0% found this document useful (0 votes)

205 views41 pages

Vocabulary Learning from Context

The document discusses learning vocabulary through context in three main ways: 1) Incidental learning from context is the most important way that native speakers acquire their first language vocabulary, but second language learners often lack the conditions needed for this type of learning. 2) Both intentional vocabulary study and incidental learning from context are complementary and important, with the latter making up about three quarters of effective language learning programs. 3) Several studies showed that inferring word meanings from context resulted in some vocabulary retention, but allowing learners to verify their inferences and memorize words led to higher retention rates than inferring alone. Directly providing word meanings led to even higher retention than other methods.

Uploaded by

nfish1046

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

205 views41 pages

Vocabulary Learning from Context

Uploaded by

nfish1046

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 41

8 Learning words from context

Incidental learning from context is the most important of all the

sources of vocabulary learning. This is particularly true for native
speakers learning their ﬁrst language. It should also be true for second
language learners, but many do not experience the conditions that are
needed for this kind of learning to occur. A major goal of this chapter
is to look at these conditions and see how they can be established. We
will look at how successful learners can be at guessing from context,
how much and what kind of learning can occur from this guessing,
and the kinds of clues available for guessing. We will then look at how
learners can be helped to become skilful at guessing from context.

Which is best: intentional or incidental learning?

Learning vocabulary from context is often seen as something opposed
to the direct intentional learning and teaching of vocabulary (Kelly,
1990). This is an unfortunate viewpoint and the position taken in this
book is that they are complementary activities, each one enhancing the
learning that comes from the other. A well-balanced language-learning
programme has an appropriate range of opportunities to learn from
message-focused activities and from direct study of language items,
with direct study of language items occupying no more than a quarter
of the total learning programme.
In this chapter, learning from context is taken to mean the incidental
learning of vocabulary from reading or listening to normal language
use while the main focus of learners’ attention is on the message of the
text. The texts may be short or long. Learning from context thus
includes learning from extensive reading, learning from taking part in
conversations, and learning from listening to stories, ﬁlms, television
or the radio.
In this chapter, learning from context does not include deliberately
learning words and their deﬁnitions or translations even if these words
are presented in isolated sentence contexts (see, for example, Gipe and
Arnold, 1979). This kind of learning is looked at in Chapter 11.

348
Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
Learning words from context 349

Context sentences and phrases are valuable aids in intentional,

language-focused vocabulary learning, and part of the confusion
behind the argument about learning from context versus learning
from lists stems from seeing the difference as relying on the presence or
absence of context, rather than the distinction made in this chapter
between message-focused, incidental learning and language-focused
intentional learning. As we shall see, however, this distinction between
incidental and intentional is not easy to maintain, particularly if we
accept that most learning involves conscious attention.
Hulstijn (2003) argues that the terms ‘intentional’ (learners are
aware that they will be tested on particular items) and ‘incidental’
(learners are not aware of a later test) are not particularly relevant to
studies of vocabulary learning. What is more important is the quality
of the mental processing that takes place during learning.
Although learning vocabulary from context should be largely inci-
dental learning, there should be a deliberate, intentional focus on
developing the skills and strategies needed to carry out such learning.
Because of the importance of guessing from context, it is worthwhile
for both teachers and learners to spend time working on guessing
strategies.
In a very carefully designed study, Mondria (2003) compared infer-
ring only with inferring and verifying on a two week delayed L2→L1
translation test, finding that learning occurred in both methods, but
inferring and verifying resulted in higher scores. When these were
compared with inferring plus verifying plus memorising, Mondria
found 6% retention for inferring, a further 9% as a result of verifying
(being told the right answer), and a further 32% as a result of a chance
to memorise. Thus, all steps resulted in some level of retention, with
deliberate learning predictably having the strongest effect. The oppor-
tunity to infer with subsequent memorisation did not result in a sig-
nificantly higher retention score than simply being given the meaning
and the chance to memorise. That is, inferring does not add greater
depth of processing that is reflected in a higher retention score. Students
spent 27% more time on the inferring method compared to the
meaning-given method with no extra gain in retention.
The meaning-given method (memorisation without inferring)
resulted in a score eight times higher than the inferring score. The
meaning-given method took 1.7 times longer, showing that it is not
only much more effective in terms of retention, but also more effi-
cient in terms of time. Mondria’s calculation of the achievement rate
was that the meaning-given method had a learning rate of .32 words
per minute, while inferring had a learning rate of .06 words per
minute.

In the meaning-inferred method, where guesses were veriﬁed or cor-

rected, correctly inferred words were retained better than incorrectly
inferred words.

How are reading and vocabulary growth related to

each other?
Perfetti and Hart’s (2002) Lexical Quality Hypothesis sees word
knowledge as central to skilled L1 reading. Good knowledge of words
involves strong knowledge of the spelling, pronunciation and meaning
aspects of knowing a word. This however is not necessarily a call for
teachers to teach more about words, but largely to ensure that learners
have large amounts of experience using language, particularly in
reading.
The simplest causal circle is that learners with good vocabulary
knowledge and skills achieve better comprehension of text. Better
comprehension of text allows learners to process more input (do more
reading). Increased practice in processing input develops good vocabu-
lary knowledge and skills (Perfetti and Hart, 2001):

Vocabulary
knowledge Comprehension

Amount of
input and
practice

In a later paper, Perfetti (2010) elaborated on what were good vocabu-

lary knowledge and skills, namely skill at decoding and skill at
accessing word meaning. If we add skill at inferring from context and
vocabulary size to these, we have a more elaborate causal circle.

Learners with good skills at inferring from context develop larger

vocabulary sizes. A large vocabulary size supports decoding skills and
skill at accessing word meaning. These four aspects of vocabulary
skills enable learners to achieve better comprehension of text. Better
comprehension of text allows learners to process more input. Increased
input and practice in processing input allows more inferring from
context.

Decoding skills Skill at accessing

word meaning

Vocabulary size Comprehension

Inferring from Amount of input

context

Like most models, this one also simpliﬁes, and it ignores or assumes
important factors like motivation, knowledge of the world and oral
language growth. It also arranges the parts in a linear unidirectional
fashion, and things are more complicated than that. Nonetheless, the
skills, knowledge and experience that it includes are essential for
vocabulary growth through reading, and there is research support for
the parts of the model.
Van Daalen-Kapteijns et al. (2001) examined the L1 inferring skills
of high and low verbal ability 11–12-year-olds. They found that the
differences between the high and low verbal ability children were ones
of degree of skill rather than the absence of a skill – all children are
capable of inferring from context, some just do it better than others.

The circle is also relevant to L2 learners. Pulido and Hambrick

(2008) provide evidence that L2 language use (both input and output)
influences vocabulary growth, which in turn influences comprehen-
sion which in turn influences vocabulary growth. This provides strong
support for making sure that learners experience quantity of input and
output at the right levels.

What proportion of unknown words can be guessed

from context?
To see what the chances are for successful guessing, we need to look at
guessing from context which occurs under both realistic and favoura-
ble conditions. Firstly, we need to look at guessing where learners
already know a large proportion of the words in the text. This is neces-
sary for learners to be able to use the clues for guessing the unknown
words. It is likely that at least 95% of the running words need to be
already familiar to the learners for this to happen (Liu and Nation,
1985). A coverage of 95% means that there is one unknown word in
every 20 running words, or one in every two lines. This is still a heavy
load of unknown vocabulary and densities like 1 in 50 (98% coverage)
are probably optimal. Studies which use higher densities of unknown
words, for example 1 in every 10 running words, have shown little
successful guessing, and set up conditions that make successful guess-
ing unlikely (Bensoussan and Laufer, 1984; Laufer and Sim, 1985a). A
critical factor in successful guessing is the learners’ vocabulary size,
because this will affect the density of unknown words in a text and
how well the words making up the context are known (Qian, 1999).
For second language learners, getting the optimal ratio of unknown to
known running words may involve using simpliﬁed or adapted texts.
Secondly, the estimates of guessing need to be based on the actual
words not known by each learner. This means that the choice of words
to be examined needs to take account of actual learner knowledge,
and not rely on teacher intuition or the unsystematic choice of words
from a text. If the choice of words was carried out properly, then more
readily generalisable statements about the percentage of text coverage
and chances of guessing, or vocabulary size and the chances of guess-
ing could be made. Schatz and Baldwin (1986) argue that most exper-
iments on success in guessing from context are ﬂawed because they use
a mixture of high-frequency and low-frequency words most of which
are already known to the learners. To truly test the availability of
context clues, experimenters would need to focus on unknown words
at the appropriate frequency level for the learners being tested. Schatz
and Baldwin worked with native speakers aged between 16 and 17.

They found no signiﬁcant difference between learners who had context

to help them guess and learners who were tested on words in isolation.
Schatz and Baldwin’s tests were multiple choice and asked learners to
provide a definition. The forms of the tests and the ways in which they
were marked did not give credit for partial knowledge (see Nation and
Webb, 2011: 83–7, for a critique of Schatz and Baldwin).
Thirdly, learner skill is a critical factor in guessing. Gibbons (1940),
Cook et al. (1963) and many other studies have found a wide variation
in the ability of learners to guess from context. From an optimistic
viewpoint, if some learners can guess large numbers of words success-
fully, then potentially most learners can if they develop the skill.
Studies of guessing should thus report performances of the best guess-
ers as well as averages.
Fourthly, learners must be given credit for guesses that are not
100% correct but which make a small but positive contribution to
knowledge of the meaning of the word. Learning by guessing from
context is a cumulative procedure by which learners gradually develop
their knowledge of words. It is likely, at least for some words, that the
initial meetings with a word in context simply give rise to a vague
knowledge of the form of the word and the awareness that it is unfa-
miliar and thus should get some attention next time it occurs. Beck
et al. (1983), in an article subtitled ‘All contexts are not created equal’,
argue that there is a range of helpfulness in natural text contexts for
unknown words. They range from misdirective contexts where learn-
ers are likely to infer the opposite meaning, through non-directive
contexts where no help is given, to general contexts where general
aspects of word meaning are inferable, and ending with directive con-
texts which could lead learners to a specific, correct meaning for a
word. Beck et al. are probably correct in saying not all contexts are
equally informative, but by seeing the goal of one meeting as a specific
correct meaning, they underestimate what can be learned from context.
For instance, here is their example of a misdirective context (the least
helpful in the scale) for the word grudgingly:

Sandra had won the dance contest and the audience’s cheers brought her to
the stage for an encore. ‘Every step she takes is so perfect and graceful,’
Ginny said grudgingly, as she watched Sandra dance.

There is useful partial information available from this context. First,

there is the form of the word. Second, it has clear affixes and a stem form.
Third, it functions as an adverb. Fourth, it can relate to the way people
say things. Knowing these bits of information is still a long way from
knowing the word, but they are initial, useful steps in the right direction.

