LEXICO-SEMANTIC ERRORS IN THE WRITTEN ... - Egerton...
Transcript of LEXICO-SEMANTIC ERRORS IN THE WRITTEN ... - Egerton...
LEXICO-SEMANTIC ERRORS IN THE WRITTEN ENGLISH OF
HEARING IMPAIRED PUPILS: A CASE STUDY OF LEARNERS AT NGALA
SPECIAL SCHOOL, NAKURU TOWN, KENYA
MANG’OKA ANTONY SOMBA
A Thesis Submitted to Graduate School in Partial Fulfillment of the Requirement
for the award of the Degree of Masters of Arts in English Language and Linguistics
of Egerton University.
EGERTON UNIVERSITY
MARCH 2009
ii
DECLARATION AND APPROVAL
This Thesis is my original work and has not been presented for examination in any other
Institution.
Mang’oka Antony Somba Signature ___________________________________________Date_________________
APPROVAL This Report has been submitted for examination with our approval as the supervisors appointented by the University.
1. Name: Dr. James Kariuki Mutiti Department of Literature, Languages and Linguistics, Egerton University
Signature ______________________________________________Date______________
2. Name: Dr. James Ogola Onyango. Department of Literature, Languages and Linguistics, Egerton University
Signature ______________________________________________Date______________
iii
COPYRIGHT
All rights reserved. No part of this thesis may be reproduced in whole or part, or
transmitted in any form or by any means (except in the case of brief quotations embodied
in critical review for educational purposes) without prior written permission of the author
or Egerton University.
Mang’oka © 2009
iv
DEDICATION
To my mother, Rose Mwikali and my late father, Matthews Mang’oka.
To my dear wife, Anne Wachera, and my lovely children, Esther Mwikali and Eva
Wanjiru.
v
ACKNOWLEDGEMENT
I am grateful to the Almighty God through whose grace I have completed this work. I
express my gratitude to those whose prayers, encouragement and support led to the
completion of this work.
My heartfelt acknowledgement and special gratitude goes to my tireless supervisors, Dr.
James Mutiti and Dr. James Onyango for their valuable guidance and patience throughout
the research process. Your professional and personal commitment made this work a
success. May the Almighty God bless you.
I appreciate the lecturers in the department of Literature, Languages and Linguistics for
their input in this work. My special thanks go to Dr. Felicia Yieke, Dr. Walunya, Dr.
Mukuthuria and Ms Juma for their support and encouragement. I also appreciate my
classmates: Kendagor, Albert, Bett, Edith, Susan, and Florence for their assistance and
moral support.
I do acknowledge the assistance and cooperation of the head teachers and the teachers of
Ngala special school for the deaf and St. Paul’s primary school in Nakuru. A special
thank you to Mary Maina.
My special appreciation goes to my dear wife and my children for being a great source of
inspiration; my mother, my brother and my sisters for their encouragement and moral
support.
Lastly, I thank Adolphus Wagalla for introducing me to ANOVA and t-test; James Mugo
for computer services; and Anne Njeri and Judy Onyanja for typing and formatting this
work.
vi
ABSTRACT
This was a study of the lexico-semantic errors made by hearing-impaired pupils in
standard six, seven and eight. The study identified and investigated the number, types and
patterns of lexico-semantic errors of hearing-impaired pupils. A comparison group of
hearing pupils in the same classes was also used. The study compared the lexico-semantic
errors of the hearing-impaired with those of the comparison group and determined the
lexico-semantic errors that could be attributed to hearing impairment. The study was
based on Interlanguage and Error Analysis theories. Instruments for data elicitation were
tests, which included a free composition, a picture story and a cloze passage. Data was
collected at Ngala Special School for the deaf and St. Paul’s primary school, where the
comparison group was sampled. Both schools are in Nakuru. The population sample
consisted of thirty hearing-impaired and thirty hearing pupils. The tests administered to
the pupils were all written tasks from which lexico-semantic errors were extracted and
categorized using the five steps of Error Analysis.
Several t-tests were done to determine whether there was a significant difference between
the lexico-semantic errors made by the hearing-impaired pupils and the comparison
group. One-way analysis of variance was done to determine whether there was a
significant difference among the lexico-semantic errors made by the three hearing-
impaired pupils’ classes. The lexico-semantic errors identified in this study were
categorized into nine groups. The study established that both groups portrayed similar
learning strategies in Lexico-semantic competence. However, the HI learners’ errors
differed from those of the HP learners in three categories. It was noted that the HI
learners made more Lexico-semantic errors than the HP learners did. There was a
significant difference in some categories of Lexico-semantic errors. Learners with a
greater exposure to English portrayed a greater lexico-semantic competence than those
exposed to English for a shorter period. The study concluded that the HI learners have not
yet acquired enough English vocabulary to express themselves. The findings of this study
will add more knowledge to other studies done in applied linguistics. The findings will
also be of pedagogical value to educationists, teachers and the Ministry of Education in
general.
vii
TABLE OF CONTENTS
DECLARATION AND APPROVAL ................................................................................ ii
COPYRIGHT..................................................................................................................... iii
DEDICATION................................................................................................................... iv
ACKNOWLEDGEMENT ...................................................................................................v
ABSTRACT....................................................................................................................... vi
TABLE OF CONTENTS.................................................................................................. vii
LIST OF TABLES...............................................................................................................x
ABBREVIATIONS AND ACRONYMS........................................................................ xiii
CHAPTER ONE ..................................................................................................................1
INTRODUCTION ...............................................................................................................1
1.1 Background to the Study........................................................................................... 1
1.2 Statement of the Problem.......................................................................................... 2
1.3 Objectives of the Study............................................................................................. 3
1.4 Hypotheses................................................................................................................ 3
1.5 Justification and Significance of the Study............................................................... 3
1.6 Scope and Limitation of the Study............................................................................ 5
1.7 Definition of Terms................................................................................................... 6
CHAPTER TWO .................................................................................................................8
LITERATURE REVIEW ....................................................................................................8
2.1 Introduction............................................................................................................... 8
2.2 Review of Related Literature .................................................................................... 8
2.2.1 L2 learners’ Errors ..............................................................................................8 2.2.2 L2 Learners’ Errors in Kenya ...........................................................................12 2.2.3 Studies Related to the Hearing-impaired ..........................................................15
2.3 Theoretical Framework........................................................................................... 22
2.3.1 Error Analysis ...................................................................................................22 2.3.2 Interlanguage.....................................................................................................23
2.4 Conclusion .............................................................................................................. 26
CHAPTER THREE ...........................................................................................................27
RESEARCH METHODOLOGY.......................................................................................27
3.1 Introduction............................................................................................................. 27
3.2 Research Design...................................................................................................... 27
viii
3.3 Population and Location of the Study..................................................................... 27
3.4 Sampling Schools.................................................................................................... 28
3.5 Sampling Students .................................................................................................. 28
3.6 Instruments for Data Collection.............................................................................. 29
3.7 Data Analysis and Interpretation ............................................................................ 29
CHAPTER FOUR..............................................................................................................33
RESULTS AND DISCUSSION........................................................................................33
4.1 Introduction............................................................................................................. 33
4.2 Identification, Description, and Discussion of the Lexico-Semantic Errors .......... 33
4.2.1 Learning Induced Errors ...................................................................................35 4.2.2 Meaning Similarity Errors ................................................................................42 4.2.3 Semantic Contiguity Errors...............................................................................44 4.2.4 Collocation Errors.............................................................................................46 4.2.5 Sound Similarity Errors ....................................................................................49 4.2.6 Meaning Duplication Errors .............................................................................51 4.2.7 Paraphrase Errors ..............................................................................................53 4.2.8 Coinage Errors ..................................................................................................56 4.2.9 Haphazard Errors ..............................................................................................58
4.3 Summary ................................................................................................................. 61
4.4 Difference between the HP and HI Learners’ Lexico-Semantic Errors ................. 61
4.4.1 Meaning Similarity Errors ................................................................................62 4.4.2 Meaning Duplication Errors .............................................................................64 4.4.3 Semantic Contiguity Errors...............................................................................65 4.4.4 Coinage Errors ..................................................................................................67 4.4.5 Learning Induced Errors ...................................................................................68 4.4.6 Paraphrase Errors ..............................................................................................69 4.4.7 Sound Similarity Errors ....................................................................................70 4.4.8 Haphazard Errors ..............................................................................................72 4.4.9 Collocation Errors.............................................................................................73
4.5 Analysis of Variance for HI Errors......................................................................... 74
4.5.1 Meaning Similarity Errors ................................................................................75 4.5.2 Meaning Duplication Errors .............................................................................76 4.5.3 Semantic Contiguity Errors...............................................................................77 4.5.4 Coinage Errors Errors .......................................................................................78 4.5.5 Learning Induced Errors ...................................................................................79 4.5.6 Paraphrase Errors ..............................................................................................80 4.5.7 Sound Similarity Errors ....................................................................................81 4.5.8 Haphazard Errors ..............................................................................................82 4.5.9 Collocation Errors.............................................................................................83
4.6 Analysis of Variance for HP Errors ........................................................................ 85
4.6.1 Meaning Similarity Errors ................................................................................85
ix
4.6.2 Meaning Duplication Errors .............................................................................86 4.6.3 Semantic Contiguity Errors...............................................................................86 4.6.4 Coinage Errors ..................................................................................................87 4.6.5 Learning Induced Errors ...................................................................................88 4.6.6 Paraphrase Errors ..............................................................................................89 4.6.7 Sound Similarity Errors ....................................................................................90 4.6.8 Haphazard Errors ..............................................................................................91 4.6.9 Collocation Errors.............................................................................................91
4.7 Summary of Statistical Analysis............................................................................. 92
4.7.1 Summary Of T-Test Results .............................................................................92 4.7.2 Summary Of ANOVA Results..........................................................................94
4.8 Discussion Of Findings........................................................................................... 96
4.8.1 Learning Induced Errors ...................................................................................97 4.8.2 Meaning Similarity Errors ..............................................................................101 4.8.3 Meaning Duplication Errors ...........................................................................102 4.8.4 Semantic Contiguity Errors.............................................................................103 4.8.5 Collocation Errors...........................................................................................104 4.8.6 Coinage Errors ................................................................................................105 4.8.7 Haphazard Errors ............................................................................................106 4.8.8 Paraphrase Errors ............................................................................................106 4.8.9 Sound Similarity Errors ..................................................................................107
4.9 Analysis of Variance for the HI and HP Learners' Lexico-Semantic Errors ........ 108
4.9.1 Summary .........................................................................................................108
CHAPTER FIVE .............................................................................................................110
SUMMARY, CONCLUSION AND RECOMMENDATION........................................110
5.1 Introduction........................................................................................................... 110
5.2 Identification and Description of the Lexico-Semantic Errors ............................. 110
5.3 Significant Difference of the Lexico-Semantic Errors ......................................... 111
5.4 Hearing Impaired Pupils’ Lexico-Semantic Errors............................................... 111
5.5 Limitation of the Research.................................................................................... 113
5.6 Recommendation for Further Research ................................................................ 114
5.7 Remedial Measures............................................................................................... 114
5.8 Conclusion ............................................................................................................ 115
REFERENCES ................................................................................................................116
APPENDIX A: PICTURE STORY................................................................................ 125
APPENDIX B: CLOZE PASSAGE ................................................................................126
APPENDIX C: ERROR SUMMARY TABLE .............................................................. 128
APPENDIX D: SAMPLE COMPOSITIONS .................................................................130
x
LIST OF TABLES
Table 1: A Summary of the Total Number of Errors........................................................ 34
Table 2: A Summary of Learning Induced Errors ............................................................ 41
Table 3: A Summary of Group Statistics for the Meaning Similarity Errors Made by HP
and HI Learners......................................................................................................... 63
Table 4: T-Test Table for the Meaning Similarity Errors Made by HP and HI Learners. 64
Table 5: A Summary of Group Statistics for the Meaning Duplication Errors Made by
HP and HI Learners .................................................................................................. 65
Table 6: T-Test Table for the Meaning Duplication Errors Made by HP and HI Learners
................................................................................................................................... 65
Table 7: A Summary of Group Statistics for the Semantic Contiguity Errors Made by HP
and HI Learners......................................................................................................... 66
Table 8: T-Test Table for the Semantic Contiguity Errors Made by HP and HI Learners
................................................................................................................................... 66
Table 9: A Summary of Group Statistics for the Coinage Errors Made by HP and HI
Learners..................................................................................................................... 67
Table 10: T- test table for the coinage Errors Made By HP and HI Learners .................. 68
Table 11: A Summary of Group Statistics for the Learning Induced Errors Made by HP
and HI Learners......................................................................................................... 69
Table 12: T-Test Table for the Learning Induced Errors Made by HP and HI Learners 69
Table 13: A Summary of Group Statistics for the Paraphrase Errors Made by HP and HI
Learners..................................................................................................................... 70
Table 14: T-Test Table for the Paraphrase Errors Made by HP and HI Learners ............ 70
Table 15: A Summary of Group Statistics for the Sound Similarity Errors Made By HP
and HI Learners......................................................................................................... 71
Table 16: T-Test Table for the Sound Similarity Errors Made By HP and HI Learners.. 71
Table 17: A Summary of Group Statistics for the Haphazard Errors Made by HP and HI
Learners..................................................................................................................... 72
Table 18: T-Test Table for the Haphazard Errors Made by HP and HI Learners ............ 73
Table 19: A Summary of Group Statistics for the Collocation Errors Made by HP and HI
Learners..................................................................................................................... 74
Table 20: T-Test Table for the Collocation Errors Made by HP and HI Learners ........... 74
xi
Table 21: A summary of group statistics for The HI learners' Meaning Similarity errors
................................................................................................................................... 75
Table 22: ANOVA table for The HI learners' Meaning Similarity errors ........................ 76
Table 23: A summary of group statistics for The HI learners' Meaning Duplication errors
................................................................................................................................... 76
Table 24: ANOVA table for The HI learners' Meaning Duplication errors ..................... 77
Table 25: A summary of group statistics for The HI learners' Semantic Contiguity errors
................................................................................................................................... 78
Table 26: ANOVA table for The HI learners' Semantic Contiguity errors ...................... 78
Table 27: A Summary of Group Statistics for the HI Learners' Coinage Errors .............. 79
Table 28: ANOVA Table for the HI Learners' Coinage Errors ........................................ 79
Table 29: A Summary of Group Statistics for the HI Learners' Learning Induced Errors80
Table 30: ANOVA Table for the HI Learners' Learning Induced Errors ........................ 80
Table 31: A summary of group statistics for The HI learners' Paraphrase errors............ 81
Table 32: ANOVA table for The HI learners' Paraphrase errors..................................... 81
Table 33: A Summary of Group Statistics for the HI Learners' Sound Similarity Errors82
Table 34: ANOVA Table for the HI Learners' Sound Similarity Errors .......................... 82
Table 35: A Summary of Group Statistics for the HI Learners' Haphazard Errors .......... 83
Table 36: ANOVA Tables for the HI Learners' Haphazard Errors .................................. 83
Table 37: A Summary of Group Statistics for the HI Learners' Collocation Errors......... 84
Table 38: ANOVA Table for the HI Learners' Collocation Errors................................... 84
Table 39: A Summary of Group Statistics for the HP Meaning similarity errors ............ 85
Table 40: ANOVA Table for the HP Meaning similarity errors ...................................... 85
Table 41: A Summary of Group Statistics for the HP Meaning duplication errors.......... 86
Table 42: ANOVA Table for the HP Meaning duplication errors................................... 86
Table 43: A Summary of Group Statistics for the HP Semantic contiguity errors........... 87
Table 44: ANOVA Table for the HP Semantic contiguity errors..................................... 87
Table 45: A Summary of Group Statistics for the HP Coinage errors.............................. 88
Table 46: ANOVA Table for the HP Coinage errors ....................................................... 88
Table 47: A Summary of Group Statistics for the HP learning induced errors ............... 89
Table 48: ANOVA Table for the HP learning induced errors ......................................... 89
Table 49: A Summary of Group Statistics for the HP Paraphrase errors ......................... 89
xii
Table 50: ANOVA Table for the HP Paraphrase errors .................................................. 90
Table 51: A Summary of Group Statistics for the HP Sound similarity errors ................ 90
Table 52: ANOVA Table for the HP Sound similarity errors ......................................... 90
Table 53: A Summary of Group Statistics for the HP Haphazard errors.......................... 91
Table 54: ANOVA Table for the HP Haphazard errors ................................................... 91
Table 55: A Summary of Group Statistics for the HP Collocation errors ........................ 92
Table 56: ANOVA Table for the HP Collocation errors .................................................. 92
Table 57: Summary of T-Tests results.............................................................................. 93
Table 58: Summary of ANOVA for HI learners’ errors................................................... 95
Table 59: Summary of ANOVA for HP learners’ errors.................................................. 96
xiii
ABBREVIATIONS AND ACRONYMS
ANOVA-Analysis of variance
ASL- American Sign Language
BSL- British Sign Language
CA- contrastive analysis
EA- Error Analysis
ESL-English as a Second Language
Fcalc. - Calculated F value
HI- Hearing Impaired
HP- Hearing Pupil
IL- Interlanguage
KSL-Kenyan Sign Language
KNEC Kenya National Examination Council
LAD- language acquisition device
L1- Language one
L2-Language Two
MT- Mother Tongue
P-value- probability value
SD- Standard deviation
SLA- Second Language Acquisition
TL Target Language
Tcalc.-Calculated T value
UG- Universal Grammar
1
CHAPTER ONE
INTRODUCTION
1.1 Background to the Study
English is the official language in Kenya. It is not only used as the medium of instruction
in Kenyan schools from standard four to university level but it is also taught as a
compulsory subject in Kenyan primary and secondary schools. The use of English as a
medium of instruction dictates that students have a certain level of proficiency in English.
It is important for the pupils to develop competence in the language. This will help them
to use English effectively and to understand the teacher in the classroom. However,
several studies that have been done on the written English of second language learners
reveal that learners exhibit errors in their written English. Examples of such studies are as
follows; Njoroge (1987); Maina (1991); Nyamasyo (1992); Simatwo (1993); Chege
(1996); Njoroge (1996); and Ayoo (2004). The analysis of errors in these studies is useful
in designing a remedial syllabus or a programme for remedial teaching. A similar study
on the written English of the hearing impaired would be necessary to meet the same ends.
Studies on hearing-impaired students indicate that the hearing-impaired have consistently
trailed behind their hearing counterparts in academic performance (Adoyo, 1995; 2002;
Ndurumo, 1993; and Okombo, 1994). Teachers’ lack of competence in the language of
instruction has been found to be one of the major obstacles to their academic
development. Other studies claim that the hearing impaired children begin their formal
school lacking the necessary language skills and general knowledge for normal language
development among their age peers (Wilbur 2000; Toth 2002). This is another obstacle in
the hearing-impaired children’s ability to learn curriculum content in school.
There have been instances of public concern on the situation of the hearing-impaired.
Many hearing-impaired children learn up to standard eight but do not sit for the Kenya
Certificate of Primary Education because schools for the hearing-impaired are not ready
to register them for the examination since they think the hearing-impaired cannot
continue learning (Ombiro, 1998:17). This explains why we have 41 primary schools and
only 4 secondary schools for the hearing-impaired in Kenya. Even for those who finish
2
their schooling, they have been found to be semi-illiterate (Adoyo, 2002; Okombo, 1994).
For the hearing-impaired to go beyond standard eight, they need to be trained to
communicate (Nyamongo, 1996).
Learners of English at all levels are bound to make lexico- semantic errors in their
attempt to achieve a mastery of the English language; however, the types of errors made
by learners of the same level depend on their hearing ability. Hearing-impaired children
have been found to be markedly retarded in their achievement test scores. Quigley and
Paul (1984) argue that the written language of hearing impaired children, compared to
that of the hearing children, contains shorter and simpler sentences. It also displays
different distribution of the parts of speech, appears rigid and more stereotyped, and
exhibits numerous errors or deviations from Standard English use.
The English syllabus for primary education aims at learners achieving communication
competence at the end of standard eight (Kenya Institute of Education, 2002). All pupils
are required to have acquired a sufficient command of English in both spoken and written
forms through the language skills of speaking, listening, writing and reading. This is
supposed to enable them communicate fluently, follow subject courses and textbooks,
and read for pleasure and information. However, the hearing-impaired pupils are
disadvantaged in listening and speaking naturally. Although the partially hearing-
impaired use hearing aids, the profoundly hearing-impaired cannot use these aids. In spite
of their disadvantages, however, the hearing-impaired pupils share the same syllabus and
sit for the same national exam with the hearing pupils. It is against this background that
the present study analyzed the hearing-impaired pupils’ lexico-semantic errors. An
analysis of the lexico-semantic errors of a comparison group consisting of hearing pupils
was also done and comparison made.
1.2 Statement of the Problem
There are lexico–semantic errors made by the hearing-impaired pupils in their written
texts. These errors have not been analyzed. There is also a dearth of research data on how
the hearing-impaired pupils write in Kenya. Analysis of these errors can help linguists
and educators to come up with intervention strategies and measures that can aid the
3
hearing impaired in language acquisition. This study investigated the lexico-semantic
errors made by the hearing-impaired pupils in their written English. The hearing-impaired
pupils were drawn from standard six, seven and eight. Their lexico-semantic errors were
compared with those of the hearing pupils in the same classes.
1.3 Objectives of the Study
The general aim of this study was to analyze the lexico-semantic errors made by hearing-
impaired pupils and determine the errors that can be attributed to their condition. The
objectives of the study were as follows:
1. To identify and describe the lexico-semantic errors made by hearing-impaired
pupils in their written English.
2. To establish if there is a significant difference between the lexico-semantic errors
made by hearing-impaired pupils and those made by the hearing pupils in their
written English.
3. To determine if there are lexico-semantic errors that can be attributed to hearing
impairment, and establish how such errors vary across the three classes of
hearing-impaired pupils.
1.4 Hypotheses
1) Hearing-impaired pupils exhibit lexico-semantic errors in their use of the English
language in written texts.
2) There is a significant difference between the lexico-semantic errors made by
hearing-impaired and hearing pupils.
3) Hearing-impaired pupils make idiosyncratic lexico-semantic errors that can be
attributed to their impairment, and these errors vary across the three groups of
hearing-impaired pupils.
1.5 Justification and Significance of the Study
As students acquire more formal exposure to English, they internalize more rules of the
English language. They are expected to acquire and be able to correctly use lexical items
in appropriate contexts as their English lexicon increases. However, it has been noted that
the standards of English have been falling in both the Kenya Certificate of Primary
4
Education and the Kenya Certificate of Secondary Education. (K.N.E.C, 1999; 2001;
2003; 2004; and Odhiambo, 2006).
This study concentrated on the lexico-semantic errors and its finding will contribute to
the field of Applied Linguistics and Second Language Acquisition. It will also help
educationists in designing teaching and remedial courses for learners in primary schools.
Corder (1974) says that analysis of errors enables teachers know how effective their
teaching materials and techniques are.
At the level of pragmatic classroom experience, Error Analysis will continue to provide one means by which the teacher assesses learning and teaching, and determines priorities for future effort. (Richards and Sampson, 1974: 15)
In this study, the analysis of errors provided useful information on common difficulties in
language learning and will aid in teaching and the preparation of learning materials.
The investigation of errors can be at the same time diagnostic and prognostic. It is
diagnostic because it can tell us the learner's state of the language (Corder, 1967) at a
given point during the learning process, and prognostic because it can tell course
organizers to re-orient language learning materials based on the learners' current
problems. The study will therefore be useful to teachers, the Kenya Institute of Education
and the Ministry of Education in general.
Gass and Schatcher (1989) observe that there has been little attention paid by researchers
to the area of lexical acquisition. Most studies are in the area of morphology and syntax
(Dulay and Burt, 1974; Bailey et al, 1974). Gass and Schatcher (1989) claim that learners
face the greatest difficulty in learning a second language in the area of acquisition of
vocabulary and their ability to use lexis correctly in conveying a given message. In this
study, an analysis of the lexico-semantic errors of the hearing-impaired learners of L2
English and a comparison group of normal hearing learners has led to explication of the
problems that the former group encounters in the area of lexical acquisition.
Although this is a case study and has many characteristics that may not be generalisable,
this researcher feels that it has significance. One, it shows that there are patterns of L2
5
performance that can be attributed to an impairment on the hearing facet of language
cognitive mechanism. Two, these areas of performance that are affected by hearing
impairment can be used by linguists , educators, and other researchers to come up with
intervention strategies and measures to help the HI learners acquire vocabulary and
English language in general.
1.6 Scope and Limitation of the Study
This study falls within the field of applied linguistics. It has analyzed the lexico-semantic
errors that occur in the written English of standard six, seven and eight hearing-impaired
pupils learning English as a second language in Kenya. A comparison group of hearing
pupils was also used and their Lexico-semantic errors analyzed and compared with those
of the hearing-impaired pupils. The study was interested on the lexico-semantic errors
identified in written texts only. This is because in written texts, learners are able to
express themselves freely using whatever language items are at their disposal. It also
lacks some of the advantages of oral speech such as pointing at some objects,
questioning, requesting for help and using other para-linguistic features that aid
communication. Comparing the spoken English of the hearing-impaired pupils and the
comparison group (hearing pupils) would have been a disadvantage to the hearing-
impaired students. Most hearing-impaired students are not able to speak. They use sign
language.
The study does not cover the whole of Kenya. Data for the study was collected in two
schools: one for the deaf and the other for the hearing students. Both schools are in
Nakuru Municipality. The size of the sample was thirty students. Because it is not
possible to get thirty students from one class in schools for the deaf, the study was
therefore limited to standard six, seven and eight pupils only. They occupy the highest
level in primary school. They are also assumed to have obtained a particular level of
competence in the English language.
6
1.7 Definition of Terms
Error – any word or sentence structure whose usage deviates from the norm. Ellis (1994)
treats errors as a deviation in a language users’ language, which results from lack of
knowledge of the correct rules. Errors are systematic deviation from the norm.
