 level in: tense have+en progressive MV tense have+gg_ tense have+gg_ggfen MV tense modal have+§g_§gfing MV tense be+igg_§gfen MV. 01$de At the .001 level these differences occur in tense modal hsigfen MV and tense have+gg_ggfing MV. Overall the differences are especially numerous for perfect. In conclusion, the statistical data show that the student's level of proficiency in English did, indeed, affect his performance on items involving 13 structures. Analysis by Native Language ifggg; The 58% all correctness for tense by Spanish speakers com- pares with 70% by Arab speakers and 82% by Japanese speakers. Similar results can be found in the data for present s_(Spanish 56%, Arab 68%, Japanese 82%) and past (Spanish 60%, Arab 73%, Japanese 82%). Whereas little difference in the results exists among Spanish speakers between the best learned structure (59%) and the least learned structure (55%) for present s, a greater difference occurs for the Arabs (77%, 44%) and the Japanese (87%, 63%). The difference is more marked for the two past structures: gssi_+ MV (Spanish 67%, Arab 82%, Japanese 91%), the most cor- rectly answered tense structure, and pgsi_progressive passive MV (31%, 37%, 47%), the least correctly answered tense structure. On the multiple-choice test the Japanese answered every item correctly for present s_+ MV. 119 2AANA RAA 2AANV RNA 2AAAV RAA ARA AAAA RNA RAAA RNA AAAA RAA ARA: 2AAA RAA AAAA RNA 2AAA RAA >AAA AAAA RAA ARAA RAA RNAA RAA >AAA >2 Apmmqv mmco> AAAAA RNA 2AAAV RAA AAAAA RAA A>A 2AAA RAA ANAA RAA AAAA RAA AAA 2AANA RNA AANNA RAA AAANV RNA RAAz AANA RAA ANRA RAA AAAA RNA A>=z AANNV RAA AAA—A RAA AAAAA RNA ARAA ANNA RPA AARA RAA 2AAV RAA RARA 2AA>A RAA 2AAAA RAA AAAA RAN _AAA AARA RAA 2AAA RAA ARRV RAA _AAA wmcmm >>< >2 cw+mn Am acmmmgmv mmcmh 2AANA RNA AAANA RAA AARNA RAA AAA 2AAA RAA AAAA RRA RAAV RAA AAA AAARA RAA AAAAA RAA AAAV RAA RAA: AANV RAA AANA RAA RANA RAA ARA: 2AA>V RAA 2AAA RNA AAAA RAA AAAA AA_A RAA AA>A RNA >A>v RNA A_AA ARAA RAA 2AAA RAA AAAA RAA AAAA 2AA RAN AAA RAA 2AA RAN AAAA ARAAQV mmcmu —_< >2 cm+mn cm+m>A2 «w ucmmmcmv mmcmh AAAAA RNA 2AANA RAA ANANA RAA RAA 2AAA RAA 2AAA RAA AAAA RAA >>A ARARA RAA AANAV RAA AANAA RAA RAA: AANA RAA ANNA RRA RANV RAA ARA: >NN>A RAA 2AA>V RAA AAAA RAA AAAA RNNV RAA AARA RAA 2AAA RAA RRAA 2AAA RAA 2AAA RAA 2AAV RNN AAAA 2AAA RAA AAAA RAA AAA RAA AAAA aw “commgav mmcmu __< >2 m:_+mn cm+m>mg «m acmAmAmv AMAAA 2AAV RAA AANA RAA AANA RAA AAA 2AAAA RAA 2AAAA RAA 2AA>V RAA RAA 2A_2 RNA AAAA RAA RAAA RRA ARA: RNAV RAAR RAAA RAA AAAA RAA RAA: AAAA RAA AAA RAA AARV RAA AAAA 2AAA RAA 2AAA RAA 2AAA RAA ARAA 2AA RAN AAA RAN AAA RNA ANNA 2AAA RAA AAAV RAA AAAA RAA RAAA >2 cm+mn m:_+ma ARAAAV «mam» >2 am ucmmommv mmcmh ommcmqmw AAA< ;m_cmam mmmcmama AAA< gmwcmam AAAA AAAA uumggou Longs: ucm ucmugmm pumsgou Amaszz ucm ucmugmg .mmmzchA >2 mmco>11.ow mqm<> 120 In nearly every instance, the percentages on the oral test are lower than those on the fill-in test, which in turn are lower than those on the multiple-choice test. Four exceptions occur: present s_MV (Arab: fill-in 91%, multiple-choice 90%) present 5 passive (Arab: oral 41%, fill-in 41%) past MV (Japanese: fill-in 93%, multiple—choice 92%) . ast progressive passive (Spanish: fill-in 41%, multiple-choice 41%). 4:de In general the results show the Spanish with the lowest per- centages, the Japanese with the highest percentages, and the Arabs in between. The results for present s_passive do not follow this pattern since the Spanish percentages on all three tests are higher than the Arab ones, giving a 55% all accuracy for the Spanish to 44% for the Arabs. This exception occurs because the Arabs chose the wrong tense of pg, wps, for the correct present tense is, more frequently on the oral and the fill-in tests (11 times) than did the Spanish (3 times). Other exceptions include: 1. resent s_MV (fill-in: Arab 91%, Japanese 84%, multiple-choice: Spanish 90%, Arab 90%) 2. present s_perfect progressive (ora1: Arab 54%, Japanese 54%; fill-in: Spanish 75%, Arab 66%) 3. present s_perfect passive (oral: Arab 33%, Japanese 25%; fill-in: Spanish 62%, Arab 62%) 4. past MV (multiple-choice: Arab 92%, Japanese 92%) 5. past progressive passive (fill-in: Spanish 41%, Arab 33%) 6. present 5 (all) (multiple-choice: Spanish 84%, Arab 83%). Of the 216 responses for ppsi_MV 66% (143) show gg_for the six regular verbs on the oral test (Spanish 40%, 29; Arab 73%, 53; Japanese 84%, 61); 82% (178) show gg_on the fill-in test (Spanish 70%, 51; Arab 80%, 58; Japanese 95%, 69); and 98% (212) show gg_on the multiple-choice test (Spanish 97%, 70; Arab 97%, 70; Japanese 121 100%, 72). For irregular verbs 76% (55 out of 72) show gg_for sipg_and mipg_on the oral test (Spanish 54%, 13; Arab 75%, 18; Japanese 100%, 24); 80% (58) show gg_for pipg_and pipg_on the fill-in test (Spanish 66%, 16; Arab 87%, 21; Japanese 87%, 21) and only 72% (52) show gg_for sppipg_and ipipg_on the multiple- choice test (Spanish 66%, 16; Arab 79%, 19; Japanese 70%, 17). Surprisingly, the 72% for irregular verbs on the multiple-choice test is lower than the 76% on the oral test and the 80% on the fill-in test. The responses for ming, ning, and tring include 79 regular gg_forms (Spanish 26; Arab 21; Japanese 32). Fifteen irregular forms (e.g., mang, hang, and ipppg) occur (Spanish 1, Arab 10, Japanese 4). One Arab also gave pppg, More Arabs than Japanese or Spanish, thus, gave strong verb forms. Use of such forms probably reflects a generalization based on well-learned paradigms like sing, sang, sung_rather than grammatical transfer 13 from the native language. Statistically significant differences among the three lan- guage groups show up in two tense structures: present s_MV (.05) and present s_passive (.05).14 13Although vowel changes occur in some verbs in all three languages, in none of the languages is the vowel change systemati- cally used to indicate changes in tense in the way that English does. Examples of vowel change are: Arabic /yaqfimu/ "he gets up“; /qama/ "he bought"; /qumtu/ "I bought"; Spanish /dwermo/ "I sleep"; /dormimos/ "we sleep"; /dormi/ "I slept"; /dormimos/ "we slept"; Japanese /su/ "I do"; /seri/ "I did." 14See Appendix C for chi square values by native language. 122 Modal The data show a 43% correctness for all modal structures for the Spanish, 53% for the Arabs, and 75% for the Japanese. All percentages for pg_(Spanish 54%, Arab 64%, and Japanese 88%) and for M!_(Spanish 50%, Arab 73%, and Japanese 84%) remain close and much higher than those for have (Spanish 27%, Arab 33%, Japanese 56%). The least correctly answered modal structure for all three languages is modal perfect MV (Spanish 18%, Arab 26%, Japanese 44%). On the multiple-choice test for modal perfect passive all Japanese speakers gave the correct have response. Once again the percentages are the highest on the multiple- choice test and the lowest on the oral test with those on the fill-in test in the middle with these exceptions: l. modal M!_(Spanish: oral 54%, fill-in 43%, multiple- ' choice 54%; Arab 77%, 70%, 72%)- 2. modal have+en MV (Spanish: fill-in 25%, multiple- choice 25%) 3. modal have+en progressive MV (Spanish: fill-in 37%, multiple-choice 33%; Arab: 50%, 45%; Japanese: 83%, 83%) 4. modal pgfing MV (Japanese: fill-in 95%, multiple- choice 93%) 5. modal pgfen MV (Arab: fill-in 97%, multiple-choice 93% 6. pp_§all) (Japanese: fill-in 95%, multiple-choice 95% . The results for modal, like those for tense, show the Spanish with the lowest percentages, the Japanese with the highest percentages and the Arab percentages in the middle. Of the all averages, only the Spanish 46% for modal pgfing MV does not follow this pattern since the Arab average is less at 42%. Several other exceptions occur, too, especially on the oral test: 123 AAAAA RAA AAAAA RAA AAAAA RAA ARA RANRA RAA AAAA RNA 2AAA RAA >AA AAANA RAA AAARA RAA AAA_A RAA A>A2 2AAA RAA ARAA RAA RAAA RNA R>=2 AAAAA RAA >AA>A RAA AAAAA RAA _>AA 2AAA RAA AARA RAA >A_V RAA ARAA AAARA RAA 2AAA RAA 2AAA RAN AAAA 2AAA RAA AAAA RNN AAAA RAN AAAA _AAos RRA >2 AAARAA AAAAE AAAAA ANNAA RAA AAAAA RAA AAAA RAA AAA 2_AA RAA AAAA RAA AAAA RAA ARA 2AAA RAA AAAA RNA AANA RAA ARA2 AANA RAAA AAAA RAA AAAA RAA RA=2 2AAA RRA AAAA RAA AANV RAA _RAA AANA RAA 2AAA RAA ANAA RAA >>AA 2AAA RAA 2AAA RAA AANA RAA AAAA 2AA RAN ANA RA AAA RN> RAAA .mm + Raves __< >2 =m+mn cm+m>A2 —Auos mmzmh 2AANA RAA 2AA>A RAA 2AA>A RAA ARA RAAA RAA AANA RAA AANA RAN AAA ANAA RAA 2AAA RAA AAAA RAA A222 AANA RAA AAAA RAA 2AA RAA “>32 RNAA RAA AAAA RAA 2AAA RAA AAAA AANA RAA ANRA RAA AAA RAA FARR 2AAA RAA AAAA RAA 2AAA RNA AAAA 2AA RAN 2AA RNA 2AA RAA >A2A .wm + FAcoE —_< >2 mc>+mn =w+m>A2 Pmcoe wmcmh AAAAA RAA AAAA RAA 2AAA RAN AAA 2AAA RAA 2AAA RAN AANA RAA ARA 2AAA RAA AAAA RAA AAAA RAA AA32 2AAA RAA 2A>v RAA RNAA RAN AAA2 2AAA RAA 2AAA RAA 2AAA RAA AAAA AANV RAA ANAA RAN ANAA RAN ARAA ANNA RNN 2A>A RAA AARA RAA AAAA AAA RAA 2>AA RNN AAA RA AAAA w>ms + FMUOE _.