Acquisition of English and Spanish morphological rules by bilinguals

MISSING IMAGE

Material Information

Title:
Acquisition of English and Spanish morphological rules by bilinguals
Physical Description:
viii, 111 leaves : ; 28 cm.
Language:
English
Creator:
Dale, Paulette
Publication Date:

Subjects

Subjects / Keywords:
Bilingualism   ( lcsh )
Interference (Linguistics)   ( lcsh )
Language acquisition   ( lcsh )
English language -- Study and teaching (Primary)   ( lcsh )
Spanish language -- Study and teaching (Primary)   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1980.
Bibliography:
Includes bibliographical references (leaves 105-109).
Statement of Responsibility:
by Paulette Dale.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000100006
notis - AAL5466
oclc - 07252898
System ID:
AA00003465:00001

Full Text















ACQUISITION OF ENGLISH AND SPANISH MORPHOLOGICAL
RULES BY BILINGUALS











BY

PAULETTE DALE


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULF ILIEN'T OF 'THE RE.:UIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY




UNIVERSITY OF FLORIDA


1.980














ACKNOWLEDGEMENTS


I would like to express appreciation to all members of

my supervisory committee for their guidance, suggestions, and

support which made the completion of this study possible.

I am grateful to Dr. R. Scholes, Dr. T. Abbott and Dr. C.

Mercer for their academic and moral support during my graduate

studies and for instilling in me an enthusiasm for research.

Many thanks are extended to the principal of Shenendoah

Elementary School in Miami, Dr. William Renuart, for his

cooperation and assistance during the weeks that I was in his

school conducting my research.

I would like to extend my sincere thanks to Mr. Nat

Terry and his assistant Ilene for their help with the

statistical analyses of this study. I also wish to thank

the administrators at Miami Dade Community College for

granting me the permission to utilize the Testing and

Research and computer facilities at the college.

Special thanks are extended to my supervisors at

Miami Dade Community College, Mr. Ed Anderson, Department

Chairman and Dr. M. Pelton, Humanities Division Director,

Without their cooperation and willingness to grant me the

necessary professional development leaves and flexible class

and speech clinic schedules, fulfilling the requirements







for the Doctor of Philosophy Degree at the University

of Florida would have been impossible for me to achieve.

Nothing can express how much I owe my husband and my

parents for their unyielding encouragement and support

toward the successful completion of this endeavor. They

are the individuals most responsible for helping me maintain

the necessary momentum to keep going for the last several

years to complete this degree.


iii















TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS ............... ........ ...... .. ........ ii

ABSTRACT ................................................ vi

CHAPTER I: REVIEW OF THE LITERATURE AND STATEMENT OF
PROBLEM .................... 1

Introduction .............................. 1
Review of the Literature .................. 2
Effects of Bilingualism on Language
Acquisition ........................ 2
Morphology Tests .................... 15
Language Tests for Bilinguals ......... 22
Statement of the Problem ..... ................. 34

CHAPTER II: METHODS AND PROCEDURES ..................... 36

Subjects ....... ....... ....... ............ 36
Materials ................................. 39
Testing Procedures ........ ................ 41
Data Analysis ............................. 46

CHAPTER III: RESULTS AND DISCUSSION ..................... 48

Introduction ......................... ..... 48
Results and Discussion .................... 49
Bilingual Morphology Measure .......... 49
Morphological Rule Acquisition ........ 60
Effects of Bilingualism on Language
Development .......................... 75

CHAPTER IV: SUMMARY .................................... 90

Future Research .......................... 92
Implications .... ........ ................. 92

APPENDIX A: ADMINISTRATION PROCEDURES .................. 94

APPENDIX B: SCORING PROCEDURES ......................... 98

APPENDIX C: BMM-S ANSWER SHEET ..........;..............103








Page

APPENDIX D: BMM-E ANSWER SHEET .................... 104

REFERENCES .............. ........ ......... ......... 105

BIOGRAPHICAL SKETCH ................................... 110














CHAPTER I

REVIEW OF THE LITERATURE AND STATEMENT OF PROBLEM

Introduction

/ The effects of bilingualism on children's language

development have been debated over the years. There have

been many hypotheses concerning the frequently poor lin-

guistic performance observed in bilingual children. Several

investigations have been conducted to support or disprove

these hypotheses and have yielded conflicting information.

Many investigators found there are disadvantages to child-

hood bilingualism; others found the negative effects of

bilingualism on a child's language development are not very

serious; others concluded that bilingualism has many positive

advantages.

Some of the investigations have tried to control

various biological and environmental factors; others have

made little or no attempt to control many of the factors

language acquisition generally is dependent on. Additional

research has also been conducted to obtain normative data

concerning bilinguals' performance on specific linguistic

tasks; however, the population of Cuban-American bilinguals

has been largely ignored.







The present investigation was conducted to obtain

information about the effects of bilingualism on language

development in Cuban-American children while carefully

controlling such factors as intelligence, socioeconomic

status, and amount of bilingualism already exhibited by

the subjects. In addition, this investigation was designed

to obtain normative data about this population's native

and non-native language development which might be used in

future assessment of bilingual children's linguistic

proficiency.

Review of the Literature

The orientating and experimental literature relevant

to this investigation is reported in three sections. The

first section discusses the effects of bilingualism on a

child's language development. General information including

observational reports and findings from controlled invest-

igations are reviewed. The previous and present applications

of tests of morphology are reported in the second section.

.The third section reviews objective measurement techniques

which have been utilized to assess aspects of bilingual

children's receptive and expressive language proficiency.

Effects of Bilingualism on Language Acquisition

Defense of childhood bilingualism

General information and observational reports. Many

investigators insist that the reported disadvantages of

childhood bilingualism either do not exist or are not as








serious as several researchers claim. There are observers

who go a step further and contend that bilingualism has

many positive advantages. The following defense of bi-

lingualism is derived from opinions that have appeared in

the literature. These opinions were supported by experience

and a number of observations of bilingual children's

native and secondary language acquisition.

Several authors (Buell, 1946; Anastasi and Cordova,

1953; Carrow, 1957; Darcy, 1963; and Garcia, 1974) seem to

feel that bilingualism alone should not be criticized

and that other factors rather than bilingualism are to

blame for the deficient performance of bilinguals in many

studies. Most of these authors have indicated that the

language handicap of the bilingual children in a variety

of studies may be related to the presence of confused and

incorrect language patterns in the home rather than the

actual learning of a second language. For example, Buell

(1946) concluded that incorrect English in Hawaii was due

to pidgin English and not bilingualism.

Several investigators (Darcy, 1963; Thonis, 1972;

Carrow, 1972; and Garcia, 1974) defend bilingualism on the

grounds that most of the investigations since 1953 which

have concluded that bilingualism has adverse effects on

language development failed to control significant research

variables such as socio-environmental and intellectual

factors. These authors generally agree that existing







theories of negative effects of bilingualism have been

exaggerated and are unjustified. They feel that judgements

as to the total effects of bilingualism should be deferred

until studies are conducted which determine the effect

of bilingualism in socioeconomic environments in which the

quality of the language environment in both languages is

high and whether intelligence influences the ability to

learn a second language proficiently. Weinreich (1961)

quotes Haugen as saying:

The linguist who makes theories about language
influence but neglects to account for the
socio-cultural setting of the language contact
leaves his study suspended, as if it were, in
mid-air. (p.379)

In a very early report (Becky, 1942) it was stated

that bilingualism is no handicap in speech development.

Carrow (1957) reports no detrimental effects on the

bilingual child's ability in spelling, total verbal

output, clause length, or complexity of sentence structure.

Jensen (1962b) summarized other reporters' observations by

stating:

Other assertions often made are that studying a
second language will aid a person to strengthen
his acquired tongue, to become more sensitive to
nuances, to manipulate languages more effectively,
and to learn additional languages more easily.
(p. 359)

Zierer (1974) describes the experience of a child

in Peru whose bilingual education was psycholinguistically

and pedagogically planned by his Peruvian mother and German

father. At home, where the German language was the








means of communication between husband and wife and

their child (only son), the child at the age of three

had achieved a slightly better than average (for a

three-year-old) mastery of German. From the age of 2 years

10 months on, the child was systematically exposed to

Spanish. At the age of four, the child had mastered

both Spanish and German with the same degree of proficiency

(without interference) at the phonological, morphological

and syntactic levels.

Lambert (1963) briefly describes a study that he and

a colleague carried out with ten-year-olds in Montreal.

When socio-environmental variables were controlled, they

found that bilinguals performed better than monolinguals

on verbal and nonverbal intelligence tests. They further

reported that the bilingual.subjects had greater cognitive

flexibility and a greater ability in concept formation

than the monolingual. Lambert (1963) stated that because

his results were in conflict with many others, he could

not be sure that the reported bilingual advantage was

peculiar to bilinguals in Canada or any bilinguals who

are actually competent in both languages.

Controlled investigations. A study by Lambert, Just,

and Segalowitz (1970) revealed that education in a

non-native language produced no retardation in learning-

and increased competence and performance in both the

primary and secondary language, Two English-speaking and







one French-speaking control groups were chosen from

a French-Canadian school, The experimental group was

chosen from an English-speaking school. Socio-environmental

variables were accounted for in the study. The control

groups followed the method of instruction normally followed

in school; i.e. the English groups were taught in English

and the French group in French. The experimental group of

English-speaking children was prepared by attending a

French kindergarten and first grade. After these two

years of instruction the children were tested in French

and English. Results were compared with those obtained

for the English group taught in English and the French group

taught in French. Even though the bilingual children were

instructed in their weaker language, French, they

demonstrated an optimum level of language skills in

French and still maintained an excellent control of their

home language, English. Little or no interference

between the children's two languages occurred.

Lambert and Tucker (1974) carried out a five year

longitudinal study with the variables of socioeconomic

status and intelligence being carefully controlled. One

of their major purposes was to assess and evaluate the

impact of elementary schooling conducted primarily in

a second language on the linguistic development of

children.







The children chosen to participate in the experiment

were from middle class neighborhoods in parts of Canada.

There was an experimental class composed of 26 children

from exclusively English speaking homes who received

instruction in French only. There were two control groups

of children from exclusively English speaking homes who

received instruction in English. There was a control

group composed of 22 children from exclusively French

speaking homes who received instruction in their native

language.

In September 1966, all children were tested for

general intelligence with a non-verbal intelligence test,

The Raven Progressive Matrices. To check the family's

socioeconomic status, all the parents were interviewed.

The authors ensured that all subjects had a high quality

linguistic environment, that an appropriate nature and

extent of extracurricular activities were undertaken by

the family (indoor activities of the family and outdoor

excursions during weekends and vacations), that there

was a good availability and quality of guidance at

home, and that the family encouraged the use of such

educational facilities as books, periodical literature,

library, etc.

At the end of the first year, the children were

tested for various skills in the English language. The

experimental group performed at the general average for








their age level in English. At the end of their fifth

year, the English speaking children who had been taught

in French only showed the same normal, average gains

in their educational attainment and were as proficient

in English as those taught in the conventional manner.

They also exhibited a great fluency in and understanding

of the French language.

Ramirez and Politzer (1974) carried out a study

on a sample of children in a bilingual education

program in San Francisco. The results provide a further

defense of childhood bilingualism. Spanish and

English versions of grammar test were administered to

40 bilingual children at grade levels kindergarten,

one, three, and five. English scores varied only

slightly among grade levels. Subjects were questioned

about language use. The relationships between reported

language use and test results appeared to indicate that

native language (Spanish) proficiency was determined

by the use of the language in the home. The

lack of any significant correlations between the

English and Spanish version scores appears to indicate

that achievement in English is unrelated to the maintenance

of Spanish for bilingual children. More importantly, it

was concluded that if the quality of the language environ-

ment was good and continuously used in the home even










'-though only English was used in school, there was no

adverse effect on the primary language.

Negative effects of childhood bilingualism

General information and observational reports.

Many observers and investigators have concluded that

childhood bilingualism results in many disadvantages.

The following case against bilingualism is derived

from opinions that have appeared in the literature.

Many authors (Weinreich, 1961; Hickey, 1972;

Whitman and Jackson, 1972; Cornejo, 1973; Ervin-Tripp,

1973; Matluck and Mace, 1973; Politzer and Ramirez, 1973;

Saville-Troike, 1973; and Taylor, 1974) generally

agree that some linguistic interference (using character-

istics of one language while speaking the other)

will occur and the individual cannot learn both

languages equally well.

In 1962, Jensen surveyed the literature on

the ill effects of bilingualism on the individual

child's speech and language development. With regard to

speech development, he summarized that a child will develop

faulty articulation, inappropriate pronunciation and many

substitutions of sounds will result. Jensen (1962a) also








found much literature asserting that the bilingual child

will encounter many problems in language development.