Error analysis: the study and analysis of the errors made by foreign and second language
learners. (O’Grady et. al. 1993).
Hearing impaired person- one who has hearing disorders and therefore needs/requires
education by suitable methods (Strong, (1988).
Hearing students – students who have no hearing problems.
Interlingual errors - errors that result from language transfer, that is, which are related
to the learners’ native language (Dulay et. al 1982).
Intralingual error-an error that results from faulty or partial learning (Dulay et. al,
1982).
Lexical item-an item that functions as a single meaning unit, regardless of the number of
words it contains (Schmitt, 2000).
Lexico- semantic errors- the term ‘lexico-semantic’ in this study will refer to
vocabulary meaning, that is, the word and the associations of meaning it embodies which
restricts how the given word relates to other words. Lexico- semantic errors will refer to
errors because of wrong choice of vocabulary meaning (Wilkins, 1974).
Kenyan sign language- a visual-gesture language used by the Kenyan deaf community
for communication.
Mistake- Non-systematic errors that learners produce. It refers to faulty use of a
linguistic item caused by aspects of performance such as lack of attention, fatigue or
carelessness.
Partially hearing-impaired person- one who has hearing disorders, but whose language
development, even if retarded, is following the normal pattern, and requires special
facilities such as hearing aids for education (Bishop and Mogford, 1993).
Prelingually hearing-impaired– a person who was deaf at birth or who became deaf
before acquiring spoken language spontaneously (Bishop and Mogford, 1993).
Profoundly hearing- impaired person- one with great hearing disorders such that he/she
cannot benefit from hearing aids and therefore depends on vision as the main channel of
communication (Bishop and Mogford, 1993).
7
Signing: - use of gestures /sign language.
8
CHAPTER TWO
LITERATURE REVIEW
2.1 Introduction
This chapter has two sections: review of related literature and theoretical framework. It
will start with a general overview of language two learners’ errors, followed by a review
of language two learners’ errors studies done in Kenya. Lastly will be studies related to
the deaf. The theoretical framework will be divided into two sections: Error Analysis and
Interlanguage Theory.
2.2 Review of Related Literature
Literature review relevant to this study was drawn from materials in second language
learning and lexical semantics. It is divided into three sub-sections: L2 learners’ errors,
L2 learners’ errors in Kenya and Studies related to the deaf.
2.2.1 L2 learners’ Errors
Wilkins (1972) says that the ability to refer to concrete and conceptual entities in
communication is as fundamental to language as it is to the capacity provided by the
grammar to relate such entities to another. It therefore follows that knowledge of a
language demands mastery of its vocabulary as much as its grammar. This observation
stresses the importance of the acquisition of lexical competence, which is the backbone of
the present study. Another observation by Wilkins (1972, 1974) is that the meaning of a
lexical item is “the product of the way in which the item relates to other words and at the
same time to the non-linguistic reality’’ (Wilkins 1974:20). He claims that any one word
is part of a larger interdependent set of words that is related to it in the degree to which
they are substitutable for it, contrast with it or occur in similar linguistic contexts. Such
associations build up for each lexical item a network of associations essential to its
definition just like the non-linguistic associations. Wilkins (1972 ibid), suggests that it is
this intricate network of associations that makes up lexical meaning which cause
difficulties to L2 learners. He cites associations such as polysemy, synonymy, idioms and
collocation restrictions as the major causes of lexical errors. His work was therefore
9
relevant to the present study as it gave basic knowledge to lexical meaning and lexico –
semantic errors.
Gairns and Rodman (1992) say that the meaning of a word can only be understood and
learnt in terms of its relationship with other words in the language. Some of the problems
which they claim may pose difficult in lexical-semantics are sense relations such as
synonymy, hyponymy, antonymy, part-whole relationships, items associated with,
collocations and translation equivalents. Gairns and Rodman (ibid) further claim that
knowledge of the world is important in lexical semantics.
Blum-Kulka and Levenston (1978) have studied universals of lexical simplification. They
define it as a process of making do with fewer words. They claim that lexical
simplification follows universal principles which derive from certain aspects of semantic
competence, particularly the following: the awareness of hyponymy, antonymy,
converseness and other systematic relationships between lexical items; the ability to
avoid the use of specific lexical items by means of circumlocution and paraphrase; and
the ability to recognize degrees of paraphrastic equivalence. They claimed that the above
aspects of semantic competence are at the root of the errors learners make. The aim of the
above study was to examine the idea of lexical simplification that operates according to
universal principles that derive from the learner’s semantic competence in his mother
tongue. The researchers observed that learners are universally aided by the strategies of
overgeneralization, transfer, circumlocution and paraphrase, language switch, appeal to
authority, change of topic and semantic avoidance.
Blum-Kulka and Levenston (1978) link these strategies to vocabulary use. They see them
in terms of a general process of simplification. The basic approach when learners are
attempting to express themselves in speech or writing is avoidance. Learners may avoid
using a specific lexical item either because they simply do not know it or because they
are not confident about it with its correct pronunciation, spelling or grammatical form.
They, however, observed that the learner’s competence develops as he internalizes the
semantic relationships in the second language independently of the first language
equivalent. This study was relevant to the present study because it deals with lexico-
semantic errors and attempts to explain their causes. It therefore provided important
10
background information for the present study. However, the study was only limited to
hearing pupils. A similar study on the hearing-impaired pupils would also be important.
The present study differs from the above study in that, it analyzed the lexico-semantic
errors made by hearing-impaired pupils and determined the errors that could be attributed
to their condition.
Tarone (1978) carried out a study on the communicative strategies on learners of ESL.
She took a small group of intermediate-level ESL learners and presented them with three
pictures, which they were asked to describe both in their L1 and in English. The pictures
included objects such as balloon, a water pipe and a caterpillar; things of which the
learners did not know there English names. She wanted to find out how her subjects
would cope with the problem of identifying objects they could not directly name in the
TL. She found out that her subjects had five ways (or strategies) in dealing with lexical
gaps, that is, words they read which they simply do not understand, or concepts that they
cannot express as adequately as they could in their L1. (See Tarone, 1983; Read,
2000:65). Tarone presented a taxonomy of five communicative strategies some of which
were divided into subtypes.
They are as follows:
1. Paraphrase: (a). Approximation – use of a single target language vocabulary item or
structure which the learner knows is not correct, but which shares
enough features with the desired item to satisfy the speaker.
(b). Word coinage – the learner makes up a new word in order to
Communicate the desired concept.
(c). Circumlocution – the leaner describes the characteristics or the
elements of the object or the action instead of using the appropriate
target language structure.
2. Transfer: (a). Literal translation – the learner translates word for word from the native
language.
(b). Language switch – the learner uses the native language term.
3. Appeal for assistance: - the learner asks for the correct form or structure.
4. Mime: - the learner uses non-verbal strategies in place of meaning structure.
5. Avoidance: (a). Topic avoidance – occurs when the learner simply does not talk about
11
concepts for which the vocabulary or other meaning structure is not
known.
(b). Message abandonment – occurs when the learner begins to talk about
a concept but is unable to continue due to lack of meaning structure, and
stops in mid-utterance.
Though the above work dealt with communication, it was relevant in the present study in
that, it provided a background on problems that learners face in use of lexical items.
Other earlier researchers observe that learners’ errors are a clear indication of their
developing language systems. Littlewood (1984) groups learners’ errors into two broad
categories, that is, interlingual errors and intralingual errors. Interlingual errors are as a
result of transfer while intralingual errors are a result of faulty or partial learning of a
second language. The two categories of errors are further divided into four subcategories
of errors. These are transfer, overgeneralization, ambiguous errors and simplification by
omission. His study helped the present study in the classification of lexico-semantic
errors.
Another classification of errors was done by Dulay, Burt and Krashen (1982). They
classified errors under four categories; linguistic taxonomy, surface structure taxonomy,
comparative analysis taxonomy and communicative effect taxonomy. The Linguistic
category taxonomy classifies errors according to either or both the language component
or the particular linguistic constituent the error affects. The surface structure taxonomy
classifies errors according to any deviations in the surface structure of the language item,
such as, word order. The comparative taxonomy compares errors in the second language
with those of children learning a first language. The communicative effect taxonomy is
based on the effect the error has on the listener or reader. These are grouped as ‘global’ or
‘local’ depending on how much they affect communication. Although the above study
differs from the present study, it helped the present study in classification of errors. The
present study used the linguistic taxonomy in analysis of the lexico-semantic errors made
by the H.I and the hearing pupils.
12
2.2.2 L2 Learners’ Errors in Kenya
Although several studies have been carried out on learners’ errors in Kenya, what is on
record as related research is inadequate or incomplete as far as acquisition of second
language by the HI learners is concerned. The researches were concerned with the
acquisition of English as a second language by normal hearing learners cutting across
primary to university level. Findings of such researches cannot be generalized for all
second language learners of English. Some of the researches undertaken on L2 learner’s
errors in Kenya include those of Njoroge (1987), Maina (1991), Nyamasyo (1992),
Simatwo (1993), Chege (1996) and Njoroge (1996).
Njoroge (1987) carried out a study on the acquisition of six morphosyntactic structures of
the English of Kenyan children. He based his study within the Interlanguage theory. He
found out that the errors that were made in process of language acquisition reflect the
strategies and process involved in L2 learning. He concluded that language acquisition
was a developmental process. However, the above study does not cover all aspects of
morphology and syntax that SL learners acquire as it only concentrated on six
morphosyntactic structures. The above study differs from the present study, in that it
studied morphosyntactic structures in hearing learners while the present study
investigated Lexico-semantic errors in HI learners. However, its methodology was
helpful in presentation and analysis of the present study’s data.
Similar observations on language acquisition as a developmental process were made by
Maina (1991) and Nyamasyo (1992). Maina (1991) carried out a study on the
grammatical errors in standard eight pupils’ written English in four city schools in Kenya.
He found out that most errors where because of overgeneralization. His study was based
on the Error Analysis approach. He presented the grammatical errors using the Linguistic
category taxonomy. The present study benefited from this methodology in analysis and
presentation of the lexico-semantic errors of the groups under study.
The role of overgeneralization in learners’ errors was also observed by Nyamasyo (1992).
She studied the grammatical and lexical characteristics of the writing of Kenyan pre-
university students. She found out that overgeneralization was the main cause of the
13
students’ errors. Similarly, Njoroge (1996) observed that overgeneralization was the main
cause of errors. He examined the morphosyntactic errors in the written English of first
year undergraduate students in Kenya. His study was based within the EA and IL theory.
He found out that verb related errors were very common in the written work of students.
He concluded that overgeneralization was the main cause of errors.
Just like in Njoroge (1987) and Maina (1991), the methodology of the above two studies
was helpful in data presentation and analysis. However, the above researches do not pay
attention to the use of vocabulary or word meaning and the errors thereof. Unlike the
present study, they do not consider how HI learners face problems in the use of
vocabulary meaning or lexical items. The present study investigated the lexico-semantic
errors made by the HI learners in primary school.
Earlier studies on hearing pupils’ lexico-semantic errors done in Kenya showed that
learners have problems in vocabulary acquisition. Simatwo (1993) and Chege (1996)
carried out examples of such studies. Simatwo (1993) did an investigation of the lexico-
semantic errors of standard seven Nandi-speaking pupils in five primary schools in Uasin
Gishu and Nandi Districts. His study was based on the Error Analysis Approach and
aimed at investigating the nature and causes of errors. He used written composition,
translation exercises, narratives and conversations as methods of data elicitation. He
classified errors into nine categories: calques, malapropisms, ignorance, coinage,
semantic contiguity, collocation, learning- induced, language switch and anglicization.
His study was important to the present study because it was not only based on lexico-
semantic errors but it was also done in a second language environment. Simatwo’s study
was carried among speakers of Nandi language while the present study consisted of HI
pupils with different MTs (mother tongues). The subjects in the present study also had an
urban setting. His work and that of Chege (1996) and Njoroge (1996) do not provide
data that can be used to establish how exceptional circumstances such as hearing
impairment can have implications for lexical meaning and theories of language learning.
Chege (1996) observed the role of formal exposure to lexico-semantic competence. She
did a study on lexico-semantic errors as indices of developing language competence
14
among Kikuyu pupils in standard five, six and seven. Her study was based on the Error
Analysis and Interlanguage approaches. Errors in her study were seen as indices of
developing language competence in primary schools pupils. She compared the errors
among the three groups using One Way Analysis of Variance. She found out that there
was no significant difference among the lexico-semantic errors made by the three classes.
She classified the lexico-semantic errors into the following categories: collocation,
coinage, learning-induced, semantic contiguity, paraphrase, translation, L1
phonologically induced, and other errors that were characterized by illogical use of
lexical items. Her study concluded that:
1. Though the three groups made similar errors, the frequency with which they made
differed. Some errors were less frequent in advanced learners than among the less
advanced.
2. The pupils with greater formal exposure to English tended to portray greater
lexico-semantic competence than those who had a shorter period of exposure.
3. Though some error types like learning induced, collocation and translation errors
reduced with each higher level, there were other error types whose frequency was
almost constant across the three levels.
The above study was relevant to the present study as it was based on lexico-semantic
errors of primary school pupils learning English as a second language. However, the
above study did not focus on how the HI pupils use lexical meaning and the difficulties
they face. The present study analyzed the lexico-semantic errors made by hearing-
impaired pupils and determined the errors that could be attributed to their condition.
A more recent study on learners’ errors was done by Mutiti (2000). He carried out a
research on the SL acquisition of English by speakers of Gikuyu first language
background. The research was aimed at the investigation of the factors related to the
setting of the parameter of syntactic information packaging towards acquisition of
English by Gikuyu learners. The study proved that Chomsky’s principles and parameters
UG (Universal Grammar) are applicable in the acquisition of a SL. The research proved a
developmental continuum in conformity with UG principles in SLA, although not in very
strong terms. The ‘back – to- UG’ position to which a learner is said to regress in the UG
model was challenged by the results.
15
Mutiti (ibid) suggested a parasitic model in which the initial UG has been filtered through
the LI experience. This means that the unmarked LI parameters may be preferred over the
L2 parameters especially during the initial stages of language acquisition. Parameters that
are completely marked or unavailable in the structure of the L2 may be transferred from
the first language. This transfer of parameters is understood as a learnability strategy that
is related to learner–speaker’s psycholinguistic disposition to acquire language. The study
agrees with the markedness differential hypothesis: areas of the TL that are more marked
than parallel areas in the subject’s first language pose considerable problems. Although
this study is different from the present study, it gave a lot of insight into the role of UG in
language acquisition, and transfer of linguistic features from LI to the TL. However, the
subjects under study came from normal hearing speakers of Gikuyu language
background. The subjects of the present study were HI pupils from different mother
tongues in an urban setting.
2.2.3 Studies Related to the Hearing-impaired
Early studies of the language of hearing-impaired people were either descriptive or aimed
to evaluate the level of language achieved at various stages in development (Bishop and
Mogford, 1993). Researchers wanted to know if the development of language in
prelingually hearing-impaired children was similar in nature to the hearing child. It was
believed that the hearing-impaired were deprived of sound stimulation during the critical
period of development (Bishop and Mogford 1993:115). Research on animals has shown
that early sensory deprivation can critically impair auditory processing abilities. Luria
(1973), as quoted in Bishop and Mogford, 1993, suggested that the functional
organization of the brain develops differently if one sensory modality is absent. Research
also shows that the difficulties experienced by the HI in developing intelligible speech
indicate that in hearing children, the development of phonology and phonetic accuracy of
articulated segments are achieved primarily through auditory means, though visual
perception of speech movements also plays part in acquisition of speech patterns in
hearing children.
Other studies on vocabulary (Bishop and Mogford, 1993) indicate that the HI children’s
vocabulary is delayed and restricted. However, the above studies do not compare the
16
errors made by the HI learners with those of the normal hearing pupils in order to know
which errors are as a result of hearing impairment. They also do not classify the lexico-
semantic errors made by the HI learners.
Anglin (1970) (as quoted in Bishop and Mogford, 1993) devised a way of examining the
organization of the lexicon. This technique involved sorting words into relating groups
and using a hierarchical cluster to represent the way in which words are related
conceptually. The depth of the resulting hierarchy indicates the degree of abstractness
within the organization. Anglin (1970) showed that there is a progressively more abstract
structure in the subjective lexicon of children as a function of age.
The above technique was used by Tweney et al (1975) to examine the organization of the
lexicon of 63 hearing and 126 hearing–impaired adolescents of between sixteen and
eighteen years of age, using two sets of items. The first experiment involved common
nouns and words referring to sounds (e.g. rustle, meow, and roar), and the second, words
with high and low imagery. The performance between the groups differed only for the
words relating to sound. In the second study, the results for both groups were comparable
in the depth of the hierarchy for hearing–impaired and hearing adolescents. Tweney et al
(ibid) concluded that there are no qualitative differences between the lexical structures of
the hearing-impaired and of hearing persons except where difference in experience with
lexical items is an important factor. Anglin (1970) and Tweney et al (1975) did not
collect their data in a second language learning environment. The L1 of their subjects was
BSL and ASL while the present study’s subjects use KSL. The above two studies did not
classify lexico-semantic errors.
Other studies on the HI have compared the writing of the HI to that of normally hearing
controls. The studies revealed differences in performance indicative of deaf subject’s
English language deficiencies (Strong 1988). Sentences written by the hearing–impaired
tend to be shorter than those written by normally hearing controls of the same age and
contain fewer conjoined and subordinated clauses. Hearing–impaired individuals also
tend to reiterate words and phrases within a discourse and use more articles and nouns
and fewer adverbs and conjunction than normally hearing individuals matched for age
17
(Myklebust, 1964 in Strong, 1988). However, these studies did not attempt to classify the
errors or determine the possible causes of these errors. The studies were also not done in
a second language learning environment.
Greenberg and Withers (in Strong, 1988) listed some of the common grammatical errors
that the hearing-impaired make. Among these errors is the recurrent use of patterns that
do not correspond with the inflectional morphology; the misuse of function words
(articles and preposition) and various other errors like incorrect subcategorisations,
inappropriate use of coordinating and subordinating conjunctions, and anomalies in
constituent structure. Kretschmer and Kretschmer (1978) say that some of these errors are
related to the factor that, traditionally, hearing-impaired children have been taught written
language sentence by sentence and not in discourse form. This study did not classify the
errors or determine the errors that could be attributed to hearing impairment.
Similar studies by Bochner (1982), Quigley and Paul (1984), (all reported in Strong
1988), show that function words and morphology pose considerable difficulty for the
hearing-impaired. These components of grammar constitute major obstacles to the
successful acquisition of written English and attainment in the proficiency in the HI
population. Articles, preposition, conjunctions, pronouns, verbal auxiliaries and
inflectional and derivational suffixes are among the most persistent and pervasive sources
of error observed in their spoken and written English and on experimental task. These
studies dealt with HI children and adults who had already acquired speech unlike the
subjects of the present study. The present study dealt with Kenyan HI learners using
KSL.
Myklebust (1965) reported that HI children do not attain the average sentence length of
eight-year-old normal hearing children until they are seventeen years old. Marshall and
Quigley (1970) also investigated the complexity of HI children's utterances based on the
length of utterance and number of subordinate clauses. They reported that, although the
HI subjects were developmentally delayed, they did show significant increases over time
and that the measures used are reliable indicators of mature language development.
18
In an extensive study of the spoken and signed English of profoundly hearing–impaired
children, Geers et. al (1984) compared spoken and signed responses in children from oral
and Total Communication schools. Their investigation confirmed earlier findings that
children with hearing impairment have a limited ability to use English structure correctly.
It showed a significant modality difference between the groups. Both groups of five to
nine-year-olds with hearing impairment were far behind normally hearing four-year-old
children in their ability to produce English sentences. Again, they experienced particular
difficulty with quantifiers, articles, pronouns, demonstratives, verb inflections, negatives,
and the copula verb. This study does not provide enough data that could be used to
explain the HI learners’ learning strategies in the use of vocabulary meaning.
Other studies have emphasized on the study of language within a pragmatic framework.
Wilbur (1977) used data collected from writing samples in earlier studies (Quigley et al.,
1976) and discovered that the HI children's difficulties with spontaneous written
expression were pragmatic rather than syntactic. For example, she noted that the problem
with determiners in their written language was not placement of a determiner before a
noun but rather the distinction of definite from indefinite, indicating an inability to use
determiners to distinguish new from old information. The HI children were therefore
unable to write coherently. This study, although different from the present study, was
helpful in explaining some of the errors the HI learners make in their written English.
Studies in word associations indicate that older HI children’s semantic fields are not
extensive or differentiated as those of normal hearing subjects. Although the HI children
seem to understand the meaning of words, they do not appreciate the interrelationships
among words that can allow them to properly place words into large conceptual
categories (see Kretschmer and Kretschmer, 1978).
Cloze procedures have been employed to explore HI children’s understanding of
syntactic/semantic constraints that apply to English sentences. Kretschmer and
Kretschmer (1978) report that HI subjects have been found to have a problem in
componential features. The HI select lexical items that even share several important
features with appropriate category item, but too frequent the lexical entry actually used
19
differs in one or two critical features which make the resulting sentence miss the mark in
meaning. It is almost as if the HI persons learn a general meaning for words, but not all
the critical dimension that govern their use with other words. They have problems with
getting the meaning of a word from its context. This is likely to explain the causes of the
lexico-semantic errors in the HI pupils’ written English. Although the subjects in the
above study used ASL (American Sign Language), the study was relevant to the present
study in understanding how the HI pupils write. However, cloze procedures alone cannot
provide enough data that can be used to infer how the HI learners acquire and use lexical
items. The subjects in the present study use KSL.
Other researchers have attempted to evaluate the encoding strategies of ASL users
engaged in traditional semantic assessment task like word associations. Putman and
Young (1962) (in Kretschmer and Kretschmer 1978:133) evaluated the ability of normal
hearing and hearing impaired adolescents to remember lists of words. The lists were
organized in four ways:
1) List of real words that were similar to one another.
2) List of real words that were dissimilar when signed (using sign language).
3) List of nonsense words that were similar in form.
4) List of nonsense words that were dissimilar in form.
The HI subjects learned all the lists more quickly than the normally hearing subjects did.
The HI subjects mastered the sign-similar real word list more easily. The reverse was true
for the nonsense lists.
These results could be interpreted that signing accelerates word memorization in HI
subjects, but that does not explain why the signed similar word list should be learned
more easily than the sign dissimilar list. The lack of vocabulary growth in HI might have
worked to their advantage in this situation (Kretschmer and Kretschmer, 1978). They
argued that a limited vocabulary could reduce potential interference from the memory or
knowledge of too many words. A limited semantic differentiation may have aided
performance rather than deterred it in this task. Another factor might have been that
similarly signed words could be filed into memory more easily because they share similar
formation characteristics or given what we know from learning theory, that dissimilar
20
items are easier to remember than similar items. Kretschmer and Kretschmer (1978) say
that similarity should have been a more difficulty condition than dissimilarity; therefore,
words thought to be similarly signed were only superficially alike, but not closely related
in form to an ASL user.
Akachi (1991) studied sentence types of Kenyan Sign Language (KSL). He worked on
the assumption that KSL has declarative, interrogative and imperative sentences, as found
in many spoken languages. He investigated how these types of sentences were formally
differentiated in KSL grammar. He claimed that a declarative sentence such as “You are
deaf,” is expressed as “DEAF YOU,” while interrogative sentences such as, “Are you
deaf,” is expressed as “DEAF YOU.”(Akachi 1991:10). It is good to note that the words
in capitals are ordinary English representing word-signs in KSL. He explained that in
sign language, the above two sentence types could be distinguished because they are
accompanied by use of ‘non-manual signals or behaviour’. These non-manual signs are
carried out simultaneously as the manual signs in the sentences. The head and shoulders
being moved forward, and eyebrows lifted, accompany the interrogative sentence.
An imperative sentence (request, command) such as “pick up the Book” is expressed as
“BOOK PICK”. It is accompanied by compressed eyebrows and constant eye contact
with the addressee, while the head and the shoulder remain in the forward position.
Akachi (1991) argued that sign language is the native language of the hearing-impaired
created by them for purposes of communication among themselves and with others. Sign
Language has structure which is independent of spoken language. The above study,
though different from the present study, enabled the researcher to understand how the
hearing-impaired use lexical items in their written grammar.
There is no international Sign Language (Adoyo 2002). There are different national Sign
languages because signs are culturally determined. There are several Sign Languages
such as the Kenyan Sign Language, German Sign Language, American Sign Language,
Zambian Sign Language, Ugandan Sign Language, Israel Sign Language, and many
others. Similarly, regional variations are manifested in Kenyan Sign Language lexicon
due to the several spoken languages that we have in Kenya. However, these variations
21
have been able to converge into a standard variety because of sociolinguistic factors (see
Okombo and Akachi, 1997). Studies by Akachi (1991), Okombo (1994), and Adoyo
(1995) show that like other Sign Languages, KSL is a formal, socially agreed-on, rule-
governed symbol system that is generative in nature. Though different in the modes of
expression, Kenyan Sign Language and other spoken languages are equivalent in their
communicative potentials.
Recent studies done in Kenya show that the HI learners have not yet acquired important
English structures in order to communicate effectively. Such studies are those done by
Wamae (2003) and Ayoo (2004). Wamae (2003) did a study on the effects of the sign
language mode of instruction on acquisition of English suffixes by hearing–impaired
form two learners of English in Butere–Mumias District. Learners in the two schools
were made to write down the sentences that their teachers were given to sign for them
(that is, use sign language). The sentences contained word affixes such as ‘ed’, ‘-ly’ and
‘-s’. The findings indicate that less than 50% of the learners got the affixes under
investigation right. Hearing-impaired students had not acquired affixes in their
vocabulary studies. The above study focused on the acquisition of suffixes in vocabulary.
