FACULTAD DE FILOSOFÍA Y HUMANIDADES DEPARTAMENTO DE LINGÜÍSTICA
Acquisition of Lexical Competence in English as a Second Language
Informe Final de Seminario de Grado para Optar al Grado de Licenciado en Lengua yLiteratura Inglesas
Profesora Guía: Alfonsina Doddis Tutora: Ximena Tabilo Alumnos Participantes: Valeria Echeverría Gregorio Godoy Carolina Norambuena Marcela Rivera Quirian Sani Eduardo Tapia Consuelo Toro Katherine Varas Javiera Zamora Santiago, Chile 2007
Table of contents
1. Introduction 7
2. Literature review 10
3. The study 20
3.1. Theoretical and descriptive framework 20
3.2. Objectives 29 3.2.1. General objective 29 3.2.2. Specific objectives 29 3.3. Hypotheses 31 3.4. Research questions 31 3.5. Methodology 32 3.5.1. Subjects 32
3.5.2. Elicitation of the data 34 3.5.2.1. Breadth tests 34
3.5.2.1.1. Receptive Breadth test 34 3.5.2.1.2. Productive Breadth test 36
3.5.2.2. Depth 37
3.5.3. Procedure 39
4. Discussion of the results 47
4.1. Test results 47
4.1.1. Quantitative results 55 4.1.1.1. Receptive Breadth test results 55 4.1.1.2. Productive Breadth test results 57 4.1.1.3. Depth test results 59 4.1.1.4. Relationship between final marks and correct responses 61 4.1.1.5. Successful and less successful students 63 4.1.2. Qualitative results of the Productive Breadth test 65
5. Conclusions 72
6. Bibliography 78
7. Appendices 84
7.1. Appendix A. Successful and Less successful students 84
7.2. Appendix B. Tests 88
7.2.1. Test 1. Receptive Vocabulary Levels test 88 7.2.2. Test 2. Productive Vocabulary Levels test 93
7.2.3. Test 3. Word Associates test 98 7.3. Appendix C. Correlation Analysis 104
7.4. Appendix D. Test results 106 7.5. Appendix E. Qualitative results 126 7.6. Appendix F. List of Common Mistakes 131
ACKNOWLEDGMENTS
"We would like to deeply thank Professor Alfonsina Doddis for the time and effort spent guiding and reviewing this seminar study. We also would like to express our gratitude to our Tutor Ximena Tabilo, who provided us with useful and helpful assistance. Additionally, we are very grateful to all the participants who collaborated in the development of this research, allowing it to be completed.
We acknowledge our professors and classmates for their encouragement, support and patience throughout the duration of this process.
AGRADECIMIENTOS
En primer lugar, quiero agradecer a mi familia, en especial a mis padres por su apoyo, comprensión y cariño, siendo para mí un ejemplo de superación y esfuerzo, ya que sin ellos nada de esto hubiese sido posible. A mis hermanos, por alegrar cada momento con una sonrisa. También quiero agradecer a mis amigos, en especial a Katherine, María José, Daniela y Carolina por estos cuatro años de momentos inolvidables, ya que me han brindado cariño, alegría y apoyo a pesar de todo, tanto en mi vida académica como en mi vida personal. Un profundo agradecimiento a todos aquellos que hicieron este camino mas agradable y feliz.
Valeria Echeverría Santibáñez Agradezco, en primer lugar, a todos los que aportaron en mi formación académica, en especial a las y los profesores Coty Vivanco, Patricio Novoa, Marco Espinoza y Pascuala Infante por su apoyo y buenos consejos. En segundo lugar agradezco a todos los que ayudaron e hicieron aportes a esta investigación en especial a las profesoras Doddis y Tabilo. Este trabajo y las horas hipotecadas en él, están dedicadas a Beatriz.
Gregorio Godoy Báez Por sobre todas las cosas quiero agradecer a Dios ya que sin su infinito amor y ayuda nada de esto hubiese sido posible. A mis padres y a mi hermano por ser los pilares fundamentales en mi vida y por brindarme su amor y apoyo incondicional. A mi madre por su dedicación y esfuerzo para ver mis sueños hechos realidad. A Felipe, por su amor y comprensión durante todos estos años y hacer mi vida aún más feliz. Agradezco a mis profesores por entregarme sus conocimientos y formarme como profesional, especialmente a Daniel Muñoz por su buena voluntad y disposición para ayudarme en todo momento. A todas las personas del departamento de Lingüística, en especial a Patricia Osorio por estar siempre dispuesta a ayudarme. A mis amigos y compañeros por entregarme cariño y alegría y hacer de estos años una etapa inolvidable.
Carolina Norambuena Muñoz Quiero expresar mi gratitud a Dios por ayudarme en todo momento; al igual, agradezco a mi hijo Benjamín, a quien amo infinitamente, por alegrar todos los dias de mi vida; a mis padres y a mi abuela, quienes me infundieron valores, me apoyaron y creyeron siempre en mi y por supuesto a mi esposo Dankan a quien amo profundamente. También a mis profesores, especialmente a Daniel Muñoz, quienes fueron cruciales en esta etapa. Finalmente, pero no menos importante, quiero agradecer a mis amigos y compañeros, ya
Marcela Rivera Lazo Quisiera agradecer a mis padres y amigos por todo el apoyo y comprensión que me han dado en este proceso; a los profesores, en especial a Daniel Muñoz y Pascuala Infante, por su constante apoyo, ánimo y orientación; a la señorita Patricia Osorio por brindarnos su ayuda en el momento apropiado, dándonos palabras de aliento cuando más lo necesitábamos; a todos aquellos que cedieron sus horas de clases para tomar las muestras y a los alumnos que las tomaron seriamente. Este trabajo no hubiera sido lo mismo sin su colaboración.
Quirian Sani Vegas A mis padres, Nora y Eugenio, por su amor, comprensión y apoyo irrestricto. A mis
hermanas por su paciencia y tolerancia. A los Gajardo Ocampo, mi segunda familia. A mis grandes amigos Lucas, Natalia y Giovanna junto con sus respectivas familias, gracias por todos estos años y los que vendrán. A Patricio Novoa, Rosita Rodríguez y Daniel Muñoz por su apoyo más allá de lo académico. A mis compañeros, por haber hecho posible este trabajo. Finalmente, y de manera muy especial, a mi abuela.
Eduardo Tapia Fuentes Quiero, en primer lugar, agradecer a Dios por todo lo que me ha ayudado y escuchado. Tambien quiero agradecer a mi familia por todo su apoyo en esta y en todas las etapas de mi vida, en especial a mi mamá y a mi papá por todo su amor, consejo y comprensión. A David, por su incondicional amor y apoyo que me ha entregado en estos 6 años. Gracias a mis compañeros y amigos con los que pasé tantos momentos inolvidables. Finalmente, a mis profesores, gracias por su tiempo y dedicación.
A todos los quiero mucho.
Consuelo Toro Navarro Gracias a todas las personas que estuvieron conmigo durante estos años. A mi familia por su apoyo y cariño. A mis padres, sobretodo a mi madre por su amor, paciencia y preocupación. A mis amigos por sus consejos, comprensión y compañía, en especial Val, Coté, Dani, Bárbara y Carito. Y a mis profesores por inspirarme y motivarme a lo largo de este proceso, en particular Miss Coty Vivanco y Miss Rosita Rodríguez.