Fifth, in discussions of learning from context, it is important to distin-

guish guessing from natural contexts on the one hand from deliberate
learning with specially constructed or chosen contexts on the other
hand.
Most studies of guessing from context do not take account of all of
these five factors and thus tend to give misleading results.
With these five cautions in mind, let us now look at the results of
studies of second language learners’ guessing from context. Seibert
(1945) found high rates of success (around 70%) in intensive guessing
with learners who knew French guessing Spanish words in context.
The similarities between these two closely related languages clearly
helped the guessing. Bensoussan and Laufer’s (1984) learners worked
on a difficult text – around 12% of the running words were unknown
to the learners, who were able to guess only a small percentage (13%)
of the unknown words. Bensoussan and Laufer estimated that clues
were not available for around 40% of the words that they considered
to be problem words for the learners in the text.
Parry’s (1991) longitudinal study of four adult learners guessing from
context supports earlier non-native speaker studies in showing reason-
able success in guessing from context with a range of 12% to 33% of
guesses classified as correct, and a range of 51% to 69% of guesses
either partly correct or correct. Most words found to be unknown were
not particularly subject matter related but were in the register of formal
expository prose. Horst et al. (1998) found gains of around 22%.
Knight (1994) found that second language learners learned words from
context while reading, on average 6% of the unknown words on an
immediate translation test and 27% on an immediate multiple-choice
test (corrected for guessing). Comparable scores were found on a
delayed test two weeks later. The twelve unknown words in each of the
tests occurred at a density of approximately one unknown word in 20
running words, meaning that the known words gave approximately
95% coverage of the text.
The findings from the few reasonably well-conducted studies of guess-
ing by non-native speakers have not had impressive results. This may be
partly due to poor design, but it is also the effect of the cumulative nature
of such learning involving only small gains per meeting for most words.
‘What proportion of unknown words can be guessed from context?’
is probably not the right question. Rather, it should be ‘Is it possible to
use context to keep adding small amounts of information about words
that are not yet fully known?’ The answer to this question is clearly
‘Yes’ (Fukkink et al., 2001). It is likely that almost every context can
do this for almost every word, but this has not yet been investigated
experimentally.

How much vocabulary is learned from context?

There are several important factors to keep clear when trying to
answer this question. First, it is important to distinguish working out
the meaning of a word from context and remembering the meaning of
a word worked out from context. Second, it is important to see learn-
ing as involving even small increases in knowledge of a word. Learning
from context is a cumulative process where meaning and knowledge
of form are gradually enriched and strengthened. Tests of learning
from context need to be sensitive to small amounts of learning (Nagy
et al., 1985). Third, it is important to see if the contexts and conditions
for learning are typical of normal reading. Nagy et al. (1987) note that
several studies use specially created contexts, combine contexts and
definitions, or replace known words with nonsense words. These
kinds of studies may provide useful information about the nature of
learning from context but they cannot be used to estimate how much
learning occurs from normal reading.
When learners who have apparently forgotten a lot of a previously
known L2 are asked to relearn words, they do better on those words
than on learning words that they have never met before (de Bot and
Stoessel, 2000; Hansen et al., 2002). This speedier relearning shows
that some subconscious memory for words can still exist even when
they cannot register on a recall test. Relearning is thus a sensitive test
of vocabulary knowledge and is capable of revealing knowledge that
other less sensitive tests cannot pick up. As may be expected, ease of
relearning is related to the length of time since the L2 was first learned,
and the present level of proficiency in the L2.
Studies with young native speakers of English using text which has
not been specially modified (Nagy et al., 1985; Nagy et al., 1987; Shu
et al., 1995) have found that there is a chance of an unfamiliar item
being learned to some degree from one exposure of between 1 in 10 or
1 in 20 respectively. The chance of learning in the experiments
depended partly on how soon learning was measured after the reading
occurred. Nagy et al. (1985) tested vocabulary learning 15 minutes
after the reading and got a 1 in 10 rate. Nagy et al. (1987) tested
vocabulary learning six days after the reading and got a 1 in 20 rate. A
meta-analysis of 20 studies involving native speakers (Swanborn and
de Glopper, 1999) confirmed these findings with students incidentally
learning an average of 15% of the unknown words they met while
reading. In all of these studies, the unknown words made up 3% or
less of the running words. Smaller proportions of unknown words
typically resulted in more learning. As we shall see later, quantity of
reading with the opportunity for previously met items to recur within

a certain time may be an important factor in learning from context.

Even with rich specially constructed contexts, up to ten repetitions
and some pre-teaching, learning is still low (Jenkins et al., 1984).
There are several things that can happen to an item met in context.
• It is guessed correctly to some degree and at least partially learned.
This may happen to 5–10% of the words.
• It is guessed correctly to some degree but nothing about it is learned.
This probably happens to many words.
• It is guessed incorrectly.
• It is not focused on, possibly because it is not important for the
wanted message in the text. This does not exclude the possibility of
small amounts of learning occurring.

Studies with second language learners have generally not been as care-
fully conducted as the studies with native speakers (Day et al., 1991;
Dupuy and Krashen, 1993; Pitts et al., 1989; Saragi et al., 1978).
Horst et al. (1998), however, in a study using a long text (a graded
reader) and two kinds of vocabulary tests, found that about one in five
of the unknown words were learned to some degree. In terms of actual
words, this averaged about five words.
The higher gains in the Horst et al. study come partly from the effect
of the length of the text, the use of a simplified reader where the
unknown words do not occur too densely, and the conceptual knowl-
edge that learners bring from their first language. Nagy et al. (1987)
found that a major factor affecting learning from context was whether
the word represented an unfamiliar concept. Shefelbine (1990) simi-
larly found a greater difficulty for new concepts. In his study, however,
the chances of learning vocabulary from natural contexts were higher
than other studies because there was a deliberate focus on guessing
vocabulary.
The incidental vocabulary learning from context in all these experi-
ments is small, not only in the likelihood of words being learned but
also in the actual number of items learned. This low rate has to be bal-
anced against other considerations:
1. Incidental vocabulary learning is only one of the various kinds of
learning that can occur when learners read. Not only can learners
begin to learn new words and enrich known ones, they can improve
grammatical knowledge, become more familiar with text structure,
improve reading skills, learn new information and learn that
reading can be an enjoyable activity.
2. Small gains become large gains if learners do large quantities of
reading. If learners read thousands or millions of running words

per year, then considerable vocabulary learning is possible. Nagy

(1997: 75) estimates that if a learner reads a million running words
of text a year, and if two per cent of these words were unknown,
this would amount to 20,000 unknown words per year. If one in
twenty of these were learned, the annual gain would be 1,000
words a year. One million running words is roughly equivalent to
three or four undergraduate textbooks (Sutarsyah et al., 1994), or
ten to twelve novels, or 25 complete Newsweek magazines
(Kennedy, 1987), or 65 graded readers of various levels (Nation
and Wang, 1999).
3. Learning rates can be increased considerably by some deliberate
attention to vocabulary (Elley, 1989; Hulstijn, 1992).

There are several implications to be drawn from the ﬁndings on the

rate of vocabulary learning from context. First, it is important that
learners do large quantities of interesting reading, large quantities for
second language learners meaning something like a graded reader of a
suitable level every week. Second, second language learners should not
rely solely on incidental vocabulary learning from context. There
needs to be judicious attention to decontextualised learning to supple-
ment and be supplemented by learning from context. Direct vocabu-
lary learning and incidental learning are complementary activities.
The low amount of learning from normal incidental guessing from
context could be a benefit rather than a cause for concern. A single
context generally provides an inadequate source of information about
a word. It is particularly difficult to distinguish between core aspects
of the meaning and those peculiar to the particular context. It may
thus be good that learners do not quickly decide on a meaning and
remember it well. Van Daalen-Kapteijns and Elshout-Mohr (1981)
found that high ability students remained flexible in the meanings they
attached to unfamiliar words so that they were ready to make later
revisions if they proved necessary. The small, gradual increments of
learning a word from context under normal conditions of incidental
learning encourage a flexible approach to finally determining the
meaning and make it unlikely that an initial, strong but wrong inter-
pretation will be made and maintained.
It has been argued (Haastrup, 1989: 319–20) that words are likely
to be remembered better if there was some difficulty in interpreting
them. This hypothesis is based on studies by Jacoby et al. (1979) and
Cairns et al. (1981) and which suggest that decision-making difficulty
results in a more distinctive memory trace. Cairns et al. suggest that
items met in highly predictable contexts may be easily processed but
have low saliency in memory. This means that if learners read texts

that they bring a lot of world knowledge to, they may be able to easily
cope with unknown words but retain little memory for these words;
that is, guessing will be easy but learning of vocabulary will be poor. If
on the other hand learners have to rely heavily on linguistic bottom-up
interpretation of the context and have to puzzle over the interpreta-
tion, guessing may be more laborious but learning of vocabulary may
be greater. Research by Pulido (2003; 2007) with second language
learners does not support this idea. Pulido (2009) found positive
effects on guessing for L2 reading proficiency and background
knowledge.
Fraser (1999) found more vocabulary was retained from inferring
from context when:
• the inferring was followed up by consulting a dictionary (this
almost doubled retention). Dictionary use makes an important con-
tribution to vocabulary growth, and learners can benefit from
training in dictionary use.
• first-language-based word identification was used, that is, learners
retrieved an L1 synonym for the unknown word. Finding an L2
synonym was also effective but not as effective for retention as an
L1 synonym, and creating a paraphrase for the meaning was the
least effective for retention. This supports previous studies showing
that a simple expression of word meaning is most effective for
learning.
• learners remembered that they had seen the word before but they
could not recall its meaning. This shows that vocabulary learning is
best regarded as a cumulative process with subsequent meetings
building on previous meetings, even though previous meetings only
resulted in very small amounts of learning.
Fraser found a very wide range of individual differences in retention.
Paribakht (2005) found that words which had a clear L1 equivalent
were easier to guess from context than L2 words which did not have a
corresponding L1 word. Similarly, L2 words without a corresponding
L1 word tended to be less well known. Examples of such words in
English without corresponding Farsi words are metropolitan, progno-
sis, intuitively, monogamy and clone. Such words have to be expressed
by phrases or definitions in Farsi.

What can be learned from context?

A critical factor in guessing from context is what is being learned. At
the simplest level, the unknown word may represent a familiar concept
and so the new label for that familiar concept is being learned. If the

concept is an unfamiliar one, then both the concept and the label need
to be learned. There is plenty of experimental evidence to show the dif-
ficulty in learning new concepts (Nagy et al., 1987). Although the word
form and its meaning are among the most important things to know
about a word, there are many other kinds of information that can be
learned from context that are important in the receptive and productive
use of the word. These are outlined in Chapter 2 and include the part
of speech of the word, its collocates, the things it can refer to, and the
various forms the word can take. These different kinds of information
are all closely related to each other and come together to enrich a
learner’s knowledge of a word. The range of collocates that a word has
helps specify its meaning. The grammatical patterns a word takes are
closely related to its collocates. The affixes a word can take may affect
its grammatical functions, its meaning and its range of collocates.
Fukkink et al. (2001) used measures for their experimental study of
young L1 learners which considered the number of correct attributes
of the word in the learners’ guess, the presence of a false attribute and
the degree of contextualisation of the guess. This rich analysis is an
attempt to capture the incremental and multi-componential nature of
the development of word knowledge.
Anderson and his colleagues (Anderson and Ortony, 1975; Halff
et al., 1976; Anderson et al., 1978; Anderson and Shifrin, 1980) make
the point that in language comprehension readers and listeners use
their knowledge of the world and the analysis of the linguistic context
to create particular instantiations of the words and phrases they com-
prehend. That is, they think of detailed particular instances guided by
the words they read or hear and their knowledge of the world. So,
when they see the sentence, The golfer kicked the ball, they think of a
particular kind of ball, most likely a golf ball. When they see the sen-
tence, The baby kicked the ball, their instantiation of ball will be dif-
ferent. The same applies to their instantiations of kicked. Word mean-
ings are context sensitive.
The point of Anderson and his colleagues’ investigations into instan-
tiation is that knowing a fixed core meaning for a word is not sufficient
for language use. People have a range of meaning representations for
each word which they draw on with the help of context when they
comprehend. Instantiation is usually necessary for full comprehension.
One very important value of context in learning vocabulary is that
a variety of contexts will evoke a variety of enriching instantiations.
Paired-associated learning is not likely to do this. Each paired-associate
repetition is likely to strengthen but not enrich.
There is L1 experimental evidence to show that providing a sen-
tence context, or several contexts, as well as a definition when learning

words helps learning. Gipe and Arnold (1979) found contexts and
definition to be superior to synonym or short definition, a classifica-
tion task, or using the dictionary. Nist and Olejnik (1995) found that
when learners saw the word in context and then looked at a definition,
the context helped their performance on a multiple-choice test which
required them to choose a correct example of use of the word.