Jensen states:

A smaller active and passive vocabulary will
result in most instances, for the child will be
occupied learning two terms for one referent.
Even the total number of terms is likely to be
less than the total number for the monolingual.
The bilingual will develop a confused, mixed
vocabulary because of lexical borrowing. He
will use shorter sentences, more incomplete
sentences, fewer interrogative and more excla-
matory sentences. .Confused structural patterns,
unusual word order, and errors in agreement and
dependency will result. The bilingual child
will make many errors in the use of verbs and
tense, connectives, prepositions, nouns and
articles. He will make faulty use of negatives,
will fail to inflect words and will employ too
much redundancy. Virtually all authors agree
that an individual cannot learn both languages
equally well,' for linguistic interference is
bound to occur. Both languages will suffer
when judged by the highest standards. (p. 135)

Cornejo (1973) elicited spontaneous speech samples

using visual cues from twenty-four bilingual five-year

old Mexican American children. After recording and

analyzing their spontaneous expressive language abilities

in both Spanish and English, Cornejo made the following

observations:

1. English is the dominant language of the subjects.

2. Even though Spanish is used at home, its structure
and phonology are influenced by English.

3. There is interference from English to Spanish.

4. The interference from Spanish to English is highly
significant at the phonological level, but minor
at the lexical and grammatical levels.







Hepworth (1974) very vaguely describes the outcome

of many bilingual school programs in various U.S. cities.

The programs taught school subjects to Spanish American

children in the native language (Spanish), with the

purpose of introducing English as a second language. In

the move from monolingualism (Spanish) to bilingualism,

English was effortlessly acquired by the children. However,

linguistic skills in Spanish were not maintained at all

and the children became essentially monolingual again

(English).

Controlled investigations. A study carried out by

Carrow (1971) revealed that bilingualism had negative

effects on both the primary and secondary language develop-

ment of a group of preschool Mexican-American children

from a low socio-economic level in Houston. The main

purposes of the study were 1) to compare the comprehension

of English with that of Spanish in the group of

Mexican-American children and 2) to compare these develop-

mental sequences of both languages in the children under

study with the performance of a group of English-speaking

children reported earlier in the literature.

The subjects were 99 children with Mexican-American

surnames, ages 3 years 10 months to 6 years 9

months. Each child was administered the Auditory

Test for Language Comprehension in Spanish and English.

This test is an instrument that allows the assessment of

oral language comprehension in such language structures








as pronominal reference, negation, tense marker compre-

hension, adjectives, prepositions, pluralization and

lexical items. Generally, the children varied widely in

the extent to which they understood the two languages.

The greater proportion of the ninety-nine children under-

stood English better than Spanish. It was further revealed

that the bilingual children's comprehension of both English

and Spanish was delayed when compared with the comprehen-

sion abilities of the children in the monolingual

Anglo-American control group.

Another study (Carrow, 1972) compared the group of

bilingual children discussed in the previous study (Carrow,

1971) with a group of monolingual Anglo-American children

of comparable age and social status in a test of auditory

comprehension of English. Once again, the test instrument

employed was the Auditory Test for Language Comprehension

(Carrow, 1968), which permits the assessment of oral

language comprehension of English and Spanish without

requiring language expression. Both groups of children

were tested by the same examiner. The results showed that

the bilingual children made a significantly greater number

of errors than did the monolingual children in all

linguistic areas tested except verbs. Carrow (1972) further

concluded that in general a language delay is initiated

in Mexican-American children of a low socioeconomic level

during the preschool years.







Politzer and Ramirez (1973) compared the English

language expression abilities of bilingual Mexican-American

children in the early primary grades in a bilingual school

and a monolingual school. The children were asked to tell

the story of a silent movie they had watched immediately

before being interviewed. Their speech samples were

recorded on tape and transcribed. Deviations from standard

English were counted and categorized. One of the main

findings of the study was that the children in the bilingual

school did not significantly differ from those in the

monolingual school with respect to the frequency and

amount of morphological, syntactical and lexical errors

made while speaking English. This indicates that bilingual

children receiving schooling in both the native and

non-native languages are no more delayed in their spoken

English abilities than are comparable children receiving

schooling in the non-native language exclusively. The

other main finding was that interference from Spanish

caused several significant deviations from standard

English in the speech of these children. These results

indicate that bilingualism had negative effects on the

acquisition of the second language.

Summary of the review of effects of bilingualism on lan-
guage acquisition

The literature reviewed still does not answer the

questions "Is this phenomenon of bilingualism a curse or

a blessing? Or is it neither?" (Jensen, 1962a).









Bilingualism is a highly complex subject and the

answer to the question is not an easy one. Comparing

the many observations made and studies performed is like

"comparing apples and oranges." The conclusions reached

are difficult to generalize as the various observations

and investigations involved subjects with a wide variety

of characteristics and abilities. The most accurate

generalizations that seem to be indicated by the review

of the literature are the following:

1) Many of the results reported are conflicting and
contradictory. Specific examples are:

a. Cornejo (1973) states that even though Spanish
is used at home, there is interference from
English to Spanish. Ramirez and Politzer (1974)
state that there is no effect on the primary
language (Spanish) if it is continuously used
at home.

b. Many authors previously cited in this review of
literature assert there is interference from the
primary language hindering non-native language
acquisition. Zierer (1974), Lambert, Just, and
Segalowitz (1970), Lambert and Tucker (1974),
and Hepworth (1974) have reported various
results which directly contradict this.

c. Several authors including Carrow (1971), Cornejo
(1973), and Hepworth (1974) present evidence that
bilingualism has a negative effect on the main-
tenance of the native language. Other authors
including Lambert, Just and Segalowitz (1970),
Lambert and Tucker (1974), and Zierer (1974)
reported that the primary language remains intact.

d. Cornejo (1973) specifically states that interference
from Spanish to English is minor at the lexical
and grammatical levels. Several authors (Carrow,
1972; Matluck and Mace, 1973; and Politzer and
Ramirez, 1973) reveal that interference is the
greatest at these linguistic levels.





15

2) Socioenvironmental and intellectual variables have
been shown to relate to the language proficiency of
a group of bilingual children.

3) The poor linguistic development of bilingual children
reported by several observers and investigators may
be a result of exposure to poor language models and
not bilingualism, per se.

4) Bilingualism has either positive or no effects on
middle class children of average intelligence from
Canadian-American homes.

5) Bilingualism has various negative effects on the
language development of Mexican-American children
from low socioeconomic backgrounds.

6) The confusing and conflicting observations have in-
dicated that a need remains to examine bilingual
groups other than Mexican-Americans or Canadian-Americans
and to determine whether bilingualism has the same
effects in socioeconomic environments in which the
quality of the language environment is high, and
whether the ability to learn two languages well is
influenced by intelligence. Judgements as to the
total effects of bilingualism should be deferred
until such studies are available.

Morphology Tests

Berko (1958) carried out a classic study which

examined children's productive knowledge of English

morphology using nonsense words. The children she studied

ranged from four to seven years of age. The rationale

for using nonsense words rather than real words was to

enable the examiner to be certain that the task required

use of morphological rules and not just memory and imitation.

In this fashion it could be ascertained whether or not

children possessed rules they could generalize to new forms.

A correct response indicated that the child had mastered

adult morphological rules. The choice of test items was

based on the vocabulary of first grade children and a fre-

quency inventory of morphological features.








The mrphcmes which Berko tested were the plural, the

third person singular of the verb, the progressive, the

past tense, the two possessives, and the comparative and

superlative of the adjective. One of the major conclusions

Berko drew from her study was that the answers were

consistent and orderly even though they were not always

correct so far as English is concerned. She further

concluded that children in this age range operate with

clearly delimited morphological rules.

Natalicio and Natalicio (1969) indicated that Berko's

work of 1958 represented one of the most important steps

in linguistic investigation. This is evidenced by the

fact that within just the past decade, numerous investi-

gators carried out studies employing the basic method

introduced by Berko (1958). For example, Ainsfeld

and Tucker (1967) studied the productive and receptive

control of pluralization rules of six-year-old American

children; Natalicio and Natalicio (1971) studied

productive control of English pluralization rules by

native and non-native English speakers. Cooper (1967)

reported on the ability of deaf and hearing children to

apply morphological rules; Shriner and Miner (1968)

reported on the ability of economically deprived children

to apply morphological rules. Several investigators

(Newfield and Schlanger, 1968; Dever and Gardner, 1970;

Dever, 1972) used Berko's design and reported on the

ability of mentally retarded children to use morphological







structures. Ivimey (1975) used the same basic method to

study the formation and use of morphological rules in a

large sample of English school children living in London.

Many of these published studies demonstrate the

validity of Berko's test of morphology by showing its

ability to identify those children who have difficulty

in generalizing morphological rules to new word forms.

Several of the previously cited researchers reported the

ability of Berko's test to discriminate between deaf

and hearing children, normal and mentally retarded children

and native and non-native English speaking children in

their ability to apply morphological rules. Several of

the studies revealed that the ability to supply the correct

inflections of nonsense words increased with age. (These

differences produced between age groups by the Berko test

scores also seem to indicate its validity.)

Berry and Talbott (Berry, 1966) published a language

test adapted from Berko's methods designed to assess features

of English morphology most commonly represented in the

vocabulary of the first-grade child. As did Berko, Berry

and Talbott used nonsense figures to elicit the correct

inflections in sentence completion tasks. Although, based

on their preliminary studies and tests, they predicted

differences in the performance levels of children from

five to eight years; no normative scores for the various

age levels were secured.








The Grammatic Closure subtest of the Illinois Test

of Psycholinguistic Abilities (Kirk, Mc Carthy and Kirk,

1968), modelled directly on Berko's test, uses meaningful

lexical items to assess a child's competence in his

productive knowledge, of morphological rules. In this

test, the examiner describes two pictures orally. One

sentence is complete while in the second a word is

missing and has to be remembered from the first sentence

and inflected verbally by the testee. Mc Carthy and

Kirk (1963) and Kirk and Kirk (1971) cite numerous studies

providing the validity and reliability of the ITPA subtests.

Vogel (1977) found both the Grammatic Closure Subtest

of the ITPA and the Berry-Talbott Test to be extremely

valuable measures for identifying children experiencing

difficulty in generating and mastering morphological rules.

After administering both tests to twenty good readers and

twenty dyslexics, Vogel found that the dyslexics were

significantly inferior to the good readers on both

measures. Her main conclusions were that poor readers

may be deficient in morphological ability which may result

in inefficient use of semantic and syntactic clues provided

in the morphology of written language and that these children

may benefit from a language intervention program designed

to help them internalize complex morphological structures.

Vogel (1977) cites a study by Brittain (1970) which

replicated Berko's study with first and second-grade

children. Brittain found significant correlations between







inflectional performance and reading ability, It was

suggested that a child of seven or eight who has not

mastered the morphological rules of his language will be

hampered in learning to read.

Some investigators still have reservations about the

use of nonsense tests of morphology. Dever and Gardner

(1970) suggest that it might be better to use real words

and not nonsense words. They state:

It may be that nonsense syllables actually defeat
the purpose of the test. The question is, are the
nonsense syllables themselves actually a confusion
factor for some subjects? The fact is that the
scores are better and more regular on items which
used real words as stimuli than they were on the
nonsense syllables which tried to elicit the same
forms. The question arises whether this is due
to the possible memorization of the entire inflected
form as postulated by Berko, or whether this is due
to unknown factors inherent in the use of nonsense
syllables. (p.179)


Johnson (1974) suggests that the Berko nonsense test

of morphology may not really be testing what it proports

to test due to the confusion caused by the nonsense words.

Johnson gave the test to college educated adult subjects

before it was administered to the children. Several of

these adults failed to demonstrate control of the plural.

Johnson implies that there must be something inherently

wrong with the stimulus item when a college educated

subject who is a monolingual native speaker of English

fails to demonstrate proper productive control of the

English plural.








Ivimey (1975), after replicating Berko's study,

reported that a majority of the adults sampled gave

technically "incorrect" responses. As a result, he

stated, ". the widespread idea that there is a more or

less stable adult norm appears to be a myth" (p.122).

However, unlike Johnson (1974), Ivimey does not blame the

unstability of the adult reponses on an inherent fault

of the use of nonsense stimuli. He states that the reason

for these "incorrect" adult responses, "arose from what

may be described as excessive knowledge; as linguistic

sophistication increased greater deviance was demonstrated

in handling the nonsense test words" (p.122).

Several investigators have used nonsense words in

combination with meaningful words to assess children's

ability to apply morphological rules. Miller and Ervin

(1964) used familiar words and nonsense words with a

group of twenty-five children. They found that the correct

use of the inflection with meaningful words consistently

preceded correct use with the nonsense counterparts.

Newfield and Schlanger (1968) used Berko's nonsense

words and a list of meaningful words to study the

acquisition of English morphology in a group of normal

children and educable mentally retarded children. They

found a statistically significant difference between

correct responses on real words and nonsense words for

both groups of children. They summarized that there is a







definite time lag between the production of correct

morphological inflection forms within a context of

familiar words and the generalization of these forms to

unfamiliar words.

Dever and Gardner (1970) studied the performance of

normal and retarded boys on Berko's test. They reported

that the scores were better and more regular on items

where real words were used as stimuli.

Dever (1972) compared the results of a revised version

of Berko's test with the free speech of mentally retarded

children. The purpose of the study was to determine if

the test could predict the occurrence of errors in the

free speech of educable mentally retarded children. The

author concluded that real-word test stimuli were only

slightly better as predictors and that both nonsense

items and real word stimuli are poor indicators of mentally

retarded children's ability to use inflections in their

free speech.

Pettit and Gillespie (1975) compared normal children's

performance on ten selected items from the Berko test and

ten selected corresponding meaningful items from the

Grammatic Closure subtest of the ITPA. The generalized

results were that the children scored significantly

higher on the ITPA items than the Berko items at every age

level from three to eight years.