It does not however provide data that can be used to investigate how the HI learners in
primary schools use lexical items, the errors they make and possible strategies for word
meaning acquisition.
The hearing-impaired pupils were also found to be disadvantaged in the learning process
due to their impairment. Ayoo (2004) studied the morphosyntactic errors in the written
English of standard eight hearing-impaired pupils. She found out that 78% of the data
collected from standard eight hearing-impaired pupils could not be described as English
structures. Hearing–impaired pupils at standard eight have not learnt or acquired parts of
speech and grammatical rules. They had errors related to parts of speech, omission,
redundant, concordial (agreement), word order, choice of word used, double use of
words, punctuation errors and expression errors. She concluded that hearing-impaired
students had not mastered many of the basic grammar rules in English. Her study offered
important information on the writing of the hearing-impaired pupils. However, Ayoo
(2004) was only interested in morphosyntactic errors while the present study analyses
22
lexico-semantic errors. A study of how the HI learners use lexical items would give more
insight into the challenges they face and the strategies they adopt in acquisition and use
of L2 vocabulary.
2.3 Theoretical Framework
The study was based on the Error Analysis (EA) approach by Corder (1967, 1974, 1981)
and Interlanguage by Selinker (1972).
2.3.1 Error Analysis
Error analysis is a type of linguistic analysis that focuses on the errors learners make. It
was after Corder’s article entitled The significance of Learner Errors (1967) that EA took
a new turn. He presented a completely different point of view from that of Contrastive
Analysis. Other articles written by Corder (1971; 1974; 1981) helped to give the study of
errors a new direction. According to Corder, the appearance of errors in a learner’s
production was evidence that the learner was organising the knowledge available to them
at a particular point in time. He said that errors were the most important source of
information, accounting for the fact that learners have a ‘built in syllabus’ and that a
process of hypothesis formulation and reformulation was continuously occurring.
EA saw errors as indicators of the learners’ current underlying knowledge of the second
language, or as clues to the hypothesis (or strategies) that a learner may be testing about
the second language (Richards, 1974; Taylor, 1975; Dulay and Burt, 1974). Like a child
struggling to acquire L1, the L2 learner also tries out successive hypotheses about the
nature of the TL. In this sense, errors provide us with insights into the language system
that L2 learners are acquiring and using at a particular period. Such an L2 system is
called interlanguage.
Corder (1967, 1974, and 1981) claimed that EA could be used as a primary pedagogical
tool because of the following:
(i) EA does not suffer from the inherent limitations of Contrastive Analysis
(CA): restriction to errors caused by interlingual transfer. EA brings to light
many other types of errors.
23
(ii) EA, unlike CA, provides data attested problems and not hypothetical problems
and therefore, forms a more efficient and economical basis for designing
pedagogical strategies.
(iii) EA is not confronted with the complex theoretical problems encountered by
CA; for example, of equating difference => difficulty and difficulty=> error.
EA was therefore appropriate for the present study because it has explicit methodology of
identifying, categorizing and analyzing errors. According to Corder (1974), the
methodology of EA consisted of the following steps.
1. Collection of data: A corpus of language is selected, for example through
composition. This involves deciding on the size of the sample, the medium to be
sampled, and the homogeneity of the sample.
2. Identification of errors in the data: At this stage, there is need to differentiate
errors from mistakes. The errors are labeled with the exact nature of deviation
from the standard form.
3. Classification of the errors identified: A grammatical description is assigned to
each error; for example, errors of articles and errors of verb forms.
4. Explanation of possible causes of errors: Attempts are made to identify the
psycholinguistic cause of the errors. For example, Richards (1974) identifies the
following strategies associated with developmental or intralingual errors. These
are overgeneralization, ignorance of rule restrictions, incomplete application of
rules, and false hypothesis of concept. These may be some of the causes of the
lexico-semantic errors in the groups under study.
5. Evaluation and pedagogic implication: The seriousness of each error is assessed
to make principled teaching decisions.
2.3.2 Interlanguage
The term “interlanguage” (IL) was coined by Selinker (1972) to refer to what
McLaughlin (1987:60) calls ‘interim grammar’ constructed by second language learners
on their way to the target language (TL). Selinker uses the term to suggest the immediate
stages between the native language (NL) and the target language (TL). This language
produced by learners is seen as a system in its own right, obeying own rules, and as
dynamic system evolving over time. IL postulates that learners pass through a number of
24
stages with the goal of achieving target language proficiency. Nemser (1971) called these
stages ‘approximative systems’ and Corder (1971) called them ‘idiosyncratic dialects and
transitional competence’.
The IL is thought to be distinct from both the learners’ first language and the target
language. Nemser (1971:116) states clearly the assumptions underlying IL as:
At any given time, the approximative system is distinct from the L1 and the L2; the approximative systems form an evolving series; and that in any given contact situation, the approximative systems of learners at the same stage of proficiency roughly coincide.
Ellis (1994:114) asserts that IL has been dominant in second language acquisition. He
further says that IL can refer to both the internal system that a learner has constructed at a
single point in time and the series of inter-locking systems which form what Corder
(1967) called the learner’s ‘built-in syllabus’ (i.e. the Interlanguage continuum).
Selinker (1972) explains IL as the result of the learner’s creativity as he processes the
input data of the TL. It is the learner’s creativity that makes interlanguage a system in its
own right as the learner’s creativity deviates from the learner’s mother tongue and from
the target language. This creativity accounts for the structures that will be referred to as
errors in this study.
Selinker (1972) argues further that this learner’s creativity is aided by some latent
psycholinguistic structures that are activated when one attempts to learn a second
language. The learner attempts to produce meanings that he already has in a second
language that he is in the process of learning. Selinker (1972) argued that the IL, which
he saw to be a separate linguistic system resulting from the learner’s attempted
production of the TL norm, was the product of five central processes involved in second
language learning. These processes include:
1. Language transfer, which is defined as interlanguage performance that is because
of the mother tongue.
2. Transfer of training, which shows features of training procedures.
3. Strategies of second language learning, which are as a result of an identifiable
approach by the learner to the material being learnt.
25
4. Strategies of second language communication, which deal with an identifiable
approach by the learner to communicate in the target language.
5. Overgeneralization of TL linguistic materials, which involves overgeneralising
TL rules.
The errors of the hearing impaired and the comparison group could occur as a result of
the five processes mentioned above. The five processes together constitute the ways in
which the learner tries to internalize the L2 system. Ellis (1985:48) says that the above
five processes are the means by which the learner tries to reduce the learning burden to
manageable proportions. Widdowson (1975b) as quoted in Ellis (1985:48) refers to it as a
concept of ‘simplification’. The L2 learner cannot cope with the complexity of a
language system and therefore limits the number of hypothesis he tests at any one point
in time.
However, according to Selinker (1972) many L2 learners, about 95%, fail to reach TL
competence. They stop learning when their interlanguage contains at least some rules
different from those of the TL system. He refers to this process as fossilization.
Fossilization occurs in most language learners and cannot be remedied by further
instruction. The learners tend to keep in their interlanguage productive performance no
matter what the age of the learner or the amount of the instruction he receives in the TL.
The fossilization mechanism accounts for the phenomenon of the regular re-appearance
of interlanguage forms which are thought to have been learnt in earlier stages. The
learner’s interlanguage ceases to develop however long he is exposed to relevant data in
the TL (Selinker 1992).
Interlanguage theory is therefore appropriate in analyzing errors that focus on second
language users. However, the theory does not have a clearly laid down procedure on how
to identify and analyze errors. Selinker proposes the five central processes that he claims
are the main causes of error but he does not show how errors can be identified and
classified. EA will therefore be used in the identification and description of errors.
Adjemian (1976) suggests that Error Analysis approach can be first used before inferring
26
the psycholinguistic mechanisms at play in the production of a given erroneous structure.
Both EA and IL complement each other in analysis of data in the present study.
2.4 Conclusion
In this chapter, we looked at the literature review relevant to this research and the
Theoretical framework. The literature review revealed a gap in the writing of the hearing-
impaired learners that needed to be filled. There are studies in the normal hearing learners
that have dealt with lexico-semantic errors. There are also studies on KSL structure,
Morphosyntactic errors in the Written English of HI learners, and effects of the sign
language mode of instruction on acquisition of English suffixes. However, these studies
did not focus on how the hearing-impaired learners use vocabulary and the type of lexico-
semantic errors they make. There is also no study that has compared the lexico-semantic
errors made by the hearing-impaired learners with those of normally hearing learners to
identify the errors that could be attributed to hearing impairment.
The study was based on the Error Analysis (EA) approach by Corder (1967, 1974, 1981)
and Interlanguage by Selinker (1972). EA was instrumental in identification,
classification and description of the lexico-semantic errors. This is because of its explicit
methodology of identifying, categorizing and analyzing errors. Interlanguage theory
helped in explanation of the possible causes of errors by inferring the psycholinguistic
mechanisms at play in the production of a given erroneous structure.
27
CHAPTER THREE
RESEARCH METHODOLOGY
3.1 Introduction
This chapter consists of the research design, study population, location, sampling
procedures, instruments for data collection and data analysis.
3.2 Research Design
This study was conducted through causal comparative research design, which is also
referred to as Ex post Facto research design. This research design entails studying causes
after they have had their effect on another variable (Mugenda and Mugenda 2003). It also
involves treatment by natural selection rather than manipulation (Oso and Onen 2005).
The study is therefore not comparative or experimental in nature because it uses a
comparison group (HP) only to identify the lexico-semantic errors that the HI learners
make as a result of their condition. The research design was appropriate for this study
because it allows comparison of groups without having to manipulate the independent
variables.
The study was based on data collected from the written texts of 30 hearing-impaired (HI)
pupils and 30 hearing pupils in standard six, seven and eight. The hearing pupils in this
study formed the comparison group. Descriptive statistics for each group’s data (HI and
the comparison group) was computed using the SPSS (Statistical Package for Social
Science - version 11.5). The five steps of Corder’s Error Analysis were used to analyze
the Lexico-semantic errors. IL theory was used to infer the psycholinguistic mechanism
at play in the production of the erroneous lexical items in the present study. Several t-
tests were done to show whether there was a significant statistical difference between the
means of the lexico-semantic errors made by the two groups. Analysis of variance was
done to show if there was any significant difference among the lexico-semantic errors of
the three classes of the HI pupils.
3.3 Population and Location of the Study
The target population was all HI and hearing pupils in standard six, seven and eight in
Kenya. The accessible population was 60 pupils: 30 drawn from Ngala Special School
28
for the Deaf and 30 from St Paul’s primary. Both schools are in Nakuru town. The
hearing pupils’ classes formed the comparison group for the study. Standard six, seven
and eight were chosen because they have acquired/learnt English for two, three and four
years respectively. They are therefore able to write in English. The class population in the
hearing-impaired school was made of a small number of pupils (between nine and
twelve) and therefore the researcher included standard six and seven to get enough
samples for the study. Researchers in the written work of hearing impaired pupils, such as
Ayoo (2004) had problems in getting enough samples for data analysis, as most of the
written work by the HI was not readable. The inclusion of standard six and seven was
therefore necessary for the researcher in the present study to come up with enough
sample size for the study.
Both schools were chosen because they had the relevant subjects needed for data
collection. Ngala Special School for the Deaf is the only school for the hearing-impaired
in Nakuru District. St Paul’s primary school is nearer to Ngala special school than the
other primary schools. Both schools share the same urban setting.
3.4 Sampling Schools
Purposive sampling technique was used to get the location of the study as well as the
schools where data was to be collected. In this case, Ngala Special School for the Deaf in
Nakuru District was chosen. It is the only school for the hearing-impaired in the District.
St Paul’s primary school, which is a few meters away from Ngala Special School, was
purposively sampled. Both Schools had the required subjects for data collection.
3.5 Sampling Students
The present study targeted standard six, seven and eight pupils in two primary schools. 30
HI pupils and 30 hearing pupils were sampled from Ngala Special School and St Paul’s
primary school respectively. In both schools, stratified random sampling was used. The
pupils in standard six, seven and eight were divided into two groups: girls and boys.
Simple random sampling was then used to select ten pupils per class in each school with
equal gender representation. This totaled to 30 students per school.
29
3.6 Instruments for Data Collection
Three tests were given to all the students. These were a free composition, (Appendix A) a
picture story, (Appendix B) and a cloze passage (Appendix C). The three different modes
of testing were helpful in getting a broader spectrum of the learners’ lexical competence.
Corder (1974:126) says that, “We should be aware that different types of written material
may produce a different distribution of error or a different set of error types”.
He categorizes written work into spontaneous productions, like a free composition and
controlled production, like a cloze passage. Controlled production is error provoking
while spontaneous production is error avoiding since the learner can easily avoid what he
or she is not sure of. The picture story is a combination of the two, in that, it allows pupils
to create own word yet limits them to what is in the picture.
The picture story was first given to teachers two weeks before the test to discuss with the
pupils. This was to enable the pupils understand what the pictures depicted clearly and
help the pupils in making lexical choices to describe a situation which they already
understood. The pupils were tested during their English lesson to fit in the school
timetable. Teachers of English in the selected schools and specific classes were requested
to help in administering the three tests. This was to make the students more relaxed and
to take the tests more seriously.
3.7 Data Analysis and Interpretation
The three tests done by the pupils were read and the five steps of Error Analysis were
used to analyze the lexico-semantic errors:
a) To identify the errors, the researcher read the three tests done by the two groups
under study, underlined the lexico-semantic errors and counted them.
b) The identified errors were categorized into error types. The researcher adapted the
approach of “let the errors determine the categories” (Norrish 1983) and
“Linguistic Category Taxonomy” (Dulay et al 1982) to categorize and present the
lexico–semantic errors. Identification and categorization of the lexico-semantic
errors helped to achieve the first objective: Identify and describe the lexico–
semantic errors made by HI in their written texts.
30
c) The researcher established the possible causes of the errors. Selinker’s five central
processes of Interlanguage were helpful in accounting for the errors.
d) The researcher evaluated the lexico-semantic errors to determine which lexico-
semantic errors affected the learner’s performance most.
The researcher calculated the percentage of each error type in every pupil and in every
group. The SPSS computer package (Statistical Package for Social Science) was used to
compute means, frequencies, standard deviation, t-test and Analysis of variance.
Analysis of variance and t-test are statistical techniques or tests for continuous data and
are used to compare means.
The t-test is a special case of the ANOVA. This procedure tests the significance of the
difference between two sample means. There are two versions of the t-test: related or
correlated t-test and unrelated or uncorrelated t-test, which is also referred to as
Independent-Samples t-test. The related or correlated t-test is used when the two sets of
scores to be compared come from the same sample of people, while the unrelated or
uncorrelated t-test is used when the scores to be compared come from two different
samples of people (Howit and Cramer 2003).
The present study used the Independent-Samples t-test because of the number of groups
in the study. Independent-Samples t-test procedure compares means for two groups of
cases. It was therefore appropriate in comparing the lexico-semantic errors of the HI
group to those of the hearing group. ANOVA would also have been used to compare the
means of the two groups but because of the sample size, the t-test procedure was more
appropriate. Analysis of Variance (ANOVA) is a data analysis procedure that is used to
determine whether there are significant differences between two or more groups or
samples (Mugenda and Mugenda 2003:137).
The One-Way ANOVA procedure produces a one-way analysis of variance for a
quantitative dependent variable by a single factor (independent) variable. It is used to test
the hypothesis that several means are equal. This technique is an extension of the two-
sample t-test (Neter et al 1990). In the present study, one way ANOVA was used to
31
compare the means of the three classes of HI pupils (standard six, seven, and eight). This
was done to establish if there was a significant difference in the lexico-semantic errors
made by the three HI classes. It also enabled the research to know if the H.I were
acquiring vocabulary meaning. When using the SPSS, both techniques display descriptive
statistics for each test variable, a test of variance equality, and a confidence interval for
the difference between the two variables (95% for the present study).
According to Sommer and Sommer (1991), Milliken and Johnson (1992), Mugenda and
Mugenda (2003), and Howit and Cramer (2003), the above two statistical techniques have
the following assumptions:
1. The scores in each group are normally distributed, that is, they should
approximately follow the normal curve. This means that the sample has to be
randomly selected from the population. If the curve is skewed and assumption
number three is met, Analysis of Variance and t-test can still be used.
2. The variance in each group is the same. This assumption is also referred to as the
homogeneity of variance. It means that the squared values of the standard
deviation should be about the same. However, the violation of this assumption has
little effect if the sample sizes are equal. The sample sizes in the present study
were equal (30 HI and 30 hearing pupils).
3. The sample is also assumed to have come from a population in which the
dependent variable is normally distributed. This assumption was met because
random sampling was used to sample the groups of pupils.
The t-tests enabled the researcher achieve objective number two, that is, to establish if
there was a significant difference between the lexico- semantic errors made by the
hearing-impaired pupils and the hearing pupils in their written text.
The lexico-semantic errors of the HI pupils group were linguistically compared to those
of the hearing pupils’ group. What the researcher did was to find out through linguistic
analysis and description how the errors of the two groups differed. This was done to
identify the errors that are idiosyncratic to hearing impairment. ANOVA (Analysis of
Variance) was done for the errors that were made by the HI group in the three classes in
32
order to get the variation. Both the linguistic analyses and ANOVA enabled the
researcher achieve objective number three, that is, to determine the lexico-semantic errors
that can be attributed to hearing impairment, and how these errors varied across the three
HI pupils classes(six, seven and eight).
33
CHAPTER FOUR
RESULTS AND DISCUSSION
4.1 Introduction
This chapter includes the results of the pupils’ data analysis and interpretation with close
reference to the objectives of the study. The chapter is divided into four parts. The first
part describes the lexico-semantic errors identified in the data collected from the hearing-
impaired and the hearing pupils’ written English. The second part discusses the
significant difference between the lexico-semantic errors made by the two groups under
study. The third part discusses the hearing-pupils’ lexico-semantic errors and their
variation. The last part gives a general discussion of the findings of the research.
4.2 Identification, Description, and Discussion of the Lexico-Semantic Errors
The first hypothesis stated that ‘Hearing-impaired pupils exhibit lexico-semantic errors in
their use of the English language in written texts’. We are now going to identify, classify
and describe the lexico-semantic errors made by the learners under this study.
The researcher read the three tests done by the two groups of pupils under study. The
researcher used the five steps of Error analysis mentioned in chapter 3. This was done in
order to achieve the first objective. The researcher classified the collected errors into the
following error types; learning induced errors, Meaning similarity errors, Meaning
duplication errors, Semantic contiguity errors, Collocation errors, Sound similarity errors,
Coinage errors, Paraphrase errors, and Haphazard errors. After classification, the
researcher did a linguistic analysis and description of the various types of lexico-semantic
errors in order to identify errors that were idiosyncratic to hearing impairment. A brief
discussion of the errors identified has been given at the end of the identification and
description of the lexico-semantic errors made by the two groups.
The above Lexico-semantic error types were common in the three tests; however, they
varied in quantity. The table below gives a summary of the lexico-semantic errors
identified from the three tests done by the two groups under study. A more detailed table
is in appendix D
34
Table 1: A Summary of the Total Number of Errors
Key: H.P – hearing pupils
H.I – hearing impaired pupils
GROUP ERROR TYPE
HP HI
TOTAL PERCENTAGE
1. Learning induced errors 1282 1851 3133 48.07 %
2. Haphazard errors 9 1666 1675 25.70%
3. Collocation errors 340 382 722 11.08%
4. Sound similarity errors 316 34 350 5.37 %
5. Semantic contiguity errors 123 165 288 4.42 %
6. Meaning similarity errors 99 48 147 2.26%
7. Meaning duplication errors 67 35 102 1.57 %
8. Paraphrase errors 24 34 58 0.89 %
9. Coinage errors 19 23 42 0.64 %
TOTAL 2279 4238 6517 100
The above table shows that there is a significant difference in the quantity of errors
produced by the two groups. There were 6517 lexico-semantic errors in total. The HI
group led with 4238 (65.03%) followed by the HP with 2279 (34.97%). The different
types of lexico-semantic errors found in the data collected from the two groups under
study have been discussed below. All the examples given originate from the data
collected from the two groups under study. Some of the data collected is in the
35
appendices. In the examples given, the italicized sentences have the errors while the
sentences in brackets are the reconstructed correct form in English.
4.2.1 Learning Induced Errors
These errors showed improper learning or inadequate learning of the rules of the second
language. Some of these errors were because of overgeneralization, incomplete
application of rules, and ignorance of rule restrictions, system simplification, and
exploiting redundancy. There were 3133 Learning induced errors made by the two groups
under study. This accounted for 48.07% of the total number of lexico-semantic errors
collected from the two groups in this study. The HI made 1851, while the HP group made
1282 Learning induced errors.
The learners created deviant lexical items based on their experience of the lexical items
and the structures in the target language (English). Most of these errors were as a result of
the learner reducing his linguistic burden. Examples of such errors found in the data
collected from the three tests given to the two groups under study are as follows:
(The source of examples 1-6 is the data collected from the HP group, while examples 7-
20 are from the HI group.).
1) My father get out and sit outside so that she can rest.
(My father got out and sat outside so that he could rest.).
In example 1, the hearing pupil (HP) did not use the correct tense and the correct gender.
The pupil should have used the past tense of the word get to show that the action took
place in the past. Similarly, the masculine gender (He) should have been used because it
refers to the subject of the sentence (My father), which is in the masculine gender.
2) The pharmacist gave the man drugs and show her how he would using
them.
(The pharmacist gave the man some drugs and showed him how he
would be using /use them)
Example 2 has omission of the past tense morpheme –ed; wrong choice of pronoun (used
her instead of him to mark masculine gender); and failure to use the verb be before using
to mark an action that will be taking place in the future. In example 3 and 4, the pupils
36
marked the past tense twice: didn’t baked and didn’t knew instead of didn’t bake and
didn’t know. In both examples, tense was marked in both the auxiliary verbs and the
main verbs.
3) My mother didn’t baked a cake for my birth day.
(My mother didn’t bake a cake for my birthday).
4) He didn’t knew where he was.
(He didn’t know where he was)
5) She praid and the party started.
(She prayed and the party started).
In example 5, the pupil thought that all verbs that end in –ay should have the y changed
into i, then add –d to mark past tense as in the word pay, whose past tense is paid. The
pupil overgeneralised this rule for marking past tense in some irregular verbs, and applied
it on the verb pray, which is a regular verb.
6) You don’t have no malaria, said the doctor
(“You don’t have malaria,” said the doctor).
In example 6, negation was marked twice by using don’t (do not) and no. Although
example 6 is American English and is appropriate in some native dialects, it is an error
because the Kenya Primary Syllabus advocates the use of the British English as the
standard variety. “It is wrong in standard English to include more than one word in a
sentence, clause, or verb phrase that negates that element” Princeton Language Institute
(1993:96).
Learning induced errors made by the HI group were characterized by the following:
i. Omission of lexical items that marked certain semantic features.
Examples: My father get out and sit outside so that she can rest.
My birthday my happiest day
ii. Omission of tense and omission of copular verb.
Examples: He is smile
The man go the hospital
The patient taking medicine and water
Teacher thank also all mens
37
iii. Omission of determiners.
Examples: Sick man go home
Teacher thank also all mens
iv. Overgeneralization of tense-marking morphemes.
Examples: We sleeped under a tree.
Kamau cake cuted clap children
v. Failure to mark possession, gender and number.
Examples: Mother car break down
One children had no any gifts
vi. Wrong use of preposition, pronouns, adjectives, adverbs, and auxiliary verbs as
seen in the examples given below.
After for two weeks the patient was fine he was healthy
He is must go to work
From the that day the patient ate and drank the medicine
Got going to at home
…just in case the snake is poisoning.
Tense is a semantic feature and it is marked by inflection of the verb (Lyons 1977:386).
He says that tense, number, mood and gender are associated with particular kinds of
semantic function. All the above grammatical categories rely on the lexemes or lexical
items to mark such features. Failure to mark the categories correctly results to not only
the wrong lexical meaning but also the wrong sentence meaning. The HI group under
study had problems with marking gender by using the correct choice of lexical items.
In example 7 below, the HI pupil may have wanted to mean ‘one child or some children
had no gifts’ but used ‘one children… and no any’. The lexical item one, precedes a
singular countable noun when used as a determiner. No and any, are both central
determiners and cannot be used together as it leads to redundancy in meaning (see Quirk
and Greenbarm, 1973). Failure to use these lexical items correctly is an indication that
the HI pupils did not know their usage and meaning.
7) One children had no any gifts.
(One child did not have a gift. OR
38
Some children did not have any gift(s))
Overgeneralization of –ed morpheme for marking past tense was observed in the writing
of the HI pupils as in the example 8. Where the HI pupils marked tense in verbs, they
used –ed regardless of whether the verb was regular or irregular. However, there were
few such cases because the HI pupils did not mark tense in most of their work as in
examples 10 and 11. They used the bare form of the verb. In example 9, the HI learner
may not have been aware that the verb cut is an irregular verb. The HI learners did not
use conjunctions in their writing as in example 9.
8) We sleeped under a tree.
(We slept under a tree).
9) Kamau cake cuted clap children
(Kamau cut the cake and the children clapped)
10) Mother car break down
(My Mother’s car broke down)
11) Peter mango eat.
(Peter ate a mango)
The HI learners did not use determiners such as possessive pronouns, demonstrative
pronouns and articles as in example 10 and 11. They omitted the determiner my to mark
possession in the lexical item, Mother. The placement of the indefinite article a was
omitted as in example 11.
12) …just in case the snake is poisoning.
(…just in case the snake is poisonous).
Example 12 was derived from the cloze passage. The HI pupil was supposed to fill in the
blank using an adjective formed from the noun poison.
13) School good for you.
(Schooling is good for you. / The school is good for you)
14) My birthday my happiest day
(My birthday is my happiest day)
15) I happy to saw many things.