An example of aberration would be:

a having a glass of cold milk with freshly baked cookies
b going to bed every night at exactly the same time
c a 16-year-old who didn’t want her own brand new car
d an infant who woke up every four hours to eat
Prince (1996) looked at weak and advanced learners learning from
context. The learning from context condition involved specially con-
structed sentences but did not provide an accompanying translation or
deﬁnition; learners had to use the context to discover the meaning of
the word. In the translation condition, learners saw an L1 word and
its L2 translation. Learning was tested in two ways: by translation of
isolated words and by ﬁlling a blank in a sentence. The sentences were
not the same as those that acted as context during learning from
context. Learning from translations resulted in higher scores than
learning from context for both weak and advanced learners, and
overall, learners found the translation test easier than the sentence
completion test. Those in the advanced group who learned through
context however did slightly better in the sentence completion test
than in the translation test. Prince interprets this as indicating that this
group were better able to transfer their knowledge to new applica-
tions. The weak group outperformed the advanced group where no
transfer of learning was required, that is, where learning by trans-
lation was tested by translation.
Webb (2007; 2008) however found that having a context sentence
during learning provided little if any advantage when many aspects of
word knowledge were tested for each deliberately learned word.
Learners were able to draw on analogy, world knowledge and com-
monsense when demonstrating their knowledge of the words.

What clues does a context provide and how effective

are they?
The major motivation for analysing and classifying the various kinds
of context clues is to provide a checklist for training learners in the
skill in guessing from context. If teachers have a well-established list to

work from, then they can be systematic and consistent in the way they
draw learners’ attention to clues and train them in recognising and
using the clues. Furthermore, if the relative frequency and effectiveness
of the various clues have been established, then it is possible to design
a well-graded programme of work covering the range of clues.
Haastrup (1985; 1987; 1989) used think-aloud introspection and
retrospection to study L2 learners’ inferencing procedures to see what
knowledge sources they used and how they combined knowledge from
various sources. Haastrup classified the knowledge sources using
Carton’s (1971) three categories (which are not mutually exclusive):
• interlingual: cues based on L1, loan words in L1 or knowledge of
other languages
• intralingual: cues based on knowledge of English
• contextual: cues based on the text or informants’ knowledge of the
world
The most careful and systematic attempt to come up with a system of
clues for native speakers was carried out by Ames (1966). Ames used
texts with every 50th word (provided the word was a content word)
replaced by a nonsense word. Native speaking doctoral students intro-
spected while they guessed each word. Ames’s study has the strengths
of systematically sampling content words and using several readers’
performance. Its major weakness is that the majority of the words
being guessed were already very well known to the students. That is,
even though the words were replaced by nonsense words, they repre-
sented known concepts in familiar collocations. In spite of this weak-
ness, this study provides a very useful survey of available clues.
Rankin and Overholser (1969) used Ames’s (1966) classification
system of contextual clues and devised test items to test the effective-
ness of each type of clue. They found a wide range of effectiveness of
the various clues but a highly consistent rank order of difficulty among
grade levels and reading levels. Learners’ reading ability was a good
predictor of the ability to use each of the types of clues.
Care needs to be taken in using Ames’s system as the names for some
of the categories, for example ‘language experience’, and ‘tone, setting,
mood’, do not clearly reflect the types of clues included. Ames pro-
vides plenty of examples in his article.
Table 8.1 (overleaf) lists Ames’s categories with the rankings of
effectiveness obtained by Rankin and Overholser (1969). There are
other ways of classifying context clues. Sternberg and Powell (1983)
use eight functional categories which describe the type of information
conveyed rather than the devices used to convey the information.
Sternberg and Powell’s categories are temporal, spatial, stative

Table 8.1 Ames’s (1966) categories of context clues with Rankin and Overholser’s (1969) rankings of effectiveness
in providing correct responses

Number of items in % correct in Rankin and

Ames’s category Example Ames’s study Overholser

Words in series sonnets and plays of William Shakespeare 31 69

Modifying phrases slashed her repeatedly with a knife 31 62
Familiar expressions expectation was written all over their faces 26 61
Cause and effect He reads not for fun but to make his 10 59
conversation less boring.
Association All the little boys wore short pants. 19 59
Referral clues Sweden 15.3 etc. These statistics carry an 13 55
unpleasant message.
Synonym clues it provokes, and she provokes controversy 36 52
Definition or description some looked alive, though no blood flowed 22 51
beneath the skin
Preposition He sped along a freeway. 20 50
Question and answer Now, what about writing …? 9 43
Comparison or contrast Will it be a blessing or a bane? 37 39
Main idea and detail I soon found a practical use for it. I put orange 17 30
juice inside it.
Non-restrictive clauses 24 hours – hardly a significant period of time 9 26
Learning words from context 363

(physical properties), functional, value (worth or desirability),

causal / enablement, class membership and equivalence. The italicised
word is the word to be guessed, and which was replaced with a non-
sense word in Ames’s study.
Ames (1966) and Sternberg and Powell (1983) describe clues in the
linguistic context of the unknown word. A variety of other factors can
affect guessing from context. Artley (1943) includes typographical
aids such as the use of italics, quotation marks or bolding; word ele-
ments such as the stems and affixes of words; and pictures and
diagrams. Artley calls most of the kinds of clues described by Ames
‘structural clues’.
In addition to these clues, there are what Jenkins and Dixon (1983)
and others call ‘mediating variables’. These mediate between the learn-
ers and the information in the text, strengthening or weakening the
chances of guessing and learning from context. They include the
following:
1. Number of occurrences. The more often an unknown word occurs
the greater the chance of guessing and learning it (Horst et al.,
1998; Stahl and Fairbanks, 1986).
2. Proximity of recurrence. The closer the repetitions the more likely
will the clues from each occurrence be able to be integrated.
3. Variability of contexts. The more different the contexts in which a
word recurs the greater the range of clues available.
4. Presence of relevant clues. Some contexts have useful clues, some
do not.
5. Proximity of relevant clues. The nearer the clues are to the
unknown word, the more likely they are to be used (Carnine et al.,
1984).
6. Number of relevant clues. The more clues there are, the easier the
guessing.
7. Explicitness of relevant clues (Carnine et al., 1984). A clearly sig-
nalled synonym within context helps learning.
8. Density of unknown words. If many unknown words are close to
each other, the harder they will be to guess. Horst et al. (1998)
found that successful guessing related to second language learners’
vocabulary size. This is at least partly because the greater the
learners’ vocabulary size, the greater the number of known words
in the surrounding context.
9. Importance of the unknown word to understanding the text. The
more needed a word is, the more likely a learner will put effort
into the guessing.

10. Prior knowledge of the topic. Real world knowledge can play a
vital part in guessing. Learners who already have a topic-related
script or schema can use this to help guessing.
11. Familiarity of the concept. If the concept is already known, guess-
ing is easier (Nagy et al., 1987). If the concept is strange and
unusual, guessing is difficult (Daneman and Green, 1986).
12. Familiarity of the referents. If the ideas in the clues are familiar to
learners, guessing is easier (Jenkins and Dixon, 1983: 251–2).
13. Concrete vs. abstract referents. If the ideas in the clues are not
abstract, then guessing is easier. Fukkink et al. (2001) found that
unknown concrete words were easier for very young learners, but
were equally difficult to abstract words for older learners.
14. Amount of polysemy. If the word is not polysemous, then guessing
is easier (Saemen, 1970).
Studies of guessing from context have shown that there are high cor-
relations between guessing skills and vocabulary knowledge, reading
skill (Herman et al., 1987), reading comprehension and verbal IQ
(Hafner, 1967). This suggests that an alternative to a direct focus on
guessing skills would be a more general focus on improving reading
skills. This more general focus is supported by the diversity of context
clues that learners need to be able to draw on. There are so many clues
that could be specifically taught and these appear in such a variety of
forms that such a focus may be bewildering and demotivating. A more
general reading skills focus may be more effective. However, if there are
specific aspects to guessing that are not included in general reading
proficiency, then a focus on guessing could be an effective way of getting
competent readers to gain more vocabulary knowledge from context.
On evidence from the study of cloze tests (Chihara et al., 1977; Leys
et al., 1983; Rye, 1985), it seems that most of the clues for guessing
word meanings from context will come in the immediate context, that
is, within the same sentence as the unknown word. Attempts to show
that cloze items are affected by constraints across sentence boundaries
have had mixed results (Rye, 1985). At most, it seems that context clues
from other sentences are likely to account for much less than 10% of the
available clues. Cziko (1978) suggests that sensitivity to discourse clues
develops after sensitivity to syntactic clues in second language learners.

What are the causes of poor guessing?

Frantzen (2003) did a very interesting and detailed investigation of the
causes of incorrect guessing. The amount of information available about
a word in a particular context is clearly a factor affecting guessing.

