Memorizing New Words: Does Teaching Have Anything To Do With It?
Memorizing New Words: Does Teaching Have Anything To Do With It?
Memorizing New Words: Does Teaching Have Anything To Do With It?
net/publication/249768592
CITATIONS READS
164 3,234
2 authors, including:
Batia Laufer
University of Haifa
124 PUBLICATIONS 11,146 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Batia Laufer on 23 May 2014.
Published by:
http://www.sagepublications.com
Additional services and information for RELC Journal can be found at:
Subscriptions: http://rel.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://rel.sagepub.com/content/28/1/89.refs.html
What is This?
Background
It is only common sense that without words there can be no proper
communication, however basic. Expressing messages of a four-year-old
native-speaking child would require about 700 words (Kirkpatrick, quoted
in Nation, 1990), and as communication becomes more complex and
sophisticated the knowledge of vocabulary needs to be refined and
expanded to several thousands (see for example ’Cambridge English
Lexicon’ by Hindmarsh, 1980).
89
90
When words are explicitly taught, rather than learnt incidentally, the
context in which they are presented can be manipulated by the teacher.
Many teachers nowadays believe words should be presented in some kind
of context and not in isolation. The common practice is to make students
encounter them in a text. The arguments for such a method are the
following: first, it is only natural for people to encounter new words in
context. Second, the context illustrates some common usages of the word.
Third, as already mentioned, context contributes to elaborate processing
of a word and provides a cognitive foothold, which in turn reinforces
memorization. If contextual associations contribute to learning, it would
follow that the richer the context, the better the learning. Thus, presentation
of words in isolation would be less effective than presentation in context.
As for type of context, the more elaborated and richer in associations, the
better. Following this argument, we could claim that text context would
be more effective than sentence context and that ’elaborated’ text (Yano
et al. 1994) would be more effective than normal text. In elaborated texts
the original ideas are repeated, or modified from time to time. For that
purpose, often paraphrase is used which has an explanatory function. From
the point of view of learning new words, such additions provide better
clues to understanding and additional associations which could serve as a
cognitive foothold for memorizing these words.
memorizing words. When new words appear in a text and the text is read
for meaning, individual words, including the new words to be learnt, may
go unnoticed, particularly if they are not essential for understanding the
main points. And one of the conditions for transforming input into intake
is noticing and attending to the new material (Schmidt 1990 and 1992).
According to Schmidt, what must be attended to and noticed is not any
input, but the specific features one is acquiring, that is to acquire
phonology, one must notice phonology, to acquire lexis, one must attend
to lexis. There is empirical evidence for the position that &dquo;intake is the
subset of input that is attended to and noticed&dquo; (Schmidt 1992:210), that
focusing on particular structure is indeed effective (Doughty 1991). Yet
when new words appear in a text which is read for global understanding,
the learner will often skip these words rather than focus on them. If focus
91
be learnt better than words in context since in the latter case, maximum
attention would be directed to the word. Following this argument, the
prediction would be that words in lists are remembered better than words
presented in sentences, and these are remembered better than words
presented in text. The worst condition for word memorization, involving
the most distraction, would be the elaborated text.
On the other hand, the proponents of translating the new words into
the learners’ mother tongue argue that first, many words can be translated
into another language quite precisely (Dagut 1977). Since human
experience is to some extent universal, many concepts are lexicalised
similarly in different languages. If this were not the case, the task of foreign
language learning would be much more difficult. In cases where concepts
are lexicalised differently (e.g. English ’know’ and French ’savoir’ and
92
93
The Subjects
Five groups of high school pupils studying in the highest level English
classes were chosen for the experiment. The pupils’ common mother
tongue is Hebrew. The classes were of a comparable level because the
pupils were all in the same grades and in the same school system. They
were normal average pupils with a grade range in each class from
Twenty low frequency words were chosen for the experiment. The
words were checked for previous knowledge by testing a small sample of
pupils in each class. The Ministry of Education’s syllabus as well as West’s
Word List confirmed the fact that all the words were considered ’low
frequency’.