(I was happy to see many things)
39
The HI pupils rarely used copula verbs in both present and past tenses as in examples 13-
15. The verbs, is, and its past tense forms were omitted in the above examples. Either, the
HI pupils thought they had communicated enough by using content words only
(simplification), or they were ignorant in the use of the copula verbs with other verbs to
mark tense.
Hearing impaired (HI) and hearing pupils learning induced errors differed in several
ways. Although in both groups there was omission of lexical items that mark certain
semantic features, this was more witnessed in the HI errors. The sentences of the HI had
multiple errors ranging from missing articles, prepositions, conjunctions, pronouns,
inflection and derivational suffixes. In other cases, the HI used content words only such
as in the example 9 below.
(9) Kamau cake cuted clap children.
(Kamau cut the cake and the children clapped)
It can be concluded that the HI group under study had not yet learned well the meaning
and usage of some function words such as prepositions, pronouns and verb auxiliaries.
In most cases, they used the bare form of the verb as in the example below emanating
from the hearing-impaired pupils’ data.
16) I was go Nairobi.
(I went to Nairobi / I was going to Nairobi).
17) Yesterday mother bake cake to my birthday
(My mother baked a cake for my birthday).
The pupil may have failed to mark tense in ‘bake’ because of the adverb of time
‘yesterday’ and therefore reducing his linguistic burden.
18) Teacher thank also all mens
(The teacher thanked all the men)
The learner in the above example was operating in the rule that the plural s is used with
all nouns. HI learners' writing portrayed ignorance of rule restrictions or incomplete
application of rules, as in the example 10 given earlier.
Mother car break down
(My Mother’s car broke down)
40
In this example, the learner did not know how to mark possession. The word break in the
phrase break down was not marked for tense by use of –ed.
19) I feel sleep but had to opened my eyes
(I felt sleepy but I had to keep my eyes open)
In example 19, the learner did not mark tense in the verb feel (felt). He also used the noun
sleep as an adjective, instead of sleepy.
20) Man ask have problem
(The man asked, “Do you have a problem?”)
In example 20, the learner failed to use the article ‘the’ to mark known and unknown
information. Most of the HI learners did not use direct and indirect speech correctly. For
example, it is impossible to know who asked the question or who was asked the question
in example 20.
The HI learners did not use conjunctions in their writing as in example 9.
Kamau cake cuted clap children
(Kamau cut the cake and the children clapped)
41
Table 2: A Summary of Learning Induced Errors
A. Semantic features omitted in lexical items
HP HI Total
1. Tense 273 635 908
2. Possession 49 50 99
3. Gender 54 63 117
5. Number 52 111 163
B. Lexical items omitted 1. Determiners 97 329 426
2. Conjunctions 31 201 232
3. Copular verbs 53 137 190
4. Main verb 0 34 34
5. Prepositions 12 70 82
C. Double marking of semantic features 1. Double negation 220 1 221
2. Double marking of tense
188 1 189
D. Wrong lexical form
1.Derivational errors
76 71 147
2. Verb forms errors
52 100 152
3. Adjectival form errors
57 19 76
4. Adverb form errors
68 29 97
Totals
1282 1851 3133
42
4.2.2 Meaning Similarity Errors
These errors reflected the learners’ inability to use words that have similar meaning in
their appropriate context. Some of the learners in this study did not know that it is not in
all contexts that synonyms can be interchanged. Partial synonymy (near synonyms) is
sometimes responsible for such types of errors. There were 147 Meaning similarity
errors. This accounted for 2.26 % of the total number of lexico-semantic errors collected
from the two groups in this study. The HP had 99, and the HI 48 Meaning similarity
errors. Examples of such errors identified in the data collected from the two groups under
study have been given below. Examples 21 to 25 emanate from the hearing pupils’ data.
21) The birthday possess many of my friends.
(The birthday party had many of my friends)
22) When the dry season reached.
(When the dry season came/started)
23) He came home and drank two spoons of the medicine.
(He came home and took two spoons of the medicine).
24) The doctor wrote a letter and gave the sick man to take to the nurse.
(The doctor wrote a note / prescription and gave the sick man to take
to the nurse).
In example 21, the learner may have wanted to say that, there were many of his friends at
his birthday party, but used the word possess. This word has the synonyms; to have or
own. According to Longman Dictionary of Contemporary English, possess also means
“to influence someone so completely as to make them do something foolish”. The word
possess in this context can therefore not be interchangeable with the word to have.
Palmer (1981) says that only true synonyms are interchangeable in all their environments
but partial or close synonyms are interchangeable in certain environments only. This
seems to be the problem that both groups under study faced in making lexical choices to
express meaning.
In example 23, the learner used the word drank (drink) because the medicine may have
been in liquid form. Both words are close in meaning as in take/drink a soda. Although
the two words may overlap in meaning, each word covers a wide semantic area and there
43
are ways in which they may not overlap as in example 23. The verb take overlaps with
the meaning of drink and is therefore more appropriate to refer to introducing into the
body by swallowing, eating, drinking and breathing (Longman Dictionary for
contemporary English). In example 24, the pupil’s interlanguage allows him to use the
word letter to refer to anything written on a piece of paper, instead of a note or a
prescription. Similar problems noted in example 25 below. Rules and instructions have a
loose sense of synonymy but cannot be interchangeable in the example below.
25) The doctor told the sick man to follow the rules written on the
medicine.
(The doctor told the sick man to follow the instructions written on the
medicine bottle).
The examples below (26-30) are all from the data collected from the H.I group.
26) Swimming we looked cinema
(After swimming, we watched a movie)
27) We saw everything had lost.
(We found everything that we had lost).
28) We completed food that serve my sister.
(We finished the food that we were served by my sister).
29) He tell her he chest wounded inside accident.
(He told her that his chest was injured in an accident).
30) The manager talk me welcome
(The manager said to me, “Welcome.”)
HI pupils had similar problems in using words that were similar in meaning. The HI
group showed lack of understanding that synonyms are not interchangeable in all
contexts. For instance, in example 26, whereas look and watch may be interchangeable in
look after the children and watch the baby. Look cannot be used when we mean, “Watch
the movie.” Similar explanation goes for example 28. Complete and finish are
interchangeable in she completed/finished the race in good time. Nevertheless, they are
not interchangeable in A marriage that has a child is complete.
44
Saw and found may be true synonyms in the HI pupil’s interlanguage, but in English they
are not interchangeable as used in Example 27. Found is the correct word to refer to
discovering something that had gotten lost. Found and lost collocate but see and find do
not. In example 29 (He tell her he chest wounded inside accident), the pupil did not
understand the meaning and the use of the word wounded and therefore misused it in that
sentence. Wounded and injured are close in meaning but differ in that, wounded refers to
bodily harm done on purpose with a weapon, while injured means bodily harm done
accidentally (Longman Dictionary of Contemporary English). Failure to use closely
related words correctly shows that the HI pupils had not mastered the functional meaning
of these words.
4.2.3 Semantic Contiguity Errors
These errors reflected the learners’ use of a single lexical item that shares certain
semantic features with the target item. The learners used a word in the target language
that provided an approximate translation of the unknown concept by referring to a similar
but known item. There were 288 Semantic contiguity errors. This accounted for 4.42% of
the total number of lexico-semantic errors collected from the two groups in this study.
The HI had 165, and the HP 123. Examples 31-35 are from the data collected from the
HP group while 36-39 are from the data collected from the H.I group.
31) Their father and mother go them round the farm.
(Their father and mother took them round the farm).
(Their father and mother got them round the farm).
The pupil had the concept of moving round the farm and therefore used the word go
instead of took in example 31, which was taken from the cloze passage. This shows that
the pupil was aware of the semantic features of the appropriate target lexical item but he
did not have it in his interlanguage. Take someone round posed difficulty to the pupil
because it is a phrasal verb. The learner may also have wanted to use went with and used
go instead despite the fact that only one word was supposed to be put in the blanks in the
cloze passage. This is the structure the learner may be familiar with. It is also possible
that the learner may have wanted to use got but used go instead.
32) Mrs. Ng'ang'a asked her man
45
(Mrs. Ng'ang'a asked her husband)
In example 32, the learner used man instead of husband. Both words have semantic
resemblance. Husband means a man who is married to someone. This pupil was relying
on his semantic knowledge of the TL.
33) …and the doctor talk, “what is wrong?”
(…and the doctor asked, “what is wrong?”)
34) The lady talking to the man, “what is the matter?
(The lady asked the man, “What is the matter?)
35) It was enjoyable to entered an aeroplane.
(It was enjoyable to board an aeroplane).
In examples 33 and 34, the learners used the lexical items talk and enter in lieu of ask
and board’. Both pairs of words share their semantic properties but are not true synonyms
in English. The same explanation can be given for example 36, with ‘say’ and told,
example 38, for talk and told, and in example 40, for tell and ask. The HI pupils have not
yet mastered the meaning and usage of these words.
36) She say come tomorrow me.
(She told me to come tomorrow).
37) He was going better.
(He was becoming better).
38) Man talk doctor he sick.
(The man told the doctor that he was sick).
39) Class six all visit and my birthday party there.
(The whole of class six came to/ attended my birthday party).
40) The woman tell man feeling well
(The woman asked the man whether he was feeling well).
In example 37, the pupil meant to say that the subject (he) was recovering or improving
from a sickness. He used going better instead of becoming better. In all the examples
given, there was a close semantic resemblance between the erratic and the correct lexical
item, an indication that the learner was making semantic approximation to convey a
certain concept in L2.
46
4.2.4 Collocation Errors
These errors reflected the learners' lack of knowledge of collocation patterns or the
syntagmatic aspect of lexical items. Collocation is observed between lexical items in the
structure of texts. It is the meaning relation between individual lexical items and the ones
that habitually co-occur with them in a language. There were 722 collocation errors. This
accounted for 11.08% of the total number of lexico-semantic errors collected from the
two groups in this study. The HI had 382 while the HP had 340 collocation errors.
Examples of such errors from the data collected in this study are given below. Examples
41-47 are from the data collected from the HP group, while examples 48-59 are from the
data collected from the HI group. Example 41 shows improper mastery of prepositions
that collocate with verbs such as flow to show movement relationship. Other collocation
errors in the use of prepositions by the HP showed improper mastery of prepositions that
collocate with the expression of time and space in English as in examples 42 and 43.
Examples 44 and 45 show that the learners failed to use the correct lexical items to make
the similes, as beautiful as a palace and as loud as a church bell, to be complete.
41) The poison will flow up with the blood
(The poison will flow out with the blood)
42) You will be okay into hospital.
(You will be okay in hospital).
43) I was born on 1995
(I was born in 1995)
44) The house was as beautiful like a palace
(The house was as beautiful as a palace).
45) His heart started beating loudly as a church bell.
(His heart started beating as loud as a church bell).
46) We either eat and sleep.
(We either eat or sleep).
47) We heard the rang of the church bell.
(We heard the ringing of the church bell).
47
In example 46, the HP used either with and, instead of either…or. Either…Or are
correlative conjunctions that are used to join words and group of words with equal weight
in a sentence. The learner had not yet learned that either collocates with or. In example
47, the learner broke the mutual expectancies that hold between words. The learner knew
that rang and bell collocate but used them in the wrong context. Rang, the past form of
the verb ring refers to the act of causing a bell to sound and not the sound itself as the
writer used it above. The use of word hear indicates that the writer may have intended to
refer to the sound and not the act. However, the learner may have used ‘back formation’-
from verb to noun to form the sentence we heard the ring of the church bell. He may
have been confused with the verb form and used rang instead of ring.
Some of the HI pupils’ collocation errors were similar to those made by the HP group.
For instance, examples 48-53 shows that the HI pupils had not mastered the meaning of
prepositions that collocate with the expression of time, space and instrument in English
language. The HI learners used two prepositions in examples 52 and 53. One preposition
was wrong while the other was correct.
48) All people on our village community meet.
(All the people in our community were having a meeting).
49) His father soon took Mbaabu to hospital with a Matatu.
(His father soon took Mbaabu to hospital by/in a Matatu).
50) My birthday in 12th December.
(My birthday was on 12th December).
51) We travel bus and loud music.
(We traveled on a bus with loud music).
52) She go home in by car.
(She went home in a car/by car).
53) Woman prayed to for the sickman.
(The Woman prayed for the sick man).
Some lexical items used to make some grammatical expressions were left out or
substituted with another word that does not co-occur with the word that preceded. The HI
pupils made more of these types of collocation errors. In examples 54, the HI learner
48
used the conjunction and instead of the word like or the construction as…as to compare
the sweetness of the cake to that of honey. The word and was used wrongly in example
55 (full and instead of: full of) and in example 56 (and table instead of on the table). In
example 55, the learner may have used additive and to mean with. It is possible that the
use of and in the learner’s source language could be extended to with.
54) Cake sweet and honey.
(The cake was as sweet as honey).
(The cake was sweet like honey).
55) My party full and children.
(My party was full of children).
56) She put cake and table cut.
(She put the cake on the table and cut it).
In example 57, the HI learner used the wrong preposition while in example 58; there was
an omission of the preposition of after a lot.
57) She went for sleep.
(She went to sleep).
58) He feel alot pain.
(He feels/felt a lot of pain).
59) We saw every thing had lost.
(We found everything that had got lost).
60) The monkeys bark we laugh all.
(The monkeys chattered and we all laughed).
There was overlooking of co-occurrence restrictions of certain lexical items by the HI
pupils as in example 59. When an object that had got lost is recovered, the words lost and
found are used rather than lost and seen. Example 60 shows that the HI learners were not
aware of the sounds animals make. Instead of using the word chattered for the sounds
made by monkeys, the learner resorted to most general term: bark, the sound made by
dogs. The learner may have had difficulty with this word (chattered) although learning
the sound made by animals is part of their syllabus at primary four (see K.I.E 2000).
49
4.2.5 Sound Similarity Errors
These errors were reflected in the learners’ misspelling of lexical items due to their
inability to differentiate meaning and usage of similarly sounding lexical items. There
were 350 Sound Similarity errors from the two groups under study: 316 for the HP group,
and 34 for the HI group. This accounted for 5.37 % of the total number of lexico-
semantic errors collected from the two groups in this study. There were two different
manifestations of Sound similarity errors found in the data collected from the writing of
the two groups.
The first one shows that the learners were confused on the use of certain
homophonous words. The HP group made most of these errors with 323. Examples of
such errors found in the data collected from the two groups under study have been given
below. Examples 61-65 were made by the H.P group while examples 66-70 were made
by the HI group. The correct lexical item and the erroneous lexical item sound alike or
are close in pronunciation but differ in meaning as examples 61 (list, least), 62 (rite,
right), 65 (by, buy) for the HP group, and examples 66 (such, search), 67 (choose, chose),
and 68 (week, weak) for the HI group. The words in examples 63, 66 and 67 have
different pronunciations in Standard English, but to the learners, they may sound the
same in their interlanguage.
61) At list, we were not late.
(At least we were not late).
62) Mutua feels well even rite now.
(Mutua feels well even right now.)
63) Kamwira was my very best friend in the all village.
(Kamwira was my best friend in the whole village).
(Kamwira was my best friend in the village).
64) I wore my best cloths that morning.
(I wore my best clothes that morning).
65) Here is the medicine you should by.
(Here is the medicine you should buy)
66) You will such for this medicine in a chemist.
(You will search for this medicine in a chemist).
50
67) I choose the good flowers.
(I chose the good flowers).
68) The sick man very week
(The sick man was very weak)
69) …because my mother was not their.
(…because my mother was not there)
70) The nurse told him to take three tablets par day.
(The nurse told him to take three tablets per day).
The second case is one in which the words used were not necessarily the homophones of
the intended ones, but they still sound similar to the learners, perhaps owing to their
inaccurate pronunciation or lack of familiarity with the words. Examples 71-75 are from
the HP groups’ data while examples 76-80 are from the HI groups’ data.
71) I am feeling so tied
(I am feeling so tired)
72) Mr. Kamau woke up filling unwell.
(Mr. Kamau woke up feeling unwell).
73) The doctor came with a new white full scape
(The doctor came with a new white foolscap).
74) I herd to thank God for you
(I had to thank God for you).
75) So that world animals from the Mulika Game Park do not come right
to our house.
(So that wild animals from the Mulika Game Park do no come right to
our house.
The HI lexical items were more distorted than those of the HP group were. Some of the
words they used instead of the correct words do not exist in English. The HP used an
existing lexical item that was close in pronunciation with the correct word.
76) The doctor told the man he was suffering from taiphoid
(The doctor told the man he was suffering from typhoid).
77) The parent wolks house.
51
(The parent works in the house).
78) The doctor see microscope you have promble there.
(…you have a problem there)
79) Nairobi visit betiffull.
(My visit to Nairobi was beautiful)
80) Just as the sun was ricing…
(Just as the sun was rising…)
Although the HI pupils cannot hear, they had errors related to the pronunciation of lexical
items. There is a possibility that some of the HI were post-lingual or were partially deaf.
Partially deaf learners’ language development follows the normal pattern; however, they
require special facilities such as hearing aids for education. HI pupils lip-read and this
shows their ability to understand sounds (phonemes), although they cannot hear them.
This is one of the causes of their sound similarity errors. Poor spelling or confusion of
lexical items that were similarly spelt cannot be ruled out in this category of errors. Such
lexical items also had similar or close pronunciation. The HI pupils’ inability to hear
contributed to their fewer errors in this category. This is because of Sign language having
different signs for similarly pronounced words. The hearing pupils had difficulties in the
use of homophonous words or words that had close pronunciation. The errors made in
this category of lexico-semantic errors show that both groups of learners in this study had
not mastered the meaning of lexical items with similar pronunciation.
4.2.6 Meaning Duplication Errors
This category of errors was because of learners’ inability to decide on the appropriate
word to convey a meaning. The reason may be that the learners have two or more lexical
items but they are unable to decide which word to use and therefore end up using two
words. This could also be a result of concept transfer. This results to redundancy or
meaning duplication that makes the sentence ineffective. A total of 102 Meaning
duplication errors were identified in the data. This accounted for 1.57% of the total
number of lexico-semantic errors collected from the two groups in this study. The HP
group had 67, while the HI group had 35 Meaning duplication errors. Examples 81-87
are from the HP groups’ data while examples 88-97 are from the HI groups’ data.
52
81) Can you give me this medicine and the syrup?
(Can you give me this medicine?)
82) Now Ramwangwa was now feeling well and better.
(Ramwangwa was now feeling better).
83) I was very so happy on my birthday.
(I was very happy on my birthday).
84) He looked and searched for three hours
(He searched for three hours).
85) His mouth had a bad stink because of not brushing his teeth.
(His mouth had a bad smell / breath because of not brushing his teeth).
(His mouth had a bad stench because of not brushing his teeth).
In example 81, the word syrup may also be implied in the word medicine. We have
medicine in tablet form and in liquid form. We may also have cough and colds medicine
in form of thick sticky liquid called syrup. Using both words in the sentence results to
repetition or redundancy in meaning. The word now has been repeated in example 82.
This duplication is a developmental stage in SLA. The learner knows that now can be
used in different positions in a sentence but forgot to use it only once. In the same
sentence, the HP used better and well. Better is the comparative form of well and should
not be used together to avoid redundancy in meaning. Other instances of meaning
duplication errors are Very so happy, looked and searched in examples 83 and 84
respectively. Use of bad stink in example 85 is also repetitive. According to The
Longman Dictionary of Contemporary English, the word stink means unpleasant or bad
smell. The word bad is therefore implied in the word stink. The use of both words (bad
and stink) leads to redundancy in meaning.
The HI learners had similar meaning duplication errors except in examples 90 -93 where
there was repetition of the same word to mark duration of the action or state. Use of
returned back, inside and in, more better than, and happy and happy in examples 86-89
are repetitions that result to redundancy in meaning. These repetitions show that the HI
group did not know the meaning and usage of these words.
86) When he returned back home.
53
(When he returned home).
87) (Inside the car my dad was in)
(Inside / in the car, my dad was.
My dad was inside / in the car).
88) My birthday party was more better than my brother’s
(My birthday party was better than my brother’s)
89) I was very strong girl because I was happy and happy
(I was a very happy girl because I was very strong)
Examples 90 –93 differ from those of the HP group. Their errors had word duplication
that resulted to redundancy in meaning.
90) I was happy happy.
(I was very happy).
91) She walk fast fast
(She walks very fast).
92) We dance dance.
(We danced a lot).
93) It was nice nice nice.
(It was very nice)
94) The man walk go to hospitil
(The man walked / went to hospital.)
In example 94, the HI pupil used two verbs with similar meaning. Walk means “use one's
feet to advance; advance by steps”, and go ‘change location; move, travel, or proceed”. In
both words, there is an aspect of moving or changing location and therefore the two
words should not be used together to avoid redundancy in meaning.
4.2.7 Paraphrase Errors
This category of errors reflected the learners’ use of second language lexicon and
structure to give a descriptive equivalent of the first language. In some cases, the learners
described the characteristics or the elements of the object or the action instead of using
the appropriate target language structure. This indicates that the learners were not
54
familiar with the appropriate lexical item(s). Although paraphrase which involved
circumlocution may not have led to an error, it was an indication of the learners’
inadequate lexical competence. A total of 58 Paraphrase errors were identified from the
data collected from the two groups under study. This accounted for 0.89% of the total
number of lexico-semantic errors collected from the two groups in this study. The HP had
24, and the HI group, 34.
Examples 95-99 are from the data collected from the HP group while 100-107 came from
the HI group’s data. From the examples given, it is evident that the two groups of learners
used paraphrase to avoid a more difficult lexical item. For instance, the HP group did not
know the single lexical item for my aunt’s child, a male hen in examples 96 and 97
respectively. Similar explanation can be given for examples 98 and 99 where lexical
paraphrase took the form of circumlocution. The HP learners gave a description of the
desired lexical item as in example 99, or gave a definition of the unknown lexical items
as in example 98.
95) His brain was not functioning the way it was supposed to function
(His brain was not functioning well / He was insane / mad).
96) We went to the beach with my aunt’s child.
(We went to the beach with my cousin).
97) We slaughtered a male hen for my birthday party.
(We slaughtered a cockerel for my birthday party.)
98) All the teachers were in the class they stay when not teaching when
it happened.
(All the teachers were in the staffroom when it happened).
99) We made noise with our mouth so that she can hear us and stop.
(We shouted to her/whistled so that she can hear us).
The HI group used similar strategies in paraphrasing as it can be seen in examples 100-
106 derived from the data collected. In example 100, the HI pupil did not know the
correct lexical item for food tasted nice. He avoided the unknown word and paraphrased.
This is a form of lexical simplification.
55
100) They cooked the food was very testing very nice.
(They cooked a delicious meal.)
(They cooked tasty food.)
(They cooked food that tasted nice)
101) My uncle buy car tyre for use when one tyre spoil
(My uncle bought a spare wheel)
102) He saw small animals in the blood.
(He found germs in the blood).
103) We saw the game of football.
(We watched football).
104) (She left this boy without everyone else.)
(She left the boy alone).
(She deserted the boy).
105) She was speaking and talking without anybody.
(She was talking to herself)
106) He sat on the chair for resting.
(He sat on the sofa set/arm chair).
In example 101, paraphrase took the form of circumlocution by the learner giving a
description of the term spare tyre. Transfer of learning or direct translation was the cause
of the error in example 102. The HI learner referred to germs as small animals. The
learner’s poor learning in Swahili might have been transferred to the learning of English
where the learner referred to viini (germs) as wadudu wadogo (small animals).
The HI learner paraphrased by giving a description of the characteristics or function of
the unknown lexical item as in examples 104, 105, and 106. In example 104 (She left this
boy without everyone else.), the learner may have wanted to say the subject deserted the
boy, or she left him alone. The learner did not have a word for a comfortable chair for
resting like a couch, sofa set or an armchair as in example 106.
Both groups under study used paraphrase as a form of lexical simplification when faced
by an unknown word. They gave the description, characteristics or function of the
intended referent. Such paraphrases were correct in the learners’ interlanguage.
56
Knowledge of semantic relation and componential features of the intended referent was
important for both groups in order to paraphrase.
4.2.8 Coinage Errors
These errors were as a result of the learners under study creating a new lexical item
which is nonexistent in the target language to convey an intended meaning or a desired
concept. There were 42 coinage errors; 19 made by the HP group, and 23 by the HI. This
category of errors accounted for 0.64% of the total number of lexico-semantic errors in
this study. The examples below were identified in the data collected from the two groups
under study. Examples 107-113 were made by the HP group, and examples 114-125 by
the HI group.
107) She broke the sugar cup.
(She broke the sugar dish).
108) A very tall fiercing boy was caught in our class.
(A very tall brutal boy was caught in our class)
(A very tall fierce looking boy was caught in our class)
109) When the eatings was over, we all went home.
(When the eating/feasting was over, we all went home ).
110) She was sadded when she saw her sick husband sitting outside.
(She was saddened when she saw her sick husband sitting outside).
111) Be a generous boy not a selfisher.
(Be a generous boy, but not a selfish one).
112) They were going to celebrationing their holiday.
(They were going to celebrate their holiday).
113) The chemist keeper gave him the drugs.
(The chemist/ pharmacist gave him the drugs)
The HP learners constructed the words fiercing, eatings, sadded, and selfisher, in
examples 108,109,110, and 111 respectively to express a desired concept. The learners
created a L2 lexical item by selecting a conceptual feature of the target item and
incorporating it in the L2 morphological system. The learners’ interlanguage may have
allowed them to create the word fiercing as an adjective to refer to a violent, brutal or
57
ruthless boy. Eatings was used to refer to the act of consuming food, but it was
inappropriate in the context. The appropriate word would have been eating, feasting, or
banqueting. From the adjective sad, the learner intended to create an adjective with a
similar meaning sadded, in his interlanguage. Selfisher and chemist keeper in examples
111 and 113 were created by the learner out of false analogy as in the words, teacher
from the verb teach, and shopkeeper from shop and keeper.