—< >2 cm+m>ws :69: mmcob RRAAV RAA RANAA RAA A>AA RAA ARA ANNAA RAA 2AA>A RAA RAAA RAA RAA AAAA RAA AAAA RAA AAAA RAA A>32 2AAA RAA 2AAV RNA AANA RAA RA32 AAAA RAA 2AAV RAA 2AAA RAA RAAA 2AAA RAA 2AAA RAA 2_NV RAA A>AA 2AAA RAA RNAV RAA >A>A RAA AAAA 2AAV RAA 2AAA RAA AANV RAA RAAA >2 cotmm Ponce mmcmh .mm FAAAE wmcmp mAmcmamw AAA< 2ARAAAA AAAAAAAA AA>< 2A>cmam uomAAou Amasaz 22A acmuAmA “Amp uquAou Amaszz 22A ucmugma ummh II .AAAAAAAA An Fooo2--.>~ m2m<> 124 l. modal have+en MV (oral: Arab 22%, Japanese 16%; fill-in: Spanish 25%, Arab 25%) 2. modal have+en progressive MV (ora1: Spanish 16%, Arab 12%) 3. modal have+en passive MV (oral: Spanish 12%, Arab 8%) 4. modal pgfen MV (fill-in: Arab 97%, Japanese 95%). For modal Mi_and for modal psfen MV, the results show a sta- tistically significant difference (.05) in the performance of the three different language groups; for modal pgfing MV the difference is highly significant (.01). Perfect (Have+en) The 46% all correctness for perfect by the Spanish is less than the 56% by the Arabs and the 73% by the Japanese. The gp_of perfect (Spanish 53%, Arab 65%, Japanese 82%) is more correctly answered than the psyg_(Spanish 37%, Arab 44%, Japanese 62%). The Japanese answered every item correctly on the multiple-choice test for: modal have+en passive MV have+en progressive MV modal—have+pp_progressive MV (and on the fill-in test, too). (JON-4 Every percentage on the oral test is lower than the cor- responding percentage on the fill-in test. Many fill-in percentages, though, are higher than or equal to the multiple-choice percentages: 1. modal have+en MV (Spanish: fill-in 25%, multiple- choice 25%) 2. modal have+en progressive MV (Spanish: fill-in 37%, multiple-choice 33%; Arab: 50%, 45%; Japa- nese 83%, 83%) 3. have+gp_progressive MV (Spanish and Arab: fill—in 83%, multiple-choice 83%) 4. have+gp_passive MV (Japanese: fill-in 95%, multiple-choice 91%) 125 2AAA RAA 2AAV RAA 2AAA RAA _2A AANV RAA> AANV RAA RANV RAA R_32 ANNA RAA AANV RAA AANV RAA AARA 2A_v RAA 2AAA RAA AAA RAA _AAA >2 ocw+ma QM+m>m2 mAcm> RAAAA RAA 2AAAA RAA 2AAAA RAA __< AAANA RAA 2NA>V RAA RAAAV RAA AAA 2AAAA RAA AAANA RAA AAANA RAA AAA2 AAAA R_A ANAA RAA AAAV RAA R_=2 AAANA RAA 2AANA RAA 2AAAA RNA _AAA 2AAV RNA 2AAA RAA AAAA RAA _AAA 2AAAA RAA AANFA RAA 2AAA RNN AALA AAAA RAA 2AAA RAA AANA RAN _AAA uumAAmm __< >2.mw+m>mg mmcmh 2AAAA RNA AAAAV RAA 2AAAA RAA __< 2AAA RAA 2AAA RAA AAAV RAA >_A 2_A_A RAA APARA RAA 2_AAA RAA R_=2 RANA RAA ARNA RAA AANA RAA RRA2 A_A_A RAA RNAAA RAA 2AAAA RAA AAAA 2AAA RAA 2AAA RNA AARA RNA AAAA RANAA RAA ANAA RAA 2AAA RAN AAAA 2AA RAN 2AA RAA 2AA RAN _AAA mm _>< >2 2m+on cm+m>A2 mmcm> 2AANV RNA ANAAA RAA RNAAV RAA >_A AAAV RAA RRAA RAA 2AAA RAA _AA RANAA RAA 2AAV RAA 2AAV RAA p.32 RANV RAA ANNA RRA RRNV RAA RA32 2AA>V RRA 2AAV RAA 2AAA RAA AAAA ANNA RRA AARA RAA 2A_A RAA _AAA AAAA RAN 2AAA RAN AARA RAA AAAA AAAA RAA AA_A RAA 2AA RAA _AAA w>o2 __< >2 mcw+mn cw+m>A2 mmcmh 2AAA RAA RAAA RAA AAAA RAA RAA AAAA RAA AAAA RAA A_AA RAA >_A RANA RAA >A>v RAA 2AAV RAA RAAz AANV RAAA 2AAA RAA AAAA RAA R>A2 AANA RAA 2A_A RAA A>>A RAA AAAA RANV RAA 2AAA RAA AN_A RAA __AA 2A>A RAA ARRV RAA AAA RAN _AAA AAA RAN 2N2 RA 2AA RNR AAAA >2 cm+mn.mm+m>mz Favoe mmcmh >2 cm+mn cm+m>mc Pmnoe mmcm> AAAA RAA 2AAA RAA RNA. RAA >>A 2AAA RAA AANV RAA AANA RAN ARA AANV RAA. AANV RAA AARA RAA p.32 AANV RAA ARRA RAA 2A2 RAA AAAz AANV RAA> 2A_2 RAA 2AAA RAA ARAA RANA RAA ANAV RAA 2A2 RAA R>RA AA_A RAA 2AAA RAA 2AA RAN >AAA 2AA RAN 2AA RNA AAA RA_ AALA >2 mcw+mn dM+m>A2 Ponce mmcmh >2 mcw+wn cm+m>A2 Fouoe mmcm> ANAV RAA ANAV RNA AAAA RAA AAA 2AAA RAA 2AAV RAN AANV RA_ AAA ANNA R_A AANA RAA ANNA RRA RAA2 AAAA RAA 2A_A R_A RNRA RAN A_A2 AANA RAA 2AAA RAA 2A_v RAA _RAA AANA RAA ANAA RAN ANAA RAN AAAA AARA RAA AAAA RAA 2AA RAA AALA 2AA RAR AARA RNN AAA RA _AAA >2 co+mo.mm+m>mc wmcm> >2 2w+m>A2 paves mmcmh mchAAma AA2< 2AAcAam mAmcmaoo AAA< 2A>2AAA RAAA RAAA Ruwgcou gossaz 22m ucmugma uumggou Amn832 22A acmugmm .AAAAAAAA AA RAAAAAA--.NN AAAAA 126 5. modal have+§p_progressive MV (Spanish: fill-in 75%, multiple-choice 70%; Japanese 100%, 100%) 6. modal have+gp_passive MV (Arab: fill-in 75%, multiple-choice 75%; Japanese: 95%, 95%). The all percentages for each structure usually show the Spanish with the lowest percentages, the Japanese with the highest, and the Arab percentages in the middle, but these exceptions occur: 1. modal have+en MV (oral: Arab 22%, Japanese 16%; fill-in: Spanish 25%, Arab 25%) modal have+en progressive MV (oral: Spanish 16%, Arab 12%) modal have+en passive MV (oral: Spanish 12%, Arab 8%) have+en progressive MV (oral: Arab 54%, Japanese 54%; fill-in: Spanish 75%, Arab 66%) have+en passive MV (oral: Arab 33%, Japanese 25%, fill-in: Spanish 62%, Arab 62%) have+§p_progressive MV (fill-in and multiple- choice: Spanish 83%, Arab 83%) have+§p_passive MV (fill-in: Spanish 75%, Arab 70%; multiple-choice; 91%, 87%) modal have+pp progressive MV (fill-in: Spanish 75%, Arab 58%). CDVOHU'IDOON The Spanish percentages never surpass or equal the Japanese per- centages but the Arab percentages occasionally do. The Spanish, though, sometimes do better than the Arabs. Of the gg_responses 51% (111) of the 216 responses show gg_ on the oral test (Spanish 29%, 21; Arab 52%, 38; Japanese 72%, 52); 69% (151) show pp_on the fill—in test (Spanish 51%, 37; Arab 72%, 52; Japanese 86%, 62) and 84% (183) do on the multiple-choice test (Spanish 70%, 51; Arab 83%, 60; Japanese 100%, 72). Percentages for the irregular verbs are lower. 0n the oral test 41% (30) of the 72 responses show gp_for sipg_and mipg_(Spanish 25%, 6; Arab 33%, 8; Japanese 66%, 16); 47% (34) show gp_for pipg_and pipg_on the fill-in test (Spanish 25%, 6; Arab 45%, 11; Japanese 70%, 17); 127 and 58% (42) do for spring and tring on the multiple-choice test (Spanish 58%, 14; Arab 50%, 12; Japanese 66%, 16). For ming, ning, and iiipg, both gg_(e.g., mipggg) and ppg_(e.g., mgpg) were counted as correct. 0f the 56 gg_forms, the Spanish had 15, the Arabs l3, and the Japanese 28. 0f the 17 correct irregular forms, the Spanish gave 4, the Arabs 9, and the Japanese 4. In addition, 6 incorrect irregular forms (e.g., mang, hang, nong) occurred (Arab 5, Japa- nese 1). The Arabs, thus, gave irregular forms more often (14) than either the Spanish (4) or the Japanese (5). 0f the ten structures for perfect, statistically significant differences among the three language groups can be found for only three: a significant difference (.05) for have+gp_passive NV and modal have+gp_passive MV, both structures requiring pggp, and a highly significant difference (JII)for have+§p_MV. Progressive (Be+ing) The 44% all correctness for progressive by the Spanish com- pares with the 52% by the Arabs and the 71% by the Japanese. While pg_is learned equally well by the Spanish (50%) and the Arabs (50%) but much better by the Japanese (77%), the igg_is learned less well by the Spanish (31%) than by either the Arabs (56%) or the Japa- nese (56%). The Japanese answered every item correctly for two similar structures: perfect pgfing MV on the multiple-choice test and modal perfect pgfing MV on the fill-in and multiple-choice tests. The results for psfing passive MV (Spanish 31%, Arab 37%, Japanese 47%) and for perfect pgfing MV (Spanish 68%, Arab 69%, 128 .1» 2G vhfo. AAAA RAA 2AA. RAA AANA RAA AAA 2AAA RAA AANA RAA ANAA RAA AAA2 AANA RAA AA>A RAA ARAA RAA AAAA AA_A RAA AARA RAA 2AA RAN _AAA >2.mmm+mn mmcmp 2AAAA RAA AAANA RNA AANNA RAA _AA 2AAA AAA AAAA RAA ANAA RAA AAA 2AAAA RNA AANRA RAA 2AARA RAA RAAz AANA RAAA AANA RAA AARA RAA RA22 AAA—A RAA 2AAA RNA RAAA RAA _RAA AANA RAAA 2AAA RAA 2AAA RAA ___A 2AAA RAA AAAA RAN AAAA RAN AAAA 2AAA RAA AAAA RAA 2AA RAN AAAA m>AAAmAmocm __< >2 chtmm 2m+m>mz _Auos oAcm> ANAA RAA 2_AA RAA AAAA R>A AAA 2AAA RRA AAAA RAA AAAA RAA _>A R_AA RAA 2AAA RRA AANA RAA R_A2 RANA RAAA AANA RAA AANA RAA AAA2 ANAA RAA AANA RAA >A_A RAN ARAA ANNA RAA RANV RAA RANA RAA ___A AAAA RAA AAAA RAN 2AA RAA AAAA AAAA RAA AARA RAA 2A. RAA AAAA .AAH AAA >2 AAAAAA AA+A>A2 AAAAA AAANV RAA RNARA RAA AAAAA RAA ARA 2AAA RAA AANA RAA RANA RRA A_A RAAAA RAA AAAA RAA ANAA RAA A>A2 2AAA RNA 2AA2 RAA 2AAA RAA A>=2 2AAAA RAA AAAA RAA 2AAV RAA RRAA AA_V RAA AAA RAA RAAV RAA _AAA 2AAA RAA AAAA RAA AAAA RAN AAAA 2AA RAN AAA RAN AAA RNA _AAA .mm RAA >2 AA+AA AARAAA AAAAA AAAA RAA ANAA RAA 2AAA RAN >AA AANAA RAA 2_AA RNA AAAA RAA _AA 2AAA RNA AAAA RAA AN_A RAA RAA2 2AAA RAA RAAA RAA 2AAA RNA AAA2 RNAA RAA AAA RAA RAA RNA RRAA 2AAA RAA RARA RAA RA_A RAA AAAA 2AA RAN 2AA RAR 2N2 R RAAA 2AAV RAA 2A>A RNN AA_A RAN PARA >2 cm+mn.mmm+mn mA2m> >2 mcwtmm vaoa mmcm> ommcmama AAA< cmwcmam mAchAAA AAA< 2A>2AAA ummw uumggou Amasaz can acmoemm N. '11 ummh uumgeou Amasaz 22A acmoemm 11‘. 1 .I .I. '..1.1 1- I I I'lll'q‘lflnnr 1" .mmmamcm2 >2 m>AAAmAAoAa--.mN m2m<> l 1' 1 1 117 II I I7 [[1 hr. . 