Katherine Varas Armijo Quisiera agradecer a mi familia, pilar fundamental en mi vida, por su amor, apoyo y comprensión. Especialmente a mis padres por su incondicional apoyo en todos los ámbitos, por su amor y completa dedicación. A mis profesores, especialmente, a Miss Ximena Tabilo, que no solo contribuyeron en mi formación académica y profesional, sino también, personal. A mis amigos y compañeros, en especial, a los chicos de la ex - sección 3: gracias
A todos ustedes, mi mas profunda gratitud y aprecio…sin todos ustedes no habría llegado adonde estoy.
Javiera Zamora Chandia 1. Introduction
Applied Linguistics developed in Europe in the 1950’s as an interdisciplinary field of research whose object of study was different aspects of language use. Thus, it focuses on the relationship among psycholinguistic and sociolinguistic theory, together with social practice, and the acquisition and use of language in various contexts. Furthermore, it includes not only Second Language Acquisition (SLA) and fields related to it, but also first language acquisition and areas such as communication disorders, language and the media, discourse and conversational analysis, sign language research, etc.
Three aspects of foreign language study are relevant to Applied Linguistics. First and closely related to stylistics, Applied Linguistics focuses on the acquisition of communicative competence in a foreign language, in terms of the ability to decode and understand messages and the value of their respective parts. The second important aspect has to do with the social problems associated with the use of a foreign language by non-native speakers in the target country. Finally, the third aspect regards what Pennycock (1990, 1997) considered 'critical applied linguistics', i.e., a critical attitude towards the research and practice of language in discourse, pedagogy and education.
Kramsch (1999) defined Applied Linguistics as an interdisciplinary field that mediates between the theory and the practice of language study, investigating the
relationship between language forms and use and between different kinds of discourse. This discipline includes Second Language Acquisition among its fields of study, which “focuses on the acquisitional aspect of language learning and teaching, both inside and outside the classroom” (Kramsch 1999: 314). SLA research has to do with the processes that take place within learners when acquiring second or more languages, as well as with the ability to speak and read them in normal everyday situations. The term ‘second language’ (L2) acquisition generally refers to a language acquired apart from the mother tongue, either in natural or instructional settings where that language is the official language. In contrast, ‘foreign language’ (FL) acquisition refers to languages learnt only in instructional settings. Thus, Applied Linguistics includes Second Language and Foreign Language Acquisition within its field of study. Additionally, “SLA is interested in the nature of [...] learner languages and their development throughout life, as well as in bilingualism, language attrition and loss” (Kramsch 1999: 315). Nevertheless, in this study, the terms Foreign and Second Language Acquisition are used as synonyms.
Depending on the researcher’s point of view, Applied Linguistics can centre on multiple disciplines such as linguistics, anthropology, psychology, education, or literature, apart from a variety of social and critical theories. On the other hand, SLA research is based on psycholinguistics, including sociocultural and sociolinguistic theory. Because of this, the L2 or FL learner is viewed as a social and cultural being whose psychological processes are experienced with the interaction with others and then are internalized as individual cognitive processes.
longitudinal ones in the study of learner’s lexical competence because of researchers’ time restrictions. Therefore, there is scarce research work on the vocabulary acquisition process and the circumstances in which it develops. In turn, this research study is also a cross-sectional study that intends to measure the vocabulary knowledge of learners of English as a second or foreign language. Since this research had to be conducted only in one year, it was not possible to track the development of the lexical competence of the same subjects throughout the four years of their academic programme. Thus, to compensate for this limitation, students from the four different levels of the programme, i.e., Pre-intermediate, Intermediate, Post- intermediate, and Advanced, were subjects of the study. In order to assess the subjects’ vocabulary knowledge, three different vocabulary tests were applied: the Receptive, and Productive Vocabulary Levels Test, and the Word Associates Test. The assessment results obtained were analysed and described in quantitative terms to prove the hypotheses that vocabulary knowledge increases together with the level of communicative competence, and in qualitative terms in order to describe the types of errors made by the subjects in the production of vocabulary. Finally, conclusions were drawn from the discussions of the results and limitations of the study were identified.
Concerning its formal arrangement, the present research report has been organised into seven sections. This first section, introduction, is followed by the literature review, which is centred on the revision of different proposals on vocabulary acquisition. Thirdly, the study section is subdivided into the theoretical and descriptive framework of the research, the objectives, hypotheses, research questions of this report, methodology used to obtain the corpus, subjects selected, description of the tests applied, procedures followed to
obtain the data, and the different criteria for the analysis. Fourthly, the discussion of the results in quantitative and qualitative terms is presented. Finally, the following section summarises the conclusions of the study, followed by section 6 and 7, which present references and appendices, respectively.
2. Literature review
Accurate definitions of the concepts of vocabulary learning and lexical competence are necessary in order to establish a specific theoretical basis for research on vocabulary learning. For this purpose, Jiménez (2002) intends to organize and clarify approaches and definitions of vocabulary acquisition proposed by different authors, and to assess different approximations to this problem. Richards (1976) was one of the first applied linguists to propose characterizations of the concepts of vocabulary learning and of lexical competence, i.e., “knowing a word”. According to Richards, knowing a word implies knowledge of semantic and grammatical aspects related to a word, such as collocational and syntactic patterns as well as its different semantic relations.
An account of research on lexical competence can be organized in terms of four main investigative approaches. First, authors such as Beheydt (1987), Carter (1987), Robinson (1989), Nation (1990), Laufer (1991), among others, have focused on defining or listing dimensions centred on different aspects of lexical competence, covering a wide variety of features, such as psycholinguistics factors that interfere in this process. Secondly, authors such as Meara (1996a, 1996b) and Henriksen (1999) suggest other possibilities for the
competence, i.e., lexical organization and breadth of lexical competence. The proposals made by Henriksen are Partial-Precise Knowledge, Depth of Knowledge and Receptive-Productive vocabulary. Thirdly, authors such as Robinson (1989), Lennon (1990), and Schmitt (1995) intend to apply definitions and views of lexical competence to vocabulary teaching. Finally, authors such as Schmitt and Meara (1997), and Wesche and Paribakht (1996) have made empirical studies related to different aspects of lexical knowledge, like breadth or size in L1 and L2.
Most definitions of lexical competence are based on a list of dimensions that are not integrated into a semantic or psycholinguistic theory. In addition, some dimensions related to lexical competence are not taken into account in most investigations, such as the semantic dimension, which includes variables like figurative language or idiomatic expressions. This dimension is essential in order to understand not only the linguistic knowledge of a word, but also the underlying psycholinguistic processes involved in that knowledge.
Read (2000) deals with vocabulary assessment and its implications. First, he suggests three dimensions for vocabulary assessment that constitute the principal framework proposed by this linguist as a basis for a range of lexical assessment procedures. The first dimension corresponds to Discrete versus Embedded tests. A Discrete test is considered as a measure of vocabulary knowledge or use as an independent construct whereas an Embedded test is a measure that forms part of the assessment of some larger construct. The second dimension refers to Selective versus Comprehensive tests. Selective tests focus on the assessment of specific lexical items, whereas Comprehensive tests focus on the whole
vocabulary content of the input material. Finally, the third dimension is Context-Independent versus Context-Dependent tests. The first term refers to a measure in which there is no context for the test-taker to produce the expected answer. On the contrary, the second term refers to a measure in which in order to produce the appropriate answer, the test-taker must consider contextual information.