The HI pupils portrayed similar coinage strategies like their hearing counterparts. They
created new lexical items; some that were in existence in the TL and used them
inappropriately, and others that were nonexistence. In example 114, the HI created a
compound noun faretaker, by combining the noun fare and taker. In the HI pupil’s
interlanguage, this word means someone who collects fares on a public vehicle.
Similarly, the HI pupils lacked the correct lexical item to use with the nouns microscope
in example 115, and stethoscope, in example 116. The two lexical items exist only as
nouns but not as verbs as the learner has used them.
114) The faretaker forgot.
(The conductor forgot).
115) The Doctor microscope test the blood.
(The Doctor used a microscope to examine the blood)
116) Doctor stethoscope him chest
(The Doctor used a stethoscope to examine his chest).
In other instances, the HI group created new lexical items by adding an affix to words in
the TL to produce an inflected or derived form as in, happyingly, sicky, Niced, goodly,
badly and misfully, in examples 117,118,119, 120, 121 and 122 respectively.
117) I was very happyingly for my sister.
(I was very happy for my sister).
118) Man very sicky man
(The Man was very sick/ He was a very sick man
119) My brother wear very niced clothes.
(My brother wears very beautiful/smart/nice clothes).
120) Dad smile goodly.
58
(My dad smiled well / nicely).
121) Malaria is a badly disease.
(Malaria is a bad disease).
122) I love you misfully.
(I love you and miss you)
Some of the HI learners created lexical items that exist in English but their meaning was
inappropriate in the context they were used. For instance, Died (past tense of die) has
been used as an adjective in example 123. This is an indication of the HI learners’ lack
of knowledge of the parts of speech in English. Similar explanation can be given for
examples 124 and 125 below.
123) The died man not my father
(The dead man was not my father).
124) I was so happiest again for another thing.
(I was so happy again for another gift/present).
125) My uncle came to celebrate born day.
(My uncle came to celebrate my birthday).
Both groups of learners under study used similar strategy in coining new lexical items in
order to communicate a desired concept. Some of the HI coinage errors however differed
from those of their hearing counterpart. Where the HI lacked the correct lexical item, they
misused a lexical item of a different grammatical category.
4.2.9 Haphazard Errors
These errors do not have any relation with language developmental patterns. They do not
follow any laid down patterns of development. These errors showed total ignorance or no
learning at all on the side of the hearing-impaired pupils. They were characterized by
illogical use of lexical items, and use of lexical items that were neither English nor any
language that was known to the researcher. There were 1675 haphazard errors. (See table
1). This accounted for 25.70% of the total number of lexico-semantic errors collected
from the two groups in this study. The HI had 1666. There were only nine haphazard
errors made by the HP group. These nine errors came from the cloze passage. The HP
filled in the blanks in the cloze passage with lexical items that were not semantically
59
close to the correct answer. Their failure to make such lexical errors in the picture story
or in the free composition was an indication that they may have misunderstood the
meaning of the sentence in the cloze passage, and therefore ended up using the wrong
lexical item.
The HI pupils had plenty of these errors in all the three tests given to them. These errors
were an indication that they had not yet mastered vocabulary meaning and usage. The
errors made by the HI group in all the three tests were a clear indication that:
a) The HI had not yet acquired enough vocabulary for communication.
b) They could not use the few words they had acquired in the correct context.
c) They used the known lexical items in a manner that did not reflect any learning
strategy or developmental pattern known to the researcher.
Examples of errors made by the HI pupils in the data collected for the present study have
been given below. Examples 126 to 131 are from the HI data collected from the cloze
passage. When compared to the correct answer, there is no close relationship between the
incorrect word and the correct word. In other cases, it was impossible to know exactly
what the learner was communicating.
126) Mboroki, Mbaabu and Nkirote were excited about going to Embu
their holidays on their new farm.
(Mboroki, Mbaabu and Nkirote were excited about going to spend
their holidays on their new farm).
127) I will give Mbaabu some anti-snake serum, just in case the snake is
impotoweni.
(I will give Mbaabu some anti-snake serum, just in case the snake is
poisonous).
128) After Mbaabu’s arm had been Mbaroki…
(After Mbaabu’s arm had been bitten..)
129) The thick undergrowth of creepers made progress Jembes
(The thick undergrowth of creepers made progress difficult)
130) They to spot porridge and tea for their breakfast and changed into
please clothes to work on the farm led by their father, Mr.
Ng'ang'a.
60
(They took porridge and tea for their breakfast and changed into
work clothes to work on the farm led by their father, Mr. Ng'ang'a).
131) We had admite attend.
(We had to attend).
It was not possible to know what the HI learners meant in haphazard errors such as the
ones given below (132-142).
132) Previous he starnd alive
133) I am in live there your Name Kabaranet.
134) I am happiest there many joy.
135) Wait one nelh new body had.
136) Smile and family happy full friend.
137) Love many miss.
138) My class six Ngala School for the deaf headteacher work storing.
139) The parent walk house food teen bitter back wilieat by much.
140) Arrived surpurital running hug finish.
141) Well problem happiest will future.
142) I was aeoptane made to fly from safari
From the examples given it is clear that some of the HI were not only poor in writing but
also poor in comprehension. This is evident from their writing in the cloze passage. They
read the cloze passage, failed to understand and filled in the blank spaces with words that
were grammatical/semantically unrelated to the correct word.
The paragraph below drawn from one of the compositions further illustrates the HI
pupils’ ungrammatical word order, failure to adhere to verb argument structure, total
ignorance of grammatical rules and poor or inadequate acquisition of English (see
appendix E).
I have beny many good I am who worty to go mather othre people see happy kay keny on there why they holle?, i am happted to you mother sorry go to home car i my good other Then you mother that father no worry and sorry you that happy aeroplane euruplem noun a ma chine that has wing and can fly an aeroplane or plane lands kend go to euaplem.
61
The HP haphazard errors were difficult to describe, as many of them were semantically
unrelated words that were carelessly put in a sentence without any punctuation. Other
lexical items were neither in English nor in any other language known to the researcher.
4.3 Summary
It is evident from the data collected that both groups under study make lexico-semantic
errors. The lexico-semantic errors identified were classified into nine categories. These
include learning induced errors, Meaning similarity errors, Meaning duplication errors,
Semantic contiguity errors, Collocation errors, Paraphrase errors, Sound similarity errors,
Coinage errors, and Haphazard errors. The errors varied from one group to another. The
HI group made more lexico-semantic errors than the HP group. The HI learners’ errors
displayed lack of knowledge of parts of speech, wrong use of lexical items, inability to
mark tense, gender, number and possession, poor grammatical word order, omission of
determiners, prepositions, conjunctions and auxiliary verbs, and poor use of synonyms.
Their inability to hear led to late acquisition or poor acquisition of language unlike the
HP group. Transfer, ignorance, overgeneralization, and redundant use of lexical items are
some of the causes of the HI learners’ lexico-semantic errors.
4.4 Difference between the HP and HI Learners’ Lexico-Semantic Errors
This section discusses the statistical analyses that were done to enable the researcher test
the second hypothesis. The second hypothesis stated that there is a significant difference
between the lexico-semantic errors made by the hearing-impaired and the hearing pupils.
Several t-tests were done to compare the means of the lexico-semantic errors made by the
two groups. A program called SPSS was used to compute the t-tests. The t-tests produces
two types of tables: one showing a summary of the group statistics and a t-test table.
In this study, the researcher used the probability value (p-value). The column labeled
(p-value) displays a probability from the t distribution with 58 degrees of freedom (df).
In order to follow the t-test analysis in this study, it is important to understand the
following.
62
Degrees of freedom (df) is the first sample size minus one plus the second sample. There
were two samples of equal size in this study, that is, 30 HI and 30 HP. The df for this
study is therefore: (30-1) + (30-1) = 29+29= 58. Degrees of freedom (df) for this study
will be displayed by SPSS. The study used a 95% confidence interval. This means that all
the tests of significance would be executed below 0.05 significant level. If the p-value (on
the column labeled Sig. (2-tailed), is lower than 0.05 significant level or value, it means
that there is a significant difference between the means of the lexico-semantic errors
made by the HI and HP. If it is higher than the 0.05 significant level, then the p-value will
be insignificant and therefore, no significant difference between the means of the lexico-
semantic errors made by the HI and HP learners.
When SPSS presents the p-value as 0.000, we present this as 0.001 since the exact level
is not given. This p-value enables the researcher to reject or accept the hypotheses that
had been put forward for the study. We are now going to analyze the results of the t - t
est done in order to test the second hypothesis of this study. The following are the results
per each lexico-semantic error category.
4.4.1 Meaning Similarity Errors
The group statistics show that the HP group made more Meaning Similarity errors than
the HI group. The HP group also had a greater variability than the HI group. The mean
and SD for the HP was 3.30 and 2.02 respectively. The mean and SD for the HI group
was 1.60 and 1.163 respectively (see table 3). These statistics show that the HP group
was more heterogeneous in performance.
The reason for such results may be use of Sign language to teach vocabulary. In English,
words with several meanings have the same spelling but in Sign language, different signs
can be used to separate each meaning (Coryell and Holcomb, 1977; Toth, 2002; Krashen
and Terrell, 2000). Sign language can also use inflections to signify different meaning in
two words that have meaning similarity such as the words ‘huge and big’ (Small and
Cripps, 2002). This may be the reason why the HI pupils made fewer errors than the
hearing pupils did and why their meaning similarity errors did not vary very much.
63
Computation of the t-test for the Meaning Similarity errors made by the two groups under
study yielded a p-value of 0.001 (see table 4). When compared to the significance level of
0.05, it was found to be very significant. There was therefore a significant difference
between the HP and HI pupils’ Meaning Similarity errors. These results were in line with
the second hypothesis of this study that states that, there is a significant difference
between the Meaning Similarity errors made by the HI and the hearing pupils. The
researcher therefore accepted the second hypothesis. Although it had been hypothesized
that the difference would be positive for the HP, it is noted that the vice versa is true.
Table 3: A Summary of Group Statistics for the Meaning Similarity Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation
Std. Error Mean
Hearing pupil 30 3.30 2.020 .369 Meaning similarity
Hearing impaired 30 1.60 1.163 .212
64
Table 4: T-Test Table for the Meaning Similarity Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the difference
Lower Upper Equal variances assumed
12.991 .001 3.995 58 .001 1.70 .425 .848 2.552
Equal variances not assumed
3.995 46.317 .001 1.70 .425 .844 2.556
4.4.2 Meaning Duplication Errors
The mean for the HP was 2.23, and the HI group, 1.17. The SD for HP was 1.870, and for
the HI, 0.986. Like in the meaning similarity errors, the HP group made more errors, and
was the more heterogeneous group. The HI group was the more homogeneous (see table
5). The HP learners attempted more sentences than the HI learners did. This explains why
the HI learners made a smaller number of errors.
Computation of the t-test for the Meaning Duplication errors made by the two groups
under study yielded a p-value of 0.008. This was found to be very significant when
compared to the significant level of 0.05 (see table 6). There was therefore a significant
difference between the HP and HI pupils’ meaning duplication errors. These results were
in line with the second hypothesis of this study that states that, there is a significant
difference between the Meaning Duplication errors made by the HI and the hearing
pupils. The hypothesis was therefore accepted.
65
Table 5: A Summary of Group Statistics for the Meaning Duplication Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation Std. Error Mean
Hearing pupil 30 2.23 1.870 .341
Meaning duplication Hearing impaired 30 1.17 .986 .180
Table 6: T-Test Table for the Meaning Duplication Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
13.397 .001 2.764 58 .008 1.07 .386 .294 1.839
Equal variances not assumed
2.764 43.961 .008 1.07 .386 .289 1.844
4.4.3 Semantic Contiguity Errors
The HI made more semantic contiguity errors than the HP. The mean for the HP was
4.10, and for the HI, 5.50. The S.D for the H.P group was 2.578, and for the HI, 1.815
(see table 7). The SD revealed that the HP was the more heterogeneous group in this
category of semantic contiguity errors. Their scores were more widely distributed within
the group. The hearing pupils’ group might have had some pupils who were extremely
poor and others extremely good. The HI was the more homogenous group in this
category of lexico-semantic errors. Although the HP attempted more sentences than the
66
HI, they made fewer errors compared to the HI. This is an indication that the HP group is
better in lexico-semantic competence in this category of errors.
Computation of the t-test for the total number of the semantic contiguity errors made by
the two groups under study yielded a p-value of 0.018. When compared to the significant
level of 0.05, it was found to be significant (see table 8). There was therefore a significant
difference between the HP and the HI pupils’ semantic contiguity errors. These results
were in line with the second hypothesis of this study that states that, there is a significant
difference between the Semantic Contiguity errors made by the HI and the hearing
pupils. The hypothesis was therefore accepted.
Table 7: A Summary of Group Statistics for the Semantic Contiguity Errors Made by HP and HI Learners
Hearing ability
N
Mean
Std. Deviation
Std. Error Mean
Hearing pupil 30 4.10 2.578 .471
Meaning duplication
Hearing impaired 30 5.50 1.815 .331 Table 8: T-Test Table for the Semantic Contiguity Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
2.986 .089 -2.432 58 .018 -1.40 .576 -2.552 -.248
Equal variances not assumed
-2.432 52.076 .018 -1.40 .576 -2.555 -.245
67
4.4.4 Coinage Errors
The HI pupils made more coinage errors than the HP as seen on their means on table 9
below. The mean for the HP was 0.63, and for the HI, 0.77. The HP were better in
performance in this category of lexico-semantic errors despite the fact that their
compositions were longer the HI groups. The SD for the HP was 0.964, and for the HI,
0.858. The HP group was more heterogeneous in performance. The scores for the HP
group were widely distributed with some students having higher scores and others low
scores. The HI was therefore the more homogeneous group.
Computation of the t-test for the Coinage errors made by the two groups under study
yielded a p-value of .574 (see table 10). When compared to the significant level of 0.05, it
was found to be statistically insignificant. There was therefore no significant difference
between the HP and HI pupils’ Coinage errors. These results were not in line with the
second hypothesis of this study that states that, there is a significant difference between
the Coinage errors made by the HI and the hearing pupils. The hypothesis was
therefore rejected.
Table 9: A Summary of Group Statistics for the Coinage Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation Std. Error
Mean Hearing pupil 30 .63 .964 .176 Coinage errors Hearing impaired 30 .77 .858 .157
68
Table 10: T- test table for the coinage Errors Made By HP and HI Learners
4.4.5 Learning Induced Errors
The group statistics for the Learning Induced errors show that the HI made more errors
than the HP (Hearing pupils). The mean for the HI was 61.70, and for HP was 42.73. The
SD for HP group was 15.503, and for the HI group, 7.853 (see table 11). This indicates
that the HP learners were the more varied group (heterogonous group). The learners in
the HP group might have performed differently; some with very many errors and others
with very few errors.
Computation of the t-test for the Learning Induced errors made by the two groups under
study yielded a p-value of 0.001 (see table 12). When compared to the significant level of
0.05, it was found to be significant. There was therefore a significant difference between
the HP and the HI pupils’ Learning Induced errors. These results were in line with the
second hypothesis of this study that states that, there is a significant difference between
the total number of learning induced errors made by the HI and the hearing pupils.
The hypothesis was therefore accepted.
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-
value Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
.110 .742 -.566 58 .574 -.13 .236 -.605 .338
Equal variances not assumed
-.566 57.232 .574 -.13 .236 -.605 .339
69
Table 11: A Summary of Group Statistics for the Learning Induced Errors Made by HP and HI Learners
Hearing ability
N
Mean
Std. Deviation
Std. Error Mean
Hearing pupil 30 42.73 15.503 2.830
Learning induced
Hearing impaired 30 61.70 7.853 1.434
Table 12: T-Test Table for the Learning Induced Errors Made by HP and HI Learners
4.4.6 Paraphrase Errors
The group statistics for the Paraphrase errors showed that the HI group made more
Paraphrase errors than the HP group. The HI group mean was 1.13, and the HP, 0.80 (see
table 13 below). These statistics show that the HP used lexical items to express
themselves better than the HI group. They were therefore more competent than the HI
group in this category of lexico-semantic errors. The SD for the HP group was 1.064, and
for the HI group, 0.973. These results showed a close homogeneity in performance
because the SD of the two groups did not differ very much.
A p-value of 0.210 was yielded after the computation of the t-tests for the Paraphrase
errors made by the two groups under study (see table 14 below). The p-value was not
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
16.799 .000 -5.978 58 .001 -18.97 3.173 -25.318 - 12.616
Equal variances not assumed
-5.978 42.962 .001 -18.97 3.173 -25.365 -12.568
70
significant when compared to significant level of 0.05. There was therefore no significant
difference between the Paraphrase errors made by the two groups under study. These
results were not in line with the second hypothesis of this study that states that, there is a
significant difference between the Paraphrase errors made by the H.I and the hearing
pupils. The hypothesis was therefore rejected.
Table 13: A Summary of Group Statistics for the Paraphrase Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation
Std. Error Mean
Hearing pupil 30 .80 1.064 .194 Paraphrase errors
Hearing impaired 30 1.13 .973 .178
Table 14: T-Test Table for the Paraphrase Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p- value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
.035 .853 -1.266 58 .210 -.33 .263 -.860 .194
Equal variances not assumed
-1.266 57.549 .210 -.33 .263 -.860 .194
4.4.7 Sound Similarity Errors
The mean for the HP group was 10.53, and for HI group, 1.13. The HP made more Sound
Similarity errors than the HI. The SD for the HP group was actually 6.704 and for HI
group 1.008; the more varied group having the higher SD (see table 15). The HP group
71
had pupils of too extremes: some who scored very high, and others who scored too low.
The HP learners were sampled from different classes and streams taught by different
teachers. This may be the reason for the high variability of performance in the group.
The computation of the t-test for the Sound Similarity errors made by the two groups
under study yielded a p-value of 0.001 (see table16). The p-value was significant when
compared with the significant value of 0.05. There was therefore a significant difference
between the Sound Similarity errors made the two groups under study. These results were
in line with the second hypothesis of this study that states that, there is a significant
difference between the Sound Similarity errors made by the HI and the hearing pupils.
The hypothesis was therefore accepted.
Table 15: A Summary of Group Statistics for the Sound Similarity Errors Made By HP and HI Learners
Hearing ability N Mean Std. Deviation
Std. Error Mean
Hearing pupil 30 10.53 6.704 1.224 sound similarity errors
Hearing impaired 30 1.13 1.008 .184 Table 16: T-Test Table for the Sound Similarity Errors Made By HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
34.549 .000 7.594 58 .001 9.40 1.238 6.922 11.878
Equal variances not assumed
7.594 30.311 .001 9.40 1.238 6.873 11.927
72
4.4.8 Haphazard Errors
The group statistics shows that the HI group made more Haphazard errors (with a mean
of 55.53) than the HP group. The HP group's mean was 0.50. However, the HP
Haphazard errors were from only one testing made: the cloze passage. There is a
possibility that the HP group carelessly used the wrong lexical choice to fill in the blanks
in the cloze passage. This may explain why they did not make similar errors in the other
two tests. The SD for the HI group was 10.618, and for the HP group, 0.794; the more
varied group having the higher standard deviation (see table 17 below). The HP group
was more homogeneous in lexical semantic competence.
Computation of the t-test for the Haphazard errors made by the two groups under study
yielded a p-value of 0.001 (see table 18). When compared to the significant level of 0.05,
it was found to be very significant. There was therefore a significant difference between
the Haphazard errors of the HP and the HI pupils. These results were in line with the
second hypothesis of this study that states that, there is a significant difference between
the Haphazard errors made by the HI and the hearing pupils. The hypothesis was
therefore accepted
Table 17: A Summary of Group Statistics for the Haphazard Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation Std. Error Mean
Hearing pupil 30 .30 .794 .145 haphazard Hearing
impaired 30 55.53 10.618 1.939
73
Table 18: T-Test Table for the Haphazard Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper
Equal variances assumed
74.224 .000 -28.413 58 .001 -55.23 1.944 -59.125 -51.342
Equal variances not assumed
-28.413 29.325 .001 -55.23 1.944 -59.207 -51.259
4.4.9 Collocation Errors
The group statistics for the Collocation errors on Table 19 show that the HI group made
more errors than the HP. The mean for the HP group was 11.33, and for the HI group,
12.73. The SD for the total number of Collocation errors was 3.585 for the HP group, and
2.477 for the HI group. The HP group was the more varied group. Computation of the t-
test for the Collocation errors made by the two groups yielded a p-value of 0.084. (See
table 20). This p-value was found to be statistically insignificant when compared to the
significant level of 0.05. There was therefore no significant difference between the
collocation errors of the HI and the HP group. These results were not in line with the
second hypothesis of this study that states that, there is a significant difference between
the Collocation errors made by the HI and the hearing pupils. The hypothesis was
therefore rejected.
74
Table 19: A Summary of Group Statistics for the Collocation Errors Made by HP and HI Learners
Hearing ability N Mean Std. Deviation
Std. Error Mean
Hearing pupil 30 11.33 3.585 .654
collocation errors
Hearing impaired 30 12.73 2.477 .452
Table 20: T-Test Table for the Collocation Errors Made by HP and HI Learners
Levene's Test for Equality of Variances
t-test for Equality of Means
F Sig. t-calc. df p-value
Mean difference
Std. Error difference
95% Confidence Interval of the Difference
Lower Upper Equal variances assumed
3.391 .071 -1.760 58 .084 -1.40 .795 -2.992 .192
Equal variances not assumed
-1.760 51.546 .084 -1.40 .795 -2.997 .197
4.5 Analysis of Variance for HI Errors
The Analysis of variance (ANOVA) was done to test whether there was a significant
difference in the total number of lexico-semantic errors made by the three HI classes in
the three tests administered to them. To know the variation among the three classes (six,
seven, and eight), the researcher used the p-value yielded after computation of ANOVA,
the mean, and the SD (Standard Deviation).
The df (the degrees of freedom) for the ANOVA is (10-1) + (10-1) + (10-1) =27. Just like
in the T-test, the probability value (p-value) was used in interpretation of ANOVA. If the
p-value is less than the significant value of 0.05, there will be a significant difference in
75
the lexico-semantic errors made by the three hearing impaired classes (class six, seven,
and eight). If the p-value is more than 0.05, the value will be insignificant and therefore
no significant difference in the lexico-semantic errors made by the three hearing impaired
classes.
4.5.1 Meaning Similarity Errors
The group statistics on table 21 indicates that the performance of the three classes varied
slightly. Classes six and eight had a mean of 1.50, while class seven had 1.08. The SD
revealed that class seven was the most heterogonous followed by class six. This suggests
that there was a bigger variability in the scores of the Meaning Similarity Errors in class
seven than the other two classes. Some pupils obtained higher scores (had more errors);
others obtained lower scores (had less errors). Class eight was the most homogenous
class in lexico-semantic competence.
The computation of ANOVA for the total number of Meaning Similarity errors made by
the HI group in the three testing modes yielded a p-value of 0.812 (see table 22). When
compared to the significant level of 0.05, it was found to be insignificant. There was
therefore no significant difference among the Meaning Similarity Errors of the three HI
classes.
Table 21: A summary of group statistics for The HI learners' Meaning Similarity errors
N Mean Std. Deviation Minimum Maximum
class 6 10 1.50 1.269 0 3 class 7 10 1.80 1.476 0 4 class 8 10 1.50 .707 1 3 Total 30 1.60 1.163 0 4
76
Table 22: ANOVA table for The HI learners' Meaning Similarity errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups .600 2 .300 .210 .812 Within Groups 38.600 27 1.430 Total 39.200 29
4.5.2 Meaning Duplication Errors
The group statistics indicates that the means of the three classes varied slightly (see table
23). Class seven had the highest mean of 1.60, followed by class six, 1.00, and class
seven, 0.90. These statistics show that the HI pupils in class seven made more meaning
duplication errors. The reason may be that the class seven HI pupils had more problems
with lexical choices and ended up using two lexical items that were similar in meaning.
Class eight pupils seemed to be aware of synonymy and therefore avoided reduplication.
The SD for the meaning duplication errors of the HI three classes showed variation with
class six being the most heterogeneous class. Class eight was the more homogenous
group in lexical semantics. The HI meaning duplication errors varied within the groups.
However, the variation of the means was not statistically significant as seen on table 24.
The computation for the ANOVA for the total number of the HI pupils’ meaning
duplication errors yielded a p-value of 0.235 (see table 24). When compared to the
significant level of 0.05, it was found to be insignificant. There was therefore no
significant difference in the total number of meaning duplication errors made by the three
HI classes (six, seven, and eight).
Table 23: A summary of group statistics for The HI learners' Meaning Duplication errors
N Mean Std. Deviation Minimum Maximum class 6 10 1.00 1.155 0 3 class 7 10 1.60 .966 1 4 class 8 10 .90 .738 0 2 Total 30 1.17 .986 0 4
77
Table 24: ANOVA table for The HI learners' Meaning Duplication errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 2.867 2 1.433 1.530 .235 Within Groups 25.300 27 .937 Total 28.167 29
4.5.3 Semantic Contiguity Errors
The means of the three classes (see table 25) did not differ very much (5.20 for class six;
6.20 for class seven; and 5.10 for class eight). Like in the Meaning Duplication errors,
class seven made more Semantic Contiguity errors. The SD for the three HI classes
differed considerably. Class eight was the most heterogeneous class with an SD of 2.183.
Class seven was the most homogeneous class with an SD of 1.317. The SD for class six
was 1.814.
Class seven HI pupils may have had more knowledge of semantic features of lexical
items, which enabled them to use lexical items that shared semantic features with the
target lexical item. As said earlier in the discussion of semantic contiguity errors, these
pupils used lexical items that shared semantic components but were not true synonyms in
English. However, such lexical items were true synonyms in the HI pupils’ interlanguage
semantic system.