1 129 Japanese 81%) show the least differences among the three language groups. The fill-in percentages are always greater than the cor- responding oral percentages. Not every multiple-choice percentage, though, is greater than its corresponding fill—in percentage as these results show: 1. modal pgfing MV (Japanese: fill-in 95%, multiple- choice 93%) pgfing passive MV (Spanish: fill-in 41%, multiple- choice 41%) perfect psfing MV (Spanish and Arab: fill-in 83%, multiple-choice 83%) modal perfect pgfing MV (Spanish: fill-in 75%, multiple-choice 70%; Japanese 100%, 100%) be+ipg_MV (Japanese: fill-in 83%, multiple- choice 66%) ipg_(all) (Japanese: fill-in 66%, multiple- choice 64%). 0301th The pattern of low Spanish, intermediate Arab, and high Japanese percentages occurs less consistently here as the results for ps_(all) and ipg_(all) mentioned before indicate. Further exceptions‘include: 1 ppfing passive MV (fill-in: Spanish 41%, Arab 33%) 2 perfect pgfing MV (fill-in and multiple-choice: Spanish 83%, Arab 83%) 3. modal perfect pgfing MV (fill-in: Spanish 75%, Arab 58%) 4 be+ipg_MV (multiple-choice: Arab 83%, Japanese 66%; all: Arab 68%, Japanese 68%) 5 be+ipg_passive MV (multiple-choice: Arab 79%, Japanese 62%). In addition, the 46% all correctness by the Spanish for modal pgfing MV is greater than the 42% by the Arabs. The Spanish percentages never equal or surpass the Japanese percentages. The gap between Spanish and Japanese is smallest on the multiple-choice test (19%) and widest on the fill-in test (34%). 130 The only progressive structure which showed a statistically significant difference among the three language groups is modal spying NV (.01). Passive (Be+en) For passive the 54% correctness for the Spanish is less than the 67% for the Arabs and the 82% for the Japanese. §p_per- centages (Spanish 63%, Arab 70%, Japanese 91%) are higher than pg_ 4 percentages (Spanish 53%, Arab 66%, Japanese 80%). 0f the five 7 ' structures testing for pp, the results for progressive ppfen MV are particularly low (Spanish 23%, Arab 44%, Japanese 45%). No language group ever got a 100% accuracy percentage for pg_though the Arabs on the fill-in test and the Japanese on the multiple- choice test for modal psfen MV both came close at 97%. For the only §p_structure on the test, be+pp_MV, the Japanese achieved 100% on the multiple-choice test. With the exception of the 41% by the Arabs on the oral and fill-in tests for ppfen MV, all other fill-in percentages surpass the oral percentages. The multiple-choice percentages surpass the fill-in percentages except in these cases: 1. modal pgfen MV (Arab: fill-in 97%, multiple- choice 93%) perfect pgfen MV (Japanese: fill-in 95%, multiple- choice 91%) modal perfect pgfen MV (Arab: fill-in 75%, multiple-choice 75%; Japanese: 95%, 95%) be+gp_MV (and all en) (Spanish: fill-in 75%, multiple-choice 75%). DOOM The pattern of low Spanish, intermediate Arab, and high Japanese percentages occurs regularly for the passive. Tense 131 AAAV RAA ANAA RAA 2AAA RAN AAA 2AAA RNA AAAA RAA ANAA RAA AAA2 ANAA RAA AAA RAA 2AA RNA _.AA AAA RAN AAA RAA ANA R AALA >2 cmtmm m=A+mn mmco> AAAAA RNA AAAAA RAA AAANA RAA AAA AAAA RAA AAAA RAA 2AAA RAA AAA AAA. RAA AAAAA RAA ANNAA RNA RA32 AANA RAA AAAA RAA 2AAA RAA RA32 AAA RAA AAAAA RAA 2AAAA RAA AARA AANA RAA AAAA RAA >A>A RAA AAAA N>A RAA 2AAA RAA 2AAA RNA AAeA 2AAA RAA AAAA RAA 2AA RAN AAeA m>Rmmam >_< >2 cm+wm cm+m>A2 Rance mmcmh 2AAA RAA AAAA RAA 2AAA RAA AAA ANAA RAA ANAA RNA 2AAA RAA AAA AANV RAAA AAAA RAA 2AAV RAA RAa2 ANNA RAA A>NA RAA ANNA RAA RAa2 AANA RAA AAAA RAA AAAA RAA _AAA AANV RAA AARA RAA 2AAV RAA AAAA 2AAA RAA AAA. RAA AAAA RAA AAAA AAAA RAA AAAA RAA AAA RAA AAAA .mm >_< >2 cmkwm cm+m>ms «Acmp AAAAA RAA AAANA RAA AANNA RAA AAA A_A_A RAA AANAA RAA AAAA RAA AAA 2AAAA RAA AAA—A RAA AAAAA RNA ARA2 AAAV RAA 2AAA RAA AAAA RAA RAA2 AANAA RAA AAAAA RAA ANAA RAA A>AA AAAA RAA AAAA RAA 2AAA RAA PARA 2AAA RAA AAAA RAA 2AAA RAN AAAA AAAA RAA ANAA RAA AAAA RAA AaeA .AA AAA >2 AAAAA AAAAE‘AAAAA AAAA RAA AAAA RAA AAAA RAA APR 2AAA RAA ANAA RAA 2AAA RAA AAA AANA RAAA 2A_A RAA 2AAA RAA R>=2 AANV RAA ANAA RAA “AAA RNA RAA2 AANA RAA AAAA RAA 2AAA RAA AAAA ANNA RAA 2AAA RAA 2A>A RAA _RAA AAAA RAA AAAA RAA AAFA RAA AAAA AAAA RAA AAAA RAA AAAA RAA AAAA . >2.mM+mn mmcwh >2 cmxmm mmcm> oAchAAn AAA< 2AA2AAA mAchAAa AAA< 2AA2AAA RAA» RAAA uumggou Amnsnz 22A acmugma pumegou Lassaz 22A ucmoewa .ommamcm2 >n w>AAAA21-.AN m2m<> 132 pgfen MV does not follow this pattern as the Spanish 55% all per- centage is higher than the Arab 44%. Other exceptions include: 1. modal pgfen MV (fill-in: Arab 97%, Japanese 95%) 2. perfect pgfen MV (fill-in: Spanish 75%, Arab 70%; multiple-choice: Spanish 91%, Arab 87%, Japanese 91%) 3. progressive be+en MV (multiple-choice: Arab 79%, Japanese 62%)_' 4. be+pp_MV (and all pg) (fill-in: Spanish 75%, Arab 70%; multiple-choice: Spanish 75%, Arab 75%). The multiple choice results for perfect pgfen MV are especially interesting since both Spanish and Japanese have 91%. Only in this instance does the Spanish percentage equal the Japanese percentage. The Spanish percentage never surpasses the Japanese percentage on any test for any structure. Statistically significant differences among the language groups at the .05 level occur for pg_in: ppfen MV, modal ppfen MV, perfect ppfen MV, and modal perfect psfen MV. The Infinitive (To my) The 65% Spanish all response for infinitive is lower than the 81% Arab response or the 85% Japanese response. -All language groups supplied MV (Spanish 79%, Arab 84%, Japanese 90%) more cor- rectly than they did is_(Spanish 52%, Arab 79%, Japanese 80%). All fill-in percentages exceed the oral percentages, but no multiple-choice percentages surpass the corresponding fill-in percentages (Spanish: fill-in 87%, multiple-choice 83%, Arab: 87%, 87%; Japanese: 100%, 100%). The general pattern for the language groups prevailed here, too, though the Arabs did particularly well. Exceptions to the pattern included: 133 TABLE 25.--Infinitive by Language. Percent and Number Correct Test Spanish Arab Japanese Ip_MV Oral 20% (5) 54% (13) 62% (15) Fill 45% (11) 87% (21; 79% (19) Mult 91% (22) 95% (23 100% (24) All 52% (38) 79% (57) 80% (58) To M!_ Oral 66% (16) 79% (19) 70% (17) Fill 87% (21) 87% (21) 100% (24) Mult 83% (20) 87% (21) 100% (24) All 79% (57) 84% (61) 90% (65) All infinitive Oral 43% (21) 66% (32) 66% (32) Fill 66% (32) 87% (42) 89% (43) Mult 87% (42) 90% (44) 100% (48) All 65% (95) 81% (118) 85% (123) l. 2. 3. to MV (fill-in: Arab 89%, Japanese 79%) EB'MV (oral: Spafi1sh 87%, Arab 87%) all (oral: Arab 66%, Japanese 66%). Arab 79%, Japanese 70%; fill-in: Neither structure shows statistically significant results. The Gerund (Ing MV) with the 66% for the Arabs and the 52% for the Japanese. The 27% correctness by the Spanish for gerund compares The responses by language groups fail to follow the general pattern, as the Arab all percentage is better than that of the Japanese, the only time such a result occurs. Responses on the three tests, 134 TABLE 26.--Gerund by Language. Percent and Number Correct Test: Spanish Arab Japanese Oral 20% (5) 54% (13) 54% (13) Fill 37% (9) 66% (16) 54% (13) Mult 25% (6) 79% (19) 66% (16) All 27% (20) 66% (48) 52% (42) though, show the expected pattern otherwise but with two excep- tions: 1. Japanese: oral 54%, fill-in 54% 2. Spanish: fill-in 37%, multiple-choice 25%. The Spanish all percentage here is markedly lower than usual. The only lower Spanish all percentages on the test are the 18% for modal have+en MV and the 23% for be+ipg_pgfen MV. The Allomorphs The oral test data for the allomorphs of the present s_and past morphemes show a consistent pattern: the Spanish supplied the fewest correct responses; the Japanese supplied the most; the Arabs nearly always fell in the middle. For present s_the /s/ is correctly given in 39 of the 72 responses (54%) (Spanish 20%, 5; Arab 62%, 15: Japanese 79%, 19), the /z/ in 41 (56%) (Spanish 29%, 7; Arab 58%, 14; Japanese 83%, 20); and the /Iz/ in 28 (38%) (Spanish 8%, 2; Arab 33%, 8; 135 Japanese 75%, 18). The differences in the percentages between Arab ‘ and Spanish responses for /s/, /z/, and /Iz/ are 42, 29, and 25, respectively, while the differences between Japanese and Spanish responses are 59, 54, and 67, respectively. These differences are clearly large. The Spanish, in particular, did very poorly. For the past tense /t/ is correctly given in 120 of the 216 responses (55%) (Spanish 30%, 22; Arab 52%, 38; Japanese 83%, 60); /d/ in 111 of the 216 responses (Spanish 19%, 14; Arab 55%, 40; Japanese 79%, 57); and /Id/ in 56 (38%) of the 144 responses (Spanish 20%, 10; Arab 45%, 22; Japanese 50%, 24). Both the Japa- nese and the Spanish correctly supplied /t/ more often than /d/; the Arabs supplied /d/ more often than /t/. While the Arabs and the Japanese supplied /Id/ the least often, the Spanish supplied /Id/ a little more often than they did /d/. Excluding zero allomorphs, errors in past allomorph usage occurred in 54 (15%) of the 341 responses. Most were made by the Spanish (22) and the Arabs (23); the Japanese made only 9 errors. Incorrect past allomorph responses included: l. /d/ for /t/ (e.g., /jampd/) (Arab 4, Japanese 2) 2. /Id/ for /t/ (e.g., /jampId/) (Spanish 9, Arab 4, Japanese 1) 3. /t/ for /d/ after nasgls or liquids (e.g., /plint/) (Spanish 11, Arab 2) 4. /Id/ for /d/ after nasals (e.g., /klinId/) Span- ish 2, Arab 5, Japanese 2) 15 In English /t/ sometimes occurs as the past allomorph after nasals and liquids as in /fElt/, /bernt/, and /drEmt/. Consequently this use of /t/ does not need to be regarded as an error; however, it fails to follow the regular past pattern whereby /d/ follows any voiced phoneme except /d/. 