Lexical Coverage Research Articles

Evaluating lists of high-frequency words: Teachers’ and learners’ perspectives

With a number of word lists available for teachers to choose from, teachers and students need to know which list provides the best return for learning? Four well-established lists were compared and it was found that BNC/COCA2000 (British National Corpus / Corpus of Contemporary American English 2000) and the New General Service List (New-GSL) provided the greatest lexical coverage in spoken and written corpora. The present study further compared these two lists using teacher perceptions of word usefulness and learner vocabulary knowledge as the criteria. First, 78 experienced teachers of English as a second language / English as a foreign language (ESL/EFL) rated the usefulness of 973 non-overlapping items between the two lists for their learners. Second, 135 Vietnamese EFL learners completed 15 yes/no tests which measured their knowledge of the same 973 words. Teachers perceived that the BNC/COCA2000 had more useful words. Items in this list were also better known by the learners. This suggests that the BNC/COCA2000 is the more useful high-frequency wordlist for second language (L2) learners.

Extensive reading and viewing as input for academic vocabulary: A large-scale vocabulary profile coverage study of students’ reading and writing across multiple secondary school subjects

The extent to which extensive reading (ER) and extensive viewing (EV) support academic literacy in secondary school, particularly for L2 students, through relevant vocabulary input has yet to be established. This study undertakes a lexical profile coverage study providing information on how much coverage is afforded by ER/EV vocabulary constructs for reading and writing in biology, physics, chemistry, mathematics and English (as a subject area). It operationalizes ER/EV as vocabulary input constructs through corpora representing general fiction, science fiction, juvenile fiction, television and movies. It computes coverage provided by this input at different frequency bands (e.g. ER/EV's 1st 1000 most frequent word families, 2nd 1000 etc.) in secondary school textbooks and student writing. It finds ER/EV input appears very valuable for English as a subject area, and provides within the first 11–12,000 most frequent word families substantial coverage of receptive and productive vocabulary in secondary school science. Science-fiction provides more coverage, and juvenile fiction less. EV does not appear to offer impoverished coverage to ER. The study suggests that additional vocabulary pedagogy is nevertheless going to be needed, to cover from about 1 in 10 to 1 in 20 vocabulary items in the target subject areas.

How Much Knowledge of Derived Words Is Needed for Reading?

AbstractThe study explores the usefulness of the word family as the unit of counting in studies of lexical coverage and comprehension. It determines the proportion of texts covered by the various members of a word family, that is, basewords, inflected words, and derived words, and analyzes the contribution of the affixed words to lexical thresholds. This exploration was performed by a text analysis computer program called Morpholex that analyzes the entire lexis of an entered text, pulling out all words bearing prefixes and suffixes and counting the unaffixed words as basewords. We analyzed a variety of texts, academic and narrative, authentic and simplified, and calculated the number and percentage of basewords and affixes in each text. We also located the most frequent affixes in our text corpus and demonstrated which affixes and how many contributed to 95 per cent and 98 per cent text coverages. Our results show that reaching the lexical thresholds for reading does not require the knowledge of most of the derived words in a word family since a small number of frequent affixes will provide the necessary coverage together with the basewords and inflections.

The relationship between verbal form and event structure in sign languages

Whether predicates describe events as inherently bounded (telic) or unbounded (atelic) is usually understood to be an emergent property that depends on several factors; few, if any, spoken languages have dedicated morphology to mark the distinction. It is thus surprising that sign languages have been proposed to have dedicated morphology for telicity, and moreover that it takes a form which iconically reflects the underlying event structure – this is known as the “Event Visibility Hypothesis” (EVH) (Wilbur 2008). The EVH has been extended with claims about its universality in sign languages (Wilbur 2008; Malaia & Wilbur 2012), its gradient nature (Kuhn 2017), and its iconic transparency (Strickland et al. 2015). However, in this paper we argue that the status of this relationship between form and meaning remains an open question due to (a) lack of independent tests for telicity, (b) lack of lexical coverage, (c) lack of demonstration that formal expressions of telicity are morphological in nature, rather than a lexical property, and (d) inability to sufficiently dissociate telicity and perfectivity. We present new data coming from verbs that alternate in both form and meaning in ASL that is in line with the EVH, and conclude that while there is evidence supporting a morphological marker, the proposed form and telicity are not isomorphic in their distribution, significantly limiting the “visibility” of the event structure. We further propose that much of the related iconicity is the result of several independent factors also found in spoken languages, so that sign languages may be more similar to spoken languages than typically implied in this domain.