Webb (2008) looked at learning words from three contexts for each
word. The contexts were rated for the amount of information each pro-
vided about the unknown word. Learners were tested with four tests
covering recognition / recall, and form / meaning. Predictably, there was
more learning of both form and meaning from the more informative
contexts. As is consistent with other studies, multiple-choice recogni-
tion tests gave higher scores than recall tests. The difference between
more and less informative contexts on the form scores (both recall and
recognition) were small and not statistically significant, although the
small differences favoured informative contexts.
A major difficulty faced when guessing words from context is the
form of the word to be guessed. Laufer and Sim (1985a), Bensoussan
and Laufer (1984) and Nassaji (2003) found that second language
learners made many responses based on known words that had some
formal resemblance to the unknown word. Sometimes, these incorrect
form-based guesses resulted in learners reshaping the grammatical
context to fit their incorrect guess.
Laufer and Sim (1985a) looked at the errors that learners made in
trying to interpret a difficult unsimplified text, and described the faulty
approaches that learners took to interpreting the text. Step 1 was to
interpret the meanings of the words, often relying on formal similarity
to known words. Step 2 involved adding textual and extratextual
knowledge. Step 3 involved imposing a sentence structure on the parts
of the text to fit with the lexical clues and knowledge of the world
gained from Steps 1 and 2. This approach resulted in considerable
misinterpretation of the text. Laufer and Sim (1985a) argue that guess-
ing from context should not be focused on until learners have a suffi-
ciently large vocabulary to support such guessing.
Saemen (1970), in a study of young native speakers, found that
uncommonly known meanings of polysemous words were harder to
guess from context when the real word form was used compared with
the use of a nonsense word – the known form led learners towards a
known but inappropriate meaning. Fraser (1999: 239) suggests that
although word form clues can be misleading, it may be impossible to
train learners to hold off using such clues because they are accessed in
a such a fast, automatic manner.
Marks et al. (1974) found that if young native speakers read a story
containing familiar words and then read the same story again but this
time with some unfamiliar words replacing some of the familiar words,
they learned some of these unfamiliar words. The establishment of the
familiar context on the first reading seemed to make it easier to learn
the unfamiliar words on the later reading. This finding provides a
degree of support for what some call ‘diglot readers’, where an L1 text

gradually has its words replaced by L2 words until eventually it turns

into an L2 text (Burling, 1968; 1983).
An important factor affecting guessing from context is the similarity
between the learners’ ﬁrst and second languages. Palmberg (1988)
found that young Swedish speakers were able to comprehend much of
a specially prepared English text even though they knew almost no
English. This can be a dangerous strategy, however, and in general it
seems best to let context guide the guess rather than form.
Neuman and Koskinen (1992) looked at the effect of captioned tele-
vision, television alone, simultaneous listening and reading, and
reading alone on the learning of unknown vocabulary from context
for ESL learners. They found the captioned television condition to be
superior to the other conditions, and also evidence of a ‘Matthew
effect’ (the rich get richer and the poor get poorer), in this case with
learners of higher English proﬁciency learning more words.
Li (1988) compared second language learners’ guessing from
context in repeated contrived contexts through listening and reading,
and found greater successful guessing from reading. Brown et al.
(2008) found much less vocabulary learning on immediate post-tests
from listening (multiple-choice 8.2 out of 28, translation 0.56) than
from reading only (multiple-choice 12.54, translation 4.10) and
reading while listening (13.31, 4.10).
The standard used to measure success at guessing is often too high
and translation tests involving recall are among the toughest. When
scoring recall tests credit is often not given for partial knowledge.
Frantzen (2003), for example, sees rope, something like a rope or chain,
fastener as incorrect guesses for wire in a context where the wire fastens
the lid of the container. These guesses however show elements of
meaning that make at least a small contribution to knowledge of wire.

Do different learners approach guessing in the same way?

We have looked at variables affecting guessing which are a result of the
word itself and the context in which it appears. We have also looked
at ‘mediating’ variables which are related to the context such as the
number of times the word is repeated and distance between the clues
and the word to be guessed. There are also variables that relate to the
person doing the guessing. There is evidence that there are different
ways of approaching the guessing task (van Daalen-Kapteijns and
Elshout-Mohr, 1981) and different ability, knowledge and skills that
learners bring to the guessing task.
There are several studies that examine second language learners’
approaches to guessing from context (Arden-Close, 1993; Bensoussan

and Laufer, 1984; Haastrup, 1989; Haynes and Baker, 1993; Homburg
and Spaan, 1982; Huckin and Bloch, 1993; Laufer and Sim, 1985b;
McKeown, 1985; Morrison, 1996; Nassaji, 2003; Parry, 1991; van
Parreren and Schouten-van Parreren, 1981; Walker, 1983). In general,
a good guesser uses a variety of clues, checks various types of clues
against each other, does not let the form of the word play too large a
part, and does not arrive at a guess prematurely. Proficiency in L2 is a
major factor in successful guessing.
We need to be careful in interpreting the results of such studies
because it is clear that the procedures used to investigate the guessing
process influence what happens. At the very least, the investigative
procedures of introspection or writing down the cues used substan-
tially increase the amount of time that a reader would normally spend
on guessing a word from context. In addition, the investigative pro-
cedures change it from being incidental learning to become an inten-
tional, problem-solving activity, and often encourage definite guesses
instead of allowing incremental learning. These studies show that
there are substantial clues in the context that are available to the sensi-
tive reader. The studies also show that not all readers can make good
use of these clues.
Van Daalen-Kapteijns and Elshout-Mohr (1981) compared high
verbal and low verbal native speakers’ performance on a deliberately
focused guessing-from-context task. High and low verbal learners
were distinguished by measures that looked at quantity of word
knowledge. High verbal learners tended to use an analytic strategy,
choosing an initial model of the word meaning and transforming addi-
tional information to fill out and refine the initial model. The trans-
forming (reshaping) part of the process was seen as being a critical
feature of the analytic process. Low verbal learners also set up an
initial model but tended to remember the various additional cues dis-
covered from other contexts with little or no reworking or transfor-
mation of the initial model. Any final summing up of a definition then
tended to rely on memory for the model and additional clues and
required a weighing up of the various bits of information at that point.
Van Daalen-Kapteijns and Elshout-Mohr also found differences
between high and low verbal learners in the quality of the form of the
definition that they arrived at as a result of guessing. Low verbal learners
tended to use a less standard form of definition compared to the succinct
classic form of superordinate plus essential defining features. This same
difference was also found when the high and low verbal learners were
asked to define common words that were well known to them. The
study shows that learners may approach guessing in different ways and
this may result in qualitatively different outcomes. Although the study

does not discuss this, it may be that there is a causative connection

between the approach taken to guessing and vocabulary size.
Shefelbine (1990) found that native speakers with higher levels of
general vocabulary were able to guess more words than learners with
lower levels. This vocabulary size difference was both quantitative
(lower vocabulary students knew fewer words) and qualitative (they
knew some words less well than the higher vocabulary students).
Lower vocabulary size means that (a) there are more words to guess;
(b) there is less comprehensible context to support the guesses; and
(c) learners bring less background knowledge to the texts they read.
Daneman and Green (1986) argue that learners’ success in guessing
from context will vary according to the size of their working memory.
Working memory can be measured by getting learners to perform a
reading span test. In this test learners are given increasingly longer sets
of sentences to read aloud and at the end of each set they try to recall
the last word of each sentence in the set. Their reading span is the
maximum number of sentences they can read aloud while still being
able to recall all of the last words in the sentences. The size of working
memory and success in guessing from context are related because
guessing from context involves integrating the information from suc-
cessively met context clues (p. 8). If these clues are no longer available
in memory then guessing will be poor. As well as finding a significant
correlation (.69) between reading span and success at guessing from
context, Daneman and Green found a significant correlation (.58)
between skill at guessing and vocabulary knowledge. Sternberg and
Powell (1983) found a similar correlation and argued that a vocabu-
lary test measures past acquisition from context, while a learning-
from-context task measures present acquisition.
Daneman and Green (1986) suggest that the capacity of working
memory will vary according to how efficient a learner is in using the
specific processes which are needed in the task they are working on.
An optimistic view would be that training in these processes would
increase the amount of information that could be held in working
memory. Training in the processes needed for guessing from context
could increase the space available in working memory for effective
application of this skill.
Churchill (2007) has a fascinating account of how his knowledge of
a particular Japanese word developed through the varied experiential
context in which he met the word, highlighting the trial and error
nature of such learning.
Arden-Close (1993) examined the guessing-from-context strategies
of second language learners of different proficiency levels by getting
learners to write their thoughts while they guessed. He found that even

proﬁcient learners were distracted by the form of the unknown words

(contamination = contain, spas = space). Arden-Close used three kinds
of texts – texts with words underlined, texts with words left out, and
texts containing nonsense words. In both the underlined and nonsense
word texts, the forms of the words tended to distract the learners. In the
blank-filling texts, there was a higher success rate, presumably because
only context and not word form clues could be used. Learners’ guessing
was limited by their knowledge of English but where they could bring
background features to bear they could make good use of it and made
more successful guesses. The lowest proficiency students often gave the
meaning of neighbouring words as the guess for the unknown word.
Arden-Close’s analysis shows the complexity of the guessing skill and
the close relationship it has with general language proficiency and
reading proficiency. While training is likely to improve skill at guessing,
it is unlikely to adequately compensate for low language proficiency.
Nassaji (2003) found that drawing on world knowledge and mor-
phological knowledge had the highest degree of success and were the
commonest sources used. While discourse knowledge was not com-
monly used, its success rate was relatively high. Nassaji’s success rates,
combining successful and partially successful, were around 50%.
Nassaji (2004) also found a relationship between success at guessing
and depth of vocabulary knowledge, thus underlining the importance
of lexical knowledge in comprehension and related skills.

How can learners be trained to guess from context?

Meta-analyses of L1 studies in training in guessing from context have
found that training results in better guessing (Fukkink and de Glopper,
1998; Kuhn and Stahl, 1998), although there was no clear consensus
on what kind of training is the most effective.
Methods of training learners to guess can be classiﬁed into three
major types (see Walters, 2004, for a very substantial critical review):
1. general strategy training, which can be as simple as a general prin-
ciple such as look for clues, or as complicated as a series of steps as
in Clarke and Nation (1980);
2. context clue instruction where learners are alerted to a range of
useful context clues such as explicit textual deﬁnitions, synonyms,
positives, and various other conjunction type relationships (Ames,
1966; Buikema and Graves, 1993); and
3. practice and feedback with cloze exercises.
Walters (2006) compared these three methods with second language
learners using six hours of training for each method (three two-hour

sessions). All three groups improved while the control group did not.
The strategy group made the biggest increase, but the small number of
subjects (around 12 in each group) and the very large standard devia-
tions meant that the differences between the training groups were not
significant. There were also increases in comprehension scores as a
result of training in guessing. Clearly, training in guessing is worth-
while and deserves a reasonable investment of time and effort, not just
a few sessions but small sustained attention to guessing (Hafner, 1965;
Jenkins et al., 1989).
The most important ways in which teachers can help learners
improve learning from context are:
1. by helping them find and choose reading and listening material of
appropriate difficulty;
2. by encouraging them to read a lot and helping them gain a lot of
comprehensible spoken input;
3. by improving their reading skills so that they read fluently and with
good comprehension; and
4. by providing training in guessing from context, including training
in a particular strategy which encompasses giving attention to
various clues, and providing substantial focused guessing practice.
These ways are ranked in order of importance with the most impor-
tant first. The reason for this ranking is that guessing from context
seems to be a subskill of reading and seems to draw heavily on other
reading skills. Good guessers are good readers (McKeown, 1985). The
four ways described above can be more generally described as a match-
ing of learner and text approach, a quantity approach, a general skill
approach, and a particular skill approach. Nassaji (2006: 388) use-
fully sees lexical inferencing as a variety of more general inferencing,
with the implication that isolating lexical inferencing from other lan-
guage-related inferencing is to some degree misrepresenting what
happens when learners read.
It may be that training in guessing helps vocabulary learning simply
because it encourages learners to give deliberate thoughtful attention
to vocabulary items – that is, it develops word consciousness (Scott
and Nagy, 2004).

Does drawing attention to words help learning from context?