136 5. change of final stem (e.g., froo for frool) (Arab 8, Japanese 4). Besides these errors, 7 instances occurred of substitution of /d/ for /t/ in the stem with zero allomorph added (e.g., zind for the uninflected zint) (Arab 4, Japanese 3). Summary An analysis of the data by language groups reveals two trends: (1) the highest percentages occur on the multiple-choice test; the lowest, on the oral test; and (2) the Spanish consis- tently gave the fewest correct responses; the Japanese, the most. It is surprising that speakers of Japanese, a non-Indo- European language, should consistently do better than speakers of Spanish, an Indo-European language which has more similarities to English in grammar and vocabulary. English proficiency scores used in classifying students into beginning, intermediate, and advanced levels may help explain these results. Little difference occurs in the composite scores by language groups: out of 100 points the Spanish average 62; the Arabs, 65; and the Japanese, 65. Of the three areas on the proficiency test, the scores for composi- tion (Spanish 66, Arab 67, and Japanese 66) and aural comprehension (Spanish 63, Arab 67, Japanese 62) are not far apart. Grammar scores are not as close: Spanish 58, Arab 63, Japanese 67. The spread in grammar scores is most noticeable at the beginning lan- guage level (beginning Spanish 31, beginning Arab 45, beginning Japanese 50). The grammar score is the crucial one, since our 137 study mainly tested the student's knowledge of English grammar rather than his ability in aural comprehension or composition. Not one of the Spanish all percentages for the 23 struc- tures ever equaled its Japanese counterpart. Only on the multiple- choice test for have+sp_pgfen MV did the Spanish (91%) do as well as the Japanese (91%). The Spanish all percentages for ippsp_ (present s) ppfen MV (55%) and for modal psfing MV (46%) did sur- pass the Arab percentages (44% and 42%, respectively). Moreover, the 50% all correctness by the Spanish for pg_of the progressive equaled the 50% by the Arabs as did the Spanish 56% for ipg_of the progressive. In contrast the 66% by the Arabs for gerund surpassed the 52% by the Japanese, and the Arab 68% for tense be+ipg_MV equaled the Japanese 68%. On the following tests, the Arab percentages (given first) surpassed those of the Japanese: tense modal have+en MV (oral: 22%, 16%) tense modal pgfen MV (fill-in: 97%, 95%) tense be+igg_ppfen MV (multiple-choice: 79%, 62%) _t_p_MV (fill-in: 87%, 79%). to My_(ora1: 79%, 70%). mkWN-J o o o o o The 54% by the Arabs on the oral test for tense (present s) have+en be+ing MV equaled the 54% by the Japanese; the Arab 92% for igpss_ (ppsi) on the multiple-choice test also equaled the Japanese 92%. The Arabs, thus, did especially well in supplying ipg_of the pro- gressive and the gerund. Statistically significant differences in the results by language groups can be found in 8 of the 23 structures. At the .05 level they occur in: 138 tense (present s) MV tense modal py_ tense modal ppfen MV tense have+gp_psfen MV tense modal have+pp_ppfen MV tense pgfen MV. 01014:de I o o o o 0 Two occur at the .01 level: (1) tense modal ppfing MV and (2) tense have+pp_MV. Most of these differences involve the pg of perfect or the pg_of passive. CHAPTER IV SUMMARY AND CONCLUSION Statistically significant differences in the results by level of proficiency occur in 13 of the 23 verb structures while significant differences by language occur in 8. Two structures show statistically significant differences for both level and language: (1) tense have+gp_MV (.01, .01, respectively) and (2) tense have+gp_pgfen MV (.Ol, .05). Three structures show nonsignificant differences for both level and language: (1) tense (past) MV; (2) tense (present s) have+en be+en MV; (3) tense be+gg_MV. In the remaining 18 structures, the results by level and language differ: a statistically significant dif- ference in the result for one occurs for each of these structures while a nonsignificant difference occurs in the other. For instance, the is_MV results show a significant difference (.05) by level of proficiency but a nonsignificant difference by language. In brief, little overlap occurs in the statistical results by level and by language. The data from out study suggest that nonnative speakers learn the elements of the English auxiliary and the infinitive and gerund transformations in this order: 139 140 ll. infinitive 2. tense 3. passive 4. perfect 5. modal 6. progressive / 7. gerund. The order for perfect, modal, and progressive, in particular, is not clear. The data show an interesting pattern concerning the acquisition of the two transformations: the infinitive (is_My) is acquired before all elements of the auxiliary; the gerund (ipg_ 1 my), in contrast, is acquired last. Tense (past) is acquired before tense (present s); the gp_of passive, before the pp; the pp_of perfect before the psyg; and the pg of progressive before the ipg,2 0f the structures that can follow a modal, My_and ps_ are acquired much earlier than ppyg, Few comparisons of our study with other Berko-type studies can be made since they mainly investigated noun morphology. If they did investigate verb morphology, they nearly always restricted themselves to the acquisition of the present s_and the past allo- morphs. For present s_our study shows that nonnative speakers acquire /s/ and /2/ before /Iz/. This finding agrees with most previous Berko-type studies (e.g., Bellamy and Bellamy 1970; Ramer and Rees 1973). For past and the gp_of perfect our study shows that nonnative speakers acquire /t/ and /d/ before /Id/.3 1Both Arab and intermediate students acquired the gerund before they acquired perfect, modal, or progressive. 2The Japanese acquired present s_and past at the same time; the Arabs acquired the ipg_of progressive before the pg, 3The Spanish supplied /Id/ more frequently than /d/. 141 Similar findings have been found in earlier studies (e.g., Dever and Gardner 1970; Lovell and Bradbury 1971). A few studies of native English-speaking children showed that past was acquired before present s, as our findings for nonnative speakers suggest (McNeill 1970: 83 and Cooper 1967: 83). Although this study indicates that nonnative speakers acquire the pg_of progressive before ipg_(with the exception of the Arabs), several studies on native English speakers indicate the reverse (Dale 1972: 54; Vogel 1970: 60-1; and Martinez-Bernal 1972: 120).4 All 23 verb structures may be classified by the number of elements composing them. There are 2 element structures (e.g., is_My); 3 element structures (e.g., tense be+ipg_MV); 4 element structures (e.g., tense ppfing be+en MV) and 5 element structures (e.g., tense modal hpygfen be+en MV). As might be expected, the fewer the number of elements, the greater the number of correct responses. Thus total correct responses for 2 element structures generally surpass those for 3 element structures, which, in turn, surpass those for 4 element structures. The data show, though, that the total correct responses for 5 element structures may slightly exceed those for 4 element structures. Exceptions do occur. The most noticeable exception is igg_My, a 2-element structure ranking in the lowest quarter of the 23 structures. Two 4-element structures (tense modal psfen MV and tense have+§p_ 4In the Vogel and Martinez-Bernal studies other factors such as nonstandard dialects and demonstrated test question flaws must also be taken into account. 