Secondly, Read intends to elucidate the concept of vocabulary, discussing what knowledge of a lexical item means and how words interact in larger units. Besides, this author gives an account of vocabulary acquisition research and of language assessment, focusing on the problems faced when testing vocabulary. Regarding language testing, Read revises four vocabulary tests, which contain two vocabulary size tests: the Vocabulary Levels Test, and the Eurocentres Vocabulary Size Test (EVST). The Vocabulary Levels Test, first designed by Nation, is divided into five parts that correspond to five levels of word frequency in English: the first 2000 words, 3000 words, 5000 words, the University Word level (beyond 5000 words) and 10000 words. The tests involve word-definition matching. At each level, there are 36 words, selected at random, and 18 definitions. The Eurocentres Vocabulary Size Test consists of a checklist that calculates the learner’s vocabulary size through a graded sample of words that represent several frequency levels. The learner has to indicate whether he knows a series of words belonging to a checklist. Non-words are also included in this list with the intention of adjusting the test-takers’ scores if they are exaggerating their vocabulary knowledge. As a result, if the student declares to know some non-words, the final score is reduced.
Scale (VKS), and the reading comprehension section of the Test of English as a Foreign Language (TOEFL). The Vocabulary Knowledge Test consists of two scales: one for eliciting responses from the test-takers, and one used for scoring the responses. The first scale and a list of words are presented to the test-takers, who have to decide for each word on the list which category adjusts to their knowledge of the word. Category I means that the word is not familiar at all, in category II, the word is recognized but the person does not know its meaning. In categories III and IV, the test-takers need to give some attestable evidence of the knowledge of the word, with a synonym or translation equivalent. In category V, it is necessary for the test-takers to demonstrate that they can use the word in a sentence. Finally, the second scale, known as the scoring scale, converts the test-takers’ responses to each word into test scores. Finally, the Test of English as a Foreign Language measures language proficiency. It consists of one single item: the multiple-choice item. The main purpose of this test is to assess the level of proficiency in English of students who want to study in a foreign country where English is the language of instruction.
In addition, Read concentrates on the design of vocabulary tests, taking into account the goal of the instrument, the definition of design, the characteristics of test input, and candidates’ responses. He deals with the design of Discrete-point tests, i.e., those that measure vocabulary knowledge or use as an independent construct, of Comprehensive tests i.e., those that focus on the whole vocabulary content of the input material.
Finally, the author suggests new directions in the development of vocabulary assessment, determining future needs of the field. These needs include the use of computer-based corpus research, the necessity to devise new word lists (scientifically constructed,
current, and accessible to the common public), and the inclusion of multiword lexical items in vocabulary assessment. A further aspect that should be considered by vocabulary assessors, in future research, is the vocabulary of informal speech, including social and cultural variables in second language acquisition, in order to assess test-takers from all extractions.
As regards Second Language Acquisition, Segler (2001) considers vocabulary as an essential element in the acquisition of a second language; as a result, lexical items have to be learned with precision in order to understand and be understood in the target language. Segler draws attention to the fact that lexical errors are the most common ones when learning a second language; besides, they are the most troublesome ones for native speakers with respect to interpretation. Considering this, it is necessary to reflect on the importance of lexical development regarding second language acquisition, and the strategies used by learners to learn vocabulary.
Vocabulary knowledge has been a matter of research only during the last two decades. Lafford, Collentine, and Karp (2000) also suggest that the study of L2 vocabulary knowledge is important because lexical errors are the most frequent ones and, at the same time, they constitute a serious impediment to communication. Concerning the components of the L2 lexicon, the authors propose that this is composed of lexical items consisting of one or more words that correspond(s) to a single meaning unit. In relation to lexical competence assessment, the authors agree with Meara (1996, in Lafford, Collentine, and Karp 2000), who proposes that lexical competence is measured by both the size of a
vocabulary size, also known as breadth of vocabulary knowledge, is essential to predict reading success. On the other hand, the organization of these items has to do with a clear understanding of several aspects of a word, such as its spoken and written form, grammatical and collocational patterns, semantic and pragmatic aspects, among others; these aspects of a word are related to what is known as depth of vocabulary knowledge, an important aspect in vocabulary assessment.
Concerning research on depth of vocabulary knowledge, Qian (1998) investigates the relationship among reading comprehension and depth of vocabulary knowledge. In this research study, Qianapplied a series of tests to 74 adult Chinese and Korean speakers with the purpose of evaluating the influence of depth of vocabulary knowledge on their reading comprehension proficiency. The results of this research confirm the importance of depth of vocabulary knowledge in reading comprehension processes, and prove that there is a close relation between breadth and depth of vocabulary knowledge, and reading comprehension. Therefore, this study reveals that it is essential to improve the depth of L2 learners’ vocabulary knowledge for reading comprehension purposes. Qian’s work is significant for the field of vocabulary knowledge research because he has carried out several studies on the influence of vocabulary size and depth on L2 reading comprehension. In 1999, Qian studied the relationship between depth and breadth of vocabulary knowledge and reading comprehension in English as a second language (ESL). In order to examine this relationship, a group of young adults ESL learners with a minimal vocabulary size of 3000 word families was tested. In this study, depth of vocabulary knowledge for reading comprehension implies the awareness of word meaning, register, frequency, syntactic
properties, pronunciation, spelling, and morphological properties. As regards vocabulary size for reading comprehension in ESL, the threshold hypothesis suggests that “there is a threshold level below which the reader will be able to apply his or her reading strategies to help comprehension and achieve better results” (Qian 1999: 4). The threshold level corresponds to a minimum vocabulary size of 3000 word families, that is to say, words that have the same root or base sound.
Qian’s study revealed the strong relationship existing between breadth and depth of vocabulary knowledge for ESL learners with a minimum vocabulary size of 3000 word families. Besides, it considers depth of vocabulary knowledge as an essential factor in the relationship between vocabulary knowledge and academic reading comprehension. This study is important, since its results have repercussions for second language education and research. Consequently, in order to improve the reading comprehension of students, it is necessary to take into account components of vocabulary depth such as syntactic properties of words, word meaning, etc.
Another proposal on depth of vocabulary knowledge has been developed by Borer (2004), who elicits the four dimensions of word knowledge proposed by Nation (1990), i.e., form (spoken and written), position, function, and conception, which stand for depth of knowledge of a word. Based on these dimensions, the elements investigated are the following:
2) Position: grammatical and collocational patterns.
3) Function: modes of discourse.
4) Conception: word meaning related to text context and synonymous/antonymous relations.
These dimensions are essential for learners to understand L2 academic texts, being necessary not only to know the meaning of a word, but also sets of word forms, sharing a common meaning i.e., word families. In this way, learners are able to read and understand texts.