Summary of the review of the application of tests of
morphology

A review of the literature on the previous and

present application of tests of morphology indicates the

following:

1) Morphology tests using nonsense forms adapted from
the methods employed by Berko (1958) have been
demonstrated to be valid test procedures to assess a
youngster's ability to apply morphological rules
to new word forms and to discriminate between normal
(in all aspects) monolingual English speaking children
and the mentally retarded; dyslexics; the deaf;
and non-native speakers of English.

2) Morphology tests using real words have been
demonstrated to be valid test procedures to assess
a child's developmental level of morphological rule
acquisition and to discriminate between normals
and such populations as dyslexics and the mentally
retarded.

3) The ability to apply morphological rules to meaningful
words precedes the ability to apply the same rules
to new word forms; OR, the use of nonsense words in
sentence completion tests underestimates a child's
productive knowledge of morphological rules.

4) Bilingual populations have been essentially ignored
by investigators studying morphological rule acquisi-
tion. (An exception is a study by Natalicio and
Natalicio, 1971, who compared only the acquisition
of noun plurals in English in a group of native
English speakers and native Spanish speakers who
acquired English as a second language.). Information
describing and comparing bilinguals' ability to
apply morphological rules in both acquired languages
as well as information regarding specific inter-
ferences of morphological rule formulation of one
language on another would greatly enhance:
1) the information available on the application of
morphology measures and 2) the conflicting information
available on the effects of bilingualism on language
acquisition.

Language Tests for Bilingual Students

Accurate assessment of limited English speaking and

bilingual children has become a national educational concern.







An essential part of planning bilingual programs includes

assessment of language dominance and proficiency. This

requires a total description of the communication skills,

linguistic structures, and functional usage of the child's

native and non-native languages. Without adequate

assessment, the implementation of special language pro-

grams and the identification of eligible students will

be hindered (Silverman, Noa, and Russell, 1976). This

section contains a review of commercially available tests

used to assess the language proficiency and dominance of

bilingual, Spanish/English speaking children. Only those

instruments which contain information and evidence

supporting their validity and reliability are described.

Silverman, Noa, and Russell (1976) provide a concise

evaluation of twenty-four tests used to evaluate the

language dominance and competence of groups of children

speaking two or more of a variety of languages. The

descriptions of the following first two instruments were

based on the information provided by the aforementioned

investigators.

Del Rio Language Screening Test (DRLST). The DRLST

was devised to identify three-to seven-year-old children

with deviant language performance in English and Spanish

who need further evaluation. In addition, it may be used

as a diagnostic instrument and as a test of bilingual

proficiency. It is designed to group children into

three language categories: 1) English speaking







Anglo-American children, 2) predominantly English

speaking Mexican-American children, and 3) predominantly

Spanish speaking Mexican-American children.

The DRLST consists of five separate subtests. The

"Receptive Vocabulary" subtest assesses comprehension

of single nouns and verbs. The "Sentence Repetition

Length" subtest measures memory for a string of gram-

matically related words. The "Sentence Repetition-Complexity"

subtest assesses ability to repeat short sentences which

gradually increase in grammatical complexity. The "Oral

Commands" subtest evaluates memory for increasing numbers

of oral commands. The "Story Comprehension" subtest

evaluates a child's memory for information presented

in brief stories.

Validity of the DRLST: Scores on the DRLST for each

of the three main groups discussed earlier were compared.

Norms for each of the three language categories differed

significantly (.001 level of confidence). In addition,

the test developers feel that the five subtests assess

some of the basic components of language development. The

"Sentence Repetition-Complexity" test was organized using

hierarchies of syntactic development cited in research

studies.

Reliability of the DRLST: From each main group

thirty-two subjects were randomly selected for retesting

exactly two weeks after the initial administration of

the test. Reliability was computed using Spearman's







rank-order correlation procedure for matched groups. All

correlation coefficients are significant beyond the ,01

level of confidence.

'Pictorial Test of Bilingualism and Language Dominance

(PTBLD). The primary objective of the PTBLD is to determine

four-to eight-year-old children's language facility in

Spanish and English. The results obtained from Part I of

the test are used to place children in one of six diagnostic

categories: 1) English dominant, 2) Spanish dominant,

3) bilingual, 4) language deficient, 5) further testing

advised for possible learning and language disorders

or mental retardation, and 6) pseudo-bilingual. Oral

vocabulary is used as the basis for these determinations.

Part II provides information on the child's use of

language structures in both Spanish and English.

For Part I of the PTBLD the subject is shown a

picture, asked what it is, and is requested to respond

orally in both Spanish and English. For Part II of the

test, the child is shown a picture and is asked to

tell a story in English or Spanish about what he sees.

Then the child is asked to tell the story again in the

other language. All responses are recorded. Correct

answers for Part I of the PTBLD are specified in the

manual. For Part II, the rules for classifying responses

according to level, based on grammatical correctness and

thought completion, are given in the manual.








Validity of the PTBLD (Part I): Part I was

administered to a group of 176 Mexican-American children

who had already been classified into two groups: Spanish

dominant and English dominant. Results suggest that the

test does differentiate between English and Spanish

dominant language preferences. A second study supporting

the validity of Part I involves a sample of Mexican-American

children who had taken both the Peabody Picture Vocabulary

Test (PPVT) and Part I of the PTBLD within a three-week

time period. A significant correlation (.05 level of

confidence) was found between the English and total oral

vocabulary scores on Part I of the PTBLD and the PPVT.

Validity of the PTBLD (Part II): The two pictures

that comprise Part II were shown to two hundred

Mexican-American children in grades kindergarten through

third. Analysis of the children's recorded answers

indicated that the responses could be rated into four

distinct levels of oral language facility.

Reliability of the PTBLD: Two student samples

selected from kindergarten and first grade were retested

after two-month and six-month intervals, respectively,

on Part I. Coefficients were reported as high. No

reliability studies for Part II of the PTBLD were reported.

The James Language Dominance Test. The James Language

Dominance Test (James, 1974) measures both comprehension

and word production abilities in kindergarten and first

grade children in Spanish and English. It was designed to







determine language dominance. Five categories of

dominance are identified: 1) Spanish dominant, 2) bi-

lingual with Spanish as home language, 31 bilingual

with English and Spanish as home languages, 4) English

dominant, but bilingual in comprehension, and 5) English

dominant.

The test is administered to individual children. The

examiner asks questions first in Spanish, then in English;

the child responds both orally and by pointing to pictures

in the test book. All responses are recorded by the

examiner. Using the child's scores and the language

classification guide provided in the test manual, the

student may be placed into one of the five specified

categories of language dominance.

Validity of the James Language Dominance Test: James

(1974) presented evidence that the test is valid with

respect to content and predictive validity. He also

demonstrated, through a review of group norms, and

analysis of expectancy tables, and a multiple discriminant

analysis, that the differences between each of the five

linguistic categories are sufficient to accurately

classify children into any of the five mentioned categories.

Reliability of the James Language Dominance Test:

The test was administered to three first-grade classes

of Mexican-American children in Austin, Texas. Six months

later the same children were retested. The

classifications obtained by the second testing were





28

in 98% agreement with the classifications of the first

testing.

Screening Test of Spanish Grammar (STSG). The STSG

(Toronto, 1973) was designed to identify 3-to 6-years,

11-months-old Spanish-speaking children who do not

demonstrate native language (Spanish) syntactic proficiency

commensurate with their age. The test is concerned with

the grammatical proficiency of Spanish speaking children,

not with their phonological or vocabulary skills. The

STSG is meant to be used as a syntax screening device only,

not as an in-depth or complete language assessment.

The STSG consists of a receptive (comprehension) and

an expressive subtest, each testing twenty-three syntactic

structures. The structures are presented in paired

sentences which differ in only a single syntactic element.

There are four pictures for each receptive item; two

correspond to the test sentences and two are decoys. The

examiner first says both sentences and then after repeating

one of the sentences, the child is asked to point to the

picture representing that sentence. On the expressive

portion of the test there are only two pictures per item

which correspond to the sentences. The examiner first

says both of the sentences and then points to one of the

pictures. The child is asked to produce from memory the

corresponding sentence. Using the child's scores and the

normal percentiles and standard deviations provided in the

test manual for Mexican-American and Puerto-Rican-American







children, it can be determined whether further language

evaluation is necessary.

Validity of the STSG: Toronto (1973) stated that

validity was indicated by the significant differences

produced between age groups by the STSG scores. As age

increased, the scores which supposedly measured increased

grammatical proficiency did, in fact, increase signifi-

cantly.

Reliability of the STSG: Twelve subjects each from

the Mexican-American group and the Puerto Rican-American

group were retested with the STSG after a one-week period.

The test-retest reliability coefficients were .69 and

.91 for the Mexican-American children on the receptive

and expressive subtests respectively, for the

Puerto Rican-American group the reliability coefficients

were .62 and .72 on the receptive and expressive subtests

respectively.

Test for Auditory Comprehension of Language (TACL).

The TACL (Carrow, 1974) was designed to measure auditory

comprehension of language structures in Spanish and English

in children of three to six years eleven months. Perform-

ance on specific items and groups of items allows the

examiner to determine the areas of linguistic difficulty.

The test permits the assessment of oral language

comprehension without requiring language expression from

the child. The TACL consists of 101 plates of line

drawings. The pictures represent referential categories







and contrasts that can be signaled by form classes and

function words, morphological constructions, grammatical

categories, and syntactic structures. The examiner

questions the child, who responds by pointing to the

correct picture in the test booklet.

Normative data are available only for the English

language version. The TACL manual states that the child

may be assigned to a developmental level of comprehension

in each language. However, norms exist only for the

English version. Consequently, the English version norms

must be used for assigning the child to a developmental

level of comprehension in Spanish. In addition, the

Spanish version of the TACL is a direct translation of

the English test questions. The practice of merely

translating English questions to Spanish for evaluating

the bilingual child is criticized later in this section

of the review of literature.

Validity of the TACL: Carrow (1974) presents good

evidence to prove the validity of the English version of

the TACL. No studies are reported on the validity of the

Spanish version.

Reliability of the TACL: The English version of the

TACL was administered to a group of fifty-one children.

The Spanish TACL was administered to a group of thirty-two

Mexican-American children. All were retested within a

two-week interval. The reliability scores were .94 for the

English version and .93 for the Spanish version.







Other diagnostic measures for evaluation of language

in young Spanish-speaking children consist of 1) unstand-

ardized translations of tests into Spanish developed for

English-speaking children living in the United States

(Clevenger, 1974) and 2) instruments which ahve no evidence

of their reliability and/or validity (Silverman, Noa, and

Russell, 1976). Diagnostic measures of these types are

not described in this review as many investigators

(Zirkel, 1972; Garcia, 1973; Clevenger, 1974; Silverman,

Noa and Russell, 1976; Rueda and Perozzi, 1977) feel

that classification or instructional decisions based on

such test results cannot be made with confidence.

Zirkel (1972) and Garcia (1973) have stated that

the translation of English language tests into Spanish

is an inappropriate means for measuring Spanish-speaking

children's linguistic abilities because the translated

tests often reflect cultural and linguistic values which

may not be found among the Spanish speaking population.

The participants of the Bay Area Bilingual Education League

Conference (1969) stressed the importance of devising

language tests which are derived from the native language

of the testee and are not just translations from existing

measures.

Rueda and Perozzi (1977) presented a detailed critique

of the Spanish version of the TACL (Carrow, 1974). After

analyzing the test and administering it to twenty

Mexican-American children of normal intelligence, they

indicated that the test's usability is questionable






32

because of translation difficulties and ambiguous pictorial

representations of stimuli. The TACL was written first in

English, and the Spanish translation was made directly from

the English wording, "a procedure of dubious validity,"

according to Rueda and Perozzi (1977). These investigators

further report that it is often impossible to impart the

meaning intended when a word-for-word translation is made.

One example, Item 37, is the verb "catching." The TACL

Spanish translation made was "pescando" which has the

implication of catching fish. The correct stimulus picture

is of a boy catching a baseball. Nineteen of the 20

children tested in this study missed item 37. Other

examples of misleading translations and poor stimulus

pictures on the TACL were described,

SThe aspects of validity and reliability are critical

in creating accurate assessment instruments. If decisions

are to be made on the basis of students' test scores, one

must know what specifically the test measures and how well

it accomplishes its objectives (validity). It is also

important to know how stable and consistent the scores are

(reliability). The tests evaluated in this section of the

review of the literature deal with the classification of

students into language dominance categories and identifying

areas of strengths and weaknesses within specific linguistic

content areas. Without evidence supporting the validity

and reliability of such tests, confidence cannot be placed

in their results (Siverman, Noa, and Russell, 1976).







Summary of the review of language tests for Spanish-speaking
children

A review of the literature on language tests available

for Spanish-speaking children indicates the following:

1) Language tests which have not been proven to be valid
and reliable and/or are direct translations from
tests originally standardized on Anglo-American
children should not be used to evaluate language
dominance or proficiency in Spanish-speaking
children.

2) Several valid and reliable tests measuring
Mexican-American children's language dominance in
both comprehension and production are available.

3) Several valid and reliable tests for screening
Mexican-American children's ability to receptively
and expressively use various linguistic structures
in Spanish and English are available. The STSG
contains norms for Puerto Rican-American children
as well as Mexican-American children.