142 psfen MV) are learned much more thoroughly than this general pat- tern would indicate. Several trends emerge from our study. In accordance with the idea that receptive tasks are accomplished more easily than productive tasks, the correct responses on the receptive task (the multiple-choice test) nearly always outnumber those on each pro- ductive task (the oral and the fill-in tests). Similarly, in accordance with the idea that written tasks are accomplished more easily than oral tasks, the correct responses on each written task (the fill-in and multiple-choice tests) usuallyr outnumber those on the oral task (the oral test). (See Anisfeld and Tucker 1967; Koziol 1971: 111-2). The greatest number of correct responses occurs on the multiple-choice test; the fewest, on the oral test; the number of correct responses on the fill-in test usually falls in the middle. Similarly, the level of a student's proficiency in English usually gives a good indication of his performance. The advanced students gave more correct responses than beginners in every case except one: to My_(fill-in: beginner 95%, 23; advanced 91%, 22). Correct responses by intermediate students fell in between those by beginning and by advanced students. Exceptions occur now and then (e.g., tense modal have+gp.pgfen multiple choice: beginner 79%, 19; intermediate 75%, 18; and is_MV multiple choice: intermediate 100%, 24; advanced 95%, 23). The data by native language group also indicate a trend. The Spanish gave the fewest correct responses while the Japanese gave the most; the Arab responses fell in the middle. (In four 143 instances the Arab response did not fall in the middle. For ippsg_ (present s) be+en MV the Spanish 55% surpassed the Arab 44%; for tense modal ppfing MV the Spanish 46% surpassed the Arab 42%; for gerund (ipg_My) the Arab 66% surpassed the Japanese 52%; and for tense be+ipg_MV the Arab 68% equaled the Japanese 68%). Since the grammar scores on the proficiency test in English show this ascending Spanish-Arab—Japanese order, the result is not unex- pected (see Summary, Chapter III). Our data reinforce the results of the proficiency test. Because of this, our results do not necessarily mean that this pattern would have occurred if the grammar proficiency scores by language had been identical. The real differences in English learning by Arabic, Japanese, and Spanish speakers may be less than our study indicates. Analysis of variance reveals highly significant differences (.001) for performance by level of proficiency and by native lan- guage on the oral test, the fill-in test, and the multiple-choice test. Highly significant differences (.001) occur for test per- formance on nonsense items by both level and language; the same is true of test performance on real word items. Analysis of variance indicates little difference in test performance on nonsense and on real words. In a study of noun morphology Martinez-Bernal (1972: 98, 100) found: a high degree of relationship between performance on the real language items and performance on the imaginary items. Those who did well on one did well on the other . . . . The difference in performance on imaginary items when presented alone compared to imaginary items preceded by real language examples was found not to be significant at the 80% level of confidence. 144 But a British study by Herriot (1968: 273) showed: a significant difference between conditions consisting of nonsense and those consisting of sense for comprehension tasks undertaken by young children. Herriot found 1K) significant differences for the tasks undertaken by older children. The Newfield and Schlanger (1968: 705) study on the acquisition of English morphology by normal and retarded children revealed that: statistically significant differences existed between correct responses on lexicon and nonsense word items in both groups of children tested. These conflicting findings do little to clarify the effect that nonsense words have on test performance. It may be, as Herriot suggests, that age has a crucial bearing on performance on questions using nonsense words. The incorrect responses in this study closely paralleled those in previous studies (e.g., Koziol 1971 and Vogel 1970). One common error in this study, as in the others, was a stimulus repe- tition of the verb as it occurred in the test question; sometimes this was the uninflected form and sometimes the inflected form. One incorrect response not found in earlier studies was the substi- tution of a form of pp_for a form of hpyp; for example, the use of pg_in "He must ____ been watching TV too long." The reverse was true, too: the substitution of a form of hpy§_for a form of pg; for example, the use of hpg_in "Yesterday afternoon the flowers ____ being watered." Incorrect substitutions of ppyg_for pg_ occurred less frequently than substitutions of pg_for have. Such 145 Such errors indicate confusion in the use of pgyg_and pg, A thorough study of the data needs to be made to determine the role of phonological and syntactic interference of the primary language in the learning of the target language. Our data seem to indicate that phonological interference was greater than syntactic inter- ference. This preliminary investigation on the acquisition of English verb morphology by nonnative speakers indicates an orderly progression of learning. Whether this order is valid for only the subjects in this study or whether it is a wider or universal pattern is not known. Our data indicate greater differences in learning by level of proficiency than by native language. Further studies on a larger sample of nonnative speakers need to be done. The role of formalized class instruction and the order of presenta- tion of the various elements of English verb morphology also need to be explored. In addition, a comparison of native and nonnative patterns of acquisition could prove useful; little is known at present about how native speakers learn the English verb system. After gaining a better understanding of the acquisition of the English verb system by both native and nonnative speakers, we can then hopefully make improvements in second-language teaching strategies. APPENDICES 146 APPENDIX A THE VERB MORPHOLOGY TEST 147 _J O N ._| _4 _- _.| _I _a _a _a .—n _a O 1.0 00 \l 01 U1 4) 0) N --' O I O O O O C C C . osooowosmc-wm APPENDIX A THE VERB MORPHOLOGY TEST 1. The Oral Test 1 This woman knows how to cook. She is cooking now. A- She does it every day. Every day she The boys jafe their horses every day. In fact, they were their horses yesterday. The men smoke their pipes every day. In fact, they were their pipes yesterday. This girl knows how to kleach. She is kleaching now. She does it every day. Every day she This woman knows how to lunt. She is lunting now. She did the same thing yesterday. Yesterday she This boy knows how to smile. He is smiling now. He did the same thing yesterday. Yesterday he . This boy knows how to walk. He has done it recently. What has he done recently? Recently he has The people are enjoying the music. Most people enjoy music. Music enjoyed by most people. The boy malks very well. He is malking now. What can he do well? He can well. This man knows how to pleen. He has done it recently. What has he done recently? Recently he has The boy didn' t write a letter to his girl friend until yes— terday. He could written the letter earlier. This man knows how to tump. He is tumping now. He did the same thing yesterday. Yesterday he This woman works very hard. She has to do it every week. What must she do next week? Next week she must very hard. The woman picks flowers every day. She did this yesterday. Yesterday the flowers were by the woman. The girl is heezing a dress now. She has heezing it since 9 o'clock. This girl knows how to ming. She is minging now. She did the same thing yesterday. Yesterday she . This woman knows how to plant. She has done it recently. What has she done recently? Recently she has The man is cutting the tree. He will cut another tree soon. Soon another tree will cut by the man. The man is liching the car now. What is happening to the car? The car is liched by the man. The girl is pawking her garden now. She been pawking her garden since noon. 4’ 148 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31R 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 149 This boy knows how to 200k. He is zooking now. He does it every day. Every day he . The girl plays with her dog every afternoon. It is now 2 p.m. She should playing with her dog now. The boy can klant the house. In other words the house can klanted by the boy. The men vump the trees every morning. What was happening to the trees yesterday morning? Yesterday morning the trees being vumped. The man is building a boat now. He been building the boat since last week. The girl bings a ball. It is fun. What is fun? It is fun bing a ball. The man was watching TV for many hours. He became very bored. He must been watching TV too long. This girl knows how to dalk. She has done it recently. What has she done recently? Recently she has The woman is opening the letters. ‘ 5" One has opened before. The girl is piffing a car. The boy also wants to a car. This man talks very well. He is talking now. What can he do well? He can well. This woman knows how to sing. She is singing now. She did the same thing yesterday. Yesterday she . The woman usually looches a picture at noon. It 1s now noon. She could looching a picture now. The people are thoying the food. Most people thoy food. Food thoyed by most people. The boy could not dulb the dog. The dog was too far away; otherwise, the dog would been dulbed by the boy. This girl knows how to clean. She has done it recently. What has she done recently? Recently she has This woman knows how to swim. She is swimming now. She does it every day. Every day she . The girl can start the fire. In other words, the fire can started by the girl. The man was fixing his car. He should have fixing his house instead. This man knows how to ming. He has done it recently. What has he done recently? Recently he has . The girls water the flowers every afternoon. What was hap- pening yesterday afternoon? Yesterday afternoon the flowers being watered. The girl is riding a horse. The boy also wants to a horse. The boy spacks the ball every day. He did this yesterday. Yesterday the ball was by the boy. The boy studied his lesson for 3 hours. He lost all interest in the lesson. He may studied too long. This boy knows how to hunt. He is hunting now. He did the same thing yesterday. Yesterday he . 