Methods of lexicological research in research. Content Introduction

2. A case study on vocabulary learning through reading
picture books

One of the aims of the second study was to gain more information on the textual

and psychological conditions in quite different circumstances. It consisted of a long-term rather loosely organised case study with three Dutch children learning English (and some French and German) by reading and listening. The greater part of the data was collected in a period of two months in which two of the children (a 13-year old girl and a 10-year old boy) read some thirty English books: picture books, early readers written for young (native) children who are learning to read, non-fiction books and graded readers especially written for foreign language learners. A follow-up study over two years involved a third child starting with English, whereas the two older children started with French and some German. The following procedure concerning the selection of the books and the new vocabulary was used. The children could each time choose a book which they would like to read from some 5 to 10 books (from a large and varied stock).
They could inspect the books at leisure and ask questions about them; they were also encouraged to put the books in order of interest and to comment on this grading. In most cases the books were read aloud to the children. In some cases no notice was taken of unfamiliar words (if possible); in other cases attention was paid to these words in a variety of ways, such as translating the words, helping the children in guessing their meaning from the context and the word form or helping the children to memorize the words (by quickly going through them once or twice after reading - when necessary showing the context - or by using vocabulary cards).[11]
The results of this study will now be discussed in so far as they throw new light on the embedding of the words in meaningful memory systems and on the textual and psychological conditions for vocabulary learning through reading. As the results of the follow-up study did not add new results with regard to the theme of this paper, I will focus here on the results of the first part of the study. Although this study centred on the textual and psychological conditions some information on the embedding of the words in memory was gained as well. In a number of cases the children appeared to remember words through recalling the situation described in the text. Often recollections of an illustration mediated recall. Again recollections of their own emotions or experiences which the words, the text or the actions of the children had evoked, influenced recognition of words. In this study, in particular, the emotions connected with funny, surprising or exciting illustrations and texts exerted an important influence on recognizing words. To
illustrate this with an example: the word "handsome" was first seen in the ironic "Jake beckoned to the rest of his ugly crew. "Right, me handsomes (...)" ⁴which was said to a funnily drawn couple of nasty scoundrels. Although "handsome" had been seen only once, it proved to be remembered when it turned up in a totally different context two weeks later (limitations of space prevent me from including illustrations and more examples; they may be found in Dutch publications.
As to the "textual conditions" the results of the first experiment were confirmed,
but also extended, because in this study some insight - at least for this age group – was gained into the significance of text genres for vocabulary acquisition. In particular, reading picture books and to some degree reading early readers and non-fiction books appeared to have a number of advantages compared with reading foreign language readers. The children, for instance, liked to experience that they could read the same books which are read by children from countries where the target language is used, at a rather early stage. They even proved to be able to read some German picture books without any help, before their first lesson in German.[13]
More generally speaking, I found the following favourable features which some of
the picture books do possess and which some of the foreign language readers do not or only to a smaller degree. Early readers and non-fiction books mostly occupy an intermediate position.
1. Authenticity. The language is more lively, more natural, because it is not hamperedby limits of vocabulary and structure.
2. Redundancy. Some readers for foreign language learners suffer from lack of
redundancy, especially concerning the structure of the text, when compared with
authentic texts. This means that an apparently easy, adapted text may be more
difficult to understand than its authentic counterpart.
One could argue against DT as a factor of difficulty by claiming that:
(a) learners misinterpret all kinds of words, not necessarily the DT ones; therefore the number of errors induced by DT words would not be significantly higher than the number of errors induced by non-DT words; (b) when the learner thinks he knows a word, but does not, this is not due to specific word characteristics (those of DT words); therefore if tested for lack of awareness of ignorance, there would be no significant difference between DT and non-DT words. The study was designed with the above arguments in mind. It addressed the following research questions:
(a) Is the frequency of errors induced by DT words different from the error frequency induced by non-DT words?
b) Is the learners' awareness of their ignorance of DT words different from the
awareness of their ignorance of non-DT words? Question (b) was subdivided into two questions:
(i) when faced with unknown DT words, will the learners recognise them as unfamiliar more often than not recognise them as such?
(ii) when learners are unaware of their ignorance of words, is it in the case of DT words more often than with non-DT words?
Procedure The subjects were 100 first year university students taking a course in English for Academic Purposes. They were all high school graduates 5 , native speakers of Hebrew and Arabic. The study was conducted in two stages. In the first stage, learners were given an unseen text on a subject of general nature with comprehension questions 6 , to ensure that they were reading for comprehension. The tasks were
(a) to answer the questions;
(b) to underline unknown words in the text. 'Unknown' meant a word that could not be understood from text context. The answer sheets with comprehension questions and the texts with the underlined words were collected.
Stage two followed immediately. The learners were given a clean copy of the same text and a list of 40 words from it (20 Dt, 20 non-DT) and were asked to translate them in text context. The translations were collected and compared with the underlined words in the texts. In other words, for each student, a comparison was made between the words students CLAIMED they did not know (out of the 40 selected ones) and the words they ACTUALLY did not know.
The comparison between translation lists and underlined words yielded three
possible results:
(a) The learners translated some words correctly and did not underline them in the text.
(b) The learners did not know the translation of some and underlined them, i.e. they were sometimes aware of their ignorance.
(c) The learners misinterpreted some words and did not underline them in the text, i.e. they were sometimes unaware of their ignorance.
Three scores were given to each student:
(a) Error score: the number of errors on the translations list: for DT words and non-DT words.[16]