4) Although the purpose of the screening tests are to
identify the specific linguistic areas that need
a more in-depth evaluation, no valid and reliable
diagnostic measures permitting a more specialized
evaluation of the Mexican-American or
Puerto Rican-American child's linguistic
abilities are available.

5) There are no valid and reliable diagnostic measures
whatsoever for determining any of the basic
aspects of language functioning in Cuban children
available.

6) The notion of assessing bilingual students'
language performance in their native and non-native
languages is a new one. One indication of this
fact is the extremely limited number of tests
available The need for effective assessment
instruments is evident. Not only is it necessary
to develop more comprehensive instruments
than are currently available, new tests must be
adapted to the characteristics of each
language group to be evaluated. The language
development of Cuban-American children has been
ignored. Until language tests based on the
Spanish spoken by Cubah children are devised, the









evaluation of the linguistic skills of the
large number of Cuban-American children in
the United States will not be maximally effective.

Statement of the Problem

As evidenced by a review of the literature, it can

be seen that there is much conflicting evidence as to the

effects of bilingualism on language development; there is

a poverty of research in the specific area of morphological

rule acquisition; and the language development of

Cuban-American children living in the United States has

largely been ignored. Consequently, there are no

valid and reliable measures for assessing their linguistic

development.

The purposes of this research are to determine the

effects of bilingualism on the native and non-native

language development of children of normal intelligence

from economically advantaged backgrounds, to obtain informa-

tion on the acquisition of morphological rules in bilinguals,

and to develop an effective diagnostic instrument which can

be used to assess the linguistic skills of Cuban-American

children. The specific goals of this investigation are:

1. To devise a standardized instrument for use in

the evaluation of the linguistic development

of the Spanish and English spoken by bilingual

Cuban-American children.

2. To analyze the developmental sequences of morpho-

logical rule acquisition in the two languages of

these children.








3. To determine if there are significant

differences between morphological responses

to test using nonsense and meaningful

words in Spanish and English.

4. To compare the ability of the Cuban-American

children to apply morphological rules in

English and Spanish. These results will be compared

with the commercially available language

dominance measure using lexicon stimuli which

previously classified the children as being

bilingual in both Spanish and English

comprehension and production.

5. To determine the effect bilingualism has

on non-native language acquisition by

comparing the developmental sequences of

the English spoken by bilingual Cuban-American

children with the English spoken by monolingual

American children.

6. To determine the effect bilingualism has on

native language development by comparing a

sample of Spanish dominant children and

bilingual children in their ability to apply

Spanish morphological rules.

7. To look for consistencies of error production

and to gain an understanding of bilingual

interference in the language learning of the

specific bilingual group-- Cuban-American.














CHAPTER II

METHODS AND PROCEDURES

Subjects

The subjects used in this study were 122 bilingual

Spanish/English speaking elementary school aged children,

31 each in kindergarten and third grade, 30 each in the

first and second grade, and 19 Spanish dominant kinder-

garten children. All children were considered by their

individual teachers and through examination of school

records to be normal with respect to intellectual ability,

hearing, speech, vision, and language, social and physical

development. In addition, the children were required to

meet the following criteria:

a. All children were born and raised in Miami, Florida,
and were of Cuban born parents.

b. Spanish was the primary language spoken at home
by both parents.

Subjects were drawn from a Dade County Public School

located in the predominantly Cuban community of Miami,

Florida,known as "Little Havana." All academic instruction

was provided in English. The children did attend Spanish

class for hour daily. Each child came from middle and

lower middle class home environments as judged by the

school principal's knowledge of the community and types of







homes in which the children lived. This was supported by

an examination of school records which indicated the

occupations of the subjects' parents. Mothers were

predominantly housewives; fathers were mainly plumbers,

mechanics, carpenters, construction workers, cooks or

bakers, bus drivers, butchers, shop owners or salesman.

A few were teachers in the Dade County Public School

System.

The chronological age range of the 19 Spanish dominant

kindergarten children was from five years, four months to

six years, four months with a mean age of five years,

eleven months. There were 10 males and 9 females. See

Table 1 for a description of the bilingual subjects.

The degree of the children's bilingualism was carefully

controlled and measured in this study. Both teachers'

evaluations of students' bilingualism and the students'

performance on a standardized, commercially available

diagnostic test of language dominance and bilingualism

were used to categorize subjects as either bilingual or

Spanish dominant for this study.

Each subject was first administered the James Language

Dominance Test (JLDT) which is designed to assess the

language dominance of Spanish/English speaking children.

(A complete description of the JLDT including information

on its validity and reliability is on page 26 of the

Review of Literature section.) For purposes of this

investigation children were considered bilingual if their















Table 1 Description of bilingual subjects in terms of
CA and sex for each grade level tested.


MEAN CA

6-0

7-2

8-3

8-11


CA RANGE

5-4 to 6-6

6-4 to 8-2

7-4 to 8-11

8-4 to 9-9


GRADE

K

1st

2nd

3rd


MALES

16

16

15

16


FEMALES

15

14

15

15








JLDT scores categorized them either as Bilingual in

Comprehension and Production with Spanish as a Home

Language or Bilingual with English and Spanish as the

Home Language. Nineteen kindergarten children obtained

JLDT scores categorizing them as Spanish Dominant.

(Approximately 3 days elapsed between the administration

of the. JLDT and any further testing.)

The JLDT Manual provides language group norms for the

receptive and expressive vocabulary development of the

normal children used in the standardization sample of the

JLDT. Table 2 shows these means and standard deviations.

Based on these norms, any child-who scored lower than 17

and 14 on the comprehension and expression subtests,

respectively, of the JLDT in Spanish or English were ex-

cluded from the present study. This procedure further

assured that subjects met the criterion of having normal

intellectual and language development.

Materials

Two 42-item morphology tests were constructed to assess

the ability of bilingual Cuban-American children to produce

grammatical morphemes in Spanish and English. The Spanish

version is hereafter referred to as the Bilingual Morphology

Measure-Spanish Version or BMM-S and the English version

is termed the Bilingual Morphology Measure-English Version

or BMM-E. These were adapted from the techniques employed

by Berko (1958) and Berry and Talbott (Berry 1966). Each

version used 21 nonsense words to determine if

the subjects were able to produce the correct











p4r ( -I







U *' 0 M IO LAN


CN C4r N

1r-I r-4
a'Zn,


r-l






0 E
3




aP4
*rl



0 0
L
OE
44 O



0O












to -r
0 4-)








I4
C C












(0
o







0 4
C *







01





Cro
CM














(N
10= I -r-1
trMM

(Mk
r0) c
1- 1I
kL
E0~q


I S .
1 -4 p. 4- 4-





H _____


ZI O o r> L o
CN r r r-





.4. 4


10
rC
ta







(a 0D



c
S)
13 3


a)
0



- 0



too
0

3C

0 M
U H


z
0
H



0
Q


CN
| rl
W\,


--


K
CO
H







morphological forms with nonsense stimuli and 21 meaningful

words which paralleled the nonsense items morphologically,

and phonologically wherever possible. Color drawings of

cartoon-like animals and cartoon-like people performing

actions were the visual stimuli for the semantically

non-meaningful subtests. Color drawings of real animals,

objects and people representing the real words were the

visual stimuli for meaningful items. Drawings were

mounted on cards.

The choice of the inflectional morphemes to be invest-

igated in this study was based on two main factors:

a) that they be representative of the most frequently

occurring forms in the language of children in the early

elementary years and, in order to be better able to compare

the bilingual children's performance with other populations

b) that they be forms assessed in a variety of previous

studies about children's morphological development.

Table 3 describes the morphemes examined and the

corresponding nonsense and lexical items tested by the

BMM-E test instrument. Table 4 describes the cate-

gories and items examined by the BMM-S test instrument.

Testing Procedures

Two bilingual examiners completed all testing. An

examiner of Cuban descent, fluent in Spanish and English,

administered the BMM-S to all of the bilingual and Spanish

dominant children. A bilingual American born native speaker of















The following is a list of the semantically
non-meaningful and lexical items represented
in the plates of the Bilingual Morphology
Measure-English Version (BMM-E) test
instrument


GRAMMATICAL
CATEGORY

plural /s/
plural /z/



plural /az/


regular
regular


regular
present
derived
*present


past tense /t/
past tense /d/


past tense /ad/
progressive
agentive
tense


comparison of adjective
superlative of adjective
singular possessive


plural possessive


NONSENSE ITEM
TESTED

mooks
luns
wugs
cras
gutches
tasses
nizzes
ricked
bomed
showed
motted
zibbing
zibber
kleats
fonces
quirkier
quirkiest
niz' hat
wug's hat
nizzes' hats
wugs' hats


LEXICAL ITEM
TESTED

sticks
lions
dogs
keys
watches
glasses
noses
picked
climbed
cried
melted
painting
painter
eats
dances
bigger
biggest
girl's hat
dog's hat
girls' hats
dogs' hats


*third person singular


Table 3















The following is a list of the semantically
non-meaningful and lexical items represented
in the plates of the Bilingual Morphology
Measure-Spanish Version (BMIM-S) test
instrument


GRAMMATICAL
CATEGORY

plural /s/



plural /es/






*regular past tense -ar verb
*regular past tense -er verb
*regular past tense -ir verb
present progressive
.derived agentive
*present tense


comparison of the adjective
superlative of the adjective
singular possessive feminine
singular possessive masculine
plural possessive feminine
plural possessive masculine


NONSENSE ITEM
TESTED'

nerras
litos
raves
pirones
micales
mubles
tapices
deyes
banto'
molido
forrio'
fintando
fintor
lunta
dobe
/
mas mecoso
mecosisimo
de la nerra
del lito
de las nerras
de los litos


LEXICAL ITEM
TESTED

palos
perros
leaves
leones
animals
rubles
narices
reyes
lloro
recogio
derretio/
pintando
pintor
baila
come
mas grande
grandisimo
de la nina
del perro
de las ninas
de los perros


*third person singular


Table 4








English administered the Bilingual Morphology Measure-English

Version to the bilingual subjects in English. Both

examiners were thoroughly trained in the administration

of the BMM and in response recording.

Each child was seen individually by one of the

examiners. While one child received the BMM-S in a small

testing room at his school, another child was receiving

the BMM-E in a different testing area. In this way, half

the children at each grade level received the English

version before the Spanish version, The other 50% of the

subjects initially received the Spanish version. The

semantically non-meaningful subtest of each version

was always administered before the real word subtest.

The examiners always conversed with the children

prior to test administration to establish rapport. The

Cuban examiner spoke only Spanish to the children, the

American examiner spoke only English. Johnson (1974)

reports that in studies of children of two ethnic

identities, only one test administrator may well be a

liability. Before administering the BMM certain

instructions were given to the children. The examiners

told the children (in the respective languages), "I'm going

to show you some pictures of things you have never seen

before. Listen very carefully because I will ask you

some questions about them." Each child was given two

practice items (noun plurals) to ensure knowledge of the

task before beginning the actual test. If a child missed








any of the practice items, the correct answer was taught

and the question repeated until the child responded

correctly. Generally the children caught on quickly and

understood what was required of them.

Productive knowledge of the morphological rules was

assessed by requiring the children to modify nonsense or

lexicon words that cued a picture. The children had to

complete a statement with certain contextual clues spoken

by the examiner while looking at a related drawing. For

example, a subject was shown a drawing of a cartoon-like

man performing a juggling type activity. The examiner

would then say, "This man knows how to zib. What is he

doing? He is __ ." If the subject responded

"zibbing," evidence existed that he had productive

ability to apply a generative rule for the progressive

inflection. The corresponding item on the meaningful

subtest included a drawing of a real man painting a fence.

The examiner's statement to the child was, "This man

knows how to paint. What is he doing? He is _."

A more complete description of test administration

procedures is in Appendix A.

Six scores were obtained from the Bilingual Morphology

Measure: a score on the semantically non-meaningful

subtest of the BMM-E based on 21 items, a score on the

real word subtest of the BMM-E based on 21 items, a total

score on the BMM-E based on all 42 items, a score on the

non-meaningful subtest of the.BMM-S based on 21 items, a








score on the real word subtest of the BMM-S based on

21 items, and a total score on the BMM-S based on all

42 items. Some of the items testing productive knowledge

of the inflectional rules had more than one possible

answer. In such cases, credit was given to those

responses for which a grammatically correct English or

Spanish model could be found. A description of the

specified types of responses considered to be quantitative-

ly appropriate and inappropriate for both the BMM-E and

BMM-S is provided in Appendix B.

Data Analysis

One-way analysis of variance was used to test for

significant differences between grade levels for all

BMM-E and BMM-S scores. In addition, linear trend

analysis was performed on the mean scores to further

investigate significant differences across grade levels.

A priori contrasts between separate grade levels were

also performed using the t statistic.

A t test was used to test for significant differences

within grade levels for BMM-E scores vs. BMM-S scores as

well as for BMM-E nonsense vs. BMM-E meaningful and BMM-S

nonsense vs. BMM-S meaningful mean scores.

Kuder-Richardson 21 reliability estimates were

computed for all grade levels combined for all scores

(BMM-E nonsense, meaningful and total; BMM-S nonsense,

meaningful and total).







Group means, medians, standard deviations and

percentile rank equivalents for the raw scores of each

grade level were computed for all scores. Percentages

of correct responses of each item tested were also

computed for each grade level.