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 150 The girl was clinning for many hours. She became very tired. She must been clinning too long. This man knows how to him. He is bimming now. He does it every day. Every day he The man found the tisked cup on the table. The cup might have tisked by his daughter. This woman knows how to sing. She has done it recently. What has she done recently? Recently she has . This man zicks very hard. He has to do it every week. “I ~ ~ iii-.flma.“ What must he do next week? Next week he must very hard. The boy snurns the field every afternoon. It is now 2 p.m. He should snurning the field now. The boy is grishing the house. It has grished before. The woman didn't joor her clothes until yesterday. She could joored her clothes earlier. The boy usually sails his boat at noon. It is now noon. He could sailing his boat now. The man could not win the race. He was sick and didn't run well; otherwise, the race would been won by him. The boy eats the bad food. He gets sick after the food. The woman is jatting the material. It been jatted before. This man knows how to jump. He is jumping now. He did the same thing yesterday. Yesterday he The girl was nuching her hair. She should have nuching her clothes instead. The man is throwing the ball now. What is happening to the ball? The ball is thrown by the man. The boy wugged his car for 2 hours. He became tired. He may wugged his car too long. This girl knows how to teach. She is teaching now. She does it every day. Every day she . The girl saw a dead animal on the road. The animal might have hit by a car. This boy knows how to zint. He has done it recently. What has he done recently? Recently he has The man is rowing a boat now. He has rowing it since noon. The woman is choofing the plant. She will choof another plant soon. Soon another plant will choofed by the woman. The girl is playing a record. It been played before. The woman thaps the big house. She feels good after the house. This girl knows how to frool. She is frooling now. She did the same thing yesterday. Yesterday she The woman drives a car. It is fun. What is fun? It is fun drive a car. c—l O NN—aa—aa-a-u-a-aaa doomummth—a 22. 23. 24. OKOWNONU'l-bwh) 151 II. The Fill-In Test The women drive their cars every day. In fact, they were their cars yesterday. The men clouf their hair every week. In fact, they were their hair last week. The girl knows how to wash. She is washing now. She does it every week. Every week she . The man knows how to croll. He is crolling now. He did the same thing yesterday noon. Yesterday noon he . The boy knows how to stoom. He is stooming now. He does it every Saturday. Every Saturday he The woman knows how to add. She is adding now. _ She did the same thing yesterday. Yesterday she . The boy knows how to jout. He has done it recently. 1 What has he done recently? Recently he has . The man is buying a car. The boy also wants to a car. The man zeps the tree every day. He did this yesterday. Yesterday the tree was by the man. The girl is telling a story. It been told before. The boy knows how to paf. He is paffing now. He did the same thing last night. Last night he The people are watching TV. Most people watch TV. TV watched by most people. The woman deeses very hard. She has to do it every week. What must she do next week? Next week she must very hard. The man is drawing a picture now. What is happening to the picture? The picture is drawn by the man. The girl is feeding the animals now. She has feeding them since breakfast. The woman knows how to trone. She has done it yesterday. What has she done yesterday? Yesterday she has The man ate the badly burned cake. The cake might have burned by his daughter. The woman didn't plant her garden until yesterday. She could planted her garden earlier. The woman waters her plants every morning. It is now 10 a.m. She should watering her plants now. The man knows now to ning. He is ninging now. He did the same thing yesterday. Yesterday he . The girl is platching the horse. She will platch another one next week. Next week another horse will platched by the girl. The woman knows how to sleep. She is sleeping now. She does it every noon. Every noon she The man lazzes very well. He is lazzing now. What can he do well? He can well. The boy is walking to town today. He soon becomes tired. He may walked too far already. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 152 The girl is snaffing a hill. It been snaffed before. The man in the car could not see the child. The night was too dark; otherwise, the child would been seen by the man. The girl knows how to wuck. She has done it last week. What has she done last week? Last week she has . The boy gloaks his ball every day at 3 o'clock. It is now 3. He could gloaking his ball now. The man knows how to love. He has done it recently. What has he done recently? Recently he has The man was nushing his house. He should have nushing his office instead. The boy races very hard. He has to do it every week. What must he do next week? Next week he must very hard. The boy is spoving the bag. It has spoved before. The man can train the dog. In other words, the dog can trained by the man. The women make the beds every morning. What was happening to the beds yesterday morning? Yesterday morning the beds being made. The girl tigs a plant. It is fun. What is fun? It is fun tig a plant. The man knows how to naz. He is nazzing now. He does it every night. Every night he . The girl was singing for a long time. She started to sing badly. She must been singing too long. The girl is tizzing a shoe now. What is happening to the shoe? The shoe is tizzed by the girl. The angry man hits the girl. He feels bad after the girl. The man didn't stim his store until yesterday. He could stimmed his store earlier. The cook tastes the food every day. He did this yesterday. Yesterday the food was by the cook. The girl knows how to ring the bell. She is ringing it now. She did the same thing yesterday. Yesterday she the bell. The man was twunching the machine for a long time. He really became tired. He must been twunching too long. The woman is snooling the dish now. She has snooling it since supper. The woman is kleving a house. The girl also wants to a house. The man knows how to laugh. He is laughing now. He did the same thing last night. Last night he The boy is flidding the store now. He been flidding the store since yesterday. The woman is cleaning the room. It has cleaned before. The people are yibbing the game. Most people yib the game. The game yibbed by most people. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 153 The boys zeave the boxes every night. What was happening to the boxes last night? Last night the boxes being zeaved. The man knows how to run. He is running now. He does it every day. Every day he . The man is mooking his land today. He soon becomes tired. He may mooked too long already. The woman knows how to ring the bell. She has done it recently. What has she done recently? Recently she has the bell. The boy milks a cow. It is fun. What is fun? It is fun milk a cow. The girl zilts her lessons every afternoon. It is now 3. She should zilting her lessons now. The boy knows how to foop. He is fooping now. He does it every morning. Every morning he The girl knows how to dance. She has done it last night. What has she done last night? Last night she has . The man smokes his pipe every day at 5 o'clock. It is now 5. He could smoking his pipe now. The girl was _baking a pie. She should have baking bread instead. The boy knows how to breet. He is breeting now. He did the same thing Sunday. Sunday he The man is flying the plane. He will fly another one next week. Next week another plane will flown by the man. The boy looked at the croasted picture. The picture might have croasted by his father. The man knows how to ning. He has done it recently. What has he done recently? Recently he has The man fishes very well. He is fishing now. What can he do well? He can well. The girl knows how to plan. She is planning now. She did the same thing yesterday. Yesterday she The girl can tadge the flowers. In other words the flowers can tadged by the girl. The girl is writing a letter now. She been writing it since yesterday. The man knows how to shout. He has done it every night. What has he done every night? Ever night he has The good woman vins the bag. She feels good after the bag. The man could not plich the tree. The tree was too big; otherwise, the tree would been pliched by the man. 10. 11. 12. 13. 14. 15. 16. 