With relation to receptive and productive vocabulary size, Waring (1997) made a study with the purpose of discovering dissimilarities between receptive and productive vocabulary. The subjects of this research were 76 female Japanese learners of English, who were tested in order to measure their knowledge regarding receptive and productive vocabulary. The tests chosen for this experiment were based on Nation's (1990) Receptive Vocabulary Levels test, and Laufer and Nation's (1995) Productive Vocabulary Levels test. The results of the study show that receptive vocabulary is larger than productive vocabulary, receptive vocabulary being easier to access than productive vocabulary. Crow (1986) suggests that this occurs because a larger body of knowledge, including meaning, spelling, pronunciation, etc. is needed for using a word (productive vocabulary). Other authors such as Blum and Levenston (1978) state that receptive vocabulary is larger than productive, since even though the words may be recognized receptively, they are not used
by learners. Other possible reasons that could explain why receptive vocabulary is larger than productive vocabulary are that comprehension takes place before production, or that production is more difficult than reception.
In 1983, the Vocabulary Levels Test was developed as a diagnostic instrument for use by teachers. It was designed by Paul Nation, and was republished later in 1990. In 1988, Read started an initial validation work on the test, finding that the scores on the different frequency levels tended to fall into an implicational scale, that is to say, there was a tendency in which subjects that knew lower frequency words tended to know higher frequency ones. In 1993, Schmitt revised the Levels Test (Version A) and created three additional versions (Versions B, C, D). Although Schmitt was not able to validate these three new versions properly, they have been used as assessment tests in different institutions as well as in a considerable number of vocabulary research studies.
In 1999, Laufer and Nation devised a productive levels format of the test, based on tests A-D. In the same year, Beglar and Hunt started some validation work on the test, specifically concerning the 2000 and University Word List sections of versions A-D. This work found a close relation between the scores of the learners´ level sections and their TOEFL scores.
Messick (1989) suggested that validation of a test should include logical argumentation and empirical evidence, based on quantitative and qualitative information. Quantitative procedures such as item analysis, profile analysis, factor analysis, and examination of the
qualitative procedures employed may correspond to interviews with some examinees, in which they were asked about their opinions on the test and about the steps they follow when dealing with the items.
In addition, an accepted validation of a test includes an examination of the correlations of the scores with other measures, taking into account variation of degrees of similarity and differences with the test. There are convergent and discriminate patterns of relationship. The convergent evidence should address the same construct. Usually, vocabulary tests are correlated with proficiency tests, especially with TOEFL. However, it has been pointed out that the TOEFL measures proficiency rather than vocabulary; therefore, it seems to be an unsatisfactory procedure. On the contrary, a good and validated criterion of probing vocabulary knowledge is through a personal interview in order to establish whether the examinees know the target words; then, the results of the interviews should be compared to the test results. One of the standard techniques is the comparison of the responses of an individual item with the scores on the overall test. This procedure may function appropriately since all the items are related to the same construct in the same way. However, vocabulary items are Discrete; that is to say, knowledge of the majority of the words of a level does not imply knowledge of all the words of the same level. Moreover, the degrees of difficulty are different depending on the L1 of the test-takers.
Strictly speaking, vocabulary items in clusters are not independent, although independence and non-independence should be seen as a continuum rather than as a dichotomy. If a test-taker knows the words in a cluster, his answer will be independent, but if one or more words is (are) not known, the answer will be dependent to varying degrees.
This implies that lower vocabulary size test-takers have a tendency towards dependence in the levels test. Generally, test-takers do not fill in the blanks when they do not know the answers. It is very difficult to determine the degree of independence precisely, but it can be suggested that there is some degree of independence within the clusters.
Since vocabulary knowledge is multi-faceted, no currently available format of vocabulary test is able to assess all forms of lexical knowledge.
3. The study
3.1. Theoretical and descriptive framework
This research study on vocabulary acquisition is based on Henriksen’s proposals (1999) for the description of vocabulary knowledge, also known as lexical competence. There have been several attempts to define what vocabulary knowledge precisely involves.
According to Read (2000), in order to assess vocabulary knowledge, it is essential for SLA research to explore the nature of words and vocabulary. Vocabulary is usually regarded as a set of individual words with related meanings, just as a dictionary. Nevertheless, under the development of Applied Linguistics and language teaching, the nature of vocabulary seems to be much more complex than just a mental dictionary. Thus, for research purposes, it is crucial to define what a word is and what to know a word means, although this issue is complex to tackle in practical terms.
The term ‘word’ is a difficult concept to define, in spite of the fact that the basic objective of vocabulary testing is the assessment of the ‘knowledge of a word’. For this reason, it is important to clarify a basic distinction between ‘tokens’ and ‘types’in order to count words in a text. Tokens refer to individual items that are counted each time they occur. Instead, types refer to different word forms that are counted only once when they occur many times in a text. The relative proportion of types and tokens (type-token ratio) is used as a measure of language development of language learners and native speakers.
There is also another important distinction concerning vocabulary, the one between ‘function’ and ‘content’ words. On the one hand, function words are regarded as grammatical rather than semantic lexical items, such as articles, pronouns, and auxiliary verbs. This kind of words play a functional role in the language and have little meaning in isolation. On the other hand, content words are the ones that convey meanings, such as nouns, full verbs and adjectives. For this reason, content words have been the focus of vocabulary assessment. A problem that arises at this point is that content words vary in form. This variation occurs either by the addition of inflections, or by their derived forms which usually change the word class and add different meanings. Additionally, some words may or may not belong to the same word family, by being related or not in form and meaning. This is an important issue in the measurement of vocabulary size, i.e., to distinguish word forms and word families, because this is the reason for researchers’ different estimations of how many words speakers know.
In addition, when assessing vocabulary, it is not clear whether a whole word family or just one word is being assessed. Another problem is the existence of homographs, which
are single words that have so different meanings that they cannot be regarded as belonging to the same word family, thus having separate entries in dictionaries.
A further complication is that vocabulary also consists of larger lexical units, more than just single words. Examples of these expressions are phrasal verbs, compound nouns, idioms, and lexical phrases; all of them function as a unit in spoken or written discourse. In spite of their importance in the use of language, the assessment of multi-word items has not received much attention due to its practical difficulties, besides, multi-word items are likely to play a role in Embedded, Comprehensive and Context-dependent vocabulary measures. Consequently, the term ‘word’ refers to a wide variety of lexical units. This is why it is essential to propose a clear definition of this term for research on size and depth of vocabulary.
As mentioned above, describing the nature of vocabulary knowledge is not an easy task. Some applied linguists have worked on the description of lexical competence. For example, Richards (1976) made a series of assumptions about what lexical competence involves. The first assumption is that unlike grammatical competence, native speakers’ vocabulary knowledge keeps expanding through adulthood; the second assumption stated that knowing a word implies to know the degree of probability of finding that word in a written or oral text. Third, knowing a word means to be aware of the limitations in the use of the word in relation to variations of function and situation. In turn, the fourth assumption is related to the knowledge of the syntactic behaviour associated with the word; fifth, knowing a word entails the knowledge of the base form of a word and the possible
word implies the knowledge of the network of associations of that word with other words in the language. The seventh assumption is related to the knowledge of the semantic value of a word. The last assumption deals with the knowledge of the different meanings associated with a word. These eight assumptions have been treated as a general framework of vocabulary knowledge.
Nation (1990), based on Richards’ assumptions, proposed a very useful distinction, that is, receptive and productive vocabulary. According to Nation, it is the difference between the ability of recognition of a word when it is seen or heard, and the ability to use a word in a written or spoken situation. Production of a word needs a higher level of knowledge than reception does. Nation stated that productive knowledge applies not to the total vocabulary of a native speaker, but just to a small proportion of it. Nation (2001) also stated that knowing a word involves knowledge of form (spoken, written and word parts), knowledge of meaning (the form and meaning, concept and referents, and associations), and knowledge of the use (grammatical functions, collocations, and constraints on use, such as register and frequency).