94
For the ’list’ group a list of 20 words was prepared. Ten words were
translated into Hebrew (the pupils’ L 1 ) and ten were paired with English
definitions or synonyms. As in a natural classroom situation, the division
was done inconspicuously, making the LI or L2 choice of
The material for the ’sentence’ group consisted of the 20 target words
(each word either translated or defined as for the list group) with the
addition of a sentence in which each of the words was used.
For example: r-eign - govern
The people were unhappy under the reign of their cruel king.
The ’text’ group read a passage which was chosen from a British
course book not in use in public Israeli high schools. All 20 target words
appeared in the text and were glossed along the left-hand side of the
page in Hebrew or English. No clear contextual clues were supplied by
the text, thereby making inferencing extremely difficult.
The material for the ’elaborated’ text group consisted of the original
text which had undergone lexical elaboration, thereby facilitating the pupils
to understand the text and/ or the meaning of the target words while reading.
Additionally, such elaboration would make the pupils refer to the target
words. Lexical elaboration was achieved by adding either synonyms or a
brief explanation to the original sentence. Consequently, the elaborated text
was 8% longer than the original version (601 words as compared to 555
words in the original text). In addition to producing a longer text, elaboration
produced a text more advanced in readability (9.7 compared with 9.1 based
on Flesch-Kincaide grade level). As in the original text, the same 20 words
were glossed and defined at the side of the text.
95
The material for this group consisted of the list of 20 target words
sans definitions and translations.
Procedure
The ’list’ group was asked to study the word pair list for 10 minutes.
The ’sentence’ group was asked to focus on the words and read the
sentences. They also were given 10 minutes to study the worksheet. (See
appendix 1 )
The ’text’ group was asked to read a passage and answer the
The ’elaborated’ text group read the elaborated text (appendix I ) and
answered the same comprehension questions as the text group. This group
worked on the task for 55 minutes, the same amount of time as the original
text group.
At the end of the practice period the four study groups were given 15
minutes to complete a cloze exercise focusing on the 20 target words.
While all 20 words were supplied in a word bank at the bottom of the
page, only 16 were needed to fill in the gaps. The cloze passage was not
taken from the text itself. After the initial study period and consolidation
task all worksheets were collected from the four study groups.
96
The control group was asked to check the meaning of the 20 target
words and learn for a quiz that would be administered the following day.
No instructions were given as to which language the quiz would be
conducted in nor how the words should be semanticized. The pupils never
underwent the class learning period nor were they given the consolidation
task (the vocabulary cloze). The pupils’ word lists were collected before
the quiz. It became evident that all the words were looked up in a bilingual
dictionary.
97
Method effect
Our first research question addresses the issue of the quantity of context
used in presentation, i.e., the method effect. Table I (see graph 1 ) presents
the short and long-term retention scores of all 5 methods of presentation
and the statistical differences between them (means, standard deviation,
and results of ANOVA). As we can see from the results, there is a
significant difference between methods. The post-hoc Duncan test shows
exactly where the differences are. The methods marked with the different
letters are significantly different. The least effective method proved to be
the control group (non-teaching group). The most effective method for
long-term retention was sentence and list presentation.
98
99
100
Our third research question addresses the issue of the interaction effect,
i.e., method by language. Tables 3a and 3b (see graphs 3a and 3b) present
the short and long-term retention scores by methods of the languages used
in presentation (English glosses and Hebrew glosses). We can see that the
best results were obtained when glossing in the pupils’ L 1 (Hebrew).
(Language effect and method by language effect were checked for the 4
experimental groups only. As mentioned before, all learners in the control
group looked up translations for the target words.)
101
p=0.0001.
3) There was no interaction between method and language of glossing,
short-term - F(3,94) = 1.54, p=0.2, long-term - F(3,87) = 0.14, p=0.94.
This lack of interaction shows that no one method favors a specific
language.
102
Retention rate .