A t test was used to test for significant differences

between the Spanish dominant and bilingual kindergarten

children's scores on the Bilingual Morphology

Measure-Spanish Version.















CHAPTER III

RESULTS AND DISCUSSION

Introduction

The results of this investigation are reported in

three sections. The first section presents the results of

this investigator's efforts to develop a diagnostic

instrument (Bilingual Morphology Measure) to objectively

assess certain language skills of Spanish/English bilingual

children. These will be the results related to specific

goal #1 described in the Statement of the Problem section

of this paper: to develop a standardized instrument that

may be used in the evaluation of the linguistic development

of the Spanish and English spoken by bilingual Cuban-American

children. The second section presents the data pertaining

to morphological rule acquisition by these subjects. The

results discussed relate to specific goal #2: to analyze

the developmental sequences of morphological rule

acquisition in the two languages of these children and

goal #3: to determine if there are significant differences

between morphological responses to tests using nonsense

and meaningful words in Spanish and English. The third

section presents the results relevant to bilingualism and

its effects on language development (specific goals #4,5,

6,7 on page 35 in Chapter I.)







Results and Discussion

Bilingual Morphology Measure (BMM)

Nature and rationale. The Bilingual Morphology

Measure provides one of the few sets of language evalua-

tion materials and norms that have been developed directly

from the language of the Cuban population within the

United States. The BMM was designed to identify through

individual testing those bilingual children who do not

demonstrate morphological rule proficiency commensurate

with their grade level. More specifically, the BMM was

designed to determine the various morphological rules

the bilingual child has and does not have in Spanish and

English. It will help the teacher of the Spanish-speaking

child in American schools determine whether a child's

poor verbal performance in English might be due to

language differences (this would be indicated if the

child scored within one standard deviation of the mean

for his grade level on the BMM-S but more than one

standard deviation below the corresponding mean on the

BMM-E) or language deficiencies (indicated by a score

of more than one SD below the mean on both the BMM-S and

BMM-E).

If a child exhibited difficulties in English,

beginning to teach him those rules in English that he

shows a deficit in or lacks completely in Spanish may be

frustrating for both teacher and child. The wiser procedure

would be to administer both versions of the BMM and to





50

first proceed by teaching the child the morphological

rules that are intact in Spanish but lacking in English.

If the BMM-S determines that the child is performing below

grade level norms in Spanish, the appropriate school per-

sonnel should consider providing language remediation in

the native language first.

Group differences. Analysis of variance revealed

that for both the nonsense and meaningful subtests of both

the BMM-E and BMM-S scores increased significantly (.001

level of confidence) across grade levels from kindergarten

through third grade, indicating that the BMM was measuring

a developmental.phenomenon. Table 5 describes the results

of this statistical analysis including the linear trend

analysis performed.

Reliability. Thirty-eight subjects, 19 Spanish

dominant children and 19 of the bilingual subjects,were

tested twice to obtain measure of test-retest reliability.

The Spanish dominant children were retested with the BMM-S

exactly one week after receiving the test the first time.

The retest was administered by the same bilingual Cuban

examiner. The same American examiner administered a

retest of the BMM-E one week after the first administration

to 19 of the bilingual subjects. The scores of the

bilingual children receiving the retest of the English

version of the BMM were compared with the original test

scores of the same 19 children.












0 0 00
0 0 00
0 0 0 0 -4
0 0 00 0 0


0

0







( 0
CO









4H



002

001
40


CO











001 0
r-4














0 o
oC



(0


14-l






















rin
01 0)
0C










C 1





I Mr
44 -






C +


0 0
0 0
0o
o o


in tp
0)01


CO0
C C;


oo
00
00
0o
o o


mm mm n mn


0) 4- )



0 w 0
C C


o 0
o 0

\o in
(M \D


4-4



r0


0
P


_ _


i I







Test-retest reliability Was computed using Pearson

Correlation Coefficients. All test-retest correlation

coefficients indicated high test-retest reliability

(p <.0001). The results are presented in Table 6. The

means and standard deviations for each subtest at the two

testing are quite consistent as shown in Table 7.

Kuder Richardson-21 reliability coefficients were

also computed. This is used as an estimate of test-retest

reliability. Coefficients ranged from .66 to .90. This

is also an indicator of high internal consistency (the

degree to which all the items measure the same thing).

KR coefficients are reported in Table 8.

Validity. The validity of the Bilingual Morphology

Measure has been demonstrated in two ways. First, since

the instrument is a test of some early stages of expressive

language ability, it should follow the same trends that

language follows in children. That is, it should be

developmental; increase in age should be paralleled by

increasing score. Validity was indicated by the significant

differences produced between grade levels by the BMM scores

and the presence of significant linear trends. As age

increased, the scores which purported to measure increased

morphological proficiency increased significantly,

indicating that morphological development was, in fact

being measured.

Second, there is preliminary evidence to show that

the BMM distinguishes between children who have delayed












Pearson Correlation Coefficients and P-values
for children given test and retest.


BMM-S Nonsense
BMM-S Meaningful
BMM-S Total
BMM-E Nonsense
BMM-E Meaningful
BMM-E Total


Table 7


Correlation
0.9205
0.8837
0.9256
0.9114
0.9688
0.9419


p-value
..0001
.0001
.0001
.0001
.0001
.0001


Means and standard deviations for children
given test and retest.


Test


BMM-S Nonsense
BMM-S Meaningful
BMM-S Total
BMM-E Nonsense
BMM-E Meaningful
BMM-E Total


x
6.94
12.89
19.83
7.47
11.89
19.32


Retest


SD
3.43
3.04
6.06
2.82
2.67
4.87


x
7.53
13.05
20.58
8.68
12.10
20.78


SD
2.68
2.68
5.02
2.57
2.31
4.30


Table 6


_ _















Kuder Richardson Reliability Coefficients


TEST

BMM-S Nonsense

BMM-S Meaningful

BMM-S Total

BMM-E Nonsense

BMM-E Meaningful

BMM-E Total


COEFFICIENT

.70

.66

.81

.83

.81


.90


Table 8








language and those who do not. This investigator

administered the BMM-S to nine kindergarten

children whose scores ranged from 10-16 and

8-13 on the comprehension and expression subtests,

respectively, of the James Language Dominance

Test. These scores are below grade level

norms for kindergarten on the JLDT. In addition,

these children's teachers had initially identified them

as being generally slow to comprehend instructions

and directions for kindergarten activities.

The mean scores on the BMM-S of these children

are contrasted with those of the Spanish

dominant and normal bilingual kindergarten

children used in the standardization of the BMM.

(See Table 9.)

Norms. Grade level medians, means, and

standard deviations are presented in Table 10.

Table 11 gives percentile rank equivalents for

the total raw scores of the BMM-E by grade

level. Table 12 gives the corresponding per-

centile rank equivalents for the BMM-S.









Table 9


Means of language delayed, normal Spanish
dominant, and normal bilingual kindergarten
children tested with the BMM-S.


SUBJECTS BMM-S Nonsense x BMM-S Meaningful x

Delayed Language 5.22 8.00

Spanish Dominant 7.47 11.89

Bilingual 11.26 14.77






- e H io r- t- 57
'o Do 0 D1 > cn o un


4C ,-lM o u m C ,
C')

n M m o
c Cn r co n n o0 mn
Q) 0 *
L4 H-I ) r 0O o 0c o m
1o C'I C C- (m C'4 C( CN m )
O1 0


Ec -
r-H 4 H 0Lf 0 ul 0 0ia

0 N C C

SI-













0 -- t O *
OH -- -- ------------. ---------
1I m f- o o
-3 4 O OI r4 r4 P -4 r-l -1


m m .w 0 m 0 N

r < a 4 >..


SO)
ff 1 1 ( 1
. ri r-4 H ,- 4 8 ,-i H ,-4 ,-i
0 U


C ~0 O o o o










0
C
0 r4 l rl CM l rOl >l rO



















C-4 O D 0 0 0
S Z H( o N M H w
-4I r-I b I H rI
t0 (14 to U) 0 c M c (N m nI 4
H H
tom o

(a 4J 4- 'O i&aO C'4 rOH
10 Q .





40 10 0 $
SC 0000 0000C f c
10 (D O f- O rl l r-I N H.-4HH






0 C ro-l r
HO 1-I HOOH



H tooooo





0 10 .H 't 4Jto3o
rl 0 UjM






58


(N r-i o Q^ No '~ ifl (N i- O (?' co un
qT ^r xT Cfl ro g' ro (m (N cMj C'4







oo o-i Q'LA> in O CHO n LA (N
(r' C oL o to (N (N H H


O L COO w w NO n oO n 1V O O m. N
o O' <> N0o r 0 a c H


O o o CO c r- oCO NMn rO L NH O' CO N %o
T% m 1 0 co en c rN (% (N N (N N N m ,H H H Hr, ,







C MO o' NC aO A (N LA (N O iN N O LM A (N iN N Ni


OO N, ,O C' NH. O0 0' N 0 L O N
C (N N (N vN (N N H H- ,- I Hw -








0 O w0 I m men (NH


a)I

L4


0

S<
OC
IN
W H
QM
K "
H C













8

1/M 1 o CT> O vo 1 V n cm N 0 w %0 in M
m en m m m n m m mmn c N cN
< 4


r a
H 4
03: '. ". O O iLA m m A -4 co H -

I


ko LO IT m Nr r-I O m w mr CN



m LAO N emmm N enN (N N en N


0w m0 0 m 0 w m N (N wm w
0 3 0 w r- tom gr m "r- H


s-I

U-


H
M TO


Ln NOO % 0 NO
m mn n n N N N N N N






0'- 0'c Co i N~ '.D ens- ?o o 4


0 o C o N a Mn i-i o
n n (N(N (N N (N (N (N (N (N (N





Co r tn L t o 4 .O% r-i Ln co LA
0CT 0 o w m n m CN rN -I








Interpretation. Children who obtain scores below the

first standard deviation from the mean of the BMM-S Total

score or BMM-E Total score are probably in need of remedial

language teaching. However, it must be emphasized that

using a single test to judge language performance is not

a sound practice, and information from other tests, case

histories and observation should be considered in

recommending language remediation.

BMM test items seem to be valid not only in terms of

their contribution to the total test score, but also in

terms of their individual contribution to the understanding

of the child's language. Several items are sufficiently

easy for all the children in each of the age groups to

pass. Most of the items show increasing percentages of

children passing with increasing grade level. Items were

not selected or retained on the basis of their correlation

to the total test score or their ability to discriminate

between grade levels. Items were selected and retained

because they evaluated the children's knowledge of

specific morphological structures.

Morphological Rule Acquisition

To analyze the developmental sequences of morphological

rule acquisition of all subjects to both the nonsense and

meaningful stimuli presented in both Spanish and English,

percentages of correct responses were calculated in-

dependently for each grade level for each inflection tested

as well as for the allomorphic subdivisions within each







major morphological category of nouns, verbs, possessives,

adjective, etc. Table 13 summarizes the percentages of

correct responses for lexicon and nonsense words obtained

for all bilingual subjects on the BMM-English Version.

Table 14 summarizes the corresponding percentages for the

subjects on the BMM-Spanish Version.

For ease of discussion, the items are grouped-into

three frequency levels of responding: 1) Items correctly

responded to by 75% to 100% of the subjects at any grade

level are considered to be mastered by the children; 2)

items having a response frequency of 50% to 74% are

considered to be in a state of learning by the children;

and 3) items having a response frequency of below 50%

indicate those inflections that the children have very

little control of or have just not acquired.