Meara (1996) noted the difficulty involved in the creation of a construct that could measure vocabulary knowledge in depth, an assumption which was supported by Schmitt and Meara (1997), and Schmitt (1998). An alternative approach is proposed by Dale (1965), who produced scales in order to represent the different degrees of partial knowledge of words. Dale defined four stages related to the knowledge of a word. Aditionally, a similar proposal was made by Paribakht and Wesche (1993).
Besides, Meara (1996) made a distinction between assessing the whole state of a learner’s vocabulary and how well individual words are known; for general proficiency purposes, the author favour the first one. In turn, Chapelle (1994) proposed a definition of vocabulary ability (based on a Bachman (1990) construct), which involves language knowledge and the ability to use it in context, including the following components: the context of vocabulary use, vocabulary knowledge and fundamental processes, and metacognitive strategies for vocabulary use. However, applied linguists and second language teachers have paid more attention to the second component of Chapelle's (1994) framework of vocabulary ability, which has four dimensions:
a) Vocabulary size: it refers to the total amount of words that a person knows.
b) Knowledge of word characteristics: it has to do with L2 learners’ knowledge of each word range from vague to precise.
c) Lexicon organization: it deals with the manner in which words are stored in the mind.
d) Fundamental vocabulary processes: they are used in order to facilitate access to vocabulary knowledge when writing and/or speaking, which work faster in native speakers than in L2 learners.
As a summary, it is necessary to bear in mind all the complexities implied in knowing a lexical item in order to delimit the assesment of vocabulary knowledge. Nation’s (1990)
of what vocabulary knowledge is, is useful for the purposes of this study. Furthermore, the difficulties noted in the creation of constructs to measure vocabulary ability, and the distinction made by Meara (1996) between assessing individual items and assessing the whole state of a learner’s vocabulary knowledge are also useful for the present study.
In order to give an account of the complexity of vocabulary knowledge, linguists such as Meara (1996) have proposed a global description of such knowledge, proposing only two dimensions, size, i.e., the number of words a subject seems to know, and organization, i.e., how these words are related among them. Other applied linguists have suggested general characterizations with several separate traits as different aspects of word knowledge. However, Henriksen (1999) proposed a more specific definition of lexical competence, which is in the middle of the global and the separate traits proposals. She suggested the existence of three lexical dimensions, as well as the need for specifying each one of them and their relationships, and explained how important they are in the word learning and use processes.
The first dimension, Partial - Precise Knowledge, deals with breadth or size of vocabulary knowledge. Meara (1996) defined it as precise comprehension, that is to say, to know the meaning of a word implies the capacity to, for example, translate it into the L1, to identify its appropriate definition, or to paraphrase it in the target language. For some researchers, it is a matter of time for learners to go from a partial to a precise comprehension of vocabulary, even though sometimes it is not necessary to know the exact meaning of a word as this can be inferred from the immediate context. The transition from Partial to Precise Knowledge is a long process in which learners gradually move from
simple word recognition to a precise comprehension level, after having passed through several stages of partial knowledge. In this process, the understanding of a lexical item is changed and increased as a result of personal experiences. At this point, learners have already created their own hedges when establishing lexical categories and networks. Wesche (1996) proposed the term ‘mature lexical entry’ for describing the mental representation that comes after the initial knowledge. This mature lexical entry is gained through the process of mapping and it refers to a wider knowledge base of the kind adult native speakers develop.
The second dimension, Depth of Knowledge, is the quality of the learner’s lexical competence (Read 1993), or as Henriksen (1999) defined it, “the knowledge of a word’s different sense relations to other words in the lexicon, e.g., paradigmatic (antonymy, synonymy, hyponymy, gradation) and syntagmatic (collocational restrictions)”. Because of its complexity, many types of knowledge comprise full understanding or rich meaning representation of a word. Following Cronbach (1942), rich meaning representation or knowledge of a word meaning implies (a) extensional relations between concept and referent; (b) intensional relations to other words in the vocabulary (paradigmatic and syntagmatic relations); and (c) being aware of the syntactic and morphological limitations and characteristics of a word, together with levels of accessibility to this knowledge. However, this characterization does not consider aspects such as spelling, pronunciation, and collocation since these have always been taken for granted.
signifier and signified are connected and, as a result, extensional links are established by means of labelling, i.e., linking concept, sign and referent, and packaging, i.e., grouping together different meanings for the same word. Within network building, also known as organizational dimension (Meara 1996), intensional links are created and sense relations between lexical items are developed. According to Beheydt (1987), the process of semantization of a new word is complete when the learner has identified its morphological, syntactic, collocational characteristics, and its possible meanings.
The third dimension, Receptive – Productive, is divided into receptive and productive vocabulary. Receptive vocabulary is defined as the ability to understand a lexical item; and productive vocabulary, as the ability to use it in production. Henriksen (1999) proposed that just a limited number of lexical items that are available receptively would become productive. Therefore, to know a word receptively does not imply to know how to use it productively; thus, the amount of receptive vocabulary is bigger than the productive one. According to Nation (1990), productive knowledge contains and expands receptive knowledge. Furthermore, as mentioned before, it is assumed that both types of vocabulary are operating on a continuum instead of on a dichotomy, though it is unclear if there is an evident division between them. Segler (2001), on the contrary, stated a set of assumptions regarding this division. First, there are overlapping phases when going from receptive to productive vocabulary knowledge. Second, their relationship is not static. Third, according to Crow (1986), productive vocabulary entails more information related to denotations, connotations, derivations, register and syntactic constraints. Fourth, productive vocabulary usually follows receptive vocabulary. Finally, the distance in size between them
would decrease as the learning process develops, but receptive vocabulary would continue being larger.
Melka (1997) attempted to clarify this difference under the concept of ‘word familiarity’, by means of which it could be established at what point the existing knowledge of a word becomes productive. In turn, Meara (1996) thought that attention should focus on the hidden lexical competence of ‘automaticity’ as this serves to develop both receptive and productive skills. Following this idea, N. Ellis (1995) added that learners have to acquire both the input and output features of a lexical item, together with mapping them to semantic and conceptual meanings (mapping meaning to form).
Concerning the relationship between dimensions 1 and 2, when acquiring the sense of a word and specifying its field of reference, i.e., labelling and packaging, respectively, learners use and develop knowledge of paradigmatic relations, i.e., network building, creating intensional and extensional relations. Thus, dimension 2 is an important factor in the development of dimension 1. In the process of understanding a word, learners have to develop a link between sign and referent. Later on, in order to establish differences among several lexical items, they have to link sign and referent, and, in addition, to rearrange the intensional relations between the items in the lexical set. For a clear understanding of each particular item within a lexical set, it is necessary for learners to clearly know the relationship between lexical items. In turn, Qian (1999) proposed that breadth and depth are highly interconnected dimensions of vocabulary knowledge, whose development is interdependent to a considerable extent. This proposal is supported by recent studies on this
issue, which have found a high correlation between the word association and the vocabulary size tests (Nurweni and Read, in press; Schmitt and Meara 1997).