Étendue du vocabulaire et compréhension écrite – le français à l'université en Suède

This paper investigates the amount of French vocabulary needed to read (with reasonable comprehension) three authentic novels used in a first semester course of French at a Swedish university. The analysis applies the concept of lexical coverage and uses the frequency lists of Lonsdale & Le Bras (2009). The counting unit is the lemma. This study also examines the common French word list used in all Swedish universities offering French degrees. The results show that 7 000 to 9 000 lemmas will provide the 98 % coverage needed for reasonable comprehension. This is at least twice the 4 000 lemmas contained in the university list. The results therefore point to the necessity of including a vocabulary component also in higher-level French courses.

The relationship between vocabulary and viewing comprehension

The present study explores the relationship between lexical coverage and viewing comprehension. The study also addresses the role of imagery in viewing comprehension by distinguishing between audio-based, audio plus imagery-based, and imagery-based comprehension questions. Learners' lexical coverage was determined by measuring participants’ knowledge of the words occurring in the documentary in an aural-meaning recall test, whereas viewing comprehension was measured by means of literal and inferential comprehension questions. One hundred and fourteen EFL learners took part in the study. Results showed an almost medium-sized correlation between lexical coverage and viewing comprehension (rs (94) = .39). Results also showed a positive almost medium-sized correlation between audio-based questions and lexical coverage (rs (94) = .36), and a small-sized correlation between imagery-based questions and lexical coverage (rs (94) = .29). No relationship was observed between lexical coverage and imagery plus audio-based questions. No threshold for minimum comprehension was observed. The results indicate that the lexical demands for viewing are lower than those for reading but similar to those for listening. Implications for pedagogy and future research are discussed.

중등교육과정 중국어 기본어휘의 텍스트 점유율에 관한 연구

The basic Chinese vocabulary of secondary curriculum in Korea is 880 items and is smaller than general vocabulary list. This is the result of a number of researchers'' efforts to achieve two goals which are ‘beginner-level learning vocabulary coverage’ and ‘curriculum content adequacy’, and this is rather contradictory. However, it is somewhat questionable whether 880 items will be enough for beginner level learning vocabulary. The purpose of this study was to examine whether the basic vocabulary of the secondary curriculum is appropriate to the elementary level learning vocabulary. Ten graded readers were selected and the lexical coverage in the basic vocabulary of secondary curriculum from each books were examined. For comparison, 1,342 items with high frequency among the first grade vocabulary from the ≪等級劃分≫ were used. Although the high difficulty of the text, it did not reach 95%, which is a reasonable level of understanding, but it was found that it was quite high despite of the small vocabulary amount. However, there are many words in the secondary curriculum that are considered to be included even if they are not listed. The difference in the coverage of these words depending on whether they are included is considerably large at around 7%.

Using a Listening Vocabulary Levels Test to Explore the Effect of Vocabulary Knowledge on GEPT Listening Comprehension Performance

ABSTRACTThe Ministry of Education in Taiwan encourages universities to adopt English ability as an exit requirement for graduates. Nearly 65% of universities use the General English Proficiency Test (GEPT) or other compatible standardized tests as an exit test for university graduates. However, it remains unclear what level of vocabulary knowledge a second language (L2) learner needs in order to achieve a passing score for listening on the exit test. The L2 vocabulary assessment literature also clearly shows a need for a test that measures aural English vocabulary knowledge. To address the gaps, this study explored the effect of vocabulary knowledge on performance on the GEPT listening subtest by using a validated Listening Vocabulary Levels Test to provide a more accurate estimate of the receptive vocabulary of L2 learners. The findings showed (1) a strong positive correlation between vocabulary knowledge and overall listening skills, with comprehension of short talks (Part III) most closely related to vocabulary knowledge and most challenging to participants; (2) no absolute vocabulary threshold above which L2 learners passed the GEPT listening subtest, although a substantial increase in passing rate was observed at the 3,000-word vocabulary level; and (3) a minimum vocabulary level of 3,000-word families is required for 98% lexical coverage in the overall GEPT listening subtest, and a higher level of vocabulary knowledge is needed for L2 learners to comprehend conversations and short talks, compared with comprehension and answering of questions. The study findings provide insight into the role of aural receptive vocabulary knowledge in listening comprehension, and the implications for language teachers, materials writers, and test developers are also discussed.