Plurals-English. From. the data given in Table 13,

it appears that all the subjects possess the allomorph

/-s/ and can generalize it to unfamiliar forms. The

allomorph /-z/ is also mastered at the kindergarten level

in lexicon words but its application to unfamiliar forms

shows a decrease to 61%. By first grade /-z/ is mastered

with unfamiliar forms. Incorrect responses consistently

included a repetition of the nonsense or lexicon noun

in its singular form. Three children over generalized the

rule and responded [kizaz] for "keys." The allomorph /-az/

for noun plural formations on meaningful lexicon words is

not controlled by even the third graders but does appear

























asuasuo o



M TrnyuTueaw 0


ooo00 4 0o 1 V- r *
o oo m n vr in o o r- r-
r4-H4


4)
+H
14

04
0




-4
a

cu
O0













0

*4
o








a
i-I

*H
u
*O











a)
,-4

S*O
un













H r




S0







,-4
a0

o







R


C O01
-r-o
C'lo%0


0 OZ 'n 0
H0~O


*W 00
00 00
a o-o


Nv r Vr- 0% -W pi
0 0%O -I OC r-


00 0
00 0
V-4 0-4 ,H-


o rI-ro- e)frn t- 0oo C mo r- 0 or 0 o r- Cm o M



goom o mmoo p o on
CD- 0 0r- 00.C O r0cM M 0 0 ON C% 00C- cohm
M ooor- m a r. o r- m C o fn o0 en 0 a n oq r

S- r-p fi- 4mo m 0- o r- o o o o oo o 0 rI
asusuoN v-Io

I r- r- y e cn 0 0 cnm 0 fn 0 00 ca 0 r. (n
m C a o7% a r- on D r o oV M o ON m as oo co a O
M .f,,._ or~mo r w~ < 0 fi r oc oo eo ^ o


ssuasuogN Co



r U
Exqu~j~re~ a-s


0% CA co
r < M
as as a


040 O i



mn o> m 4
Nr N es N


%D 0

~o0
H


o0
in


0 COW
o me
t.' Uo


0'
I) 10 W W ) -ri- 4J 4)

asuasuc 0 03 3 WB -N4 0 O 40 a 0 0
oa ato w.=w4 C-4 o a 0 *4) o
E >m t 4J0%0 $ 202 : g N m


102 0 *o
u a m m C 0 .0 E' 0) P
(tr9. I G, oum inW C B u -4 A 0 C U
4U 0 BO >-r (o0 .a 4 *4 W *) (M C
3 0 V-40 0' 130 la, U E 4 4
a VHM 03 c a 00u E Qa 4) V


I S
N N


H W
i I


O 00 O



Or 4
1o o r-
p-I


k a
0) S
- o M M )
aN 4 $4 .0
0.N ** *r< *Q
t V4 -'.4




.J4
- 0) ON 0 r.
0 6)6)4)



14 W 0










>
4 >
41 "4


fa
a -
- V-
UN 0$4


N
N N
I I


00C
00 C
HHH- l


m w



en N









000 M %D kD 'O
0 0 0 ano Co mo
asuasuoN H o v H vi

1ooo 0-0 r-r- W



000 om 4-



cooo Wr-onn
000 HN




SoosuasuoN r- o r-




0 0 o'c moooo
0 c 00 HN N
Sr su asuom 00 0 C
S0 0 n r- r-iC


N- R q q v' ,r-4 rq- ,- r"r- 'q r c co 0 CO ONo 0a o0 r- q r


wo000
O'0000
HHH


r- 0 0o 00 0 -i 0
m 0 00 00 0 rN 0
-4 r1- r-4 -4 rH


r- r- o r. r- r- m o
N N 0 N N N -0
%0NNNNcl cOQ0
Hl


m 0 rO 0 0
0H c o
r-q r-4 r-4


r- m r- m o


r- 00 00
m o0 o0
HHr HHr


r rI r- m r- 0 C
r- r- r- co r- in oo
,--H ,-.H


M o r- o
Ho H


c.) N 00
o r- oo
,- ,oo


oo r- r-



N. Noo00


0)
41
-4
$4
04
0








O
4.

*tr
>1








0-4


CU
#P
t-










*H
.0 d
O
r-

0

40


H 1-
tO 0
.rQ0




0)U
U0
.(0

c c
Ird 0






() O





PL|(X


r- r- o
o0 oa co o


esutasuom ) 0) ) a) C \ o\
d cm r- M u L z) 0o -H 0
4 0 ) r 0 )-r O''-4 -) 4 -rl
$4 > U > 4Mo ..Q r O H-
Q) -ri m a*H) .-r : (e 0 0
r. 4M Ea+ie O A -i E


0


(1) 0
*rl

m \0 0
{N~-u .u m O in n m o o m
O U$4 ) r-i (-0 0 \ -

o > E V-H C \r4 4 o 0
r-4 k fu -H >4 o0a o u 9:
t 0) H-I (0) 3 r (0 a -Oi
40

A
U)




,1
0)





m

P4


o0 0 00
HHr


$4

9 -

*r)
H


m \o tn
C-i LAr


, fr- r-0 -o -
m 0 H CM r~



ao 0
4 4J 0 0

a) r-4 o 0 -
C u U

OH 0 0
H r- 0 .P
i 0 Lt 0 .0
a) a) ( 0) r.-4
So E Er 4-4


a 0
(00

1C 0 *0 0
*U M Ca
CM

0 l H H
O rie rl r


0)



' 5
0d m



M< $4 *H
m tn a


4J
0


0





0

--


H <
cc 4
tO >
r-l *r
r( +J(
CUQ
3 t
10


r-i P r- r- N r- r- P-
kD r- r r r aen m O co<





64

to be in a state of learning, Incorrect responses at all

levels consisted of repetitions of the nonsense and meaningful

nouns in their singular form. By second grade, more of the

incorrect responses consisted of using the /-s/ allomorph

instead of /-az/. (Many children stated [gAtjs] instead of

"gutches" and [watSs] instead of "watches.") This indicates

a basic awareness on part of the children that something must

be done to the singular noun to make it plural; they added

the most familiar inflection they knew--/-s/ which is also

the first one acquired in the sequence.

Plurals-Spanish. All subjects possess the allomorph

/-s/ and are able to apply it to both nonsense and meaningful

forms at the kindergarten level. The /-es/ allomorph is not

mastered to any degree until the third grade. By third grade,

the children control the rule that /-es/ is added to a noun

ending in a consonant sound. They have not acquired the rule

that /-es/ is also added to singular nouns ending in the

letter "y" or in a stressed vowel sound. The ability to

apply the /-es/ allomorph to any unfamiliar forms is not

acquired by even the third graders. Incorrect responses

at all grade levels include a repetition of the Spanish

nonsense or lexicon noun in its singular form or the

supplying of /-s/ instead of /-es/ to the singular noun.

Several children at each grade level supplied the English

allomorph /-z/ in the attempt to make the Spanish nouns

plural. As /-z/ is not an allomorph in Spanish, this would

indicate some interference from English.





65

Past tense-English. Only the third graders have mas-

tered all the past tense allomorphs /-t/, /-d/, and /-ad/

with lexicon and nonsense verbs. Kindergarten children have

not acquired control over any of the allomorphs even with

familiar forms. In first grade /-t/ and /-d/ are in a state

of learning and are mastered in meaningful lexicon verbs by

the second grade. The allomorph /-ad/ for preterite verb

formations is not applied to either meaningful or

unfamiliar forms until the third grade. Incorrect responses

at all grade levels for all allomorphs examined include a

repetition of the verb infinitive with no attempt at conjugation

or the supplying of the progressive allomorph rickingng" for

rickedd") or the 3rd person present tense allomorphs [s] or

[z] ("ricks" for rickedd" or "bomes" for "bomed"). Only a few

children overgeneralized the rule and stated [pIktad] for

"picked," [klaImdad] for "climbed" or [bomad] for "bomed."

Past tense-Spanish. Kindergarten subjects have mastered

all the past tense Spanish allomorphs and are able to apply

them to meaningful forms. At the kindergarten level the

children are still learning to generalize the rules to un-

familiar forms. By first grade, the children have mastered

the ability to apply the past tense allomorphs to nonsense

forms as well. Incorrect responses include repetitions of the

verb infinitive with no attempt at conjugation or the supplying

of the Spanish progressive allomorphs /-ando/ or /-iendo/ to

the verb stem.

-Present progressive. All subjects have mastered the

English and Spanish present progressive allomorphs and







are able to apply them to both nonsense and meaningful

forms at the kindergarten level.

Present third person singular-English. The

allomorph /-s/ for the third person singular (kleats/eats)

was mastered by the kindergarten children for the meaning-

ful item and was in a state of learning for the nonsense

item. By second grade, mastery of this allomorph was

demonstrated in the unfamiliar form. The allomorph /-az/

(fonces/dances) poses more problems. In second grade, it

is in a state of learning and is not mastered until

third grade. The nonsense form begins to be in a state of

learning in third grade. Errors include repetitions of the

verb infinitive ("dance" for "dances" and "kleat" for

"kleats") or the stating of the progressive ("eating" for

"eats"). One first grade child overgeneralized the rule

and said [klitsaz] for [klits].

Present third person singular-Spanish. The Spanish

present tense allomorph /-a/ used with /ar/ verbs is mastered

with both meaningful and unfamiliar forms in kindergarten.

The /-e/ allomorph used with /er/ and /ir/ verbs is mastered

with meaningful forms in kindergarten but the children

have not mastered the ability to apply this allomorph to

unfamiliar forms until the third grade. In addition to

repeating the infinitive nonsense verb "dober" or responding

with the progressive form "dobiendo," incorrect responses

included the supplying of /-a/, the /ar/ verb's third person

singular allomorph.







Possessives-English. The scores of each of the plural

possessives for both meaningful and nonsense forms was

higher than for any of the singular possessives. Newfield

and Schlanger (1968) feel that the plural possessive does

not actually assess the use of any morphological ability.

In their study, for example,40% of the retarded subjects

correctly used the singular form "wug's," 77% used the

plural form "wugs'." The same pattern emerged with the

bilingual children in the present investigation: 58%

of the youngest children correctly used the singular

form "wug's," 90% used the plural form. Berko (1958)

noted that the wrong answers to the singular possessive

items consisted of doing nothing to the word given. If

the children just repeated the plural.possessive item

as stated by the examiner, they would be correct as the

plural possessive is already inflected. It appears,

therefore, that the plural possessives are not reliable

measures of morphological ability.

Kindergarten children are still learning to apply the

singular possessive allomorph /-z/ to both meaningful and

nonsense words. By first grade they have mastered the use

of the allomorph /-z/ in meaningful words and by second

grade the children control it in nonsense forms. The

allomorph /-az/ has not been acquired by any of the children.

Only 32% of the third graders were able to respond [nIzaz]

("niz's"). Incorrect responses consisted of not doing

anything to the word given. ("It is the niz hat.")







Possessives-Spanish. Both the singular and plural,

masculine and feminine possessives are acquired early.

Kindergarten children demonstrated mastery over the

ability to use all possessives with both meaningful and

unfamiliar forms.

Derived agentive. The bilingual kindergarten children

have mastered both the English and Spanish agentive allomorphs

/-er/ and /-or/ with the meaningful items "painter"/"pintor."

It isn't until third,grade that the children generalize

the use of these allomorphs to both English and Spanish

nonsense forms. Incorrect responses to the derived

agentive nonsense items generally consisted of a shrug of

the shoulders or an "I don't know"/"No se" response.,

Adjectives-English. The kindergarten children have

mastered the superlative of the adjective allomorph /-est/

with the familiar form "biggest." It wasn't until the

third grade that they demonstrated control of /-est/ with

the nonsense form "quirkiest." The comparison of the

adjective allomorph /-er/ was not as easy for the children.

The meaningful form "bigger" was mastered by the first

graders, however; not even the third grade children were

able to apply this allomorph to the nonsense form. Only

19% of the third graders correctly responded quirkierr"

to the examiner's stimulus. Incorrect responses at

each grade level included "quirker," a repetition of the

word "quirky," no response and "more quirky." The "more

quirky" response may indicate some interference from

Spanish.







Adjectives-Spanish, The Spanish rule for forming the

superlative of the adjective is not mastered even by the

third graders, but does appear to be in a state of learning

at this level for the meaningful form. The nonsense form

is poorly handled even by these older children. Many of

the children responded to this item by repeating the

comparative form "mas mecoso." In several instances there

was a definite vocal emphasis placed on the word "mas."

This indicates that although the children couldn't think

of any of the possible grammatical ways the superlative

could be formed in Spanish, they were aware that the

semantic implication for the superlative is different

from the comparative. The comparison of the adjective rule

was easier for the children. This was mastered by first

grade for the meaningful item; the ability to apply this

rule to an unfamiliar form is not acquired until third

grade. Incorrect responses generally consisted of a

repetition of the examiner's uninflected adjective or no

response.

Discussion of results pertaining to the-developmental

sequences of morphological rule acquisition in Spanish and

English. Examination of Tables 13 and 14 comparing the

development of morphological constructions and the general-

ization of rules to Spanish and English forms by the

bilinguals indicates a parallelism in the acquisition of

shared structures. It appears that those morphological

rules which are similar in both Spanish and English are








acquired in approximately the same order and at the same

rate. The rules which are unique to each language are

acquired at different times; in many instances the Spanish

rules are acquired first, This most likely reflects the

linguistic complexity of the specific structures involved.

For example, the iules for forming the present progressive,

agentive, and noun plurals are similar in both Spanish and

English. Examination of the data reveals that the present

progressive allomorphs are among the first mastered in

Spanish and English. The nonsyllabic plural allomorphs

(/-s/ in Spanish and /-s/ and /-z/ in English) are also

among the first mastered in both languages. The syllabic

allomorphic form of the plural (/-es/ in Spanish and

/-az/ in English) are among the last to be acquired by

the children in both languages. The agentive allomorphs

/-or/ and /-er/ in Spanish and English,respectively, are

acquired at approximately the same time.

On the other hand, the differences in acquisition

order reflect language contrasts. For example, the

Spanish past tense allomorphs are mastered by the kinder-

garten children. The English past tense allomorphs /-t/

and /-d/ are not mastered until second grade and /-ad/ is

not controlled until the children are in the third grade.

The morphological rules for forming past tense in Spanish

are quite different from English (see Appendix B) and

appear to be less complex linguistically.








Kessler (1972) stated that "contrasts in the

simultaneous acquisition of two languages in bilingual

children support the theory that languages share

deep-structures and that differences derive from language

specific rules" (p.221). She initially hypothesized that

"morphological structures shared by Italian and English

develop in the same order and at the same rate in the bi-

lingual child since these shared structures find their

source in a common underlying base and are realized by

the same set of transformational rules" (p.229). Results of

research conducted by Kessler (1972) revealed that the

acquisition of inflectional categories follows approximately

the same sequencing in Italian and English with an

exception made for those structures requiring language-specific

case consideration. She stated, "The Italian system is

generally acquired somewhat later, possibly because of the

morphophonemics of the language" (p.229). In the present investi-

gation, it appears that the English system is generally

acquired somewhat later, probably due to the morphophonemics

of the language as well as the effects of bilingualism on

second language development in the Cuban/American population.