According to Henriksen (1999), these three dimensions should not be seen as separate, but as a continuum by means of which learners go through in the vocabulary learning process. The first and second dimensions, Partial - Precise Knowledge and Depth of Knowledge, respectively, can be considered as a knowledge continuum, related to acquiring word meaning, and developing and understanding sense relations. The third dimension, Receptive – Productive, in turn, is a control continuum that refers to use ability. Therefore, the Depth of Knowledge of a word (dimension 2) is important for an accurate understanding of a word (dimension 1), being also a relevant element for a word to become productive (dimension 3). In this way, an increase in the range of accessibility of a lexical item (dimension 3) is due to the quality of the semantization process (dimensions 1 and 2).
3.2. Objectives
3.2.1. General objective
To describe the vocabulary knowledge of learners of English as a foreign language at four levels of communicative competence: Pre-intermediate, Intermediate, Post-intermediate, and Advanced.
a. To assess the receptive breadth of lexical competence of learners of English as a foreign language at four levels of communicative competence: Pre-intermediate, Intermediate, Post-intermediate, and Advanced.
b. To assess the productive breadth of lexical competence of learners of English as a foreign language at four levels of communicative competence: Pre-intermediate, Intermediate, Post-intermediate, and Advanced.
c. To assess the depth of lexical competence of learners of English as a foreign language at four levels of communicative competence: Pre-intermediate, Intermediate, Post-intermediate, and Advanced.
d. To analyse the results of the assessment of depth, and the receptive and productive breadth of lexical competence of subjects at all levels of competence, in quantitative terms.
e. To analyse the results of the assessment of the productive breadth of lexical competence of the subjects, in qualitative terms.
f. To compare the results of the assessment of lexical competence of learners at the different levels of competence.
g. To identify the relationship between depth and breadth of the lexical competence of learners, and their general proficiency in the target language.
h. To identify the relationship between successful and less successful learners in terms of their general proficiency at all levels of communicative competence, and their lexical knowledge.
i. To compare the results of the assessment of lexical competence of successful and less successful learners in terms of their general proficiency at all levels of competence, and their lexical knowledge.
3.3. Hypotheses
1. The higher the level of communicative competence, i.e., Pre-intermediate, Intermediate, Post-intermediate and Advanced, the higher the receptive breadth of vocabulary knowledge.
2. The higher the level of communicative competence, i.e., Pre-intermediate, Intermediate, Post-intermediate, and Advanced, the higher the productive breadth of vocabulary knowledge.
3. The higher the level of communicative competence, i.e., Pre-intermediate, Intermediate, Post-intermediate, and Advanced, the higher the depth of vocabulary knowledge.
4. The more successful a learner is at any level of communicative competence, the higher the breadth and depth of vocabulary knowledge.
3.4. Research questions
Concerning the productive breadth of vocabulary knowledge, the following research questions were asked:
a. What kinds of word classes were unanswered at each level of competence, i.e., Pre-intermediate, Intermediate, Post-intermediate, and Advanced?
b. What kinds of mistakes were committed at each level of competence, i.e., Pre-intermediate, Intermediate, Post-Pre-intermediate, and Advanced?
3.5. Methodology
3.5.1. Subjects
The subjects who provided the data for the present study were students of the Academic Undergraduate Programme of “Licenciatura en Lengua y Literatura Inglesas” of the Department of Linguistics, Facultad de Filosofía y Humanidades, Universidad de Chile. This course of studies consists of four years in which students have a wide variety of courses on general and English linguistics, English literature and English language. Some of these courses are Structural Linguistics, Generative Linguistics, Sociolinguistics, Psycholinguistics, English Text Grammar, English Phonology, English Phonetics, Applied Linguistics, Semantics, Pragmatics, Discourse Analysis, and History of the English Language. Apart from these, the programme includes four English Language courses (one
course per year), in which students learn the language, practise it and interact with each other, thus developing their linguistic and communicative skills in the target language.
The subjects who took part in this research study were chosen neither at random nor on purpose. Since these tests were taken by them during their lessons at the university, the students who actually sat for the tests were those who were present at the time the tests were given.
The number of subjects who participated in the study was 80; their age ranged from 17 to 25. From these 80 subjects, 19 belonged to the Pre-intermediate level, i.e., first year, 6 males and 13 females; 22, to the Intermediate level, i.e., second year, 7 males and 15 females; 19, to the Post-intermediate level, i.e., third year, 3 males and 16 females; and 20, to the Advanced level, i.e., fourth year, 4 males and 16 females. The research group expected to have 20 subjects per level; however, this number eventually varied according to the availability of the students to sit for the tests.
The subjects belonging to the four levels of communicative competence, i.e., Pre-intermediate, Intermediate, Post-Pre-intermediate, and Advanced, were divided into two different groups, according to their final marks obtained in their corresponding English Language courses in the present year. One group consisted of "successful" learners and the other, of "less successful students" The successful subjects were those who, in a scale ranging from 1 to 7, attained from 5.1 to 7.0 as their final marks in the course. The less successful learners were those who obtained marks below 5.1 as their final marks. This distinction was made with the purpose of comparing these different subjects´ final marks to
their performance on the vocabulary tests, by running a statistical correlation test between these variables (See Appendix A, Tables 1 to 8).
3.5.2. Elicitation of the data
The data for the research study were elicited by applying three tests which intend to measure the dimensions of the subjects’ vocabulary knowledge: Receptive and Productive Breadth, and Depth.
3.5.2.1. Breadth tests
The Vocabulary Levels Test, designed by Nation (1983; 1990), was used to measure breadth vocabulary knowledge. Its both formats, Receptive and Productive, consist of five sections, which represent five different vocabulary size levels in the English language. These are: level one, 2000 words, corresponding to basic everyday oral language (Schonell et al. 1956); level two, 3000 words, related to the words needed to begin reading authentic texts. In turn, level three, 5000 words, corresponding to what is required for reading authentic texts; level four, 10000 words, which has to do with the necessary vocabulary to deal with the challenges of university study in an L2; and finally, University Word Level, which is associated to academic vocabulary in pedagogical contexts (Schmitt et al. 2001). In the Receptive and Productive formats, only three word categories are used: nouns, verbs, and adjectives, which have a different frequency in each format.
3.5.2.1.1. Receptive Breadth test
The test used to assess the receptive breadth of the learners’ vocabulary knowledge was the one developed by Paul Nation (1983; 1990) (see Appendix B, Test 1). In this test, the test designers took the lexemes for each level of frequency randomly from the Thorndike and Lorge word frequency list (1944). Concerning the words in the University Word Level test, the words were selected from the University Word List (Xue and Nation 1984).
The Receptive Breadth format of the Levels Test involves word-definition matching items, i.e., the test-takers have to match words to short definitions. At each level, there are 36 words and 18 definitions. An example of this test format is quoted below:
1. apply
2. elect ____ choose by voting 3. jump ____ become like water 4. manufacture ____ make
5. melt
6. threaten
As can be observed in this example, all the words in each item belong to the same word category, with the purpose of, on the one hand, not providing any grammatical evidence as a clue for the right answer; and, on the other, grouping together words and definitions with no meaning relations (Read 2000). Regarding word class, there is a
consistent number of items belonging to each of them, distributed into a 3 (noun): 2 (verb): 1 (adjective) ratio (Schmitt et al. 2001).