The computation of ANOVA for the total number of Semantic Contiguity errors made by
the HI pupils in the three classes yielded a p-value of 0.337. (See table 26). When
compared to the significant level of 0.05, it was found to be statistically insignificant.
There was therefore no significant difference in the Semantic Contiguity errors among
the three classes of the HI group.
78
Table 25: A summary of group statistics for The HI learners' Semantic Contiguity errors
N Mean Std. Deviation Minimum Maximum
class 6 10 5.20 1.814 2 8 class 7 10 6.20 1.317 3 8 class 8 10 5.10 2.183 2 9 Total 30 5.50 1.815 2 9
Table 26: ANOVA table for The HI learners' Semantic Contiguity errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 7.400 2 3.700 1.134 .337 Within Groups 88.100 27 3.263 Total 95.500 29
4.5.4 Coinage Errors
The computation of ANOVA for the total number of Coinage errors made by the HI
pupils in the three classes (6, 7 and 8) yielded a p-value of 0.959 (see table 28). When
compared to the significant level of 0.05, it was found to be insignificant. The means of
the coinage errors in the three classes did not differ much. Class eight, and six had the
same means of 0.80 while seven had 0.70 (see table 27). Class 8 was the most
heterogeneous class with an S.D of 1.135, while class seven was the most homogeneous
class with an S.D of 0.675. The S.D for class six was 0.789. This showed that there was a
great variability in the coinage errors of HI pupils in class 8 followed by class 6, and
lastly, class seven. Class 7 and 8 HI pupils made more use of word coinage strategy in
creating a new interlanguage word in order to communicate a desired concept.
79
Table 27: A Summary of Group Statistics for the HI Learners' Coinage Errors
N Mean Std. Deviation Minimum Maximum CLASS 6 10 .80 .789 0 2 CLASS 7 10 .70 .675 0 2 CLASS 8 10 .80 1.135 0 3 Total 30 .77 .858 0 3
Table 28: ANOVA Table for the HI Learners' Coinage Errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups .067 2 .033 .042 .959
Within Groups 21.300 27 .789 Total 21.367 29
4.5.5 Learning Induced Errors
The descriptive statistics show that the HI Learning Induced errors decreased from class
six to class eight. This shows that as the HI pupils advanced to a higher class, they gained
more lexico-semantic competence (by making fewer errors).The mean for the HI learners
Learning Induced errors were 0.69.20 for class six, 60.60 for class seven, and 55.30 for
class eight. The SD for class six was 7.146, class seven, 4.169 and class eight, 4.572 (see
table 29). Class six had the highest variability; therefore, it was the most heterogeneous
class in performance.
Some of the HI pupils in class six made many errors while others made few errors. The
HI pupils Learning Induced errors varied within the classes. These group statistics
indicate that the length of exposure to English language played a big role in lexico-
semantic competence as far as this category of lexico-semantic errors is concerned. Class
eight had the most lexico-semantic competence.
The computation for ANOVA for the total number of Learning Induced errors made by
the HI pupils in the three classes yielded a p-value of 0.001 When compared to the
80
significant level of 0.05, it was found to be very significant (see table 30). There was
therefore a significant difference in the Learning Induced errors made by the three
classes.
Table 29: A Summary of Group Statistics for the HI Learners' Learning Induced Errors
N Mean Std. Deviation Minimum Maximum
CLASS 6 10 69.20 7.146 60 81 CLASS 7 10 60.60 4.169 51 65 CLASS 8 10 55.30 4.572 49 62 Total 30 61.70 7.853 49 81
Table 30: ANOVA Table for the HI Learners' Learning Induced Errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 984.200 2 492.100 16.524 .001 Within Groups 804.100 27 29.781 Total 1788.300 29
4.5.6 Paraphrase Errors
There was a slight difference in the means of the HI paraphrase errors with class seven
having the most errors, while class six had the least number of Paraphrase errors. The
mean for class six was 1.00; for class seven, 1.30; and class eight, 1.10. The SD for class
six was 0.943, for class seven, 1.252, and for class eight, 0.738 (see table 31). Class
seven was the most varied class (the most heterogeneous) while class eight was the most
homogeneous. These results show that class six made little use of this second language
communicative strategy. They may have had a limited lexical semantic knowledge, and
therefore avoided giving a descriptive equivalent of a difficult lexical item. Instead, they
may have coined new words, given a direct translation or made haphazard errors. They
had the most errors in these three categories of lexico-semantic errors.
81
The computation of the ANOVA for the total number of HI pupils’ Paraphrase errors
yielded of p-value of 0.793. When compared to the significant level of 0.05, it was found
to be insignificant (see table 32). There was therefore no significant difference in the
Paraphrase errors of the three HI classes.
Table 31: A summary of group statistics for The HI learners' Paraphrase errors
N Mean Std. Deviation Minimum Maximum
CLASS 6 10 1.00 .943 0 2 CLASS 7 10 1.30 1.252 0 3 CLASS 8 10 1.10 .738 0 2 Total 30 1.13 .973 0 3
Table 32: ANOVA table for The HI learners' Paraphrase errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups .467 2 .233 .233 .793 Within Groups 27.000 27 1.000 Total 27.467 29
4.5.7 Sound Similarity Errors
The means of the three HI classes varied considerably. The values ranged from 0.80 to
1.60, to 1.00 in class six, seven and eight respectively (see table 33). The SD for class six
was 0.789, for class seven, 1.350, and for class eight, 0.667; the most varied group with
the highest standard deviation. Class seven HI pupils may have had two extremes of
pupils: those who had very high number of sound similarity errors, and those who had a
very low number of sound similarity errors.
Like in many of the other categories of lexico semantic errors, class seven had the highest
mean. This indicates that learners in class seven had the most problems in differentiating
lexical meaning, in usage of similarity sounding lexical items, and in lexical items with
82
similar spelling. The HI pupils lip-read. This might have been the possible cause of
sound similarity errors because they cannot hear. Another reason may be presence of HI
pupils who became impaired after acquiring spoken language, and therefore made use of
their phonetic and phonological knowledge acquired earlier.
The computation of the ANOVA for the total number of Sound Similarity errors made by
the HI pupils yielded a p-value of 0.185 (see table 34). When compared with the
significant level of 0.05, it was found to be insignificant. There was therefore no
significant difference among the sound similarity errors made by the three H.I classes.
Table 33: A Summary of Group Statistics for the HI Learners' Sound Similarity Errors
N Mean Std. Deviation Minimum Maximum CLASS 6 10 .80 .789 0 2 CLASS 7 10 1.60 1.350 0 4 CLASS 8 10 1.00 .667 0 2 Total 30 1.13 1.008 0 4
Table 34: ANOVA Table for the HI Learners' Sound Similarity Errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 3.467 2 1.733 1.800 .185 Within Groups 26.000 27 .963 Total 29.467 29
4.5.8 Haphazard Errors
The means showed a tendency to decrease from 64.80, to 56.90 to 44.90 from class six to
class eight (see table 35). These results suggest that at each higher level, the Haphazard
errors decreased. The SD for class six was 6.303, for class seven, 9.243, and for class
eight, 3.872; the most varied group with the highest SD. Class seven was the most
heterogeneous while class eight was the most homogeneous.
83
The reason for such observation may be that class seven had extremes of pupils hence the
wide range of scores from the mean. It was also noted that class seven had the most
number of pupils (16 pupils) unlike class six and eight, which had 11 and 12 pupils
respectively. Class seven pupils consisted of repeaters whom the teachers thought were
unfit to proceed to class eight. Sampling of ten pupils from a population of 16 pupils may
therefore have given a better sample than the other two classes.
The computation of the ANOVA for the total number of Haphazard errors of the HI
pupils in the three classes yielded a p-value of 0.001. When compared to the significant
level of 0.05, it was found to be significant. There was therefore a significant difference
in the total number of Haphazard errors made by the three HI classes (see table 36).
Table 35: A Summary of Group Statistics for the HI Learners' Haphazard Errors
N Mean Std. Deviation Minimum Maximum CLASS 6 10 64.80 6.303 57 74 CLASS 7 10 56.90 9.243 39 71 CLASS 8 10 44.90 3.872 38 51 Total 30 55.53 10.618 38 74
Table 36: ANOVA Tables for the HI Learners' Haphazard Errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 2008.067 2 1004.033 21.491 .000 Within Groups 1261.400 27 46.719 Total 3269.467 29
4.5.9 Collocation Errors
The means of the three classes decreased from 14.10 to 12.10 to 10.00 from class six,
seven, and eight respectively (see table 37). This was an indication that errors reduced
from the lowest level (class six) to the highest level (class eight).
84
The SD revealed a similar trend. The SD for class six was 2.767, class seven, 2.514, and
class eight, 1.633. This was an indication that the HI pupils became consistently more
homogeneous from one level to the other. The reasons for such results may be that the
learners acquired more collocation restrictions from one level to the next, hence
achieving lexico semantic competence. Class six portrayed the lowest lexico semantic
competence and the highest variability while class eight portrayed the highest lexico
semantic competence. Class eight made fewer collocation errors and was the most
homogeneous in performance.
Computation of ANOVA for the total number of Collocation errors made by the three H.I
classes yielded a p-value of 0.098 (see table 38). This p-value was found to be
insignificant when compared to the significant level of 0.05. There was therefore no
significant difference among the total collocation errors made by the three H.I classes.
Table 37: A Summary of Group Statistics for the HI Learners' Collocation Errors
N Mean Std. Deviation Minimum Maximum CLASS 6 10 14.10 2.767 9 18 CLASS 7 10 12.10 2.514 8 16 CLASS 8 10 12.00 1.633 10 15 Total 30 12.73 2.477 8 18
Table 38: ANOVA Table for the HI Learners' Collocation Errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 28.067 2 14.033 2.529 .098 Within Groups 149.800 27 5.548 Total 177.867 29
85
4.6 Analysis of Variance for HP Errors
ANOVA for the HP group was done to determine if there was a significant difference in
the total number of the lexico-semantic errors made by the three HP classes. As in the
other the HI group, the p-value yielded after computation of ANOVA was used to
determine the difference.
4.6.1 Meaning Similarity errors
The means of the three classes differed with class six and seven having 3.90 and class
eight, 2.10 (see table 39).The number of errors decreased from the lower to the upper
level. This difference in means was however not significant. The SD. revealed that class
six was the most heterogeneous class followed by class seven. There was a bigger
variability in the scores of meaning similarity errors from class six to class eight. Class
eight was however the most homogeneous in lexico-semantic competence.
The Computation of ANOVA for the total number of Meaning Similarity errors made by
the three HP classes yielded a P-value of 0.066 (see table 40). When compared to the
significant level of 0.05. It was found to be insignificant. There was no significant
difference in the meaning similarity errors made by class six, seven, and eight.
Table 39: A Summary of Group Statistics for the HP Meaning similarity errors
N Mean Std. Deviation Minimum Maximum Class 6 10 3.90 2.424 1 8 Class 7 10 3.90 1.663 1 6 Class 8 10 2.10 1.449 0 5 Total 30 3.30 2.020 0 8
Table 40: ANOVA Table for the HP Meaning similarity errors
Sum of Squares df Mean Square Fcalc. p-value
Between Groups 21.600 2 10.800 3.016 .066 Within Groups 96.700 27 3.581 Total 118.300 29
86
4.6.2 Meaning duplication Errors
The group statistics show that class eight made the least number of errors followed by
class seven. This shows that the longer the exposure to English, the better the
performance, that is, the fewer the errors. The SD revealed that class six was the most
varied class in performance with a SD of 2.573, followed by class seven with a SD of
1.434. Class 8 was the most homogeneous class (see table 41).
The computation for the ANOVA for the total number of meaning duplication errors
made by the three HP classes yielded a p-value of 0.216 (see table 42). When compared
to the significant value of 0.05, it was found to be insignificant. There was therefore no
Significant difference in the Meaning duplication errors made by class six, seven and
eight.
Table 41: A Summary of Group Statistics for the HP Meaning duplication errors
N Mean Std. Deviation Minimum Maximum Class 6 10 2.80 2.573 0 7 Class 7 10 2.50 1.434 1 5 Class 8 10 1.40 1.174 0 3 Total 30 2.23 1.870 0 7
Table 42: ANOVA Table for the HP Meaning duplication errors
Sum of Squares df Mean Square F-calc. P-value
Between Groups 10.867 2 5.433 1.621 .216 Within Groups 90.500 27 3.352 Total 101.367 29
4.6.3 Semantic Contiguity Errors
The group statistics indicates that the means of the three classes varied. Class six had the
highest number of errors with a mean of 5.90, followed by class seven with 3.80, and
class eight with 2.60 (see table 43). Class eight therefore performed better than the other
87
two classes. This was also supported by the SD. Class six had 2.998, class seven 1.932
and class eight 1.578. Class six was the most heterogeneous class followed by class
seven. Class eight was the most homogenous class.
The computation for the ANOVA for the total number of the HP pupils' Semantic
Contiguity errors yielded a P-value of 0.010 (see table 44). When compared to the
significant value of 0.05, it was found to be significant. There was therefore a significant
difference in the total number of semantic contiguity errors made by the three HP classes.
This is an indication that the interlanguage of the three classes differed significantly.
Table 43: A Summary of Group Statistics for the HP Semantic contiguity errors
N Mean Std. Deviation Minimum Maximum Class 6 10 5.90 2.998 1 10 Class 7 10 3.80 1.932 1 6 Class 8 10 2.60 1.578 0 5 Total 30 4.10 2.578 0 10
Table 44: ANOVA Table for the HP Semantic contiguity errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 55.800 2 27.900 5.503 .010 Within Groups 136.900 27 5.070 Total 192.700 29
4.6.4 Coinage Errors
Class six made more errors with a mean of 1.00, followed by class seven, 0.70 and class
eight, 0.20. The SD for class six was 1.333, class seven, 0.823, and class eight, 0.422.
Class six was the most heterogeneous group. Class eight was the most homogenous group
(see table 45).
88
The computation of the ANOVA for the total number of coinage errors made by the HP
group yielded a p-value of 0.175 (see table 46). When compared the significant level of
0.05, it was found to be insignificant. The ANOVA results indicated that the
interlanguage of the three classes did not differ significantly, although the errors reduced
as one moved from class six to class eight.
Table 45: A Summary of Group Statistics for the HP Coinage errors
N Mean Std. Deviation Minimum Maximum Class 6 10 1.00 1.333 0 4 Class 7 10 .70 .823 0 2 Class 8 10 .20 .422 0 1 Total 30 .63 .964 0 4
Table 46: ANOVA Table for the HP Coinage errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 3.267 2 1.633 1.861 .175 Within Groups 23.700 27 .878 Total 26.967 29
4.6.5 Learning Induced Errors
Class seven had the highest number of Learning Induced errors. The mean for the HP
learners’ Learning Induced errors was 44.10 for class six, 50.70 for class seven, and
33.40 for class eight. The SD for class six was 15.77, class seven, 13.62 and class eight,
13.057 (see table 47). Class six had the highest variability; therefore, it was the most
heterogeneous class in performance. The HP pupils Learning Induced errors varied
within the classes.
The computation for ANOVA for the total number of Learning Induced errors made by
the three HP classes yielded a p-value of 0.036. When compared to the significant level
of 0.05, it was found to be significant (see table 48). There was therefore a significant
difference in the means of the Learning Induced errors made by the three HP classes.
89
Table 47: A Summary of Group Statistics for the HP learning induced errors
N Mean Std. Deviation Minimum Maximum Class 6 10 44.10 15.779 22 71 Class 7 10 50.70 13.622 28 66 Class 8 10 33.40 13.057 17 53 Total 30 42.73 15.503 17 71
Table 48: ANOVA Table for the HP learning induced errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 1524.467 2 762.233 3.779 .036 Within Groups 5445.400 27 201.681 Total 6969.867 29
4.6.6 Paraphrase Errors
It was noted that class eight made the least number of errors followed by class six. The
means were 0.63 for class six, 1.00 for class seven and 0.50 for class eight (see table 49).
The SD revealed that class seven was the most heterogeneous group followed by class
six. The SD for class six was 1.101, class seven 1.247, and class eight 0.850. Class eight
was therefore the most homogeneous group.
The computation of the ANOVA for the total number of paraphrase errors made by the
HP group yielded a p-value of 0.555 (see table 50). When compared to the significant
level of 0.05, it was found to be insignificant.
Table 49: A Summary of Group Statistics for the HP Paraphrase errors
N Mean Std. Deviation Minimum Maximum
Class 6 10 .90 1.101 0 3 Class 7 10 1.00 1.247 0 4 Class 8 10 .50 .850 0 2 Total 30 .80 1.064 0 4
90
Table 50: ANOVA Table for the HP Paraphrase errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 1.400 2 .700 .602 .555 Within Groups 31.400 27 1.163 Total 32.800 29
4.6.7 Sound Similarity Errors
Class seven made the highest number of sound similarity errors followed by class six.
The mean for class six was 10.90, for class seven 13.10, and for class eight, 7.60. The SD
for class seven was 7.752, for class six 6.839, and class seven 4.575 (see table 51). Class
seven portrayed the lowest lexico-semantic competence and the highest variability (it was
the most heterogeneous class). Class eight made fewer sound similarity errors and it was
the most homogeneous in performance.
The computation for total number of sound similarity errors made by the HP. group
yielded a p-value of 0.185 (see table 52). This was not significant when compared to the
significant level of 0.05.
Table 51: A Summary of Group Statistics for the HP Sound similarity errors
N Mean Std. Deviation Minimum MaximumClass 6 10 10.90 6.839 2 24 Class 7 10 13.10 7.752 1 26 Class 8 10 7.60 4.575 1 16 Total 30 10.53 6.704 1 26
Table 52: ANOVA Table for the HP Sound similarity errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 153.267 2 76.633 1.799 .185 Within Groups 1150.200 27 42.600 Total 1303.467 29
91
4.6.8 Haphazard Errors
Class seven made the highest number of errors. Both class eight and six made the same
number of haphazard errors, with a mean of 0.10 (see table 53). Class seven portrayed
the lowest variability while class eight portrayed the highest lexico-semantic competence.
Computation of ANOVA for the total number of haphazard errors made by the three HP
classes yielded a P-value of 0.150 (see table 54). This was found to be insignificant when
compared to the significant level of 0.05. There was therefore no significant difference in
the haphazard errors made by the three HP classes.
Table 53: A Summary of Group Statistics for the HP Haphazard errors
N Mean Std. Deviation Minimum Maximum
Class 6 10 .10 .316 0 1 Class 7 10 .70 1.252 0 4 Class 8 10 .10 .316 0 1 Total 30 .30 .794 0 4
Table 54: ANOVA Table for the HP Haphazard errors
Sum of Squares df Mean Square F Sig.
Between Groups 2.400 2 1.200 2.038 .150 Within Groups 15.900 27 .589 Total 18.300 29
4.6.9 Collocation Errors
The group statistics shows that the means decreased from class six to class eight. The
mean for class six was 12.50, for class seven 12.30, and for class eight, 9.20 (see table
55). This shows that as the HP learners advanced to a higher class, they gained more
lexico-semantic competence. The SD for class six was 3.659, class seven 3.592 and class
eight 2.741. Class six had the highest variability; therefore, it was the most heterogeneous
class in performance. Class eight was the most homogeneous class.
92
Computation of ANOVA for the total number of collocation errors made by the three HP
classes yielded a P-value of 0.065 (see table 56). When compared to the significant value
of 0.05, it was found to be insignificant. There was therefore no significant difference in
the total number of collocation errors made by the three HP classes.
Table 55: A Summary of Group Statistics for the HP Collocation errors
N Mean Std. Deviation Minimum Maximum Class 6 10 12.50 3.659 7 18 Class 7 10 12.30 3.592 7 19 Class 8 10 9.20 2.741 4 13 Total 30 11.33 3.585 4 19
Table 56: ANOVA Table for the HP Collocation errors
Sum of Squares df Mean Square Fcalc. P-value
Between Groups 68.467 2 34.233 3.038 .065 Within Groups 304.200 27 11.267 Total 372.667 29
4.7 Summary of Statistical Analysis
This subsection gives a summary of the statistical analysis done in the previous section.
Two statistical analyses were done: t-tests and Analysis of variance.
4.7.1 Summary of T-test results
The computation of t-tests was done in order to compare the means of the lexico-
semantic errors made by the HP and HI learners. This was to establish whether there was
a significant difference between the lexico-semantic errors made by the two groups. The
results from the t-tests enabled the researcher to reject or accept the second hypothesis of
this study that stated that there is a significant difference between the lexico-semantic
errors made by the hearing-impaired and the hearing pupils.
93
The t-test results revealed that there was a significant difference between the lexico-
semantic errors made by the hearing-impaired and the hearing pupils in the six out of the
nine categories of lexico-semantic errors identified. In the remaining three categories,
there was no significant difference. Although there were cases where the HI made fewer
errors than the HP group, this is not an indication that they were better than the HP group.
In most cases it was impossible to identify the errors because most of the words they used
were not in English. Their compositions were also shorter than the HP group. The table
below gives a summary the t-test results.
Table 57: Summary of T-Tests results
GROUP
ERROR TYPE HP
HI TOTAL
SIGNIFICANT DIFFERENCE
1. Learning induced errors 1282 1851 3133 √
2. Haphazard errors 9 1666 1675 √
3. Collocation errors 340 382 722 x
4. Sound similarity errors 316 34 350 √
5. Semantic contiguity errors 123 165 288 √
6. Meaning similarity errors 99 48 147 √
7. Meaning duplication errors 67 35 102 √
8. Paraphrase errors 24 34 58 x
9. Coinage errors 19 23 42 x
TOTAL 2279 4238 6517
Key: √= there is a significant difference.
X= no significant difference.
94
4.7.2 Summary of ANOVA results
Analysis of variance (ANOVA) for the HI learners’ lexico-semantic errors made in the
three classes showed that there was a significant difference in the means of two of the
categories of lexico-semantic errors (see table 58). These are learning induced and
haphazard errors. In the rest of the categories, there was no significant difference. This
means that the HI learners’ lexico-semantic competence was almost at the same level in
the following categories; Meaning similarity errors, Meaning duplication errors,
Semantic contiguity errors, Collocation errors, Sound similarity errors, Coinage errors,
and paraphrase errors. However, there was a difference in lexico-semantic competence in
learning induced errors and haphazard errors. Length of exposure to language had played
a role in lexico-semantic competence. The HI learners’ errors were also found to be
varied across the three classes as seen in the table below. There was a general decrease of
the number of errors made by the HI learners from standard six to seven to eight in some
categories of lexico-semantic errors. This was observed in learning induced errors,
Haphazard errors, and Collocation errors. It was however noted that the frequency of
errors in the other categories (Sound similarity, Semantic contiguity, Meaning similarity,
Meaning duplication and Paraphrase errors) increased from class six to class seven, and
then decreased in class eight.
Analysis of variance (ANOVA) for the HP learners’ lexico-semantic errors made in the
three classes showed that there was a significant difference in the means of two
categories of lexico-semantic errors (Learning induced errors and Semantic contiguity
errors). In the rest of the categories, there was no significant difference (see table 59).
This means that exposure to English did not play a significant role in the Lexico-semantic
competence of these categories. However, the quantity of the errors differed. There was a
general decrease in the number of errors made by the HP learners from class six to class
seven to class eight. This tendency was observed in Coinage, Collocation, Semantic
contiguity, Meaning similarity errors and Meaning duplication errors. It was also noted
that there was an increase of some categories of errors from class six to seven, but a
decrease in class eight. This tendency was found in Learning induced, Haphazard Sound
similarity and paraphrase errors.
95
Table 58: Summary of ANOVA for HI learners’ errors
HI CLASS
ERROR TYPE 6 7 8 Total
SIGNIFICANT DIFFERENCE
1. Learning induced errors 692 606 553 1851 √ 2. Haphazard errors 648 569 449 1666 √
3. Collocation errors 141 121 120 382 x 4. Sound similarity errors 8 16 10 34 x
5. Semantic contiguity errors 52 62 51 165 x
6. Meaning similarity errors 15 18 15 48 x
7. Meaning duplication errors 10 16 9 35 x
8. Paraphrase errors 10 13 11 34 x
9. Coinage errors 8 7 8 23 x
TOTAL 1584 1428
1226
4238
Key: √= there is a significant difference/ there is variation
X= no significant difference/ no variation
96
Table 59: Summary of ANOVA for HP learners’ errors
HP CLASS
ERROR TYPE 6 7 8 Total
SIGNIFICANT DIFFERENCE
1. Learning induced errors 441 507 334 1282 √ 2. Haphazard errors 1 7 1 9 x
3. Collocation errors 125 123 92 340 x 4. Sound similarity errors 109 131 76 316 x
5. Semantic contiguity errors 59 38 26 123 √
6. Meaning similarity errors 39 39 21 99 x
7. Meaning duplication errors 28 25 14 67 x
8. Paraphrase errors 9 10 5 24 x
9. Coinage errors 10 7 2 19 x
TOTAL 821 887 571 2279
Key: √= there is a significant difference/ there is variation
X= no significant difference/ no variation
4.8 Discussion of findings
This section gives a discussion of the nine categories of lexico-semantic errors that were
identified in the data collected from the two groups under study. The discussion will
entail a quantitative and qualitative analysis of each of the nine categories of lexico-
semantic errors identified. It will also involve the difference of the lexico-semantic errors
made by the HI and HP group.
97
4.8.1 Learning Induced Errors
There were 3133 Learning induced errors made by the two groups under study. This
accounted for 48.07% of the total number of lexico-semantic errors collected from the
two groups under study. The HI made 1851 (59.08%), while the HP group made 1282
(40.92%) Learning induced errors. The computation of the t-test for the total number of
Learning induced errors made by the two groups revealed that there was a significant
difference between the errors of the two groups.
Learning induced errors made by the HI group were characterized by: omission of lexical
items that marked certain semantic features, omission of tense, omission of determiners,
overgeneralization of tense-marking morphemes, failure to mark possession, gender and
number, omission of copular verbs, wrong lexical form, and wrong use of lexical items
such as preposition, pronouns, adjectives, adverbs, and verb auxiliaries.