There is some evidence that a combination of attention-drawing activ-
ities, such as presenting words to learners before reading (Jenkins et al.,
1984) and deﬁning words as they occur in context (Elley, 1989)
increases the amount of vocabulary learning. Swanborn and de Glopper

(1999), in a meta-analysis of 20 studies of learning from context,

found that the nature of the vocabulary pre-test affected the amount of
words learned. Laufer and Hill (2000) suggest that having words high-
lighted in their computerised text probably increased dictionary
look-up and therefore learning. Drawing attention to words increases
the chance of them being learned. It is important to distinguish between
the effects of these kinds of activity on vocabulary learning and on
comprehension of the text. Jenkins et al. (1984), studying young native
speakers of English, found no direct effect of pre-teaching on compre-
hension, but there was a marked effect on the learning of the words
from context.
Attention-drawing could be done in the following ways (several of
these methods have been tested in experimental studies, but many
have not):

1. Drawing attention to the word

pre-testing
pre-teaching
seeing a list before reading
highlighting (colour, bold, italics) in the text
having a list while reading
2. Providing access to the meaning
glossing
teacher defining through pre-teaching
teacher defining while listening to the text
hypertext look-up
dictionary look-up
3. Motivating attention to the word
warning of a test
providing follow-up exercises
noting contexts while reading (e.g. filling in a notebook)

Do glossing and dictionary use help vocabulary learning?

There is now considerable evidence that when learners’ attention is
drawn towards unfamiliar words and there is a clear indication of
their meaning, vocabulary learning is much greater than when learners
read without deliberately focusing on new vocabulary.
Nist and Olejnik (1995) examined the procedure of meeting a word
in context and then looking up its meaning in the dictionary. Four dif-
ferent kinds of tests were used to measure the learning of each word.
They found that there was no interaction between the meeting in context

and the looking up of the word, and that the quality of the dictionary
definition determined the quality of learning. Nist and Olejnik argue
that dictionaries can be substantial contributors to the process of
vocabulary learning. Hulstijn’s (1993) study of inferencing and dictio-
nary look-up behaviour found that learners who were good at inferring
preferred to confirm their guesses by consulting a dictionary. Learners
differed greatly in their skill at inferring. There was a modest correlation
(.50) between inferring ability and overall vocabulary size.
Watanabe (1997) compared three forms of vocabulary glossing in
texts on second language learners’ vocabulary learning. The three forms
of glossing were: (1) inserting a brief explanation of the word in the text
immediately after the word (Bramki and Williams, 1984, call this ‘lexical
familiarization’); (2) glossing the word in the margin ‘[crib = baby’s
bed]’; and (3) providing two-choice glosses in the margin. Hulstijn
(1992) has suggested that multiple-choice glosses supplement contex-
tual information, encourage mental effort by having to choose, and
avoid incorrect inferences by providing a meaning. Glossing appeared
to improve comprehension. The two conditions involving glosses in the
margin of the text resulted in higher scores in the immediate and delayed
post-tests compared to providing the meaning in the text and having no
glosses or meaning provided in the text. Learning from the single-gloss
treatment was higher than the multiple-choice gloss treatment in all
post-tests but not significantly so. The slightly lower scores for multiple
choice may have come from learners making the wrong choice. Glossing
almost doubled the learning (17 words) compared to learning from the
text with no glosses or lexical familiarisation (10 words).
Mondria and Wit-de Boer (1991) used specially constructed, iso-
lated sentences to investigate second language learners’ learning from
context. The experience involved three stages: (1) a guessing stage
where the context sentences were shown and learners guessed the
translation of the target words; (2) a learning stage where learners saw
the correct translations of the target words and had to learn them; and
(3) a testing stage where learners saw the words in new non-informative
contexts and had to translate them. It seems that the testing stage
immediately followed the learning stage. Mondria and Wit-de Boer
found no relationship between success at guessing and retention. It is
likely that the learning stage overwhelmed the effects of guessing.

What formats should be used for testing or

practising guessing?
Researchers have used a variety of formats for testing or practising
guessing. These range from ﬁxed deletion cloze procedure where the

missing item is a blank, to unaltered texts where learners guess words

with the real word form present.
There are several factors that need to be considered when deciding
on a format for guessing:
The effect of the word form. Several studies (Bensoussan and Laufer,
1984; Laufer and Sim, 1985a; Nassaji, 2003) have shown that learn-
ers are often inﬂuenced by the actual form of the word. If the word
resembles a known word, the form may lead them to a wrong guess. If
the form contains familiar parts, then these may be used to guide the
guess. One of the most difficult things to learn when becoming proﬁ-
cient at guessing is to let the context rather than the form guide the
guess. Formats which use a blank remove this distraction. This may be
useful at the early stages of developing a guessing strategy but it is
important at some stage that learners get practice in suspending form-
based guesses while they use the context to guess. When learners’
guessing skill is tested, it is useful to see if they have control of this
aspect of the strategy. It may also be useful to see if learners can deal
with homographs when a different member is known.
Previous knowledge of the word to be guessed. When testing the
guessing skill, it is necessary to be sure that learners do not already
know the word that is to be guessed. One way of solving this problem
is to replace the words to be guessed with nonsense words. This then
means that any answer is truly a guess. Leaving blanks also achieves
this purpose. However, there are several context clues that are avail-
able for known words that are not available for unknown words.
These clues make guessing easier and not representative of guessing
truly unknown words. For example, there are Ames’s (1966) clues of
familiar expressions (collocations), as in: Who spends one evening a
week thacing the fat with the boys? Because the collocation is known,
the word is easily guessed, but if chewing the fat was not known
before, then it would not be guessable.
There are thus two kinds of previous knowledge to consider: the
knowledge of the form itself, and the collocational, grammatical
knowledge. Using nonsense words deals with the problem of knowl-
edge of the form but it does not deal with the other kinds of knowl-
edge. The validity of Ames’s study is thus severely compromised by not
taking account of this kind of knowledge. Nation and Webb (2011:
265–268) look more closely at the use of nonsense words in research.
The density of unknown words and the size of the context. An
important factor affecting success at guessing is the ratio of known
words to unknown words. Liu and Nation (1985) suggest that a ratio
of one unknown to 24 known words is needed for successful guessing.
That is, at least 95% of the words in the text must be familiar to the

reader. If the density of unknown words is too great then learners do

not have a chance to show their guessing skill.
Guessing could be tested or practised with isolated sentences or
with continuous text. As we have seen, it seems that only a small pro-
portion of the clues needed for guessing occur outside the sentence
containing the unknown word. It thus may be acceptable for practical-
ity reasons to practise or test some guessing in isolated sentences.
However, at some stage in a learner’s development of the guessing
strategy it is important that the few clues from the wider context are
given attention.
The types of words that are guessed. Words that represent unfamil-
iar concepts are more difficult to guess than words that represent
known concepts (Nagy et al., 1987). It is likely, especially for second
language learners, that the majority of words to be guessed represent
known concepts. However, some, especially technical words, will also
require learners to develop new concepts. Second language learners in
their later meetings with words in context will also need to see distinc-
tions between the L2 word and the nearest L1 equivalent. When testing
the guessing skill it is thus important to see if learners are able to deal
with unfamiliar concepts.
The different parts of speech are not equally represented at the
various frequency levels. There tend to be more nouns among the
lower frequency words, for example. If a true measure of the learners’
guessing skill is needed, it is important that the kinds of words to be
guessed represent the kinds of words that a learner with a given
vocabulary size would need to guess. A strength of Ames’s (1966)
study was that he tried to get a representative sample of words to guess
by using a cloze procedure. Unfortunately he did not take the vocabu-
lary size of his learners into account and so did not restrict his sample
to words outside their level of vocabulary knowledge.
It should be clear from the discussion of these four factors affecting
guessing from context that the validity of a practice or testing format
for guessing from context would be enhanced if: (a) the actual word
form appeared in the context; (b) learners did not already know the
word; (c) there was a low density of unknown words; (d) the unknown
words were in a continuous text; and (e) the unknown words were
typical of those a learner of that vocabulary size would meet. This is
the ideal and for a variety of reasons, many of them related to peda-
gogical, reliability and practicality issues, other formats have been
used. Table 8.2 lists the possibilities.
When practising and testing guessing from context it may be effec-
tive to draw on a variety of formats to focus attention on particular
aspects of the guessing skill.

Table 8.2 Features of formats for

testing or practising guessing

1. Word form
a. a blank space instead of the word
b. a nonsense word
c. a real word
2. Selection of words and contexts
a. real randomly sampled contexts
b. real selected contexts
c. contrived contexts
3. Size and relationship of contexts
a. isolated sentence contexts
b. isolated paragraph contexts
c. continuous text contexts

Dunmore (1989) reviewed exercise types in ﬁve different coursebooks

for practising guessing from context and found four major exercise
types:
• matching a given synonym with a word in the text;
• filling a blank with a suitable word;
• providing words before reading and then seeing if the learner can
use context to find the meanings of the words; and
• developing awareness of text features that could help guessing.
Dunmore is critical of the various exercise types because they tend to
test rather than train guessing, and encourage a belief that synonyms
are sufficient to express the meanings of unknown words. This last
criticism may be a little harsh as finding a first language translation or
a second language synonym may be a reasonable first approximation
of the meaning of a word.

What are the steps in the guessing-from-context strategy?

There is no one procedure for guessing from context, but most proce-
dures draw on the same kinds of clues. Some procedures work towards
the guess in an inductive approach. Others work out deductively from
the guess, justifying the guess. A deductive approach is more suited
to younger learners who will be less analytical in their approach and to
advanced learners who are familiar with the various clues and wish
to concentrate on developing ﬂuency in guessing. An inductive
approach, such as that described by Clarke and Nation (1980) is useful
for making learners aware of the range of clues available and for

developing the subskills that may be needed to make use of the clues.
The aim of all guessing procedures is to help learners become fluent
and skillful at guessing from context so that the guessing does not
interrupt the normal flow of reading too much.
Let us look first at Clarke and Nation’s (1980) five-step procedure.
Further discussion of it can be found in (Nation, 1990) and Nation
and Coady (1988).
Step 1. Decide on the part of speech of the unknown word.
Step 2. Look at the immediate context of the word, simplifying
it grammatically if necessary.
Step 3. Look at the wider context of the word, that is, the
relationship with adjoining sentences or clauses.
Step 4. Guess.
Step 5. Check the guess.
Is the guess the same part of speech as the unknown
word?
Substitute the guess for the unknown word. Does it fit
comfortably into the context?
Break the unknown word into parts. Does the meaning
of the parts support the guess?
Look up the word in the dictionary.
This procedure is strongly based on language clues and does not draw
on background content knowledge. Linguistic clues will be present in
every context, background clues will not. This procedure aims at being
as generalisable as possible.
The procedure moves from a narrow focus on the word in Step 1 to
a broader view in Step 3. Van Parreren and Schouten-van Parreren
(1981) suggest that there are various levels of information with the
grammar level being lower than the meaning level. The higher meaning
level can only be used if the lower grammar level does not cause prob-
lems. Making a guess involves choosing the appropriate level at which
to seek information and moving to another level if this proves to be the
wrong one (p. 240).
Step 1 in Clarke and Nation’s procedure encourages the learner to
focus on the unknown word and ensures that the right word is focused
on. Note that word-part analysis does not occur at this step. Arriving
at a correct guess from word-part analysis is less sure than using
context clues. Getting learners to delay using word-part clues is the
most difficult thing to learn when developing skills in guessing from
context.
Step 2 looks at the immediate context, that is, the clause containing
the unknown word. This source of information will contain most of