Restoring Arabic vowels through omission-tolerant dictionary lookup

Vowels in Arabic are optional orthographic symbols written as diacritics above or below letters. In Arabic texts, typically more than 97 percent of written words do not explicitly show any of the vowels they contain; that is to say, depending on the author, genre and field, less than 3 percent of words include any explicit vowel. Although numerous studies have been published on the issue of restoring the omitted vowels in speech technologies, little attention has been given to this problem in papers dedicated to written Arabic technologies.f In this research, we present Arabic-Unitex, an Arabic Language Resource, with emphasis on vowel representation and encoding. Specifically, we present two dozens of rules formalizing a detailed description of vowel omission in written text. They are typographical rules integrated into large-coverage resources for morphological annotation. For restoring vowels, our resources are capable of identifying words in which the vowels are not shown, as well as words in which the vowels are partially or fully included. By taking into account these rules, our resources are able to compute and restore for each word form a list of compatible fully vowelized candidates through omission-tolerant dictionary lookup. Our program performs the analysis of 5000 words/second for running text (20 pages/second). Based on these comprehensive linguistic resources, we created a spell checker that detects any invalid/misplaced vowel in a fully or partially vowelized form. Finally, our resources provide a lexical coverage of more than 99 percent of the words used in popular newspapers, and restore vowels in words (out of context) simply and efficiently.

Lexical coverage and profiling

Lexical coverage and profiling

E2SAM: Evolutionary ensemble of sentiment analysis methods for domain adaptation

Currently, a plethora of industrial and academic sentiment analysis methods for classifying the opinion polarity of a text are available and ready to use. However, each of those methods have their strengths and weaknesses, due mainly to the approach followed in their design (supervised/unsupervised) or the domain of text used in their development. The weaknesses are usually related to the capacity of generalisation of machine learning algorithms, and the lexical coverage of linguistic resources. Those issues are two of the main causes of one of the challenges of Sentiment Analysis, namely the domain adaptation problem. We argue that the right ensemble of a set of heterogeneous Sentiment Analysis Methods will lessen the domain adaptation problem. Thus, we propose a new methodology for optimising the contribution of a set of off-the-shelf Sentiment Analysis Methods in an ensemble classifier depending on the domain of the input text. The results clearly show that our claim holds.

Construct validity of six sentiment analysis methods in the text of encounter notes of patients with critical illness.

Sentiment analysis may offer insights into patient outcomes through the subjective expressions made by clinicians in the text of encounter notes. We analyzed the predictive, concurrent, convergent, and content validity of six sentiment methods in a sample of 793,725 multidisciplinary clinical notes among 41,283 hospitalizations associated with an intensive care unit stay. None of these approaches improved early prediction of in-hospital mortality using logistic regression models, but did improve both discrimination and calibration when using random forests. Additionally, positive sentiment measured by the CoreNLP (OR 0.04, 95% CI 0.002-0.55), Pattern (OR 0.09, 95% CI 0.04-0.17), sentimentr (OR 0.37, 95% CI 0.25-0.63), and Opinion (OR 0.25, 95% CI 0.07-0.89) methods were inversely associated with death on the concurrent day after adjustment for demographic characteristics and illness severity. Median daily lexical coverage ranged from 5.4% to 20.1%. While sentiment between all methods was positively correlated, their agreement was weak. Sentiment analysis holds promise for clinical applications but will require a novel domain-specific method applicable to clinical text.

The Relationship Between Lexis and Reading Comprehension: A Review

It has been commonly reported by teachers and learners the strong and positive relationship between lexis and reading comprehension. This claim has been usually grounded on experience. In consequence, this paper attempts to find research-based evidence to support the connection between lexis and reading comprehension. To do this, a review of 128 studies was conducted using a set of criteria, after which thirteen studies were selected and screened to look for specific patterns. Salient themes emerged from the analysis: lexical coverage, direct instruction, lexical knowledge (breadth and depth), incidental and intentional learning and receptive and productive lexical knowledge. The analysis supports the connection between lexis and reading comprehension not only in EFL, but also in ESL and L1 contexts. In line with it, the research suggests that teachers should focus on systematic lexis instruction since it aids learners to increase both lexical knowledge and reading comprehension gains.

Improving lexical coverage of text simplification systems for Spanish

The current bottleneck of all data-driven lexical simplification (LS) systems is scarcity and small size of parallel corpora (original sentences and their manually simplified versions) used for training. This is especially pronounced for languages other than English. We address this problem, taking Spanish as an example of such a language, by building new simplification-specific datasets of synonyms and paraphrases using freely available resources. We test their usefulness in the LS task by adding them, in various combinations, to the existing text simplification (TS) training dataset in a phrase-based statistical machine translation (PBSMT) approach. Our best systems significantly outperform the state-of-the-art LS systems for Spanish, by the number of transformations performed and the grammaticality, simplicity and meaning preservation of the output sentences. The results of a detailed manual analysis show that some of the newly built TS resources, although they have a good lexical coverage and lead to a high number of transformations, often change the original meaning and do not generate simpler output when used in this PBSMT setup. The good combinations of these additional resources with the TS training dataset and a good choice of language model, in contrast, improve the lexical coverage and produce sentences which are grammatical, simpler than the original, and preserve the original meaning well.

