Learning about Learning from Dictionaries

 

Jim Ronald

Hiroshima Shudo University

 

Introduction

There seem to be two almost opposing views regarding the use of monolingual learnersf dictionaries as instruments or means of language learning:

a) That they are an invaluable tool for language learners, whether in aiding comprehension or as a resource or guide when writing in the target language. In addition, it is believed that using these dictionaries promotes vocabulary acquisition.

b) That whatever to qualities of monolingual learner dictionaries, their use is beyond the capabilities of all but the most advanced of foreign language learners. In other words, for most learners using these dictionaries is a waste of time and will lead to unnecessary errors of comprehension or production.

I will begin by suggesting why such diverging views may be held, even by teachers with the same students, with one cause being the lack of evidence to support either view. This will be followed by a report of research, summarized below, undertaken to compare how effective are two types of language resource, dictionary definitions and encounters in context, in facilitating comprehension and retention of unknown words.

Dictionary Definitions vs. Example Sentences: Comprehension and Recall
 

Design

Following a pretest to ensure that target words were unknown, Japanese intermediate learners of English were given a set of 40 unknown words: 20 verbs and 20 adjectives. Together with the target words, one group (the Dictionary Definitions group) were given a set of monolingual dictionary definitions for the target words. The other group (the Example Sentences group) were given a set of typical corpus-drawn example sentences. The subjects were instructed to study the materials and then write Japanese equivalents for each of the target words. Two weeks later, the subjects were given a test of vocabulary retention, a kind of gap-fill exercise in which they had to match the sentences or definitions with the correct word.

 

Subjects.

A total of 82 Japanese university students participated in the first part of the experiment. For the retention test, however, only 78 of the students were available. The participants were 1st year (n. 29) and 2nd year (n. 53) students, aged between 18 and 20, majoring in English at a private middle-ranking university. They had all received between six and seven years of formal instruction in English. Despite this, most were intermediate level students and their TOEFL scores would range between about 440 and 520.

 

Target words

There were 40 target words, 20 verbs and 20 adjectives, all unknown to the participants. The words were selected according to the following criteria: that they were in the learner dictionaries being used, that there was only a single sense given for the words, that there were at least 30 occurrences of the word in the Cobuild Direct 50 million word corpus, and that the words were unknown to the participants. These were the target words:

 

Verbs: abbreviate, amputate, appal, blab, bode, cackle, chomp, coerce, dilate, elope, feign, jilt, perspire, pooh-pooh, raze, sulk, suss, trounce, waft, whinge

Adjectives: afoot, akin, averse, bereft, blatant, callous, colossal, defunct, dilapidated, eerie, fleeting, furtive, galore, gaudy, hoarse, illicit, morbid, obese, ostensible, poignant

Learning materials

The participants were divided into two groups:  the Dictionary Examples group and the Example Sentences group:

Dictionary examples group: Participants in this group received two dictionary entries for each of the target words, taken from the Longman Dictionary of Contemporary English 3rd Edition (1995) and the Collins COBUILD English Dictionary 2nd Edition (1995). The dictionary entries were stripped of any example sentences but included the definitions and grammatical information.

Example Sentences Group: This group received three example sentences for each of the target words, drawn from the 50 million word COBUILD Direct corpus. Sentences were chosen for displaying typical syntactic patterns and collocations, and for their comprehensibility. Wherever possible, sentences were taken directly from the corpus without changing them in any way.

Vocabulary Retention Test

Three weeks after the main vocabulary learning session in which the above materials were used, a test of vocabulary retention was conducted. For this test, subjects were given an answer sheet, together with the same materials as three weeks before, except that the target words were separated into verbs and adjectives and put into four ten-word sets. The target words were deleted wherever they occurred in the definitions or example sentences and the test items were randomly reordered within each set

Rating

Two pairs of highly proficient Japanese users of English rated the subjectsf translation equivalents for the target words. They were provided with dictionary entries from two monolingual English and one English-Japanese dictionary for each target word. Each translation equivalent for the target words was judged as Correct, Partially Correct, or Wrong. Interrater agreement overall was 79.3 %. Agreement for translation equivalents for verbs was 76.6%, for adjectives it was 82%. Differences were resolved at a joint meeting of the two raters.

 
Results

There were two sets of results from the experiment: for the subjectsf Japanese translation equivalents of the target words, and for the retention test. These results were also broken down according to the part of speech of the target words: correct or partially correct translation equivalents for verbs and for adjectives. In this presentation, only the results for the two groups will be discussed.