154 III. The Multiple-Choice Test The women serve milk every day. In fact they were milk yesterday. a. serve b. serves c. served d. serving The boy knows how to gurr. He is gurring now. He does it every afternoon. Every afternoon he a. gurr b. gurrs c. gurred d. gurring The man knows how to dress. He is dressing now. He does it every day. Every day he . a. dress b. dresses c. dressed d. dressing The woman knows how to frad. She is fradding now. She did the same thing last month. Last month she a. frad b. frads c. fradded d. fradding The girls zish hills every Sunday. In fact they were hills last Sunday. a. zish b. zishes c. zished d. zishing The man knows how to count. He is counting now. He did the same thing yesterday. Yesterday he . a. count b. counts c. counted d. counting The girl knows how to zub. She has done it recently. What has she done recently? Recently she has a. zub b. zubs c. zubbed d. zubbing The boy is dawling his car now. He has dawling it since 5 o'clock. a. been b. being c. had d. having Two cars are passing a very slow car now. Most cars pass the slow one. The slow car passed by most cars. a. be b. had c. has d. is The boy knows how to moash. He has done it every week. What has he done every week? Every week he has . a. moash b. moashes c. moashed d. moashing The man grools his family every Saturday. He did this last Saturday. Last Saturday the family was by the man. a. grool b. grools c. grooled d. grooling The girl doops very well. She is dooping now. What can she do well? She can well. a. doop b. doops c. dooped d. dooping The man knows how to tring. He is tringing now. He did the same thing last night. Last night he a. tring b. trings c. tringed d. trang The man is guarding his house now. What is happening to the house? The house is guarded by the man. a. been b. being c. has d. have The woman is brushing her hair. It been brushed before. a. be b. has c. have d. was The boy knows how to theet. He is theeting now. He does it every noon. Every noon he . a. theet b. theets c. theeted d. theeting 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 155 The woman usually bakes bread in the morning. It is now morning. She could baking bread now. a. be b. being c. had d. have The boy hilches a river. It is fun. What is fun? It is fun hilch a river. a. be b. has c. have d. to The man fights very hard. He has to do it every week. What must he do next week? Next week he must very hard. a. fight b. fights c. fought d. fighting The men fly the planes every night. What was happening to the planes last night? Last night the planes being flown. a. been b. had c. have d. were The woman knows how to jud. She has done it last Sunday. What has she done last Sunday? Last Sunday she has . a. jud b. juds c. judded d. judding The girl knows how to paint. She is painting now. She does it every day. Every day she a. paint b. paints c. painted d. painting The man sails his boat every day. He did this yesterday. Yesterday the boat was by the man. a. sail b. sails c. sailed d. sailing The man didn't meeve his horse until yesterday. He could meeved his horse earlier. a. be b. been c. had d. have The man knows how to jeal. He is jealing now. He did the same thing last week. Last week he . a. jeal b. jeals c. jealed d. jealing The girl picked flowers for 2 hours. She became tired. She may picked flowers too long. a. be b. been c. has d. have The girl is plimming the sea now. What is happening to the sea? The sea is plimmed by the girl. a. been b. being c. has d. have The boy is cutting the grass. It has cut before. a. been b. being c. had d. have The man can sign the letter. In other words, the letter can signed by the man. a. be b. being c. had d. have The boys are blooting the game now. Most boys bloot the game. The game blooted by most boys. a. be b. had c. has d is The woman knows how to call. She is calling now. She does it every day. Every day she . a. call b. calls c. called d. calling The boy knows how to relf. He is relfing now. He did the same thing last night. Last night he a. relf b. relfs c. relfed d. relfing The boy is joaming the box now. He been joaming the box since noon. a. be b. has c. have d. was 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 156 The woman is learning French. The girl also wants to French. a. learn b. learns c. learned d. learning The man knows how to vote. He has done it last week. What has he done last week? Last week he has . a. vote b. votes c. voted d. votin The boy knows how to spring up in the air. He is springing up now. He did the same thing yesterday morning. Yesterday morning he up in the air. a. spring b. springs c. springed d. sprang The woman is gliving the child today. She will glive another one tomorrow. Tomorrow another child will glived by the woman. a. be b. been c. had d. have The girl didn't invite her friend until yesteday. She could invited her friend earlier. a. be b. been c. had d. have The girl discovered the dassed box. The box might have dassed by her mother. a. be b. been c. had d. having The man knows how to help. He is helping now. He did the same thing last Sunday. Last Sunday he a. help b. helps c. helped d. helping The boy usually flokes a ball in the evening. It is now evening. He could floking a ball now. a. be b. being c. had d. have The women glip the pictures every afternoon. What was hap- pening to the pictures yesterday afternoon? Yesterday afternoon the pictures being glipped. a. been b. had c. have d. were The man knows how to stuch. He is stuching now. He does it every afternoon. Every afternoon he a. stuch b. stuches c. stuched d. stuching The man was painting his boat. He should have painting his house instead. a. be b. been c. had d. having The girl knows how to spring up in the air. She has done it last night. What has she done last night? Last night she has up in the air. a. spring b. springs c. springed d. sprung The boy was breeving the ball for many hours. He became tired. He must been breeving too long. a. be b. has c. have d. was The man hides very well. He is hiding now. What can he do well? He can well. a. hide b. hides c. hid d. hiding The man is lanning his watch. It been lanned before a. be b. has c. have d. was The girl drinks milk at breakfast each day. It is now 8 a.m. She should drinking milk now. a. be b. been c. has d. have 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 157 The man knows how to kill. He has done it last week. What has he done last week? Last week he has a. kill b. kills c. killed d. killing The woman could not thamp her office this morning. The door was locked; otherwise, the office would been thamped by now. a. be b. have c. having d. was The woman beats her big dog. She becomes tired after her dog. a. beat b. beats c. beaten d. beating The woman whotted her kitchen for 2 hours. She became tired. She may whotted the kitchen too long. a. be b. been c. has d. have The man is buying a record today. He will buy another one next month. Next month another record will bought by the man. a. be b. been c. had d. have The woman is tooging the bridge. The girl also wants to the bridge. a. tooge b. tooges c. tooged d. tooging The girl knows how to type. She has done it recently. What has she done recently? Recently she has a. type b. types c. typed d. typing The man is truffing the machine. It has truffed before. a. been b. being c. had d. have The boy found the murdered man. The man might have murdered a week earlier. a. be b. been c. had d. having The woman is washing the clothes now. She has washing them since noon. a. been b. being c. had d. having The man noves very hard. He has to do it every week. What must he do next week? Next week he must very hard. a. nove b. noves c. noved d. noving The man was slooming his friend's dog. He should have slooming his own dog instead. a. be. b. been c. had d. having The boy can lawp the dog. In other words the dog can lawped by the boy. a. be b. being c. had d. have The woman shakes a box. It is fun. What is fun? It is fun shake a box. a. be b. has c. have d. to The man knows how to farm. He is farming now. he did the same thing yesterday. Yesterday he a. farm b. farms c. farmed d. farming . The boy could not read his lesson. Someone had taken his book; otherwise, the lesson would been read. a. be b. have c. having d. was 66. 67. 69. 70. 158 The boy himps the machine. He gets tired after the machine. a. himp b. himps c. himped d. himping The woman is making a dress now. She been making the dress since last week. a. be b. has c. have d. was The woman klests her store every afternoon. It is now 3 p.m. She should klesting her store now. a. be b. been c. has d. have The man knows how to tring. He has done it yesterday noon. What has he done yesterday noon? Yesterday noon he has a. tring b. trings c. tringed d. trung The woman was reading the book for many hours. She became bored. She must been reading too long. a. be b. has c. have d. was APPENDIX B STATISTICAL RESULTS BY LEVEL OF PROFICIENCY 159 APPENDIX B STATISTICAL RESULTS BY LEVEL OF PROFICIENCY Structure Chi-Square Significance1 l tense (present s) MV 3.000 n.s. 2 tense (past) MV 6.857 n.s. 3 tense modal py_ 9.414 n.s. 4 tense modal have+en MV 19.643 .001 a/S tense modal have+en be+ing MV 10.500 .05 6 tense modal have+en be+en MV 12.252 .05 7 tense modal pgfing MV 4.500 n.s. 8 tense modal be+en MV 9.270 n.s. 9 tense (present s) have+en be+ing MV 17.871 .01 10 tense (present s) pgyg+en be+en MV 9.075 n.s. ll tense have+en MV 14.169 .01 12 tense have+en be+ing MV 21.027 .001 13 tense have+en be+en MV 16.455 .01 ,44 tense modal have+en be+ing MV 16.643 .01 15 tense modal have+en be+en MV 4.257 n.s. l6 tense (past) pgfing be+en MV 10.900 .05 17 tense be+ipg_MV 8.687 n.s. 18 tense be+i_g_be+en MV 17.171 .Ol 19 tense (present s) be+en MV 5.061 n.s. 20 tense e+gp_MV 9.464 n.s. 21 is_MV 10.717 .05 22 to py_ 9.786 .05 23 ipg_My_ 11.733 .05 1n.s. indicates no significance statistically. 