3.5.2.1.2. Productive Breadth test
The Productive Breadth Test used in the research study was the one designed by Laufer and Nation (1999) as a further development of the Receptive Levels Test (see Appendix B, Test 2). The same target words of the Receptive Breadth Test are used; however, each level consists of 18 sentences that have blanks to be filled in with the missing segment of the target word. In order to elicit the appropriate target word, its initial letters are provided, ranging from one to five letters. The following example illustrates the format of this test:
The picture looks nice; the colours bl_______ really well.
In relation to word category distribution, there is no consistent number of words belonging to each category per frequency level in the different sections of the test. Nevertheless, there is a tendency for nouns to be more frequent than verbs and, in turn, verbs are more frequent than adjectives.
As regards the validation of the Vocabulary Levels Test, Read (2000) states that a study for such a purpose has not been properly carried out since there is no other construct for comparing its effectiveness. However, there have been several attempts to validate the test. For instance, Read, in 1998, attempted to validate the test and found it reliable on the basis of an implicational scale, that is to say, it is likely for a subject to know more higher-frequency words than lower-higher-frequency ones. Besides, this test has been considered as an appropriate measure of vocabulary size by different L2 researches (e.g. Laufer 1992, 1996; Laufer and Paribakht 1998; Qian 1998, 1999; Yu, 1996).
3.5.2.2. Depth
Finally, the test used to assess the depth, or quality of vocabulary knowledge, is the Word Associates Test (See Appendix B, Test 3). This test was originally designed by Meara (1983), and was afterwards modified by Read (1993). It is based on the concept of word association, and the task requires the learner to select -among distractors- associates or words related in meaning to the target word. The associates are related to the target word in terms of three different relationships:
• Paradigmatic: the words are synonyms or at least similar in meaning.
• Syntagmatic: the two words often occur together in a phrase.
The Word Associates Test, intended to assess the depth of vocabulary knowledge, consists of 40 adjectives chosen as target words. A set of eight words, associates and distractors, are presented in a box divided into two groups. In the group on the left side, the associates are words that have a paradigmatic relation with the target word, that is, synonyms or words that represent one aspect of its meaning. In the group on the right side, the associates are words that have a syntagmatic relation with the target word, that is, its possible collocations. An example is quoted below:
Sudden
X X X X
beautiful quick surprising thirsty change doctor noise school
Thus, the subjects have to select the words that they believe are related to the target word, in terms of synonyms and collocations. There is no consistent number of correct answers on the left or on the right side. They have to choose four options, but not necessarily two on each side.
According to Read (2000), the following characteristics of this format would provide a complete profile of the learners’ vocabulary knowledge:
a. Its guiding principle is the selection of high-frequency words.
d. Words are presented in isolation, which allows subjects to cover a higher number of words to be assessed.
e. It is a recognition task rather than a recalling task.
The Word Associates Test was improved through a repeated piloting process as a means of validation. With this measure, the reliability obtained reached .93 (N=94; Read 1995). Additionally, Messick (1989) suggested that the best way to validate this test was by analysing the data not only in quantitative terms but also qualitatively.
3.5.3. Procedure
The data were collected in several sessions, according to the subjects' levels and the availability of the teachers in giving permission to administer the tests in their classes. The tests were given to the students in sessions of approximately 40 minutes, depending on the time provided for this purpose. In this way, the Productive Breadth Tests and the Receptive Breadth Tests were taken by learners during the same sessions, or, in some cases, the Receptive Breadth Test was taken a week after the Productive one, depending on the availability of the subjects. The Depth Test was given approximately three weeks after the other tests. In several cases, extra sessions had to be held in order to test the subjects that were not present in one of the previous sessions.
Every session started with the teacher in charge of the course explaining that these tests had research purposes for seminar students, without informing the subjects of the
objectives of this study. Afterwards, the members of the research group gave the subjects the corresponding instructions for the tests.
3.5.4. Criteria for the analysis of the data
For marking the Productive Breadth Test, the criteria taken into account were the following:
1. Every correct response counted as 1 point. A correct response meant that the subject filled in the blank in the sentence with the expected answer. For example, in the sentence “Lack of rain led to shortage of water in the city”, ‘lack’ is the target word. Additionally, mistakes regarding grammatical tenses (past or present) or no –s ending in 3rd person singular in present tense were not relevant to the objectives of the study; therefore, they were not considered as mistakes. For example, in the sentence “The dog cringed when it saw the snake”, the word ‘cringes’ was accepted as correct, thus counted as 1 point.
2. A response was worth 0.5 in the following cases:
a. Whenever the subject added or omitted a letter that did not affect the pronunciation of the target word,
e.g. mortgage* instead of ‘mortgage’ as in “People manage to buy houses by raising a mortgage from a bank.”
b. When the subject pluralized the target word in its adjectival form (negative transfer from Spanish into English),
e.g. obsoletes* instead of ‘obsolete’ in the sentence “Computers have made typewriters old-fashioned and obsolete."
c. When there was no grammatical agreement between the target word and the sentence,
e.g. In the sentence “Since he is unskilled, he earns low wages”, the target word is ‘wages’ and not wage*.
3. The following responses got 0 points:
a. When another word with a different meaning was the result of a misspelling, that is to say, another English word was provided as the answer,
e.g. ‘relay’ instead of ‘rely’ in the sentence “He’s irresponsible. You cannot rely on him for help.”
b. When the misspelling altered the grammatical function of the target word, since to know a word also means to know its function,
e.g. ‘chilly’ instead of ‘chill’ in the sentence “It was a cold day. There was a chill in the air.”
c. When the misspelling affected the pronunciation of the target word,
e.g. fragant* instead of ‘fragrant’ as in “Many gardens are full of fragrant flowers.”
d. When the subjects provided a word which is not a lexical item of the English language,
e.g. vicary* instead of ‘vicar’ as in “The new vicar was appointed by the bishop.”
e.g. inspectioned* instead of ‘inspected’ as in “They inspected all products before sending them out to stores.”
e. When the misspelling altered the target word, which is part of an idiomatic expression,
e.g. aprons* and not ‘apron’ as in “Some people find it hard to be independent. They prefer to be tied to their mother’s apron strings.”
Whenever the misspelled target word contained two or more of the mistakes stated in the second criterion, it got 0 point. On the other hand, if a sentence allowed two or more options as target words, any of the possible answers was counted as correct if they made sense. For example, in the sentence “The pro_________ of failing the test scared him”, the blank could be filled in with either 'probability' or 'prospect'; and in the sentence “Despite
his physical condition, his int__________ was unaffected”, 'intelligence' or 'intellect' were considered correct responses.
Thus, the idea of scoring the answers using these criteria was to give an account of the vocabulary acquisition process by which subjects go through when learning a new language. Nevertheless, because these criteria proved to be time-consuming, a second marking was carried out in order to prove if it was worth applying them. Within the new criteria, only those answers filled in with the target word were considered correct, and wrong those which contained any kind of mistake, as in the Levels Tests (Nation, 1990.)
By comparing the results of the tests using both criteria, the score of every subject in the different levels of the test varied in 1% approximately. Therefore, the second criterion was chosen as it was concluded that there were no considerable differences between them.