Let us look first at long term vocabulary retention scores. The learners
in the ’list’ and ’sentence’ methods retained 75% of words (15 out of 20):
in the elaborated text condition - 61%, in the text condition - 60%. In
Mondria 1993 (a study that also compared retention of words presented
in isolation, in sentences and in a text), retention rates after 4 weeks were
I
as follows: 52% (5.2 out of 10) in the ’list’ condition, 62% in ’sentence’
condition and 76% in ’text’ condition. The results. of the two studies, as
far as retention figures are concerned, show good retention. In Hulstijn et
al. (forthcoming), however, learners were found to retain 18% of the words
that were glossed in the text they were reading and 25% of the words they
had looked up. These results were obtained immediately after completing
the reading task. The conclusion of the authors was that learners pick up
very little vocabulary from texts. Such differences between the last study
and the results from text conditions in Mondria and our own study is not
surprising. Vocabulary learning in Hulstijn et al. was purely incidental.
The learners were given a reading task, not a vocabulary task. In Mondria’s
study, the learners were specifically asked to learn the target words from
the text, i.e., they were engaged in an explicit vocabulary learning task. In
our study, though the learners were reading the text for comprehension,
Language of glossing
Glosses in L 1 proved more beneficial for retention than glosses in
English. This was the case with the mean retention scores when the
language of glossing conditions were compared. As there was no ’language
by method’ interaction, we could not claim that one of the methods lent
itself better to glossing in English. The reason for the superiority of L 1
glosses may be explained by the fact that in this condition (glossing in
L I ), maximum attention is directed to the new L2 word since the L 1
equivalent is fully familiar to the learner and consists of only one word.
103
Looking at the Post-Hoc ANOVA tables, we can see that the control
group which studied the words by itself fared worst. Yet it is still unclear,
from this experiment, whether this difference is due solely to the effect of
presentation method, or whether the consolidation exercise played a part
in it as well. A follow up study will examine self-learning with a
consolidation task and compare it with the four experimental methods.
As for these four methods checked in the present study, we can see that an
identical consolidation task did not abolish the differences among them,
even though it may have reduced them. The ’list’ and ’minimal context’
methods were superior to the ’text’ and ’elaborated text’ methods. In other
words, the focus oriented methods were more effective for memorization
than context oriented methods. Moreover, they took up less teaching time,
25 minutes as opposed to 70 minutes. This importance of focus may also
have been responsible for the better scores in the ’elaborated text’ group
than in the ’text’ group. The nature of the elaborated text was such that
the additional phrases referred back and forth to the target words (they
included synonyms, paraphrases, etc.) and might have made learners look
at the target words once more.
104
Our results also raise the question how real is the common belief that
learning words in lists is done by rote only. There is no reason why lists of
words with their L translations should preclude deep processing on the
part of the learner who can embed the word in some kind of mental context.
In addition to simple rehearsal of the target word with its translation
(shallow processing), the learner may be learning the words by self-
generated imagery and semantic mediation (linking them to keywords,
making up phrases with them, etc.). It may be precisely this kind of learning
that explains the good results of learning words in lists which are surveyed
in Nation ( 1982).
Concluding remarks
The conclusion of this study is that focus oriented methods of
presenting new vocabulary are more effective than context oriented reading
methods as the former yield better long term retention scores while at the
105
Note
A preliminary version of this paper was presented at the SEAMEO
RELC 31 st Seminar in Singapore.
References
Atkins, B.T. and F.F. Knowles. 1990. Interim report on the EURALEX/
AILA Research project into dictionary use. In BudaLEX 88
Proceedings, eds. I. Magay and J. Zigany. pp.381-392.
Atkins, R.C. 1975. Mneumonics in second language learning. American
Psychologist 30:821-828.
Coady, J., J. Magott, P. Hubbard, J. Graney and K. Mokhtari. 1993. High
frequency vocabulary and reading proficiency in ESL readers.
In Second Language Reading and Vocabulary Acquisition, eds.
T. Hucking, M. Haynes and J. Coady. ABLEX. pp.217-228.
106
107
-
1990. Teaching and Learning Vocabulary. Newbury House.
Saragi, T., I.S.P. Nation and G.F. Meister. 1978. Vocabulary learning and
reading. System 6: 72-78.
Schmidt, R. 1990. The role of consciousness in second language learning.
Applied Linguistics 11:129-158.
Yano, Y., M. Long and S. Ross. 1994. The effects of simplified and
elaboratedtexts on foreign language reading comprehension.
Language Learning 44:189-219.
108