(b) Awareness score: the total number of instances in which the learner was aware of his ignorance: in the case of DT words; and non-DT words.
(c) Reading comprehension score on the basis of his answers to comprehension
questions. The following grid illustrates how the information about the two first scores was coded.
Results Error scores in DT and non-DT words were compared by a matched t-test. The number of errors in DT words was significantly higher than in non-DT words (1=1.67 p<0.05). The number of +aware instances was compared with the number of -awareinstances for DT words. The latter was significantly higher than the former (t=8.46 p<0.0005). This means that DT words are not recognised as unfamiliar more often than they are recognised as such.[20]
The number of -aware instances was compared for DT and non-DT words. The
frequency of -aware for DT words was significantly higher than that for the non-DT words (t=10.31 p<0.0005). This means that unawareness of ignorance is more frequent with DT words than with other words.
The relationship between awareness of unknown DT words and reading
comprehension was measured by correlating the 'awareness' scores with reading
comprehension scores. Pearson product moment correlation was .65, significant at .0001 level.
Discussion The study compared DT and non-DT words with regard to:
(a) number of errors induces by each group of words.
(b) the extent to which learners were conscious of their ignorance of words in each
group (the DT and non-DT).
(c) the relationship between awareness of unknown DT words and success in reading comprehension.
The results showed that
(a) Errors were more frequent with DT words.
Students were less aware of their ignorance with DT words than with non-DT ones. There was a significant correlation between reading comprehension and learners' awareness of unknown DT words.
It can be argued, therefore, that deceptive transparency is indeed a factor which has an effect on comprehension.
DT Words And Reading Comprehension
As mentioned before, the correlation between awareness of DT words and reading
scores was .65, significant at .0001 level, though correlations do not show cause-effect relationships between the variables, they do indicate the degree of common variance. In our case, it seems that about .4 of variance in reading could be accounted for by the degree of awareness of DT words. However, I will try to argue for a possible cause-effect relationship between the two.
When a foreign learner does not understand a word in the text, he has the following options: ignore it (if he considers it unimportant), look it up in a dictionary, ask
someone who knows its meaning, or try to guess it from con-text. Many researchers of reading and pedagogues have emphasized the importance of guessing as a strategy of successful reading presupposes awareness, on the part of the learner, that he is facing an unknown word. If such awareness is not there, no attempt is made to infer the missing meaning. This is precisely the case with deceptively transparent words. The learner thinks he knows then and assigns the wrong meaning to them, distorting the immediate context on this way.
But this may not be the end of the distortion process. The misinterpreted words will sometimes serve as clues for guessing words which the learner recognises as unknown, which may lead to larger distortions. Graphically, the process can be represented in the following manner:
awareness of ignorance of DT words -> misinterpretation of DT words -> distortion of immediate context -> using distorted context for further inter-pretation -> distortion of larger context. Here is an example of a distorted sentence resulting from misinterpretation of three words. The original sen-tence was:
'This nurturing behaviour, this fending for females instead of leaving them to fend
for themselves may take many different forms.' 'nurturing' was confused with 'natural', 'fending' with 'finding', 'leaving' with 'living'.
The result was the following:
'Instead of living natural life, natural behaviour, females and children find many
different forms of life.' One might wonder about the lack of syntactic resemblance between the original and the misinterpreted sentences. Such incongruencies in sentence structure show that students are willing to rely on lexical clues more than on syntactic ones; they are even prepared to impose a sentence structure on the idea they have already arrived at via lexis. If the learner had recognised 'nurture, fend, leave' as unknown words in the given example, he might have looked up or guessed their meaning and arrived at a different interpretation.
The suggested cause-effect relationship between awareness of DT words and
reading comprehension can therefore be explained as follows. A better awareness of DT words is necessary for attempting to find their meaning. Such an attempt will result in a larger number of correctly interpreted words. These will in turn reduce the density of unknown words. Such reduction will result in an increase in contextual clues which are necessary for outstanding additional new words. This understanding will increase the total number of correctly interpreted words. A larger number of known words will be an asset to global comprehension of the text.
DT Words And Vocabulary Testing. One way of checking the vocabulary size of a learner is by giving him a list of words (real and non-existing) and asking him to reply 'yes' to every known word and 'no' to every unknown one. (E.g. Meara and Buxton, 1987). The calculation of the score takes into account the number of real words identified as familiar and the number of 'identifications' of the nonsense words. The more nonsense words are claimed to be known, the lower is the credibility of the subject. However, if the text includes many DT words among its items, the learner may be making 'yes' responses to unknown words not be-cause of low acceptance threshold, but due to genuine unawareness of his ignorance. How this might affect the test score is not clear. It could be simpler to exclude such words from the test than to find a suitable correction formula.
Errors In DT Words And The Mental Lexicon
The study did not attempt to investigate the characteristics of the mental lexicon.
However, some of me errors caused by DT words can provide some information about the organisation of L2 words in the memory. Investigations have indicated that while in the native speaker's mental lexicon there are strong semantic links between the words, the connections between words in additional languages are primarily phonological.
Synformic errors, particularly confusion of phonologically similar words (cute/acute, valuable/available), provide additional evidence for such organisation. In searching for the right word, the learner selects its neighbour in the lexicon which sounds similar, but is erroneous.

Another interesting issue that has been debated is whether words com-posed of

root and affixes are stored as single units or whether the stems and affixes are stored separately. Errors in the words with a deceptively morphological structure seem to support the latter (lexical decompositions hypothesis).
The learner might store the prefix 'dis-' separately and therefore interpret 'discourse' as 'without direction', combining what looks like two separate units if meaning. Also the confusions of morphological synforms (industrial/industrious, economic/economical) might result from storing the suffixes separately and substituting one by an-other.