the clues needed to guess most words correctly. Sometimes the imme-
diate context is difficult to interpret because it is in the form of a
passive with a missing agent, because the subject and verb are sepa-
rated by a relative clause or through nominalisation, or pronouns are
present which need to be interpreted. Learners can practise clarifying
the immediate context by unpacking nominalisations, turning the
passive construction into an active one, and by interpreting reference
words. There is an exercise called ‘What does what?’ which gives this
practice. Here is an example of the exercise applied to a text. The exer-
cise is very easy to prepare.
The teacher chooses an appropriate text, preferably with line
numbers. The teacher then writes the line number and word and learn-
ers have to ask ‘What does what?’ about the word. In the example
below, the ‘What does what?’ questions have been added to clarify the
procedure. Usually learners will have to make the questions them-
selves. More information on ‘What does what?’ can be found in
Nation (1979; 2009: 39–43).
We live in a style that most of our grandparents could not even have imag-
ined. Medicine has cured diseases that terriﬁed them. Most of us live in
better and more spacious homes. We eat more, we grow taller, we are even
born larger than they were. Our parents are amazed at the matter-of-fact
way we handle computers. We casually use products – microwave ovens,
graphite tennis rackets, digital watches – that did not exist in their youth.
Economic growth has made us richer than our parents and grandparents.
But economic growth and technical change, and the wealth they bestow,
have not liberated us from scarcity. Why not? Why, despite our immense
wealth, do we still have to face costs? (Parkin, 1990: Chapter 3)
Who imagines what?
What terriﬁes who?
Who produces what?
What did not exist?
What grows?
What bestows what?
What liberates who?
Step 3 involves looking at the wider context. A conjunction relation-
ship activity can be used to practise this part of the procedure. In this
activity, the learners have to see what joining word can be put between
the clause containing the unknown word and the adjoining clauses.
Sometimes the relationship will already be marked by a conjunction,
adverbial or some other sign of the relationship, but these will still
have to be interpreted. Learners can be helped with this by having a list
of prototypical conjunction relationship markers like those in the left-
hand column of Table 8.3 (overleaf). The learners may need to become

Table 8.3 Conjunction relationships and their meaning

Relationship and
prototypical marker Other markers The meaning of the relationship between the clauses
Inclusion furthermore, also, in addition, The classes joined together are in a list and share similar
and similarly … information.
Contrast however, although, nevertheless, The clauses are in contrast to each other. One may be negative and
but yet … the other positive. They may contain opposing information.
Time sequence next, after, before, when, first … The clauses are steps in a sequence of events. They might not be in
then the order in which they happened.
Cause–effect thus, so, since, as a result, so that, One clause is the cause and the other is the effect.
because in order to, if …
Exemplification and e.g., such as, for instance … The following clause is an example of the preceding more general
amplification statement, or the following clauses describe the general statement in
for example more detail.
Alternative nor, alternatively … The clauses are choices and they will share similar information.
or
Restatement that is (to say), namely … The following clause has the same meaning as the preceding clause.
in other words
Summary to sum up, in a word … The following clause summarises what has gone before.
in short
Exclusion rather than, on the contrary … The following clause excludes what has just been said. That is, it
instead has the opposite meaning.
Learning words from context 379

familiar with the kinds of information that each of these markers pro-
vides, which is outlined in the right-hand column of the table, and they
need to know the range of words that signal these relationships (see
Nation, 1979; 1984; 1990: Appendix 6; and Halliday and Hasan,
1976, for further information on conjunction relationships).
Learners can practise this step by interpreting the relationship
between pairs of sentences in a text. This is usefully done by learners
working in pairs or small groups initially.
Step 4 is the guess – the moment of truth. When this is done as a
class activity, the teacher can award percentage points for the guesses
with 100% (or 110%) for a fully correct guess, 90% for a very good
guess, 80% for a guess that comes close to the meaning, and so on.
This is not the last step.
Step 5 involves checking the guess to see if it is on the right track.
Comparing the part of speech of the guess with the part of speech
decided on at Step 1 makes sure that the learner is focusing on the
unknown word. Sometimes incorrect guesses are simply the meaning
of an adjoining word. The second way of checking, substitution,
makes sure that the context has been considered, because the word
will not ﬁt if it has not been considered. The third way of checking
involves word part analysis. We will look at this in detail in Chapter 9.
It comes at this stage to make sure that the learner does not twist the
interpretation of the context on the basis of what the word looks like.
Laufer’s (1988) and Laufer-Dvorkin’s (1991) study of synforms shows
that this is a very common problem. The learner analyses the word
parts and sees if the meaning of the parts relates to the guess. If they
do, the learner can feel happy. Looking up the word in a dictionary is
the last way of checking. It should be easy to choose the appropriate
meaning from the dictionary if several meanings are listed there,
because the guess will have given a good indication of which one to
choose.
The deductive procedure (see, for example, Bruton and Samuda,
1981), involves the following steps:
Step 1. Guess the meaning of the word.
Step 2. Justify the guess using a variety of clues.
Step 3. Readjust the guess if necessary.
The advantage of this procedure is that it places the guess at the fore-
front of the activity and allows for intuition to play a part. It also
works well as a group and class activity. Whichever approach learners
tend to favour, they need not follow a rigid procedure when guessing
but they should be aware of the range of possible clues and should
have the skills to draw on them.

How should we plan the training of learners in the strategy

of guessing from context?
Guessing from context is a complex activity drawing on a range of
skills and types of knowledge. It is worth bearing in mind that it is a
subskill of reading and listening and depends heavily on learners’
ability to read and listen with a good level of proficiency. Learning a
complex guessing strategy will not adequately compensate for poor
reading or listening skills and low proficiency. Developing these
reading and listening skills is the first priority.
When learners are given training in guessing from context, they
should work with texts where at least 95% of the running words are
familiar to them. This will allow them to have access to the clues that
are there. In addition, the words chosen for guessing should be able to
be guessed. Not all words have enough clues; adjectives are usually
difficult to guess because they enter into few relationships with other
words, while nouns and verbs are usually easier.
Training in guessing should be given plenty of time. In a pre-univer-
sity course, it could be practised three or four times a week for about
ten minutes each time for at least six weeks, and preferably longer. The
aim of the practice is to get learners guessing quickly without having
to deliberately go through all the steps. Fraser (1999) found that
making learners familiar with the strategies of ‘ignore, consult (a dic-
tionary) and infer’, involving about eight hours of instruction, resulted
in a decrease in the amount of ignoring and an increase in the amount
of inferring. The success rates were over 70% for consulting a diction-
ary and inferring from context if partially correct inferences were
included. A further eight hours of instruction on linguistic context
clues may have helped maintain the success rate of inferring, especially
for inferring where learners created a paraphrase for the meaning of
the unknown word.
Involving the class working together with the teacher, in groups,
pairs and then individually, training can focus on the subskills:
determining part of speech; doing ‘What does what?’; interpreting
conjunction relationships; and doing word part analysis. Training
should also involve going through all the steps, gradually getting
faster and faster. The teacher can model the procedure first, gradu-
ally handing over control to the learners. Learners can report on
guessing that they did in their outside reading and listening and
others can comment on their attempts. There can be regular guessing
from context tests using isolated sentences and corrected texts.
Learner improvement on these tests can be recorded as a means of
increasing motivation.

Teachers should be able to justify the time and effort spent on the
guessing strategy to themselves, their learners and other teachers.
These justifications could include:
• the value of the strategy for high-frequency, mid-frequency and
low-frequency words;
• the fact that the strategy accounts for most vocabulary learning by
native speakers;
• the enormous number of words that can be dealt with and perhaps
learned through this strategy;
• the effectiveness of the strategy;
• the benefits of the strategy in contributing to reading and listening
comprehension;
• the fact that learners differ widely in their control of this skill, and
training can narrow these differences; and
• the need for this skill in dictionary use.
Teachers should also be able to look critically at the various activities
suggested for improving guessing (Dunmore, 1989; Honeyfield, 1977;
Walters, 2004). Yosuke Sasao (http://ysasaojp.info) has developed a
test of the guessing skill that can be used diagnostically and to measure
improvement in guessing.
In any list of vocabulary-learning strategies, guessing from context
would have to come at the top of the list. Although it has the disadvan-
tages of being a form of incidental learning (and therefore less certain)
and of not always being successful (because of lack of clues), it is still
the most important way that language users can increase their vocabu-
lary. It deserves teaching time and learning time. A well-planned
vocabulary development programme gives spaced, repeated attention
to this most important strategy.
In the following chapters we will look at the strategies of using word
parts, using dictionaries and using word cards.