Hearing impaired (HI) and hearing pupils learning induced errors differed in several
ways. Although in both groups there was omission of lexical items that marks certain
semantic features, this was highly prominent in HI learners' writing. The sentences of the
HI had multiple errors ranging from missing articles, prepositions, conjunctions,
pronouns, inflection and derivational suffixes. In other cases, the HI used content words
only as in the example below.
Kamau cake cut clap children.
(Kamau cut the cake and the children clapped)
The HI group under study had not yet mastered the meaning and usage of prepositions,
pronouns and verb auxiliaries. It may be because they have not acquired them properly or
they hardly use some of the above lexical items in their KSL. Not all lexical items in a
sentence are written or signed (Akachi, 1991).
Omission of important lexical articles such as those marking semantic features as tense
and possession was more in the HI writing than in the hearing pupils. Such errors of
omission are characteristic of language two learners in their early stages of languages
acquisition (Dulay et al 1982). The hearing-impaired pupils’ lexical semantic errors in
98
this category had omissions of grammatical morphemes for tense as seen in the examples
10, 11, 14 in section 4.2.1.
Hearing pupils’ errors in this category indicated that they had acquired rules better than
the HI pupils had. The reason is that the HI learners have delayed exposure to English
language, and therefore have not mastered lexico-semantic relations. (Mayberry 1992,
1994). Because of this, the HI pupils have a delayed acquisition of grammatical
morphemes that mark tense, possession, and participles in lexical items to make them
meaningful. Hearing pupils’ Learning induced errors were characterized by additions
which indicate that they had acquired some rules but were overgeneralising. In some of
the examples given earlier in 4.2.1, they marked some semantic features such as tense
and negation twice. Addition errors usually occur in the later stages of language two
acquisitions, when the learner has acquired some target language rules (Dulay et al 1982).
The HI pupils’ lexico-semantic errors could possibly be attributed to their delayed
learning or acquisition of language, which is caused by hearing impairment. (See Conrad
1979, Akachi 1991, Mayberry 1992, 1994). Research has shown that HI children of
hearing parents constitute 95% of the deaf community. The HI children do not have
access to the acquisition of a first language early enough due to their parents' inability to
communicate to them in a natural language. Conrad (1979) argues that such children
reach school with poor or no linguistic preparation at all. HI learners therefore lack the
necessary language skills and general knowledge for normal language development.
Wilbur (2000) argues that lack of general knowledge in the HI learners is mainly because
of limited input in a language they cannot understand fully. A strong first language base
provides children with an easier transition to learn a second language such as English
(Krashen and Terrell, 2000). However, if the HI learners are exposed to L1 (Sign
language) early enough, they can still learn a language. The current study attributes most
of the lexico-semantic errors made by the HI learners to delayed language acquisition,
which is because of hearing impairment.
99
The HI learners’ learning induced errors related to tense and omission might be a result
of the pupils mapping their written language in KSL (Kenya Sign Language) syntactic
base, similar to other children who are simultaneously acquiring two languages (Bishop
and Mogford, 1993). Learners acquiring two languages seem to go through a stage of
language mixing as argued by Bishop and Mogford (ibid).
Other causes of learning induced errors in this study may be language transfer. Akachi
(1991) says that in KSL, past tense is marked at the beginning of a sentence. The rest of
the manual word signs in a sentence are in their present tense form. Past tense is marked
by a flat hand-configuration moving from the front of the head. On paper, past tense is
represented as [PST] at the beginning of a sentence.
[PST] MAN STEAL BOOK
“The man stole the book”
In written English, the above sentence would be mal-formed because of omission of the
definite article “The” before “man” and before “book,” and failure to mark tense on the
verb “stole”. This explains why the HI failed to mark tense on lexical verbs. It is also an
explanation to why they omitted determiners such as the definite and the indefinite
articles. The learners used the bare form of the verb as in the example below.
I was go Nairobi.
(I went to Nairobi / I was going to Nairobi).
Akachi (1991:65) says that, “What is regular in spoken language may not be regular in
sign language”. This may have influenced the HI learners not to differentiate irregular
verbs from regular verbs and therefore used the same marker for past tense. Irregular
verbs such as put, sleep, cut, and tell were used with –ed. This affected the meaning of
the lexical item and the sentence in which the lexical item appeared because tense is a
semantic feature.
Quigley and Paul (1984) noted that the HI people have difficulties with inflections. This
may explain why the HI failed to mark some semantic features in their writing. In most
cases, they used the bare form of the verb as in the example below emanating from the
hearing-impaired pupils’ data.
The doctor write in paper
100
(The doctor wrote on a paper)
Unlike the HP group, the HI learners used the wrong parts of speech as in the example
below. The HI learner used the noun prayer as a verb, instead of pray.
Do I want to prayer God?
(Do I want to pray God?)
This observation agrees with Ayoo (2004) and Akachi (1991). Ayoo (2004) found out
that the HI learners were poor in marking tense, and used the wrong choice of
grammatical category.
Overgeneralization, incomplete application of rules, ignorance of rule restrictions, system
simplification and exploiting redundancy may have played a role on the HI learners'
errors. Richards (1974:174) says that overgeneralization is associated with redundancy
reduction. It covers instances where the learner creates a deviant structure based on his
experience of other structures in the target language. It may be the result of the learner
reducing his linguistic burden. For example
Yesterday mother bake cake to my birthday
(My mother baked a cake for my birthday).
The HI learner failed to mark tense in ‘bake’ because of the adverb of time ‘yesterday’, a
form of simplification in order to reduce his linguistic burden.
Teacher thank also all mens
The teacher thanked all the men)
The HI learner in the above example was operating in the rule that the plural s is used
with all nouns. Other causes for the lexico-semantic errors in this category may be
ignorance of rule restrictions or incomplete application of rules, as in example 10 given
earlier.
Mary party good
(Mary’s party was good)
In this example, the HI learner did not know how to mark possession by use of the
apostrophe between y and s in the word Mary. The learner also omitted the verb was.
The HI learners in this study did not use determiners. Omission of determiners may be
because of KSL, which lays a lot of emphasis on content words other than function words
101
Demonstrative pronouns and articles were omitted in most of their writing. HI learner
failed to use the article ‘the’ in the example below, to mark known and unknown
information.
Man ask have problem
(The man asked, “Do you have a problem?”)
This observation agrees with Wilbur (1977) findings. She noted that the problem with
determiners in the HI written language was not the placement of a determiner before a
noun but rather the distinction of definite from indefinite, indicating an inability to use
determiners to distinguish new from old information.
4.8.2 Meaning Similarity Errors
As said earlier, these errors were characterized by the learners’ inability to use words that
have similar meaning in their appropriate context. The learners in this study did not know
that it is not in all contexts that synonyms can be interchanged. Both groups made the
same type of meaning similarity errors. The total occurrence of Meaning similarity errors
was 147. This accounted for 2.26% of the total number of errors made by the two groups
under study. The HI group made 48 (32.65%), and the HP group, 99 (67.35%). There
was a significant difference between the Meaning similarity errors made by the HI and
the HP groups. However, the HP group made more meaning similarity errors than the HI
group.
There are lexical items that share a general sense and may be interchangeable in a limited
number of contexts, but which on closer inspection reveal conceptual differences. Both
groups of learners faced difficulties with such lexical items. Palmer (1981) says that only
true synonyms are interchangeable in all their environments but partial or close synonyms
are interchangeable in certain environments only. Both groups under study had problems
with lexical items whose meaning overlapped as in the words drink and take in the
sentence below.
The man drank two spoons of the medicine.
The man took two spoons of the medicine.
The HI learners may have made few lexico-semantic errors because of use of sign
language. Studies by Toth, (2002), Krashen, and Terrell (2000) show that sign language
102
can be used in teaching synonymous words. Both researchers argue that sign language
can use 'inflections' to signify different meaning in words that have close meaning such
as huge and big. This may have given the HI pupils advantage when making lexical
choices.
4.8.3 Meaning Duplication Errors
The total occurrence of Meaning duplication errors was 102. This accounts for 1.57% of
the total number of errors in this study. The HP had 67 (66%), and the HI, 35 (34%) of
the total number of Meaning duplication errors. Computation of the t-test for meaning
duplication errors showed that there was a significant difference between the meaning
duplication errors made by the HI and the hearing pupils.
Meaning duplication errors were caused by partial synonymy. However, for the HI
pupils, such errors were because of reduplication. Their errors had word duplication that
resulted to redundancy in meaning. This reduplication made the HI meaning duplication
errors unique. Earlier studies on sign language found that the HI children use
reduplication to express plurality or emphasis by using or saying the same word twice.
Prinz and Prize (1979), (in Bishop and Mogford 1993) blame such reduplication on the
influence of sign language. Akachi (1991:65) says that in KSL, there is reduplication that
is used to denote plural forms. For example, the word CHILD is reduplicated to become
CHILDREN, and PERSON to become PERSONS. He further argues that in KSL, this
type of reduplication has been used to function on plurals irrespective of whether they are
regular or not.
There is a difference in meaning in reduplicated forms, which is due to lexical meaning
or the aspectual character of word-signs. For instance, a reduplicated verb like WALK
can be translated into English phrase as Walk and walk and walk. According to the data
collected, we have found that this is the reason why the HI pupils duplicated words such
as; happy to mean very happy, fast to mean very fast, dance to mean dancing a lot or
dancing for a long time, and nice to mean very nice (refer to the examples 90,91,92,93
and 94). This means that they had not yet mastered how to express plurality and emphasis
103
in English. They transferred what they had learned using KSL into written English. These
types of errors caused by reduplication are idiosyncratic to the writing of the HI learners.
In other examples given earlier (in 4.2.6), the HI learners were not able to use the
comparative form of adjectives and adverbs correctly. The adverb more was frequently
used with an adjective with the comparative suffix –er as in the example below.
She was more happier
(She was happier)
The HI learners also used two verbs with similar meaning. This is a form of word
duplication, which is redundant in meaning because the two words share semantic
features. The HP used words such as now and ever twice in the same sentence. This is a
developmental pattern in SLA where learners know that certain words can be used in
different positions in a sentence. However, the learners used the same word twice in a
sentence in their interlanguage.
4.8.4 Semantic Contiguity errors
The total occurrence of Semantic contiguity errors was 288. This accounted for 4.42% of
the total number of lexico-semantic errors identified in this study. Out of the 288
Semantic contiguity errors, the HI group had 165 (57.29%), and the HP group, 123
(42.71%). Statistical analysis showed that there was a significant difference between the
total number of Semantic contiguity errors made by the two groups of learners under
study.
The Semantic contiguity errors made by the HI group had the same developmental
pattern with those made by the HP group. Both learners used a single lexical item that
shares semantic features with the target item. This showed that the learners were aware of
the semantic fields in second language. This knowledge enabled the learners to choose a
lexical item that provided an approximate translation of the unknown concept by
referring to a similar known item. Lexical items such as asked, talk, and told were used
interchangeably as if they were true synonyms in English. Other lexical items that were
used interchangeably are pairs such as; boy and son, husband and man, visitors and
104
guests, my uncle's son and my cousin. To both groups of learners, such words became
true synonyms in their interlanguage.
4.8.5 Collocation Errors
The total occurrence of Collocation errors was 722. This accounted for 11.08% of the
total number of lexico-semantic errors in this study. The HI had 382 (52.91%), and the
HP, 340 (47.09%). Statistical analysis showed that there was no significant difference
between the total number of Collocation errors made by the two groups of learners under
study.
Both groups of learners showed lack of knowledge for collocation restriction of certain
lexical items. Collocation errors resulted from wrong use of preposition, wrong choice of
preposition and linking of lexical items that do not habitually co-occur together. As
discussed earlier in section 4.2.5, the HI learners' collocation errors were characterized by
the following.
a. Overlooking of co-occurrence restrictions of certain lexical items; for example,
saw and lost instead of found and lost
b. Substitution of Lexical items used to make some grammatical expressions with
other words that do not co-occur.
She put cake and table cut.
(She put the cake on the table then cut it).
c. Failure to master the meaning of prepositions that collocate with expression of
time, space and instrument in English language propositions.
d. Use of two prepositions as in example below. One preposition is wrong while the
other is correct.
All people in on our village community meet.
It was noted that the HP were better in usage and meaning of prepositions than the HI
learners, although this difference is not statistically significant. The HI did not understand
the locative meaning expressed by certain prepositions. They therefore used the wrong
prepositions, double prepositions, and illogically used preposition (used them where they
were not needed at all). The results of this study agree with Ayoo 2004. In her study, she
105
found out that use of prepositions by HI created difficulties. The HI used incorrect
prepositions, double preposition, omitted preposition, and in other cases used other
grammatical items instead of a preposition.
4.8.6 Coinage Errors
The total occurrence of Coinage errors was 42. This accounted for 0.64% of the total
number of lexico-semantic errors identified in this study. Out of the 42 Coinage errors,
the HI group had 23 (54.76%), and the HP group, 19 (45.24%). Statistical analysis
showed that there was a significant difference between the total number of Coinage errors
made by the two groups of learners under study.
The learners in this study were involved in creative construction of a new lexical item to
enable them communicate a concept they desired but lacked the appropriate words to use.
The learners used a strategy of lexical and semantic innovation to come up with lexical
items that neither were in the LI nor in TL as in the examples discussed in section 4.2.8.
Both groups showed the same strategy in coining words. Some of the HI learners created
lexical items that exist in English but their meaning were inappropriate in the context
they were used. For instance, Died (past tense of die) was used as an adjective (see
example 123). In other instances, the HI group created new lexical items by adding an
affix to existing words in the TL to produce an inflected or derived form such as
happyingly, sicky, Nice, badly, goodly and misfully.
The HI learners also created compound nouns such as faretaker, by combining the noun
fare and a new word coined from the verb taker. In the HI learner's interlanguage, this
word means someone who collects fares on a public vehicle. Where the HI learners
lacked the correct verb to use, they coined a verb from a noun. Such examples are words
like microscope in example 115, and stethoscope, in example 116. The two lexical items
exist only as nouns but not as verbs as the learners have used them.
106
4.8.7 Haphazard Errors
The total occurrence of Haphazard errors was 1675. This accounted for 25.70% of the
total number of lexico-semantic errors identified in this study. Out of the 1675 Haphazard
errors, the HI group had 1666 (99.46%), and the HP group, 9 (0.54%). Statistical analysis
showed that there was a significant difference between the total number of Haphazard
errors made by the two groups of learners under study.
These errors do not have any relation with developmental patterns. They do not follow
any laid down patterns of language development. They showed total ignorance or no
learning at all on the side of hearing impaired pupils. They included illogical use of
lexical items and use of lexical items that were neither English nor any language that the
researcher was aware of. In the illogical use of lexical items, the HI pupils lacked
syntactic means to express their thoughts and resulted to stringing together associated
words and phrases.
The HI haphazard errors were difficult to describe, as many of them were semantically
unrelated words that were carelessly put in a sentence without any punctuation. Other
lexical items were in neither English nor any other language known by the researcher.
It was evident from the data collected in this study that the HI learners had not yet
acquired enough vocabulary for communication in written English. The way they
carelessly put words (with differently meaning or with no meaning) together showed that
they were ignorant of word order and meaning.
4.8.8 Paraphrase Errors
The total occurrence of Paraphrase errors was 58. This accounted for 0.89% of the total
number of lexico-semantic errors identified in this study. Out of the 58 Paraphrase errors,
the HI group had 34 (58.62%), and the HP group, 24 (41.38%). Statistical analysis
showed that there was no significant difference between the total number of Paraphrase
errors made by the two groups of learners under study.
107
This category of errors reflected the learners’ use of second language lexicon and
structure to give a descriptive equivalent of first language. In some cases, the learners
described the characteristics or the elements of the object or the action instead of using
the appropriate target language structure. This indicates that the learners were not
familiar with the appropriate lexical items in language two. Although paraphrase which
involved circumlocution may not have led to an error, it was an indication of the learners’
inadequate lexical competence.
Both groups under study used paraphrase as a form of lexical simplification when they
encountered unknown lexical item. They gave the description, characteristics or function
of the intended referent. Such paraphrases were correct in the learners’ interlanguage.
However, in the TL, they were an indication that the learner had not yet acquired
semantic competence.
4.8.9 Sound Similarity Errors
The total occurrence of Sound similarity errors was 350. This accounted for 5.37% of the
total number of lexico-semantic errors identified in this study. Out of the 350 Sound
similarity errors, the HI group had 34 (9.71%), and the HP group, 316 (90.29%).
Statistical analysis showed that there was a significant difference between the total
number of Sound similarity errors made by the two groups of learners under study.
Although the HI pupils cannot hear, they had errors related to the pronunciation of lexical
items. There is a possibility that some of the HI were post-lingual or partially hearing
impaired. Such students may still have had knowledge of the sounds of the spoken
language they had acquired before becoming deaf. HI pupils' lip-read and this shows
their ability to understand sounds (phonemes), though they cannot hear. This may be one
of the causes of their sound similarity errors. Poor spelling or confusion of lexical items
that were similarly spelt cannot be ruled out in this category of errors. Such lexical items
also had similar or close pronunciation.
The HI pupils’ inability to hear contributed to their fewer errors in this category. Words
with close pronunciation were not a problem to them. The hearing pupils may have had
108
some difficulties in the use of homophonous words or words that had close
pronunciation. There is still a possibility that they may have misspelled such lexical items
because of influence of their language one (KSL) as in the examples given earlier in
4.2.6.
4.9 Analysis of Variance for the HI and HP Learners' Lexico-Semantic Errors
Analysis of variance for the HI lexico-semantic errors made by class six, seven, and eight
revealed that there was a significant difference in Learning induced errors and Haphazard
errors made by the three classes. There was no significant difference in the in the
following error categories: Meaning Similarity, Meaning Duplication, Semantic
Contiguity, Coinage, Sound Similarity, Collocation, and Paraphrase errors. These results
put the three HI classes at the same level of Lexico-semantic competence. It means that
the linguistic exposure given to class seven and eight is not different from that of class
six. However, there was a tendency for the errors to decrease from class six to class eight.
A similar trend, though in different categories of lexico-semantic errors, was observed in
the ANOVA for the HP learners’ errors. There was a significant difference in two
categories of lexico-semantic errors: Learning induced and Semantic Contiguity errors.
There was no significant difference in the rest of the categories although the quantity of
the errors differed. Class eight made fewer errors than the other two classes. They
portrayed a greater lexico-semantic competence than the other two classes. This is
because they have been exposed to English language longer than class six and class
seven.
4.9.1 Summary
The results of this study show that the two groups made Lexico-semantic errors. These
Lexico-semantic errors were categorized into to nine types. The HI group made more
errors in total than the HP group. However, in some categories that include Sound
similarity errors and Meaning similarity errors the HP made more errors than the HI
learners did. The nine categories displayed similar developmental patterns apart from
Haphazard errors, Meaning duplication errors and Learning induced errors where some of
the HI errors differed.
109
HI learners’ haphazard errors did not follow any laid down patterns of development. The
HI pupils put lexical items and phrases together in a haphazard manner. They lacked
syntactic means to express their thoughts and describe events or states. This showed that
they did not understand the meaning of the lexical items they used. They made the wrong
lexical choices. In other cases, they used lexical items from language unknown to the
researcher. Kenyan sign language structure was also evident in the written English of the
HI. This contributed to the Lexico-semantic errors identified in the HI learners’ data.
Learning induced errors were characterized by omission or failure to mark semantic
features such as tense, possession, number and gender. The HI misused many
grammatical categories such as preposition, nouns, conjunctions and determiners. They
omitted the use of determiners to mark known from unknown information.
Statistical analysis of the total number of lexico-semantic errors showed that there was a
significant difference between the lexico-semantic errors of the HI and HP in six
categories out of the nine categories identified. These were, Meaning similarity, Meaning
duplication, Semantic contiguity, Learning induced, Sound similarity, and Haphazard
errors. There was no significant difference between the collocation, Coinage, and
paraphrase errors made by the two groups under study.
Analysis of variance for the HI lexico-semantic errors made by class six, seven, and eight
revealed that there was a significant difference in the following error categories: Learning
induced errors and Haphazard errors. There was no significant difference in the following
error categories: Meaning similarity, Meaning duplication, Semantic contiguity, Coinage,
Sound similarity, collocation, and paraphrase errors. This shows that the HI learners were
making the same errors statistically and little learning was taking place from class six to
class eight. There is a possibility that some elements had become fossilized in the HI
group. This shows that the HI pupils in this study may have been delayed to language
exposure at their early childhood. Research has shown that individuals born deaf and
isolated from language at their early age grow up being linguistically dysfunctional
(Mayberry, 1992, 1994). The reading ability of the HI child becomes very impaired and
they have slow acquisition of lexical items.
110
CHAPTER FIVE
SUMMARY, CONCLUSION AND RECOMMENDATION
5.1 Introduction
This chapter includes a summary of the study’s objectives and their attainment, the
findings, conclusion and recommendation for further studies. The chapter will be based
on the results and discussion in chapter 4.
5.2 Identification and Description of the Lexico-Semantic Errors
From the three tests administered to the two groups of learners, the researcher found out
that the two groups of learners made various lexico-semantic errors. The lexico-semantic
errors were identified and classified using the Error Analysis Theory. Selinker’s
Interlanguage theory was used to give a possible explanation of the lexico-semantic
errors. The errors were classified into nine groups. These are as follows: Learning
induced errors, Meaning Similarity errors, Meaning Duplication errors, Semantic
Contiguity errors, Collocation errors, Coinage errors, Paraphrase errors, Sound Similarity
errors and Haphazard Errors.
The findings indicated that learning induced errors were the most common, with a
percentage of 48.07. The lexico-semantic errors of the two groups differed in frequency.
HI pupils made more Learning induced, Collocation, Haphazard, Semantic contiguity,
Paraphrase and Coinage errors than the hearing pupils.
The errors of the HI pupils were because of over generalization, incomplete application
of rules, ignorance of rules restriction and language transfer. The haphazard errors did not
show or follow any laid down patterns of development. They may be because of
ignorance or no learning at all. The HI may have been exposed to language late because
of their impairment, hence poor acquisition of lexico-semantic competence. This is in
agreement with earlier studies on HI pupils’ written language by Wilbur (1977) and
Mayberry (1992).
111
Identification and classification of the lexico-semantic errors in the two groups under
study enabled the researcher to achieve objective number one: Identify and describe
Lexico-semantic errors made by HI pupils in their written English.
5.3 Significant Difference of the Lexico-Semantic Errors
To establish if there is a significant difference between the Lexico-semantic errors of the
HI and HP, several t-test procedures were done to compare the means of the errors of the
HI group with those of the HP group. The findings were as follows.
T-tests for the means of the total number of errors from the three modes of testing
revealed that:
1. There was a significant difference between the means of the total number of
Lexico-semantic errors in the following categories: Meaning Similarity, Meaning
Duplication, Semantic Contiguity, Coinage, Learning Induced, Sound Similarity,
and Haphazard errors. It is worthy noting that, the HI made more errors in six
categories: Learning Induced Haphazard errors, Semantic Contiguity errors,
Collocation errors, Paraphrase errors, and Coinage errors. These six groups of
errors formed a bigger percentage (63.23 %.) of the total number of Lexico-
semantic errors made by the two groups in this study (4121 out of 6517).
2. There was no significant difference between the means of the total number of
Collocation, Coinage and Paraphrase errors in the two groups. However, the HI
made more Collocation, Coinage and Paraphrase errors than the Hearing pupils
did.
5.4 Hearing Impaired Pupils’ Lexico-Semantic Errors
From the data analyzed in this study, it is evident that both HI and HP make Lexico-
semantic errors. These errors, however, differ in quality and quantity. The HI made more
Lexico-semantic errors than their hearing counterparts did. However, in some categories
of lexico-semantic errors, the HP made more errors. These categories include Sound
similarity errors and Meaning similarity errors.
112
The two groups portrayed similar developmental patterns with the exception of learning
induced, Meaning duplication, and Haphazard errors. A linguistic analysis and
description showed that the learning induced errors of the HI were characterized by
omission of lexical items that marked certain semantic features. In the learning induced
error category, the HI learners' errors were characteristic of errors made by learners in
their early stages of language acquisition. Unlike the HP, the HI learners did not mark
tense, number, possession, and in some cases, gender. They rarely used conjunctions in
their writing. They used parts of speech such as determiners, verbs, adverbs, and
adjectives incorrectly.
Their written English showed ignorance of the meaning of most of the words they used.
Unlike the HP learners, they used the wrong parts of speech; for example, noun as verb.
This shows that the HI group lagged behind the HP group in language acquisition. Such
delay is attributable to hearing impairment because the learners were not exposed to
natural language early enough, like the HP group.
There was also reduplication of lexical items in their writing. This is normally reflected
in KSL that reduplicates words to mark plural and emphasis. This is the reason why some
of the HI did not mark number in nouns. The HI learners also made redundant use of
lexical items such as two verbs that shared semantic features. This pattern of writing
(reduplication) was not witnessed in the writing of the HP group. The HI learners have
not yet mastered lexical meaning to enable them express emphasis, plurality, and use of
the comparative form of adjectives.
The written English of the HI learners revealed that many of the HI learners were unable
to generate any connected written English. They carelessly put words that were unrelated
together forming a group of words that did not make any sense. Some of the lexical items
that were carelessly put together were in a language unknown to the researcher. This was
reflected in their Haphazard errors. These errors were characterized by severe structural
anomalies, including English word order errors and numerous syntactical and semantic
errors. The HI lacked the proper syntactic and semantic means to express themselves, and
describe events.
113
Analysis of variance for the HI lexico-semantic errors made by class six, seven, and eight
revealed that there was a significant difference in the following error categories: Learning
induced errors and Haphazard errors. There was no significant difference in the three
classes in the following error categories: Meaning similarity, Meaning duplication,
Semantic contiguity, Coinage, Sound similarity, Collocation, and Paraphrase errors.