Consequently, the wrong answers were classified into unanswered items and mistakes. The former deals with omitted items, and they were classified according their grammatical category, i.e., nouns, adjectives, and verbs. The latter is related to the different kinds of mistakes that were classified as follows:
1. Grammatical mistakes: They were subclassified as follows:
1.1. Grammatical disagreements of number between the subject and the verb:
b. ‘diapers’ instead of ‘diaper’ as in “The baby is wet. Her diaper needs changing.”
1.2. Pluralization of adjectives:
c. obsoletes* instead of ‘obsolete’ as in “Computers have made typewriters old-fashioned and obsolete.”
1.3. Incoherent use of past tense, participle –en and progressive –ing forms according to the context:
d. ‘stretch’ instead of ‘stretched’ as in “This sweater is too tight. It needs to be stretched.”
e. ‘hug’ instead of ‘hugging’ as in “The child was holding a doll in her arms and hugging it.”
2. Spelling mistakes: the addition or elision of one or two letters that led to the use of a non-existent word in the target language.
e.g. a. squimed* instead of ‘squirmed’ as in “The wounded man squirmed on the floor in agony.”
b. clim* instead of ‘climb’ as in “They have to climb a steep mountain to reach the cabin.”
c. mes* instead of ‘mess’ as in “The workmen cleaned up the mess before they left.”
d. lovly*instead of ‘lovely’ as in “The dress you are wearing is lovely.”
e. relly* instead of ‘rely’ as in “He is irresponsible. You cannot rely on him for help.”
f. mortage* instead of ‘mortgage’ as in “People manage to buy houses by raising a mortgage from a bank.”
3. Lexical mistakes: they were subdivided into three categories:
3.1. Morphological Creativity: use of L2 morphological rules to create incomprehensible L2 lexes (Poulisse 1990).
e.g. a) vicary* instead of ‘vicar’ as in “The new vicar was appointed by the bishop.”
b) fragant* instead of ‘fragrant’ as in “Many gardens are full of fragrant flowers.”
c) satured*instead of ‘saturated’ as in “The victim’s shirt was saturated with blood.”
d) anomality* instead of ‘anomaly’ as in “The anomaly of his position is that he is the chairman of the committee, but isn’t allowed to vote.”
3.2. Another existent word in the TL: replacement of the target word by another existent word in the TL.
e.g. a) ‘impossible’ instead of ‘impeded’ as in “The rescue attempt could not proceed quickly. It was impeded by bad weather.”
b) ‘messy’instead of ‘mess’ as in “The workmen cleaned up the mess before they left.”
c) ‘drawings’ instead of ‘drafts’ as in “Before writing the final version, the student wrote several drafts.”
d) ‘evening’instead of ‘Eve’ as in “We decided to celebrate New Year’s Eve together.”
e) ‘cried’instead of ‘cringed’ as in “The dog cringed when it saw the snake.”
3.3 Borrowing: the target word is replaced by a word from the L1.
e.g. a) fragil* instead of ‘fragrant’ as in “Many gardens are full of fragrant flowers”.
b) vicario* instead of ‘vicar’ as in “The new vicar was appointed by the bishop.”
c) diarrea* instead of ‘diaper’ as in “The baby is wet. Her diaper needs changing.”
d) imposible* instead of ‘impeded’ as in “The rescue attempt could not proceed quickly. It was impeded by bad weather.”
4. Discussion of results
In this section, we present and discuss the results of the vocabulary tests and analyse them in quantitative and qualitative terms. Additionally, a statistical analysis is included in order to validate the study.
4.1 Test results
The results of the Receptive and Productive Breadth Tests, and the Depth Tests per communicative competence level, and the subjects’ performance at each word frequency level in the tests are presented in Tables 1, 2, and 3.
A) Receptive Breadth Test results
Table 1 shows the performance of subjects in the Receptive Breadth Test per level of communicative competence, in which an increase of correct answers in the higher levels of communicative competence at the different levels of word frequency can be clearly appreciated (See Figure 1). Additionally, this difference is minimal between the last three levels of competence, being the Pre-intermediate level the most significant one.
Figure 1. Receptive Breadth Test average results per communicative competence level. Level
Final
Mark 2000 3000 5000 10000 UWL AVERAGE Pre-intermediate 4,7 83,51% 65,96% 62,11% 31,05% 81,23% 64,77% Intermediate 4,5 93,94% 84,09% 75,91% 41,82% 89,55% 77,06% Post-intermediate 4,9 96,49% 86,49% 79,74% 44,04% 94,04% 80,16% Advanced 5,3 96,96% 89,31% 82,97% 51,30% 97,47% 83,60% Average 4,8 92,73% 81,46% 75,18% 42,05% 90,57% 76,40%
Receptive Breadth Test 0,00% 10,00% 20,00% 30,00% 40,00% 50,00% 60,00% 70,00% 80,00% 90,00%
Communicative Competence Level
A
ver
ag
e
Pre-intermediate Level Intermediate Level Post-intermediate Level Advanced Level
B) Productive Breadth Test results
Table 2. Productive Breadth Test results per communicative competence level.
Concerning Productive Breadth Test results, they show an increase in the amount of correct answers of students of higher levels of communicative competence. However, the Level Final Mark 2000 3000 5000 10000 UWL Average Pre-intermediate 4,7 43,27% 22,51% 7,60% 3,80% 11,40% 17,72% Intermediate 4,5 66,92% 37,63% 19,95% 11,11% 26,26% 32,37% Post-intermediate 4,9 76,56% 48,85% 29,82% 9,94% 37,13% 40,46% Advanced 5,3 71,11% 52,78% 31,94% 14,44% 38,44% 41,74% Average 4,8 64,47% 40,44% 22,33% 9,82% 28,31% 33,07%
final scores were lower than those of the Receptive Breadth Test. Moreover, this difference is minimal between the three higher levels of competence, being the Pre-intermediate level the most significant one. (See Figure 2)
Figure 2. Average results of Productive Breadth Test per communicative competence level. Productive Breadth Test
0,00% 5,00% 10,00% 15,00% 20,00% 25,00% 30,00% 35,00% 40,00% 45,00%
Communicative Competence Level
Av
er
ag
e
Pre-intermediate Level Intermediate Level Post-intermediate Level Advanced Level
Concerning the Breadth Test results, there is a tendency in the production of correct answers according to the level of word frequency. In this way, 2000 and University word level obtain the highest scores in Receptive Breadth test in all levels of communicative competence, followed by 3000, then 5000, and finally 10000 word levels (See Figure 3). Regarding Productive Breadth tests, the 2000 word level presents the highest score of correct responses in all levels of communicative competence, followed by 3000, then University, 5000, and finally 10000 word levels (See Figure 4). This tendency, i.e., lower
levels of word frequency present higher scores at all levels of communicative competence, can be explained by the fact that words specific in register are more problematic than general ones, since general words are more common than specific lexical items, therefore more frequent (Segler 2001).
Figure 3. Receptive Breadth Test results per frequency level at each level of communicative competence.
Breadth Receptive Test
0,00% 20,00% 40,00% 60,00% 80,00% 100,00% 120,00%
Pre-internediate Level Intermediate Level Post intermediate Level Advanced Level
A
ver
ag
e
2000 3000 5000 10000 Academic
Receptive Breadth Test
0,00% 20,00% 40,00% 60,00% 80,00% 100,00% 120,00%
Pre-intermediate Level Intermediate Level Post intermediate Level Advanced Level
Av
er
ag