160 APPENDIX C STATISTICAL RESULTS BY NATIVE LANGUAGE 161 APPENDIX C STATISTICAL RESULTS BY NATIVE LANGUAGE 1 Structure Chi-Square Significance 1 tense (present s) MV 13.000 .05 2 tense (past) MV 9.429 n.s. 3 tense modal MV_ 10.038 .05 4 tense modal have+en MV 6.000 n.s. 5 tense modal have+en be+ing MV 6.500 n.s. 6 tense modal have+en be+en MV 5.971 n.s. 7 tense modal Le+ing MV 16.000 .01 8 tense modal be+en MV 9.990 .05 9 tense (present s) have+en be+ing MV 5.014 n.s. 10 tense (present s) have+en be+en MV 2.100 n.s. ll tense have+en MV 13.554 .01 12 tense have+En Le+ing MV 3.682 n.s. 13 tense have+En be+en MV 10.012 .05 14 tense modal have+en Le+ing MV 4.982 n.s. 15 tense modal have+En be+en MV 9.829 .05 16 tense (past) Le+ing be+en MV 3.100 n.s. l7 tense be+jpg_MV 4.372 n.s. 18 tense be+__g Le+en MV 5.343 n.s. 19 tense (present s) Le+en MV 12.727 .05 20 tense be+Ln MV 8.918 n.s. 21 §p_MV 8.346 n.s. 22 to MV_ 4.286 n.s. 23 jpg_MV_ 8.000 n.s. 1 n.s. indicates no significance statistically. 162 L I ST OF REFERENCES 163 LIST OF REFERENCES Anisfeld, Moshe, and G. Richard Tucker. 1967. English Pluraliza- tion Rules of Six-Year-Old Children. Child Development 38.1201-17. Baird, Raymond. 1973. Children's Phonological Rules: A Failure to Replicate. Language Learning 23.223-30. Bateson, Mary. 1967. Arabic Language Handbook. Washington, 0.6.: The Center for Applied Linguistics. Bellamy, Martha, and Sidney Bellamy. 1970. The Acquisition of Morphological Inflections by Children Four to Ten. Lan- guage Learning 20.199-211. Bereiter, Carl, and Siegfried Engelmann. 1966. Teaching Disad- vantaged Children in the Preschool. Englewood Cliffs, N.J.: Prentice-Hall. Berko, Jean. 1958. The Child's Learning of English Morphology. Word 14.150-77. Blalock, Hubert M. 1972. Social Statistics. New York: McGraw- Hill. Carroll, John B. 1961. Language Development in Children. In Psycholinguistics: A Book of Readings, ed. by $01 Saporta, 331-45. New York: Holt, Rinehart and Winston. . 1971. Current Issues in Psycholinguistics and Second Language Teaching. TESOL Quarterly 5.101-15. Cazden, Courtney. 1968. The Acquisition of Noun and Verb Inflec- tions. Child Development 39.433-48. Chomsky, Noam. 1970. Formal Discussion of "The Development of Grammar in Child Language" by Wick Miller and Susan Ervin. In Readings in Applied Transformational Grammar, ed. by Mark Lester, 41-50. New York: Holt, Rinehart and Winston. Cooper, Robert. 1967. The Ability of Deaf and Hearing Children to Apply Morphological Rules. Journal of Speech and Hearing Research 10.77-86. 164 165 Dale, Philip. 1972. Language Development. Hinsdale, Illinois: Dryden Press. Dever, Richard. 1972. A Comparison of the Results of a Revised Version of Berko's Test of Morphology with the Free Speech of Mentally Retarded Children. Journal of Speech and Hearing Research 15.169-78. Dever, Richard, and William Gardner. 1970. Performance of Normal and Retarded Boys on Berko's Test of Morphology. Lan- guage and Speech 13.162-81. Dulay, Heidi, and Marina Burt. 1972. Goofing: An Indicator of Children's Second Language Learning Strategies. Language Learning 22.235-53. Dunn, C. J., and S. Yanada. 1958. Teach Yourself Japanese. London: The English Universities Press. Ervin, Susan. 1964. Imitation and Structural Change in Children's Language. In New Directions in the Study of Language, ed. by Eric Lenneberg, 163-89. Cambridge, Mass.: MIT Press. Falk, Julia. 1974. Second Language Learning and First Language Acquisition: Providing Comparable Data. Prepublication copy. Foley, Helen, and John Locke. 1971. Young Children's Knowledge of Morphological and Phonological Rules. Journal of Com- munication Disorders 4.259-62. Fries, Charles C. 1952. The Structure of English. New York: Harcourt, Brace and Co. Gesell, Arnold. 1941. Wolf Child and Human Child. New York: Harpers. Goodglass, Harold, and Jean Berko. 1960. Agrammatism and Inflec- tional Morphology in English. Journal of Speech and Hearing Research 3.257-67. Graves, Michael, and Stephen Koziol. 1971. Noun Plural Develop- ment in Primary Grade Children. Child Development 42.1165-74. Herriot, Peter. 1968. The Comprehension of Syntax. Child Development 39.273-83. 166 Jakobovits, Leon. 1968. Implications of Recent Psycholinguistic Developments for the Teaching of a Second Language. In Readings in Applied Transformational Grammar, ed. Mark Lester, 253-76. New York: Holt, Rinehart and Winston. 1969. Second Language Learning and Transfer Theory: A Theoretical Assessment. Language Learning 19.55-87. Johnson, Nancy. 1974. Zombies and Other Problems: Theory and Method in Research on Bilingualism. Language Learning 24.105-33. Kelly, L. G. 1969. Description and Measurement of Bilingualism: An International Seminar. Toronto: University of Toronto Press. Kernan, Keith, and B. G. Blount. 1966. The Acquisition of Spanish Grammar by Mexican Children. Anthropological Linguistics 8.1-14. Koziol, Stephen. 1971. AniAnalysis of the Development of Noun Plural Rules in the Primary Grades. Stanford Univer- sity, dissertation. Lambert, Wallace. 1972. Language, Psychology, and Culture. Stanford, Ca1.: Stanford University Press. Lenneberg, Eric. 1964. New Directions in the Study of Language. Cambridge, Mass.: MIT Press. . 1967a. Biological Foundations of Language. New York: John Wiley and Sons. 1967b. The Biological Foundations of Language. In Readings in Applied Transformational Grammar, ed. by Mark Lester, 3-21. New York: Holt, Rinehart and Winston. Lester, Mark. 1971. Introductory Transformational Grammar of English. New York: Holt, Rinehart and Winston. Linden, Eugene. 1974. Apes, Men, and Language. New York: Dutton. Lovell, K., and B. Bradbury. 1971. The Learning of English Mor- phology in Educationally Subnormal Special School Children. American Journal of Mental Deficiency 71.609-15. Mackey, William. 1965. Bilingual Interference: Its Analysis and Measurement. Journal of Communication 15.239-50. Martinez-Bernal, Janet. 1972. Children's Acquisition of Spanish and English Morphological Systems and Noun Phrases. Georgetown University, dissertation. 167 McNeill, David. 1970. 'The Acquisition of Language: The Study of Developmental Psycholinguistics. New York: Harper and Row. Menyuk, Paula. 1969. Sentences Children Use. Cambridge, Mass.: MIT Press. Natalicio, Diana. 1969. Formation of the Plural in English: A Study of Native Speakers of English and Native Speakers of Spanish. University of Texas, Austin, dissertation. Natalicio, Diana, and Luiz Natalicio. 1969. "The Child's Learn- ing of English Morphology" Revisited. Language Learning 19.205-15. Newfield, Melanie, and Bernard Schlanger. 1968. The Acquisition of English Morphology by Normal and Educable Mentally Retarded Children. Journal of Speech and Hearing Research 11.693-706. Ramer, Andrya, and Norma Rees. 1973. Selected Aspects of the Development of English Morphology in Black American Children of Low Socioeconomic Background. Journal of Speech and Hearing Research 16.569-79. Shriner, Thomas, and Lynn Miner. 1968. Morphological Structures in the Language of Disadvantaged and Advantaged Children. Journal of Speech and Hearing Research 11.605-10. Smith, Frank, and George Miller. 1966. The Genesis of Language: A Psycholinguistic Approach. Cambridge, Mass.: MIT Press. Solomon, Martha. 1972. Stem Endings and the Acquisition of Inflections. Language Learning 22.43-51. Stockwell, Robert; J. 0. Bowen; and J. Martin. 1965. The Gram- matical Structures of English and Spanish. Chicago: University of Chicago Press. Tabakowska, Elzbieta. 1969. The Notions of Competence and Per- formance in Language Teaching. Language Learning 19.41-55. Thorndike, Edward, and Irving Lorge. 1944. The Teacher's Word Book of 30,000 Words. New York: Columbia University Press. Vogel, Phyllis. 1970. A Comparative Survey of the Morphology of Lower Class Rural Kindergarten Students in Alchua County, Florida. University of Florida, Gainesville, dissertation. 168 Weinreich, Uriel. 1970. Languages in Contact. The Hague: Mouton. Weir, Ruth. 1962. Language in the Crib. The Hague: Mouton. Whitman, Randal, and Kenneth Jackson. 1972. The Unpredict- ability of Contrastive Analysis. Language Learning 22.29-43. Williams, Frederick. 1971. Language and Poverty: Perspectives on a Theme. Chicago: Markham. HICH mmmm