The most frequent BNC/COCA mid- and low-frequency word families in English-medium traditional Chinese medicine (TCM) textbooks

This research set out to examine the vocabulary used in English-medium traditional Chinese medicine (TCM) textbooks and aimed to establish a TCM English word list (EWL) as a reference for English for Chinese Medicine purposes. The researcher built a 13-million-token TCM Textbook Corpus and measured the vocabulary levels of the books from within along the word-frequency scale of the British National Corpus (BNC) and the Corpus of Contemporary American English (COCA). Results show that knowledge of the first BNC/COCA 10,000 word families plus proper nouns would provide 98% lexical coverage of a TCM text. The first 3,000 word families were not considered for inclusion in the TCM EWL, but the next 605 most frequent word families were selected to form the word list. The TCM EWL accounted for 11.42% of running words of the corpus and encompassed various technical vocabularies—crypto-technical, lay-technical and sub-technical words. For Chinese medicine students, it can provide a window to the TCM English register, and for English for Chinese Medicine teachers, it can serve as a guide in curriculum preparation, particularly in setting vocabulary goals at the initial phase of medical English instruction.

Vocabulary knowledge and listening comprehension at an intermediate level in English and French as foreign languages

Abstract The present study aims to replicate Stæhr’s (2009) study on the relationship between vocabulary knowledge and listening comprehension. To explore the generalizability of Stæhr’s findings, two experiments were conducted. However, the study changed the population (Flemish learners) and proficiency level of the materials in the first experiment (intermediate) and also the language in the second experiment (French). Our results generally confirm Stæhr’s findings. We also found a positive correlation between vocabulary knowledge and listening comprehension. Although our findings reveal that learners probably need fewer words for adequate listening comprehension at an intermediate proficiency level, our study also showed that learners who knew more words had higher listening comprehension scores. Finally, our study seems to suggest that learners might need less lexical coverage for intermediate listening.

Potentials of English Children’s Literature Books for Extensive Reading with a Focus on Vocabulary Size and Word Repetition

The primary purpose of the current study is to explore the potentials of children’s literature books for extensive reading by measuring the vocabulary size necessary for comprehension and estimating the number of words that could be learned incidentally through reading the books. For the analysis, the texts of 218 books in the ten popular English children’s book series in Korea were compiled and analyzed utilizing the Range, one of the comprehensive vocabulary analysis tools. The results revealed that about 3,000 word families were necessary for 95% lexical text coverage of the books and 6,000 word families for 98% lexical text coverage in general, although the vocabulary size necessary for the text coverage largely vary according to the book series. It was also found that readers may be able to learn about 776 word families incidentally through reading the books extensively, assuming that they know about 6,000 word families which can lead to 98% lexical coverage of the texts.

The Influence of Unknown Words Density on Listening Comprehension in Chinese English Learners

Although the relationship between lexical coverage and L2 reading comprehension has been well documented, it is far from the case for L2 listening. This study has investigated the relationship between unknown words density (UWD) and listening comprehension among Chinese English learners. It shows that listening comprehension is negatively correlated with UWD in the listening texts. Furthermore, UWD threshold is approximately 3 percent for the adequate listening comprehension. The result will help college English teachers to choose more targeted listening materials and encourage students to improve listening efficiency.

A framework for enriching lexical semantic resources with distributional semantics

AbstractWe present an approach to combining distributional semantic representations induced from text corpora with manually constructed lexical semantic networks. While both kinds of semantic resources are available with high lexical coverage, our aligned resource combines the domain specificity and availability of contextual information from distributional models with the conciseness and high quality of manually crafted lexical networks. We start with a distributional representation of induced senses of vocabulary terms, which are accompanied with rich context information given by related lexical items. We then automatically disambiguate such representations to obtain a full-fledged proto-conceptualization, i.e. a typed graph of induced word senses. In a final step, this proto-conceptualization is aligned to a lexical ontology, resulting in a hybrid aligned resource. Moreover, unmapped induced senses are associated with a semantic type in order to connect them to the core resource. Manual evaluations against ground-truth judgments for different stages of our method as well as an extrinsic evaluation on a knowledge-based Word Sense Disambiguation benchmark all indicate the high quality of the new hybrid resource. Additionally, we show the benefits of enriching top-down lexical knowledge resources with bottom-up distributional information from text for addressing high-end knowledge acquisition tasks such as cleaning hypernym graphs and learning taxonomies from scratch.

Applying the Lexical Coverage Hypothesis to Establish the Suitability of EFL Reading Materials: A Case Study of the TOEFL (ITP)

Applying the Lexical Coverage Hypothesis to Establish the Suitability of EFL Reading Materials: A Case Study of the TOEFL (ITP)