References
Ames, W. S. (1966). The development of a classiﬁcation scheme of contextual
aids. Reading Research Quarterly, 2, 1, 57–82.
Anderson, R. C. and Ortony, A. (1975). On putting apples into bottles: A problem
of polysemy. Cognitive Psychology, 7, 167–80.
Anderson, R. C. and Shifrin, Z. (1980). The meaning of words in context. In
Spiro, R. J., Bruce, B. C. and Brewer, W. F. (eds.), Theoretical Issues in Reading
Comprehension (pp. 330–48). Mahwah, NJ: Lawrence Erlbaum Associates.
Anderson, R. C., Stevens, K. C., Shifrin, Z. and Osborn, J. (1978). Instantiation
of word meanings in children. Journal of Reading Behavior, 10, 2, 149–57.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
382 Learning Vocabulary in Another Language
Arden-Close, C. (1993). NNS readers’ strategies for inferring the meanings of
unknown words. Reading in a Foreign Language, 9, 2, 867–93.
Artley, A. S. (1943). Teaching word-meaning through context. Elementary
English Review, 20, 1, 68–74.
Beck, I. L., McKeown, M. G. and McCaslin, E. S. (1983). Vocabulary: All contexts
are not created equal. Elementary School Journal, 83, 3, 177–81.
Bensoussan, M. and Laufer, B. (1984). Lexical guessing in context in EFL reading
comprehension. Journal of Research in Reading, 7, 1, 15–32.
Bramki, D. and Williams, R. C. (1984). Lexical familiarization in economics text,
and its pedagogic implications in reading comprehension. Reading in a
Foreign Language, 2, 1, 169–81.
Brown, R., Waring, R. and Donkaewbua, S. (2008). Incidental vocabulary
acquisition from reading, reading-while-listening, and listening to stories.
Reading in a Foreign Language, 20, 2, 136–63.
Bruton, A. and Samuda, V. (1981). Guessing words. Modern English Teacher, 8,
3, 18–21.
Buikema, J. L. and Graves, M. F. (1993). Teaching students to use context cues to
infer word meanings. Journal of Reading, 36, 6, 450–57.
Burling, R. (1968). Some outlandish proposals for the teaching of foreign
languages. Language Learning, 18, 1, 61–75.
Burling, R. (1983). A proposal for computer-assisted instruction in vocabulary.
System, 11, 2, 181–70.
Cairns, H. S., Cowart, W. and Jablon, A. D. (1981). Effects of prior context upon
the integration of lexical information during sentence processing. Journal of
Verbal Learning and Verbal Behavior, 20, 445–53.
Carnine, D., Kameenui, E. J. and Coyle, G. (1984). Utilization of contextual
information in determining the meaning of unfamiliar words. Reading
Research Quarterly, 19, 2, 188–204.
Carton, A. S. (1971). Inferencing: a process in using and learning language. In
Pimsleur, P. and Quinn, T. (eds.), The Psychology of Second Language
Learning (pp. 45–58). Cambridge: Cambridge University Press.
Chihara, T., Oller, J., Weaver, K. and Chavez-Oller, M. A. (1977). Are cloze items
sensitive to discourse constraints? Language Learning, 27, 63–73.
Churchill, E. (2007). A dynamic systems account of learning a word: From ecology
to form relations. Applied Linguistics, 29, 3, 339–58.
Clarke, D. F. and Nation, I. S. P. (1980). Guessing the meanings of words from
context: Strategy and techniques. System, 8, 3, 211–20.
Cook, J. M., Heim, A. W. and Watts, K. P. (1963). The word-in-context: A new
type of verbal reasoning test. British Journal of Psychology, 54, 3, 227–37.
Cziko, G. A. (1978). Differences in ﬁrst- and second-language reading: The use of
syntactic, semantic and discourse constraints. Canadian Modern Language
Review, 34, 473–89.
Daneman, M. and Green, I. (1986). Individual differences in comprehending
and producing words in context. Journal of Memory and Language, 25,
1–18.
Day, R. R., Omura, C. and Hiramatsu, M. (1991). Incidental EFL vocabulary
learning and reading. Reading in a Foreign Language, 7, 2, 541–51.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
Learning words from context 383
de Bot, K. and Stoessel, S. (2000). In search of yesterday’s words: Reactivating a
long-forgotten language. Applied Linguistics, 21, 3, 333–53.
Dunmore, D. (1989). Using contextual clues to infer word meaning: an evaluation
of current exercise types. Reading in a Foreign Language, 6, 1, 337–47.
Dupuy, B. and Krashen, S. D. (1993). Incidental vocabulary acquisition in French
as a foreign language. Applied Language Learning, 4, 1&2, 55–63.
Elley, W. B. (1989). Vocabulary acquisition from listening to stories. Reading
Research Quarterly, 24, 2, 174–87.
Frantzen, D. (2003). Factors affecting how second language Spanish students
derive meaning from context. Modern Language Journal, 87, 168–99.
Fraser, C. A. (1999). Lexical processing strategy use and vocabulary learning
through reading. Studies in Second Language Acquisition, 21, 225–41.
Fukkink, R., Blok, H. and de Glopper, K. (2001). Deriving word meaning from
written context: A multicomponential skill. Language Learning, 51, 3, 477–96.
Fukkink, R. G. and de Glopper, K. (1998). Effects of instruction in deriving word
meaning from context: A meta-analysis. Review of Educational Research,
68, 4, 450–69.
Gibbons, H. (1940). The ability of college freshmen to construct the meaning of a
strange word from the context in which it appears. Journal of Experimental
Education, 9, 1, 29–33.
Gipe, J. P. and Arnold, R. D. (1979). Teaching vocabulary through familiar
associations and contexts. Journal of Reading Behavior, 11, 3, 282–5.
Haastrup, K. (1985). Lexical inferencing: A study of procedures in reception.
Scandinavian Working Papers on Bilingualism, 5, 63–87.
Haastrup, K. (1987). Using thinking aloud and retrospection to uncover learners’
lexical inferencing procedures. In Faerch, C. and Kasper, G. (eds.),
Introspection in Second Language Research (pp. 197–212). Clevedon:
Multilingual Matters.
Haastrup, K. (1989). Lexical Inferencing Procedures, vols 1 & 2. Copenhagen:
Handelshojskolen i Kobenhavn.
Hafner, L. E. (1965). A one-month experiment in teaching context aids in ﬁfth
grade. Journal of Educational Research, 58, 10, 472–4.
Hafner, L. E. (1967). Using context to determine meanings in high school and
college. Journal of Reading, 10, 7, 491–8.
Halff, H. M., Ortony, A. and Anderson, R. C. (1976). A context-sensitive
representation of word meanings. Memory and Cognition, 4, 4, 378–83.
Halliday, M. A. K. and Hasan, R. (1976). Cohesion in English. London: Longman.
Hansen, L., Umeda, Y. and McKinney, M. (2002). Savings in the relearning of
second language vocabulary: The effects of time and proﬁciency. Language
Learning, 52, 4, 653–78.
Haynes, M. and Baker, I. (1993). American and Chinese readers learning from
lexical familiarization in English text. In Huckin, T., Haynes, M. and Coady,
J. (eds.), Second Language Reading and Vocabulary (pp. 130–52). Norwood,
NJ: Ablex.
Herman, P., Anderson, R. C., Pearson, P. D. and Nagy, W. E. (1987). Incidental
acquisition of word meaning from expositions with varied text features.
Reading Research Quarterly, 22, 3, 263–84.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
384 Learning Vocabulary in Another Language
Homburg, T. J. and Spaan, M. C. (1982). ESL reading proficiency assessment:
Testing strategies. In Hines, M. and Rutherford, W. (eds.), On TESOL ’81
(pp. 25–33). Washington: TESOL.
Honeyfield, J. (1977). Word frequency and the importance of context in vocabulary
learning. RELC Journal, 8, 2, 35–42.
Horst, M., Cobb, T. and Meara, P. (1998). Beyond a Clockwork Orange: Acquiring
second language vocabulary through reading. Reading in a Foreign Language,
11, 2, 207–23.
Huckin, T. and Bloch, J. (1993). Strategies for inferring word meanings: A
cognitive model. In Huckin, T., Haynes, M. and Coady, J. (eds.), Second
Language Reading and Vocabulary (pp. 153–78). Norwood, NJ: Ablex.
Hulstijn, J. (2003). Incidental and intentional learning. In Doughty, C. and Long,
M. (eds.), Handbook of Second Language Acquisition (pp. 349–81). Oxford:
Blackwell.
Hulstijn, J. H. (1992). Retention of inferred and given word meanings: Experiments
in incidental vocabulary learning. In Arnaud, P. J. L. and Bejoint, H. (eds.),
Vocabulary and Applied Linguistics (pp. 113–25). London: Macmillan.
Hulstijn, J. H. (1993). When do foreign-language readers look up the meaning of
unfamiliar words? The influence of task and learner variables. Modern
Language Journal, 77, 2, 139–47.
Jacoby, L. L., Craik, F. J. M. and Begg, J. (1979). Effects of decision difficulty on
recognition and recall. Journal of Verbal Learning and Verbal Behavior, 18,
585–600.
Jenkins, J. R. and Dixon, R. (1983). Vocabulary learning. Contemporary
Educational Psychology, 8, 237–60.
Jenkins, J. R., Matlock, B. and Slocum, T. A. (1989). Two approaches to
vocabulary instruction: The teaching of individual word meanings and
practice in deriving word meanings from context. Reading Research
Quarterly, 24, 2, 215–35.
Jenkins, J. R., Stein, M. L. and Wysocki, K. (1984). Learning vocabulary through
reading. American Educational Research Journal, 21, 4, 767–87.
Kelly, P. (1990). Guessing: no substitute for systematic learning of lexis. System,
18, 2, 199–208.
Kennedy, G. (1987). Expressing temporal frequency in academic English. TESOL
Quarterly, 21, 1, 69–86.
Knight, S. M. (1994). Dictionary use while reading: The effects on comprehension
and vocabulary acquisition for students of different verbal abilities. Modern
Language Journal, 78, 3, 285–99.
Kuhn, M. R. and Stahl, S. A. (1998). Teaching children to learn word meanings
from context. Journal of Literacy Research, 30, 1, 119–38.
Laufer, B. (1988). The concept of ‘synforms’ (similar lexical forms) in vocabulary
acquisition. Language and Education, 2, 2, 113–32.
Laufer, B. and Hill, M. (2000). What lexical information do L2 learners select in
a CALL dictionary and how does it affect word retention? Language Learning
&Technology, 3, 2, 58–76.
Laufer, B. and Sim, D. D. (1985a). Taking the easy way out: non-use and misuse
of clues in EFL reading. English Teaching Forum, 23, 2, 7–10, 20.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
Learning words from context 385
Laufer, B. and Sim, D. D. (1985b). Measuring and explaining the reading threshold
needed for English for academic purposes texts. Foreign Language Annals,
18, 5, 405–11.
Laufer-Dvorkin, B. (1991). Similar Lexical Forms in Interlanguage. Tübingen:
Gunter Narr Verlag.
Leys, M., Fielding, L., Herman, P. and Pearson, P. D. (1983). Does cloze measure
intersentence comprehension? A modiﬁed replication of Shanahan, Kamil,
and Tobin. In Niles, J. A. and Harris, L. A. (eds.), New Enquiries in Reading
(pp. 111–14). Rochester, NY: National Reading Conference.
Li, X. (1988). Effects of contextual cues on inferring and remembering meanings.
Applied Linguistics, 9, 4, 402–13.
Liu, N. and Nation, I. S. P. (1985). Factors affecting guessing vocabulary in
context. RELC Journal, 16, 1, 33–42.
Marks, C. B., Doctorow, M. J. and Wittrock, M. C. (1974). Word frequency and
reading comprehension. Journal of Educational Research, 67, 259–62.
McKeown, M. G. (1985). The acquisition of word meaning from context
by children of high and low ability. Reading Research Quarterly, 20, 4,
482–96.
Mondria, J. A. (2003). The effects of inferring, verifying and memorising on the
retention of L2 word meanings. Studies in Second Language Acquisition, 25,
4, 473–99.
Mondria, J. A. and Wit-de Boer, M. (1991). The effects of contextual richness on
the guessability and the retention of words in a foreign language. Applied
Linguistics, 12, 3, 249–67.
Morrison, L. (1996). Talking about words: A study of French as a second language
learners’ lexical inferencing procedures. Canadian Modern Language
Journal, 53, 1, 41–75.
Nagy, W. E. (1997). On the role of context in ﬁrst- and second-language learning.
In Schmitt, N. and McCarthy, M. (eds.), Vocabulary: Description, Acquisition
and Pedagogy (pp. 64–83). Cambridge: Cambridge University Press.
Nagy, W. E., Anderson, R. C. and Herman, P. A. (1987). Learning word meanings
from context during normal reading. American Educational Research
Journal, 24, 2, 237–70.
Nagy, W. E., Herman, P. and Anderson, R. C. (1985). Learning words from
context. Reading Research Quarterly, 20, 2, 233–53.
Nassaji, H. (2003). L2 vocabulary learning from context: Strategies, knowledge
sources, and their relationship with success in L2 lexical inferencing. TESOL
Quarterly, 37, 4, 645–70.
Nassaji, H. (2006). The relationship between depth of vocabulary knowledge and
L2 learners’ lexical inferencing strategy use and success. Canadian Modern
Language Review, 90, 3, 387–401.
Nation, I. S. P. (1979). The curse of the comprehension question: Some alternatives.
Guidelines: RELC Journal Supplement, 2, 85–103.
Nation, I. S. P. (1984). Understanding paragraphs. Language Learning and
Communication, 3, 1, 61–8.
Nation, I. S. P. (1990). Teaching and Learning Vocabulary. Rowley, MA: Newbury
House.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
386 Learning Vocabulary in Another Language
Nation, I. S. P. (2009). Teaching ESL/EFL Reading and Writing. New York:
Routledge.
Nation, I. S. P. and Coady, J. (1988). Vocabulary and reading. In Carter, R. and
McCarthy, M. (eds.), Vocabulary and Language Teaching (pp. 97–110).
London: Longman.
Nation, I. S. P. and Webb, S. (2011). Researching and Analyzing Vocabulary.
Boston: Heinle Cengage Learning.
Nation, P. and Wang, K. (1999). Graded readers and vocabulary. Reading in a
Foreign Language, 12, 2, 355–80.
Neuman, S. B. and Koskinen, P. (1992). Captioned television as comprehensible
input: Effects of incidental word learning from context for language minority
students. Reading Research Quarterly, 27, 1, 95–106.
Nist, S. L. and Olejnik, S. (1995). The role of context and dictionary definitions on
varying levels of word knowledge. Reading Research Quarterly, 30, 2, 172–93.
Palmberg, R. (1988). On lexical inferencing and language distance. Journal of
Pragmatics, 12, 207–14.
Paribakht, S. (2005). The influence of first language lexicalization on second
language lexical inferencing: A study of Farsi-speaking learners of English as
a foreign language. Language Learning, 55, 4, 701–48.
Parkin, M. (1990). Macroeconomics. Boston, MA: Addison-Wesley.
Parry, K. (1991). Building a vocabulary through academic reading. TESOL
Quarterly, 25, 4, 629–53.
Perfetti, C. (2010). Decoding, vocabulary, and comprehension. In McKeown, M.
G. and Kucan, L. (eds.), Bringing Reading Research to Life (pp. 291–303).
New York: Guilford Press.
Perfetti, C. and Hart, L. (2001). The lexical basis of comprehension skill. In
Gorfien, D. S. (ed.), On the Consequences of Meaning Selection: Perspectives
on Resolving Lexical Ambiguity (pp. 67–86). Washington, D.C.: American
Psychological Association.
Perfetti, C. A. and Hart, L. (2002). The lexical quality hypothesis. In Verhoeven,
L., Elbro, C. and Reitsma, P. (eds.), Precursors of Functional Literacy
(pp. 189–213). Amsterdam: John Benjamin.
Pitts, M., White, H. and Krashen, S. (1989). Acquiring second language vocabulary
through reading: A replication of the Clockwork Orange study using second
language acquirers. Reading in a Foreign Language, 5, 2, 271–5.
Prince, P. (1996). Second language vocabulary learning: The role of context versus
translations as a function of proficiency. Modern Language Journal, 80, 4,
478–93.
Pulido, D. (2003). Modelling the role of second language proficiency and topic
familiarity in second language incidental vocabulary acquisition. Language
Learning, 53, 2, 233–84.
Pulido, D. (2007). The effects of topic familiarity and passage sight vocabulary on
L2 lexical inferencing and retention through reading. Applied Linguistics, 28,
1, 66–86.
Pulido, D. (2009). How involved are American L2 learners of Spanish in lexical
input processing tasks during reading? Studies in Second Language
Acquisition, 31, 31–58.