Discussion of results pertaining to the bilingual's

developmental sequence of morphological rule acquisition

in English compared with existing data on monolingual English

speakers' acquisition of morphological rules. The acquisition

order of the various allomorphs is basically similar to the

order reported by such investigators as Anisfield and Tucker




72
(1967), Berko (1958), and Newfield and Schlanger (1968) for

monolingual English speaking children. For example both the

bilinguals and monolinguals master the /-In/ allomorph very

early as well as the allomorph /-z/ which occurs correctly

first with noun plurals and second with the possessive

singular. Words adding syllables such as allomorphs /-az/

for plurals and present tense third person singular and /-ad/

for past tense are mastered much later in development by the bilinguals

as well as the monolinguals. In general the results seem to indicate

a similar sequence of acquisition for both language groups with the

bilinguals exhibiting a slower learning Dace. This finding is

supported by the information reported in Table 21 (page 84). These

results indicate that by approximately 9-years of age the bilingual

children's performance is on par with monolingual 8-year-olds.

Differences between lexicon and nonsense word items.

A t test for correlated samples was used to determine if

there were significant differences between the lexicon

and nonsense word subtest scores for the BMM-E and the

BMM-S. Table 15 contains the results for the English

Version, Table 16 for the Spanish Version. For both the

BMM-E and BMM-S, the children performed significantly

better (p<.0001) on lexicon words at all grade levels.

Discussion of results pertaining to the differences

between lexicon and nonsense word items. These results

support the conclusion arrived at by Newfield and Schlanger

(1968) that there is an undefined time lapse between the

production of correct morphological inflection forms within

a context of familiar words and the generalizations of

these forms to unfamiliar words. Several investigators








Table 15


Summary of the t values testing the differences
between means of the total raw scores of the
lexicon and nonsense word subtests of the BMM-E.


Grade t value df p


Kindergarten -12.02 30 <.0001

First -11.96 29 <.0001

Second 415.64 29 <.0001

Third -8.81 30 <.0001






Table 16 Summary of the t values testing the differences
between means of the total raw scores of the
lexicon and nonsense word subtests of the BMM-S.


Grade t value df p


Kindergarten -12.16 30 4.0001

First -10.51 29 <.0001

Second -13.33 29 <.0001

Third -9.66 30 <.0001







such as Dever and Gardner (1970), Johnson (1974) and

Pettit and Gillespie (1975) contend that this indicates

that nonsense tests of morphology may not be valid measures

of children's ability in this area, that there are unknown

confusion factors inherent in the use of nonsense syllables,

and/or the use of nonsense words in sentence completion

tests underestimates a child's productive knowledge of

morphological rules.

Another more likely explanation pertains to the children's

competency level in understanding and applying the morpho-

logical rules to new word forms. The difference in scores

between the nonsense and meaningful subtests suggests that

the acquisition of these rules may not be complete in

children of this age. If the children's ability to control

the morphological forms is better in familiar words, as

demonstrated, then these results suggest that a child's

production of a morphological rule in a rehearsed response

need not necessarily require complete comprehension of that

rule (Larson, Summers and Jacquot, 1976). Larson et al.

further states "Conditioning of these rules may result first

in correct production using automatic responses followed

in later stages by developing comprehension" (p.266). The

present investigation which included this comparison study

between spontaneous automatic responses using common words

and spontaneous responses using nonsense forms on matched

grammatical items seems to clarify this conclusion.







Larson, Summers, and Jacquot's (1976) contentions

as well as Berko's (1958) original theory that a child

supplying correct morphological endings to common words may

only indicate that he has merely memorized the forms and does

not actually have the working system of morphological

forms internalized are further supported by the following

observations. In the present investigation, several second and

third grade children immediately realized their errors and

appropriately self corrected their responses after incorrectly

responding to one or more of the lexicon items. After

testing was completed and the examiner asked, for example,

"Why did you change your answer to 'keys' after saying

'keyses'" or "why did you say 'more bigger' and then say.

'no that's wrong, I meant biggest,'" the children would

say something to the effect "My first answer didn't sound

right after I said it." This observation,that several

of the subjects spontaneously corrected their responses

to the meaningful stimuli whereas none of the subjects

self corrected their responses to the nonsense stimuli

tends to confirm the "memorized rule" theories.

Effects of Bilingualism on Language Development

One of the goals of this investigation was to compare

a sample of Spanish dominant children with bilingual child-

ren on their ability to apply Spanish morphological rules

in the hopes that information yielded by such a comparison

would help determine what effect bilingualism has on native

language development. Results indicate that the bilingual







kindergarten children performed significantly better

(p(.0001) on the Bilingual Morphology Measure-Spanish

Version than the Spanish dominant children of the same

age. Table 17 summarizes the results of the t test used

to determine if there were significant differences between

the Spanish dominant and bilingual kindergarten children's

scores on the BMM-S.



Table 17 Summary of the t values testing the differences
between mean scores of Spanish dominant and
bilingual children on the BMM-S.



TEST t VALUE df p


Nonsense 5.26 48 .0001

Meaningful 4.54 48 .0001

Total Combined 5.39 48 .0001


Incorrect responses by the Spanish dominant children

were similar to those made by the bilingual children. There

were, however, many more instances of no response, shrugs

of the shoulders and "No se" (I don't know) responses by

the Spanish dominant children. In addition, none of the

Spanish dominant children incorrectly supplied the English

allomorph /-z/ in an attempt to make the Spanish singular








nouns plural as did several of the bilingual children.

There was no indication of any interference from English

in any of the answers given by the Spanish dominant

children.

Discussion. An explanation for the bilinguals'

better ability to apply morphological rules in Spanish

might be that bilingualism encourages a more complete

and rapid development of certain linguistic forms than

does monolingualism. It is likely that the bilingual

children have had experience with inflectional rules

in two languages; this repetition enhances their ability

to use these rules in their native language. Ben-Zeev

(1977) brings out that generalization of particular

rules does not depend on mastery of particular words or

symbols but on a grasp of the basic idea that the structure

of language is subject to change. The experience that

bilinguals have in learning two different language structures

with the repetitiveness and redundancy in grammar may foster

this type of awareness. The bilinguals, therefore, are

better able to analyze language as an abstract system. This

seems to be indicated by the present investigation as far

as primary language development is concerned.

Effects of bilingualism on the children's non-native

language development can be seen by comparing aspects of

the English spoken by the bilinguals with comparable

aspects of English spoken' by monolingual American children.

Berko (1958) obtained percentages of combined preschool and







first grade children on their ability to correctly produce

a variety of morphological forms. For purposes of

comparison with Berko's data, mean percentages of the

combined bilingual kindergarten and first grade children's

correct responses to corresponding items were computed.

It should be noted that the Berko children ranged from

four to seven years of age as compared to the combined

kindergarten and first grade bilingual group of this

study who were approximately a year older at both the

low and high age range. Table 18 provides a comparison of

percentages reported by Berko for combined preschool and

first grade children with those obtained for the combined

kindergarten and first grade bilingual children used in

the present investigation on equivalent BMM-E items.

Newfield and Schlanger (1968) report percentages for

a sample of normal monolingual English speaking kindergarten,

first and second grade children combined, on the similar

task. For purposes of comparison with Newfield and

Schlanger's data, mean percentages of the combined bilingual

kindergarten through second grade children's correct

responses were computed. The children used in Newfield and

Schlanger's study ranged from 5 years, 8 months to 8 years,

4 months. The ages of the combined kindergarten, 1st, and

2nd grade bilingual children ranged from 5 years, 4 months

to 8 years, 11 months. Table 19 provides a comparison of

percentages reported by Newfield and Schlanger for combined

kindergarten, 1st and 2nd grade normal monolingual English









Table 18


Comparison of percentages reported by Berko for
combined preschool and first-grade children with
those obtained for combined kindergarten and
first-grade bilingual children on equivalent BMM-E
items.


CORRECT RESPONSES


MEAN PERCENTAGE CORRECT
BMM-E BERKO


/-s/ mooks*
/-z/ luns
wugs
cras
/-az/ gutches
tasses
nizzes
glasses


/-t/


ricked
showed
bomed**


/-ad/ motted
melted


/-I /
/-az/


Possessives


zibbing
fonces***
wug's


/-az/ niz's
/-/ wugs'
nizzes'


Adjectives


Agentive


/-ar/
/-est/
/-ar/


quirkier
quirkiest
zibber


*heafs was Berko's equivalent
bingeded was Berko's equivalent
***nazzes was Berko's equivalent


CLASS


Plurals


Verbs


S 2









r-.4a r-lf 1 or ON m< 0 0 in
p[aTjMaN ON\ cy o'OY m m o.

c" cor- %D>o c P-4 co e
a-mWa oc C co Nr- ri-rl r- m CN -l


m r- O r-
ON o aO mr


-1 r--I %O
00 i- O


O1 -i r
ito
'8


rsl
S1 rn

















O'0
-
B'





5 B
io c- *

g '0 C71i















aQ'
6S'


PIGT~'MYN


01 om C0\ o0 co co


r- O r- o 0 -i 'o
a-0wwa oC o01'o Ln 0\


m03:















0 30
0 0 0
'oD 4


0 0 CN 0
0 0 0 0
1-1 .-I

'0 Lf C 0
rlr( r l


r- 0
CO4
rfl


U)
OU
Sra
r. Ol


03
.0
5.


0 N 0
4J N0
3 -r4 (
C 4c


(0
44 N
(0 3


0) 0
roa


r- rI
93

z z
2Z
* *l
*ua








speaking children with those obtained for the combined

kindergarten, 1st and 2nd grade bilingual children on

corresponding BMM-E items.

Pettit and Gillespie (1975) report percentages for

6, 7, and 8-year-old children on their ability to correctly

produce inflectional endings to ten nonsense and ten

meaningful words. The bilingual kindergarten children of

the present investigation (mean age = 6 years) were compared

with the monolingual 6-year-olds; the bilingual 1st graders

(mean age = 7 years 2 months) were compared with the

monolingual 7-year-olds; the bilingual 2nd graders (mean

age = 8 years, 3 months) were compared with the monolingual

8-year-olds of Pettit and Gillespie's study on twenty

corresponding items. Table 20 provides a comparison of

percentages reported by Pettit and Gillespie for normal

6, 7, and 8-year-old monolingual English speaking children

correctly producing a variety of morphological forms with

those percentages obtained for kindergarten, first, and

second grade bilingual children on equivalent BMM-E items.

Examination of Tables 18, 19, and 20 indicates that

the findings for the bilingual subjects were compatible

with the results found for monolinguals by all other

investigators for meaningful and nonsense nouns requiring

the short form plural allomorphs /-s/ or /-z/. The

bilinguals, as evident by the comparisons with monolinguals,

possess much less ability to apply the syllabic plural

allomorph /-az/ to both meaningful and unfamiliar forms.










C
0)

0 E


C 0)
0 4
44

0 aV
0 0

o o





00 C
-4 H 0 -H
U4-)













r o 0 w
0- 0


V0H0 U
4- )- 0 0-


au V














a Q) M .
'0Q) "







w o
0)U0

j 0 z






00-
0 0 (a









aH 0
P4 c 0)
c 0 0



000 0E U
0 r= (,-t


U) 0 -H



Wm rU












0 04 0
r- 0O (0











to 0 Hr o


0 ri atO
H o eb u


z a
H
HZ







Of
<0
Di-4

00



ZI















4













H
CQ4]










H n


-r- o N 000 NO o
OD NI CN 0I CN r- Oi n -i L








se be be be a t ae ab be be be













Hi %o 0 r-l M NVr kL O Ln
S-I CO Ln r- m


0
1c
, O












S0) 0 0 0
C o to ii





0 O 0 mU 0
Z 7 0 E
0 3400
Za -+


0e
0






ro


0
rI

4-1


00000
0~ HeA'0




0 0 0 0 0




t o *H 0 to

M *H e (U <
01 .0Q .Q -H -
3 -H -H J 3


o 0 0 0' O C ) 0
r o mo m r omm, o
a) 'r 0 ir-i r~, w
r-i


a-

O
0



0
o 0)
n -
0 Li
O U


Oe b
00
o o
c



to to
0-I
4 -0)




) to
Om
vO 0-
0) U
H-) 4
01to


ra
oo



OH
00







00
0-
0 m




+ -


0) to
-I-)-

0)0l


0000
0 C 0 O









0)




u o 4
P 4-4 a) a0
ft ft.Q .I


O O m O N O O wO ON
o r-



















o I O ONO O O O wO
be be ae be be be be be be be
000000 0 0 0









r 4 o i o r -4 P-4 ro i -4
'o o'm ai 0 co oo'









0 o0 00 0











be se be se be b9 be be be
r n LO IV %0 w A r- i-i O
O ON LO 0 NO O w
r-i


0000 0 0 0 0 0
o L O LO CO o O
,-, -- H--


0
0
H

Z

E








The data reveal that the bilinguals remain behind the

monolinguals in their ability to apply the regular past

tense English allomorphs. The bilinguals do as well as the

monolinguals in applying the progressive allomorph and

agentive allomorph to familiar as well as nonsense verbs.

In fact, compared with the data in Table 19, the kindergarten

and first grade bilingual children surpass the monolinguals

in their ability to generalize these allomorphs to new

forms. The bilinguals are quite behind the monolinguals

in their ability to use the third person singular present

tense allomorph /-az/.