Lack of statistical significance in the errors made by the three classes shows that despite
exposure to English language, the HI learners were not achieving much. It was however
noted that some errors decreased from class six, seven to class eight.
5.5 Limitation of the Research
Eight out of the thirty scripts from the HI were partially unreadable. Some sentences were
in KSL structure and others in English. To reconstruct meaning from such data, the
researcher made use of available literature for KSL, such as Akachi (1991). Out of the
eight scripts, four were for class six, three for class seven, and one for class eight. Basic
knowledge of the KSL structure was therefore important for the researcher to reconstruct
the meaning of the sentences in such scripts. Ayoo (2004) faced similar problems.
In some error categories, the HI made fewer errors than the HI group. As earlier said, this
was not an indication that the HI group was better in such categories. It was impossible to
identify errors from the HI texts because they were not in English. Their compositions
were also shorter compared to those of the HP learners.
There was also uniformity in the data collected from the picture story. Five scripts had
the first two paragraphs similar. The learners may have written exactly what the teacher
had discussed with them about the picture story. Initially, the researcher had planned to
use class eight only, but due to low enrollment of HI pupils, class seven and six were
used to get enough samples.
114
5.6 Recommendation for Further Research
This study concentrated on Lexico-semantic errors in the written English of standard six,
seven and eight HI Learners. It compared the HI lexico-semantic errors with those of
hearing pupils in the same classes.
Further studies can be done in the following areas:
1. An investigation on the effect of KSL on the second language acquisition of
English by HI learners. This will enable educationist to identify errors as a result
of cross-linguistic influence.
2. A follow up study can also be done on the written English of HI learners in
secondary school or colleges to know how they write. A pragmatic approach such
as giving of spontaneous tests can be used in the mode of data collection and
instrumentation.
3. A longitudinal study comparing the Lexico-semantic errors of the two groups can
be done to get the learners’ interlanguage over a long time such as five years.
4. A comparison of the HI written English sentence structure and the KSL sentence
structure can be done.
5.7 Remedial Measures
This research recommends the following remedial measures in order to improve the HI
English.
1. Teachers should concentrate more on the teaching of vocabulary. This is because
vocabulary forms foundation of the grammar of every language. Learners also
face the greatest difficulty in learning a second language in the area of acquisition
of vocabulary and their ability to use lexis correctly in conveying a given
message.
2. Use of bilingual approach in the teaching of the HI learners. This means that the
young HI learners can be taught KSL as their first language. When they gain their
first language competence, they can use it to learn English effectively. Research
has shown that a strong base of LI can enable learners learn a second language
without much problems.
115
3. There is need to have teachers who are competent in sign language to teach the HI
learners KSL. This will reduce the transfer of wrong learning from the teacher to
the learners.
4. Use of Total Communication. This is the use of oral and manual approaches to
teach the hearing-impaired learners. It includes use of sign language such as KSL,
lip reading, finger spelling, sound, eye contact, English, Pantomime, drawing and
mouth hand system.
5.8 Conclusion
From the findings of the study, it is evident that HI learners have not acquired lexico-
semantic competence in written English. Their written English shows that a great
majority of them have not acquired enough English language in order to express
themselves. They hardly understand the meaning of many of the lexical items they use.
They do not understand the semantic relations between words. The written English of
many of the HI could not be understood.
Although the HP and HI learners portrayed similar learning strategies in lexico-semantic
competence, the HI group made more Lexico-semantic errors. In six categories of the
lexico-semantic errors, there was a significant difference between the HI and HP lexico-
semantic competence. There is need to put more effort in teaching vocabulary meaning.
The mode of instruction in teaching of the HI pupils should be revisited to ensure all
teachers are competent in KSL. The HI learners have the capacity to acquire English
language as much as the hearing learners can.
116
REFERENCES Adams, G., and Schvaneveldt, (1985) Understanding Research Methods. New York,
Longman Inc.
Adjemian, C. (1976)”On the nature of interlanguage systems.” Language Learning
26:297-26:297-320
Adoyo, P.O. (1995) “An investigation of Kenyan Sign Language development”.
Unpublished PGDE Dissertation, University of Bristol, U.K.
Adoyo, P.O. (2002) “A Comparative investigation on the differential effects of Kenya
sign language and simultaneous communication on the memory and
comprehension of deaf children in Kenya”. Unpublished Ph.D Thesis, University
of Hamburg, Germany.
Adoyo, P.O. (2002) “Emergent Approaches towards Sign Bilingualism in Deaf Education
in Kenya.” Stichproben. Wiener Zeitschrift für kritische Afrikastudien 3/2002,
Jg.2
Akachi, P.O. (1991) “Sentence Types in Kenyan Sign Language; A Structuralism
Approach.” Unpublished M.A Thesis, University of Nairobi.
Ayoo, E.A. (2004) “Analysis of the Morphosyntactic Errors in the Written English
of Standard Eight Hearing Impaired Pupils”. Unpublished M.A Thesis,
Kenyatta University.
Bailey, N., Madden, C., and Krashen, S. (1974) “Is there a ‘natural sequence’ in adult
Second Language Acquisition?” Language learning 24:235-243
Bishop, D., and Mogford, K. (1993) Language Development in Exceptional
Circumstances. East Sussex: U.K. Lawrence Erlbaum Associates Publishers
Bialystok, E. (1978) A theoretical model of second language learning. Oxford: Basil
Blackwell.
Blum-Kulka, S., and Levenston, E. (1978) “Universals of Lexical Simplification.”
Language Learning.28: 399-415.
Botvin, G.J., and Sutton-Smith, B. (1977). “The Development of Structural Complexity
in Children Narratives”. Developmental Psychology, 13, 377-388.
Chappell, G.E. (1980). “Oral language performance of upper elementary school students
obtained via story reformulation”. Language, Speech, and Hearing Services in
117
Schools 11,236-250.
Chege, B.K. (1996) “Lexico-Semantic Errors as Indices of Developing Language
Competence: A Case Study of Ol-Kalou Primary School”. Unpublished M.A.
Philosophy Thesis. Moi University.
Conrad, R. (1979). The Deaf School Child. Language and Cognitive function. London:
Harper and Row.
Cooper, R.L. (1967). “The ability of deaf and hearing children to apply morphological
Rules.” Journal of Speech and Hearing Research, 10, 77-82.
Corder, S.P. (1967) “The Significance of Learners’ errors”. International Review of
Applied Linguistics. (5:161-70). London: Oxford Press
_________. (1971) ‘‘Idiosyncratic dialects and error analysis” International Review of
Linguistics (IX: 149-59). London: Oxford Press
_________. (1974) “Error Analysis”. In Allen. J and Corder. P (Eds) The Edinburgh
Course in Applied Linguistics. (Vol.3, pp 122-154). Oxford: Oxford University
Press.
_________. (1974) Error Analysis. Edinburgh. London: Oxford University Press.
_________. (1981) Error Analysis and Interlanguage. Oxford: Oxford University Press.
Coryell, J., and Holcomb, T. (1997). “The use of sign language systems in facilitating
language / communication Acquisition of students with deafness”. Language
Speech and Hearing Services in Schools, 28(4), 384-94
Difrancesca, S. (1972). "Academic achievement test results of a national testing program
for hearing-impaired students. Washington, DC: Gallaudet College, Office of
Demographic Studies". In Journal of Speech and Hearing Disorders, Volume 55
628 -634November 1990
Dulay, H., and Burt, M. (1974) Natural Frequency in Child Second Language
Acquisition; Language learning. Oxford: Oxford University Press.
Dulay, H., and Burt, M. (1974). “You can't learn without goofing” In Error analysis.
(Ed.), J. C. Richards. London: Longman
Dulay, H., Burt, M., and Krashen, S. (1982) Language Two. Oxford: Oxford University
press.
Ellis, R. (1985) Understanding Second Language Acquisition. Oxford: Oxford University
Pre Press.
118
______. (1990) Instructed Second Language Acquisition. Oxford: Basil Blackwell
______. (1994) The Study of Second Language Acquisition. Oxford: Oxford University
Press.
Gairns, R., and Redman, S. (1992) Working with words: A guide to Teaching and
Learning Vocabulary. Cambridge: Cambridge University Press.
Gass, S., and Schachter, J (eds.) (1989) Linguistic Perspective on Second Language
Acquisition. . New York: Cambridge University Press.
Gass, S. (1984). “A Review of Interlanguage Syntax: Language Transfer Universal”
Language Learning 34:115-132
Geers, A., Moog, J., and Schick, B. (1984). “Acquisition of spoken and signed English by
Profoundly Deaf Children.” In Journal of Speech and Hearing Disorders, 49, 378-
388
Glenn, C.G. (1978). “The role of episodic structure and story length in children's recall of
Simple stories.” In Journal of Verbal Learning and Verbal Behavior, 17, 229-247
Goda, S. (1964). “Spoken syntax of normal, deaf and retarded adolescent children.” In
Journal of Verbal Learning and Verbal Behavior, 3, 401-405.
Halliday, M.A.K., and Hasan, R. (1976). Cohesion in English. London: Longman Group.
Haslett, B.J. (1983). Children's strategies for maintaining cohesion in their written and
oral stories. Communication Education 32, 91-104.
Howit, D., and Cramer, D. (2003). A Guide to computing Statistics with SPSS 11 for
Windows. Revised Edition. Great Britain: Dorset Press.
Kasper, G., and Kellerman, E. (1997). Communication Strategies: Psycholinguistic and
Sociolinguistics Perspectives. London: Longman
Kellerman, E., and Sharwood, M. (1986) Cross-Linguistic Influence in Second
Language Acquisition. Oxford: Pergamon Institute of English.
Kenya Institute of Education. (1994) Primary English Syllabus. Nairobi: Kenya Institute
of Education.
Kenya Institute of Education. (2002) Primary Education Syllabus. Nairobi: Kenya
Institute of Education
Kenya Institute of Education. (2002) Primary English Syllabus. Nairobi: Kenya Institute
of Education.
119
Kenya Institute of Education. (2004) Kenya Sign language Syllabus. Nairobi: Kenya
Institute of Education.
Kenya Literature Bureau. (2004). 2nd Edition. Let’s Learn English. Nairobi: Kenya
Literature Bureau
Kenya National Examination Council. (1999). K.C.P.E Report. Nairobi: Kenya National
Examination Council.
Kenya National Examination Council. (2001). K.C.P.E Report. Nairobi: Kenya National
Examination Council.
Kenya National Examination Council. (2003). K.C.P.E Report. Nairobi: Kenya National
Examination Council.
Kenya National Examination Council. (2004). K.C.P.E Report. Nairobi: Kenya National
Examination Council.
Krashen, S.D., and Terrell, T.D. (2000). The Natural Approach: Language Acquisition in
the Classroom. Essex, England: Pearson Education
Kretschmer, R., and Kretschmer, L. (1978). Language development and intervention with
the hearing-impaired. Baltimore: University Park.
Lenneberg, E. (1967) Biological Foundation of Language. New York: Wiley.
Littlewood, W. (1984) Foreign and Second Language Learning. Oxford: Oxford
University Press.
Longman Publishers. (1992). Dictionary of Contemporary English (New Edition).
Longman Group UK Limited: England
Lyons, J. (1977). Semantics Vol. 1 London: Cambridge University Press.
Lyons, J. (1977). Semantics Vol. 2 London: Cambridge University Press.
Maina, J. (1991) “A Study of The Grammatical Errors in Standard Eight Pupils Written
Work in English in Four City Schools”. Unpublished M.A Thesis Kenyatta
University.
Marshall, W., and Quigley, S.P. (1970). “Quantitative and qualitative analysis of
syntactic structures of written language of deaf students”. Urbana, IL: Institute
of Research on Exceptional Children. In Journal of Speech and Hearing
Disorders, Volume 55, 628-634, November 1990
120
Marshark, M. (1997) Raising and educating a Deaf Child. Oxford: Oxford University
Press.
Mayberry, R.L. (1992) “The cognitive development of deaf children: Recent insights” in
Handbook of Neuropsychology, F. Boiler and J. Grafman (Eds) 51-68
Amsterdam: Elsvier.
Mayberry, R.L. (1993) “1st Language Acquisition after Childhood differs from Second
Language Acquisition: The case of American Sign Language” Journal of Speech
and Hearing Research, 36, 1258-1270.
Mayberry R.L. (1994) “The importance of Childhood to Language Acquisition: Insights
from American Sign Language.” In J.C. Goodman and H.C. Nusbaum (Eds.), The
Development of Speech Perception: The transition from Speech sounds to Words
(pg 57-90). Cambridge: MIT Press.
McLaughlin, B. (1987) Theories of Second Language Learning. London: Edward Arnold.
Milliken, G., and Johnson, D. 1992. Analysis of Messy Data: Designed Experiments.
Volume 1. New York: Chapman & Hall.
Mugenda, A. and Mugenda, O. (2003) Research Methods: Quantitative and Qualitative
Approaches: Nairobi: Acts Press.
Mutai, K.B. (2000) How to Write Quality Language Research Proposal. Edinburgh, U.K:
Thelley Publications.
Mutiti, J.K. (2000). “The Parameters of Syntactic Information Packaging in the Second
Language Acquisition of English by Gikuyu First Language”. Unpublished Ph. D
Thesis. Egerton University.
Myklebust (1965) “Journal of Speech and Hearing Disorders” Volume 55, 628-634,
November 1990
Nation, I.S.P. (1990). Teaching and Learning Vocabulary. New York: Heinle and Heinle.
Ndurumo, M.M. (1993) Exceptional children. Development consequence and
interventions. Nairobi: Longman Kenya Limited
Nemser, W. (1971) “Approximate Systems of foreign Language Learners” in Richards, J
(Ed.) Error Analysis: Perspectives on Second Language Acquisition (pp.55-64)
London: Longman.
Neter, J., Wasserman, W., and Kutner, M. H. 1990. Applied Linear Statistical Models.
Homewood, IL: Irwin.
121
Njoroge, K. (1987) “The Acquisition of Six Morphosyntactic Structures of English by
Kenyan Children”. Unpublished Doctoral Dissertation University of Edinburgh,
Edinburgh.
Njoroge, M. (1996) “Morphosyntactic Errors in the Written English of First Year
Undergraduate Students in Kenya”. Unpublished M.A Thesis Kenyatta
University, Nairobi.
Norrish, J. (1983) Language learners and their errors. London: Macmillan Press Ltd.
Nyamasyo, E. (1992) “A Corpus Based Study of Grammatical and Lexical
Characteristics of the Writing of Kenyan Pre-University Students”. Unpublished
Doctoral Dissertation University of Lancaster, Lancaster.
Nyamongo, M. N. (1996) November 1 “National Sign Language needed”. Daily Nation.
Odlin, T. (1989) Language Transfer: Cross-linguistic influence in language learning.
Cambridge: Cambridge University Press
Odhiambo, O. (2006) March 2nd. “Low English grades puzzle experts”. Daily Nation.
O’Grady, W., Michael, D., and Aranoff, M. (1993) Contemporary Linguistics:
An introduction .2nd Edition. New York: St. Martins
Okombo, O. (1994) “Kenyan Sign Language: Some attitudinal and cognitive issues in the
Evolution of a language community”. In Ahlgren and Hyltenstam I.K. (Eds)
Bilingualism in Deaf Education. Hamburg: Signum, 37-54.
Okombo, O., and Akach, P.O.A. (1997) “Language convergence and wave phenomena in
the growth of a national Sign Language in Kenya.” In De Gruyter (1997) (Eds).
Linguistic issues in Sub-Saharan Africa. International Journal of the Sociology of
Language: 125:131-144.
Ombiro, E. (1988. April 17) “Lack of Clear policy handicap for the deaf”. Daily Nation.
Oso, W.Y.; and Onen, D. (2005) A General Guide to Writing Research Proposal and
Report: A Handbook for Beginning Researchers. Kisumu: Options Press and
Publishers
Palmer F.R. (1991) Semantics Cambridge: Cambridge University Press.
Princeton Language Institute. (1993) 21st Century Grammar Book New York: Dell
Publishing Company.
122
Quigley, S., Wilbur, R., and Montanelli, D. (1976) “Complement structures in the
language of deaf students”. In Journal of Speech and Hearing Research 19:448—
457
Quigley, S.P., Power, D.J., and Steinkamp, M.W. (1977) “The language structure of deaf
children.” The Volta Review 79(80):72--84.
Quigley, S., and Paul, P. (1984) Language and Deafness. San Diego, California: College-
Hill Press.
Quirk, R., and Greenbarm (1973) A University Grammar of English. Essex: Longman
Read, J. (2000) Assessing Vocabulary. Cambridge: Cambridge University Press
Read, J. (2004). “Research in Teaching Vocabulary.” Annual Review of Applied
Linguistics 24: 146–161, Cambridge University Press
Richards, J.C. (1974) Error Analysis London: Longman
Richards, J.C. (ed.) (1974). Error Analysis: Perspectives on second language acquisition.
London: Longman.
Richards, J.C., and Sampson, G.P. (1974). “The study of learner English”. In J.C.
Richards (ed.) Error Analysis. Perspectives on second language acquisition, pp. 3-
18.
Richards, J.C. et al. (1992). Dictionary of Language Teaching & Applied Linguistics.
Second Edition. Essex: Longman Group UK Limited.
Schmitt, N. (2000) Vocabulary in Language Teaching. Cambridge: Cambridge University
Press.
Seliger, H.W., and Shohamy, E. (1989) Second Language Research Methods. Oxford:
Oxford University Press.
Selinker, L. (1972) “Interlanguage.” International Review of Applied Linguistics.10:
209-231
Selinker, L and Gas, S (1994) Workbook on Second Language Acquisition. Rowley,
M.A: Newbury house.
Simatwo, S. (1993) “An Investigation of the Lexico-Semantic Errors of the Learners of
ESL: A Case Study of Nandi Speaking Pupils in Primary Schools”. Unpublished
M.Phil Thesis Moi University
Small, A., and Cripps, J. (2002). "Questions Parents ask". Toronto, ON: The Canadian
Hearing Society and the Canadian Cultural Society of the Deaf.
123
Sommer, B., and Sommer, R. (1991) A Practical Guide to Behavioral Research: Tools
and Techniques.3rd Edition. New York: Oxford University Press.
Spolsky, B. (1989) Conditions for second language acquisition Oxford: Oxford
University Press
Strong, M. (1988). Language learning and Deafness. Cambridge: Cambridge University
Press.
Suri, L.Z. (1991). Language transfer: A foundation for correcting the written English of
ASL signers. Technical Report 91-19, Department of Computer and Information
Sciences, University of Delaware, Newark, DE.
Tarone, E. (1978). “Conscious Communication Strategies in Interlanguage: a progress
report”. In H.D. Brown, C.A. Yorio and R. Crymes (Eds), on TESOL ’77.
Teaching and Learning English as a Second Language (pp. 194-203).
Washington, DC: TESOL.
Tarone, E. (1983). “Some Thoughts on the Notion of Communication Strategy”. In
Afresh and Kasper (Eds), (1983), 61-74.
Taylor, B.P. (1975). “The use of overgeneralization and transfer learning strategies by
Elementary and Intermediate Students of ESL”. Language Learning, 25:73 -107
Toth, A. (2002). Communication: Bridging the worlds of the deaf and the hearing.
Sarnia, ON: Clear Design.
Tweney, R.D., Hoemann, H W., Andrews, C. E. (1975) “Semantic Organization in Deaf
and Hearing Subjects”. Journal of psycholinguistic research 4:61-73.
Wamae, G.M. (2003) “Effects of Sign Language Mode of Instruction on Acquisition
Hearing Impaired form Two Learners”. Unpublished M.A. Thesis (Kenyatta
University.
Wario, H. (1991) Practical Primary English Pupil’s Book. Nairobi: Longman Kenya.
Wilbur, R.B. (1977) “An explanation of deaf children’s difficulty with certain syntactic
Structures in English.” In Journal of Speech and Hearing Disorders, Volume 55,
November 1990.
Wilbur, R.B. (2000). ‘The use of ASL to support the development of English and
literacy”. Journal of Deaf Studies and Deaf Education, 5(1), 81-104.
Wilkins, D. (1972) Linguistics and Language Teaching. London: Edward Arnold.
124
Wilkins, D. (1974) Second Language Learning and Teaching. London: Edward
Arnoldsing the Pictures below, write a story.
125
APPENDIX A: PICTURE STORY
Using the Pictures below, write a story
(Adopted from Kenya Literature Bureau (2004) 2nd Edition. Let’s Learn English)
126
APPENDIX B: CLOZE PASSAGE Please read the passage below and fill in the blanks using appropriate words.
A snake bite
Mboroki, Mbaabu and Nkirote were excited about going to ____________ their holidays
on their new farm. They were going to live in a new house and help their parents on the
farm.
When they arrived, their father and mother ____________them round the farm. The farm
was still covered by thick bush and needed clearing. On this first day, the children were
tired from their long journey from boarding school and they simply took ____________.
The next day, the children ___________up very early, just as the sun was ______.
They ________ porridge and tea for their breakfast and changed into
____________clothes to work on the farm led by their father, Mr. Ng'ang'a. Armed with
pangas and jembes, the children went to ____________ the bush.
‘Let us begin from here’, Mr. Ng'ang'a said. ‘We must get the bush cleared from near the
house first so that _________animals from the Mulika Game Park do no come right to
our house’
The family attacked the bush with their jembes and pangas swinging high above
their heads. The thick undergrowth of creepers made progress _______. Soon they were
covered in sweat.
‘This bush is full of ________ snakes and other animals’, Mr. Ng'ang'a told his
children. ‘You must therefore watch out as you work’.
As if it heard this warning, a black snake appeared near Mbaabu’s right arm, ready to
bite.
‘Mbaabu, move back! There is a snake’, Mboroki shouted.
It was too late. Before he had finished the sentence, the snake had ____________.
It dug its fangs into Mbaabu’s raised arm near the elbow. Mbaabu dropped his panga and
coiled in pain holding his injured arm with the other. Then the snake
______________immediately in the bush
Mboroki, Nkirote and ____________ father quickly went to Mbaabu.
127
‘Hold this joint firmly with both of your hands’, Mr. Ng'ang'a instructed Mboroki. I will
then tie his arm tightly with my handkerchief above the point that has been bitten. This
will stop the flow of poison to the rest of the body’
After Mbaabu’s arm had been _________, his father said, ‘Nkirote, run to the house and
ask your mother to give you a razor blade’
In a short while Nkirote was back together with her mother.
‘Did you see the snake?’ Mrs. Ng'ang'a asked her __________.
‘Yes’, he ___________
‘What colour was it?’
‘Black’
‘Oh black snakes are _____________. My poor son!’
‘Did you kill it?’
‘No. We couldn’t. It soon slithered away into the thick bush. We had ____________
attend to Mbaabu first before thinking of killing the snake’, Mr. Ng'ang'a explained.
‘The hospital. Please rush him to the hospital’, said Mrs. Ng'ang'a anxiously.
With the razor blade Mr. Ng'ang'a then made a cut on Mbaabu’s arm and let it bleed.
‘This is good for you. If the snake is poisonous, most of the poison will flow________
with the blood. Let it bleed’, his father said.
His father soon took Mbaabu to hospital _______a Matatu. After half an ________drive,
they were at the hospital.
‘It is a snake bite’, Mr. Ng'ang'a explained to a nurse,’ here on his arm. I have tied and
bled the arm’.
‘That was very thoughtful of you, Mr. Ng'ang'a’, the nurse said. ‘It is the right thing to
do. The doctor is not here now. I will give Mbaabu some anti-snake serum, just in case
the snake is ____________. I will then admit him for tonight. The doctor will be here
tomorrow. He will then tell us whether the snake that ______ your son is poisonous or
not’.
Mbaabu was soon fast asleep. He was woken up by the doctor the __________morning.
128
APPENDIX C: ERROR SUMMARY TABLE
Key: H.P – hearing pupils
H.I – hearing impaired pupils
Error Type
Group
Level Cloze
PassagePicture
storyFree
Composition
Total Errors
Per level
TotalErrors
PerCategory
%
HP
6 7 8
533629
187224128
201247177
441 507 334 1282 19.67
Learning induced errors HI
6 7 8
736259
306279255
313265239
692 606 553 1851 28.40
HP
6 7 8
1392
12167
141412
39 39 21 99 1.52Meaning
similarity errors HI
6 7 8
485
556
654
15 18 15 48 0.74
HP
6 7 8
300
9149
16115
28 25 14 67 1.03Meaning
duplication errors.
HI 6 7 8
121
483
565
10 16 9 35 0.54
HP
6 7 8
1445
151815
30166
59 38 26 123 1.89Semantic
contiguity errors
HI 6 7 8
8119
252623
202519
52 62 51 165 2.53
HP
6 7 8
200
421
451
10 7 2 19 0.29Coinage
errors
HI 6 7 8
101
534
243
8 7 8 23 0.35
129
HP
6 7 8
300
124
581
9 10 5 24 0.37
Paraphrase errors
HI 6 7 8
233
454
454
10 13 11 34 0.52
HP
6 7 8
594
564940
487332
109 131 76 316 4.85Sound
similarity errors
HI 6 7 8
283
334
353
8 16 10 34 0.52
HP
6 7 8
15113
435330
675959
125 123 92 340 5.22Collocation
errors
HI 6 7 8
81115
594643
746462
141 121 120 382 5.86
HP
6 7 8
171
000
000
1 7 1 9 0.14
Haphazard errors
HI 6 7 8
10510376
246230191
297236182
648 569 449
1666 25.56
Sub Total
813 2734 2970 6517
6517 100%
130
APPENDIX D: SAMPLE COMPOSITIONS H.I class 8 picture story
131
132
H.I composition class 6
133
H.I free composition class 8
134
135
H.I free composition class 8