 

The Correct and Partially Correct equivalents were collapsed into one category as eacceptable answersf. There was a large difference between the two groups for their production of acceptable translation equivalents. T-tests were conducted on the ratings of the Japanese translation equivalents for the two groups, confirming that this difference were significant (t = 8.8, p <.001). The results are shown in Table 1.

 

Table 1. Analysis of results for Japanese translation equivalents for target words             

Dictionary definitions   Example sentences
Acceptable answers (av.) 19.22 8.98
s.d.   6.44 3.67

Retention Test

The results for the Retention Test are shown below. In this case there was no rating of answers; answers were either right or wrong. The results, shown in Table 2, indicate that there was an average difference between the two groupsf scores of almost 4% on the retention test. While a T-test confirmed that there was a significant difference between the two groups (t = 1.63, p < .1), further research with larger groups is clearly required.

 

Table 2. Analysis of Retention Test scores according to learning materials used             

Dictionary definitions  Example sentences
Correct answers (av.)   4.81 (12.3%) 6.10
s.d.   2.40 (16.2%)       3.42 

                                                          

Discussion

Here, I will focus solely on the results of the experiment and post-test in terms of the two sets of learning materials.

Japanese translation equivalents

As with previous experiments, results for the Japanese translation equivalents simply show how much more successful the Dictionary Definitions group was than the Example Sentences group in identifying, more or less, what the target words mean. There were high levels of standard deviations for both groups, and scores within each learning material group vary widely: for the Dictionary Definitions group, between 29 (72.5%) and 4 (10%) acceptable answers, and for the Example Sentences group, between 17 (42.5%) and 1 (2.5%) acceptable answers.

 

Such differences can probably not be simply attributed to large differences in language ability. Rather, these differences seem to point to an ability that these tests require of the subjects: to make reasonable guesses based on limited information. While a larger vocabulary would make more of each definition or example sentence comprehensible, and so make the task of guessing easier, in many cases the participants seemed to adopt one of two stances: either gItfs like a jigsaw puzzle: some pieces are missing but with what I know I can try to guess what the picture ish or gWithout all the clues, itfs impossible to guess the correct answer. Itfs a pointless exerciseh.

Nation suggests that L2 proficiency is a major factor in successful guessing (2001: 247). That, undoubtedly, is true in one respect; the higher the level of proficiency, the more clues are available to the guesser. This is reflected in the answers of the highest scoring participants: those expected to get the highest scores generally did. In the case of the lowest scoring subjects, however, there had been no indication until this point that they were especially weak in terms of L2 proficiency.

Retention

For each item, participants had to choose from among 10 possible answers. The retention test does not in fact test retention of the meaning of the words encountered through the two sets of materials but retention of the materials themselves: ability to recognize the learning materials for each target word. There are two reasons for this:

i)          The two groupsf scores for comprehension in terms of translation equivalents were very different,

ii)         A sensitive measure of vocabulary retention was wanted: levels of recall of word meaning would probably be too low to be of value.

This test, which calls for the recognition of the context in which target words were previously encountered, was felt to be both more sensitive and more relevant to two groups whose comprehension of the target words differs so widely.

The result of the retention test was not as might be expected. The Example Sentences group, which performed much worse than the Dictionary Definitions group in the test requiring them to give translation equivalents for the target words, performed slightly better than them in this retention test. One possible reason for these results is that the retention test scores reflect the depth of processing required for the two groups to complete the translation equivalents test (Craik and Tulving, 1975) regardless of their success in identifying an acceptable translation equivalent.

Conclusion

Three aspects of the results of this experiment stand out. One is the general success of the Dictionary Definitions participants to identify the meaning of unknown words as compared to the Example Sentences group. Another point is that there is wide variation between participants: many were as unable to guess from the context of a definition as others were to guess from example sentences. Finally, recognition of the context in which a word was previously encountered does not appear to depend on comprehension of the word. This serves as a reminder that much learning takes place through purposeful contact with the L2, learning that may remain undetected and unvalued.

References

Craik, F.I.M and Tulving, E. 1975 Depth of processing and the retention of words in episodic memory. Journal of experimental psychology 104:268-284.

Nation, I.S.P. 2001 Learning vocabulary in another language. Cambridge: Cambridge University Press.