With regard to adjectives, the findings for the

bilingual subjects were compatible with those found for

monolinguals. Pettit and Gillespie's data indicate that

the bilinguals have less ability to control the comparison

and superlative of the adjective allomorphs with nonsense

forms. Lastly, the bilinguals acquire mastery of the

singular possessive allomorphs at an older age than the

monolinguals.

Table 21 compares the performance of the bilingual

third graders (mean age = 8 years, 11 months) with Pettit

and Gillespie's monolingual 8-year-olds. The purpose of

this contrast is to help determine if older bilingual

children finally acquire the proficiency with English

morphological rules that their younger monolingual

counterparts have or if it appears they will need special

assistance to help them develop the underlying knowledge of










Table 21 Comparison of percentages correct for each
inflectional item reported by Pettit and
Gillespie for normal monolingual 8-year-olds
with those obtained for the third grade
bilingual children (mean age = 8 years, 11
months)

NONSENSE ITEMS MONOLINGUAL BILINGUAL
8-YEAR-OLDS 3RD GRADE


wugs 100% 100%
showed 100% 77%
tasses 50% 52%
mooks (heafs)* 100% 100%
motted (bodded)* 50% 74%
wug's 100% 97%
zibbing 90% 100%
zibber 80% 71%
quirkier 80% 19%
quirkiest 90% 94%
LEXICON ITEMS
dogs 100% 100%
cried (opened)* 100% 94%
glasses (dresses)* 100% 97%
sticks (leafs)* 100% 100%
melted (planted) 100% 87%
dog's (John's) 100% 100%
painting 100% 100%
painter 90% 100%
bigger 100% 100%
biggest 100% 100%
*items used in Pettit and Gillespie's study








morphological rules that monolinguals seem to acquire

naturally as they mature.

'Discussion. With the exception of noun plural

a'llomorphs /-s/ and /-z/ and the present progressive and

agentive allomorphs, bilinguals ate delayed in their

ability to correctly apply a variety of morphological

forms when speaking their non-native language, English,

as compared to monolingual speakers of English.

Examination of Table 21 reveals that the bilingual children

do finally achieve mastery over many of the morphological

forms by approximately 9-years of age and are not far

behind monolingual 8-year-olds on other inflections. Their

greatest difficulty seems to be with their ability to generalize

the comparison of the adjective allomorph to new forms

(only 19% of the third graders could supply the nonsense

form quirkierr"). This is most probably due to the fact

that the linguistic structure of the corresponding rule

in their native language is completely different. (In

Spanish the comparative of the adjective is formed by

placing a free morph-- the word "mas" before the adjective

being compared; in English, a bound morph-- /-r/ is

tacked on the end of the adjective and is part of the

same word.)

'In general, it seems that bilingualism does not

prevent children from eventually achieving mastery of

the grammatical rules of the second language. It does

hinder them to the extent that they don't learn the





86
various linguistic rules as rapidly as monolinguals. Rather

than relying on the maturation process alone, it might be

beneficial to provide the bilingual children with some special

educational intervention. With assistance the bilingual students

might be able to achieve mastery of the grammatical morphemes

at the same time as their monolingual peers.)

To further explore the effects of bilingualism on the

children's language development,t~ests were used to test

for significant differences between separate grade levels

for all BMM-E and BMM-S scores as well as to test for

significant differences between BMM-E and BMM-S scores

within each grade level. Additionally, this analysis helped

to determine if being bilingual at the lexical level of

language is an indicator of bilingualism at the morpho-

logical level of language.

Table 22 summarizes the results of the t test used

to test for significant differences between kindergarten

and first grade, first and second grade, and second and

third grade for all Bilingual Morphology Measure scores.

Table 23 summarizes the results of the t test used to

test for significant differences between the BMM-E and

BMM-S scores for the kindergarten children, first graders,

second graders, and third graders.

Discussion. Examination of the data presented in

Table 22 reveals that the only subtest on which the bilingual

children demonstrate a significant increase in their

productive knowledge of morphological rules as they progress

from grade to grade is the BMMI-E Meaningful subtest. No












Table 22


Summary of t test results of differences between
grades for all Bilingual Morphology Measure scores


COMPARISON


t VALUE df t PROBABILITY


BMM-S Nonsense


BMM-S Meaningful


BMM-S Total


BMM-E Nonsense


BMM-E Meaningful


BMM-E Total


Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade
Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade
Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade
Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade
Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade
Grade K x Grade
Grade 1 x Grade
Grade 2 x Grade


*significant at .05 level
**significant at .01 level


-1.768
0.409
-4.188
-2.899
-1.410
-3.660
-2.457
-0.390
-4.344
-3.282
-0.944
-4.871
-5.585
-2.824
-3.189
-4.602
-1.871
-4.535


118
118
118
118
,118
118
118
118
118
118
118
118
118
118
118
118
118
118


0.080
0.683
'0.000
0.004
0.161
0.000
0.015
0.697
0.000
0.001
0.347
0.000
0.000
0.006
0.002
0.000
0.064
0.000


_ _






88








I.4
0
O



o-o o o0
o 000 .r* O *00
000 O fl0 0 0 n 00







r}I I I I I I



4v v w v m e N o m


2 m Im CM (NC I I rn m
a















OI O Xl
4 z w r :l


>> LC O -
oI ) o -,w 0 -r




w) 4D I 0 m a) t a)
U 1E I C c C T CC CC

0 0m 0 -4 0i n0 r-I o 0
I-P i 1 I 1 I 1






-O w 0 w 0 w x 0 wz 00
4-1 M)C C z2 C;l r" C
41 0( 0 0 0m0 0 0 -0
09 M2 ZEb ZE S ZCE
to >W ( W ) W44 Wa 4)






0 m m aca yl x n0 mm
S0 r r r ZM r.

4J 0 r. a n r. r lz m f w (


o 00 0 0 0 'H 0 a) 0
0-4 Z 2 E- Z X E-i IZ E- Z Z E-
0 X XI t I I 1 4-X










Q 0Hr4 -- H-4 Hu r--
m (U4- (U4-
I II II I V I 4 4




a 01O Q0) l r.
r (0 S-I~0 0l 0

E-4 C F4 ) (n 0 *







significant improvement in their productive knowledge of

Spanish morphological rules as applied to either nonsense

or meaningful forms occurs between first and second grade.

Examination of Table 23 indicates that at the kindergarten

level the children categorized as bilingual by the James

Language Dominance Test are not really bilingual at the

morphological level of language as they are significantly

better in Spanish in their ability to apply morphological

rules to both familiar and unfamiliar forms. By second

grade the children are significantly better in English with

regard to productive control of inflectional rules with

meaningful forms, and by third grade they have a significantly

greater overall ability to apply English morphological rules.

This all indicates that not only does morphological

rule development in their native language stagnate between

first and second grade, it also fails to keep pace with the

acquisition of the secondary language. By third grade they

are significantly poorer in Spanish than English whereas

at the kindergarten level they were significantly better.

In addition, it may be concluded that comprehension and

production of English vocabulary items at the kindergarten

level do not seem to indicate that the children are

bilingual in other linguistic areas. The James Language

Dominance Test is a better predictor of bilingualism at

the morphological level of language in first grade

children.














CHAPTER IV

SUMMARY

The results provided insight into the three main areas

that this study initially sought to investigate: 1) that

of obtaining normative information regarding the approximate

ages at which bilingual children acquire productive control

of various morphological rules, 2) that of obtaining*

information about English and Spanish morphological rule

acquisition sequences by bilinguals, and 3) that of

obtaining information about the effects bilingualism

has on the native and non-native language development of

Cuban-American children.

First, normative data were obtained and grade level

medians, means, standard deviations and percentile rank

equivalents for the raw scores by grade level were reported

for Spanish and English. Validity was indicated by the

finding that the children's scores increased significantly

(p<.001) as grade level increased. A test-retest

investigation indicated good reliability as well. In

addition, the finding that children with identified poor

language development did poorer than the normal children

indicates the ability of the BMM to discriminate between

different groups of children.







Second, results indicated that those morphological

rules which are linguistically similar in the two languages

are acquired at approximately the same rate and in the same

order. Rules not sharing a common linguistic structure

are acquired at different times. A comparison of the

bilingual children's performance in English with existing

data on morphological rule acquisition by monolingual

American children revealed that the bilingual children

eventually acquire English morphological rules but at a

much slower rate than the monolinguals. Error analysis

revealed that the errors in the learning of the second

language were not the result of interference from the first

language.

Finally, the various effects of bilingualism on language

development were looked at, At the kindergarten level

the bilingual children performed significantly better than

their Spanish dominant peers on the BMM-S subtests. This

may result from the bilingual children having had experience

with inflectional rules in two languages which enhanced

their ability to use these rules in their native language.

However, as the bilingual children's grade level increased,

their productive control of Spanish morphological rules did

not develop as rapidly as their control over the corresponding

rules in English. By third grade, the children had sig-

nificantly greater ability to apply inflectional rules in

English.







Future Research

Several areas of future research could provide valuable

information which would enhance already existing information

including that generated by this study. By administering

both versions of the BMM to a sample of bilingual 5-year-

olds, norms could be extended to include this younger age

group. It would be interesting to see if there would be

a significant increase in scores between the 5-year-olds

and the 6-year-olds. In addition, the inclusion of other

grammatical categories in a test battery would allow for

a broader discussion of bilingual children's linguistic

performance. Another extension of the present study would

be to administer the.BMM to a larger sample of a language

deviant population and compare their performance with that

of normals.

Other goals of future research in this area might be

to 1) determine if the BMM can predict the occurrence or

nonoccurrence of errors in the free speech of bilingual

children in Spanish and English and 2) determine the

effect bilingualism has on native language development and

maintenance by comparing the developmental sequences of

an aspect of the Spanish spoken by monolingual Spanish

speaking children of various ages with the Spanish spoken

by bilingual children of matched ages.

Implications

The results of this study may be considered with

respect to assessment and remediation procedures in classroom







and clinical situations. The finding that children with

observed delayed language performed much poorer than the

normal children suggests that these children could be

identified at an early age. The results of testing with the

BMM would enable the language specialist to make statements

regarding a child's level of language development and to

plan remediation strategies. Language instruction specifically

designed to enhance morphological development could then begin.

Teachers of the Spanish speaking child should be alert

to the types of error patterns of morphology which have been

consistently observed in the expressive language of the

bilingual children. For example, the Spanish teacher, if made

aware that even the third grade children don't have expressive

control over the /-es/ allomorph for forming plurals in

Spanish, could plan some specific lessons in this area to

be presented in Spanish class.

It might be helpful to make a direct comparison between

the Spanish and English rules for pluralization, for example.

The teacher might say something to the effect:

We say "lionz" in English because a noun that ends
in an./n/ sound in English takes the sound "Z" to
form the plural BUT we don't say "leonz" in Spanish
because IN SPANISH a noun that ends in an /n/ sound
takes the sound "ES" to form the plural. That is why
we say "lionz" in English and leoness" in Spanish.














APPENDIX A

ADMINISTRATION PROCEDURES

Nonsense Versions

Seat the child in a comfortable position in which he

can see and respond easily to the picture stimulus. Give

the child the following directions (give in Spanish when

administering the BMM-S):

I am going to show you some pictures of some
things you have never seen before. Listen
very carefully because I am going to ask you
some questions about them. Okay?

Show the child the first practice plate and ask the

stimulus question. Continue to the second practice plate.

If, in your opinion the child understands the task, proceed

to the first plate of the instrument and continue testing.

If necessary, get the child's attention by saying, "Ready"

then present the first plate and question. The practice

items at the beginning of the subtest should be repeated

until the child understands the task. If necessary, provide

the answers for the practice items and then re-present

them until the chid responds on his own.

In presenting the stimulus questions, speak clearly

and slightly slower than in normal conversation. The target

nonsense word in the stimulus should be emphasized. If it

is felt that the child did not understand or if the child








did not respond, the stimulus may be repeated. If the child

responds in an unexpected way or gives reason to believe he

has not perceived the stimulus correctly, the child should

be taught the nonsense name and be asked to repeat if before

the particular test plate and stimulus question are

re-presented.

Example: The examiner shows a cartoon-like animal
and says, "This is a wug." Then the examiner shows
two of the same cartoon-like animals and says,
"Now there are two of them. There are two ."
If the child responds "bugs" or "wubs" or "fugs,"
he may be considered to have perceived the stimulus
incorrectly.

If the above were to occur, re-present the one wug and

say "This is a wuq" and then ask "What is it?" If the child

repeats the stimulus as it was provided, he has perceived

it correctly. Continue with the test. Natalicio and

Natalicio (1969) stated:

Before a judgement is made and statistics calculated
regarding plural responses, it seems necessary to
consider the stimulus itself and its perception by
the S; we cannot judge the S's internalized rule
for pluralization if we do not know the singular
stimulus to which he responded.(p.211)

The above contentions of Natalicio and Natalicio were

generalized to include all morphological forms tested in

the present investigation, not only noun plurals.

Meaningful Versions

After completing the administration of the nonsense

version, give the child the following instructions (say in

English for the BMM-E; Spanish for the BMM-S):

"Now I am going to show you pictures of animals
and people that you have seen many times. Listen
carefully because I have more questions to ask you.
Ready?