Downloaded from https://www.cambridge.org/core. University College Birmingham, on 23 Nov 2021 at 14:31:46, subject to the Cambridge
Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139858656.010
Learning words from context 387
Pulido, D. and Hambrick, D. Z. (2008). The virtuous circle: Modeling individual
differences in L2 reading and vocabulary development. Reading in a Foreign
Language, 20, 2, 164–90.
Qian, D. (1999). Assessing the roles of depth and breadth of vocabulary knowledge
in reading comprehension. Canadian Modern Language Review, 56, 2, 282–
307.
Rankin, E. F. and Overholser, B. M. (1969). Reaction of intermediate grade
children to contextual clues. Journal of Reading Behavior, 1, 3, 50–73.
Rye, J. (1985). Are cloze items sensitive to constraints across sentences? A review.
Journal of Research in Reading (UKRA), 8, 2, 94–105.
Saemen, R. A. (1970). Effects of commonly known meanings on determining
obscure meanings of multiple-meaning words in context. Retrieved from the
Office of Education (DHEW), Washington, DC.
Saragi, T., Nation, I. S. P. and Meister, G. F. (1978). Vocabulary learning and
reading, System, 6, 2, 72–8.
Schatz, E. K. and Baldwin, R. S. (1986). Context clues are unreliable predictors of
word meaning. Reading Research Quarterly, 21, 4, 439–53.
Scott, J. A. and Nagy, W. E. (2004). Developing word consciousness. In Baumann,
J. F. and Kame’enui, E. J. (eds.), Vocabulary Instruction: Research to Practice
(pp. 201–17). Guilford Press: New York.
Seibert, L. C. (1945). A study of the practice of guessing word meanings from a
context. Modern Language Journal, 29, 4, 296–323.
Shefelbine, J. L. (1990). Student factors related to variability in learning word
meanings from context. Journal of Reading Behavior, 22, 1, 71–97.
Shu, H., Anderson, R. C. and Zhang, Z. (1995). Incidental learning of word
meanings while reading: A Chinese and American cross-cultural study.
Reading Research Quarterly, 30, 1, 76–95.
Stahl, S. A. and Fairbanks, M. M. (1986). The effects of vocabulary instruction: A
model-based meta-analysis. Review of Educational Research, 56, 1, 72–110.
Sternberg, R. J. and Powell, J. S. (1983). Comprehending verbal comprehension.
American Psychologist, 38, 878–93.
Sutarsyah, C., Nation, P. and Kennedy, G. (1994). How useful is EAP vocabulary
for ESP? A corpus based study. RELC Journal, 25, 2, 34–50.
Swanborn, M. S. L. and de Glopper, K. (1999). Incidental word learning while
reading: A meta-analysis. Review of Educational Research, 69, 3, 261–85.
van Daalen-Kapteijns, M., Elshout-Mohr, M. and de Glopper, K. (2001). Deriving
the meaning of unknown words from multiple contexts. Language Learning,
51, 1, 145–81.
van Daalen-Kapteijns, M. M. and Elshout-Mohr, M. (1981). The acquisition of
word meanings as a cognitive learning process. Journal of Verbal Learning
and Verbal Behavior, 20, 386–99.
van Parreren, C. F. and Schouten-van Parreren, M. (1981). Contextual guessing:
A trainable reader strategy. System, 9, 3, 235–41.
Walker, L. J. (1983). Word identiﬁcation strategies in reading a foreign language.
Foreign Language Annals, 16, 4, 293–9.
Walters, J. (2004). Teaching the use of context to infer meaning: A longitudinal
survey of L1 and L2 vocabulary research. Language Teaching, 37, 243–52.

Handbook of Practical Second Language Teaching And... - (Part V Vocabulary and Grammar) - 2
No ratings yet
Handbook of Practical Second Language Teaching And... - (Part V Vocabulary and Grammar) - 2
12 pages
2014 05 20 - Reading Meeting - Notes On Schmitt
No ratings yet
2014 05 20 - Reading Meeting - Notes On Schmitt
4 pages
2128-Article Text-5896-1-10-20210526
No ratings yet
2128-Article Text-5896-1-10-20210526
5 pages
Lognostics
No ratings yet
Lognostics
6 pages
154 Vocabulary in Language Teaching Teaching and Learning Vocabulary
No ratings yet
154 Vocabulary in Language Teaching Teaching and Learning Vocabulary
1 page
Guessing Meaning From Context
No ratings yet
Guessing Meaning From Context
5 pages
Teaching Children To Learn Word Meanings From Context:: A Synthesis and Some Questions
No ratings yet
Teaching Children To Learn Word Meanings From Context:: A Synthesis and Some Questions
20 pages
Why Teach Vocabulary
No ratings yet
Why Teach Vocabulary
16 pages
Vocab and Learning Cycle
No ratings yet
Vocab and Learning Cycle
15 pages
Learning Styles of Uni EFL Students
No ratings yet
Learning Styles of Uni EFL Students
11 pages
Applied Linguistics 3rd Year Edu.+Basic Ch.3 (Part 2)
No ratings yet
Applied Linguistics 3rd Year Edu.+Basic Ch.3 (Part 2)
30 pages
Vocabulary Instruction: Wake Robin Elementary August 13, 2010
No ratings yet
Vocabulary Instruction: Wake Robin Elementary August 13, 2010
15 pages
EXTRA Ila Second Language Learners Vocabulary Oral Language
No ratings yet
EXTRA Ila Second Language Learners Vocabulary Oral Language
8 pages
The Modern Language Journal - 2020 - JIN - Incidental Vocabulary Learning Through Listening To Teacher Talk
No ratings yet
The Modern Language Journal - 2020 - JIN - Incidental Vocabulary Learning Through Listening To Teacher Talk
17 pages
Effective Strategies For Second Language Vocabulary Acquisition
No ratings yet
Effective Strategies For Second Language Vocabulary Acquisition
4 pages
Deliberate Learning From Word Cards
No ratings yet
Deliberate Learning From Word Cards
42 pages
Intentional Vs Incidental Vocabulary Learning
No ratings yet
Intentional Vs Incidental Vocabulary Learning
9 pages
How Effective Is Second Language Incidental Vocabulary Learning A Meta Analysis
No ratings yet
How Effective Is Second Language Incidental Vocabulary Learning A Meta Analysis
20 pages
Context Clues Versus Word Definitions
No ratings yet
Context Clues Versus Word Definitions
32 pages
Paribakht&Wesche (1997)
No ratings yet
Paribakht&Wesche (1997)
14 pages
Vocabulary Teaching Strategies
No ratings yet
Vocabulary Teaching Strategies
44 pages
10.4324 9781315802336 Previewpdf
No ratings yet
10.4324 9781315802336 Previewpdf
34 pages
CH 4 - The Conditions That Contribute To Vocabulary Learning
No ratings yet
CH 4 - The Conditions That Contribute To Vocabulary Learning
16 pages
Chapter 7 - The Acquisition of Vocabulary
No ratings yet
Chapter 7 - The Acquisition of Vocabulary
9 pages
Nation, I. S. P. (2001) - Learning Vocabulary in Another Language. Cambridge University Press.
100% (2)
Nation, I. S. P. (2001) - Learning Vocabulary in Another Language. Cambridge University Press.
25 pages
(I. S. P. Nation) Learning Vocabulary in Another Language
100% (1)
(I. S. P. Nation) Learning Vocabulary in Another Language
0 pages
Incidental l2 Vocabulary Acquisition From and While Reading An Eye Tracking Study
No ratings yet
Incidental l2 Vocabulary Acquisition From and While Reading An Eye Tracking Study
34 pages
Chapter 7 - The Acquisition of Vocabulary
No ratings yet
Chapter 7 - The Acquisition of Vocabulary
11 pages
Handbook of Vocabulary Studies-246-270
No ratings yet
Handbook of Vocabulary Studies-246-270
25 pages
The Internet As An Educational Tool in Vocabulary Instruction
No ratings yet
The Internet As An Educational Tool in Vocabulary Instruction
5 pages
Teching and Explaining Vocabulary in A Lexical View
No ratings yet
Teching and Explaining Vocabulary in A Lexical View
27 pages
BEST - Week 11-12 Vocab Improvement
No ratings yet
BEST - Week 11-12 Vocab Improvement
27 pages
Incidental Vocab Acquisition
No ratings yet
Incidental Vocab Acquisition
1 page
Vocabulary Session 22-23 - Seminar Slides
No ratings yet
Vocabulary Session 22-23 - Seminar Slides
53 pages
Vocabulary: TE 301 Spring 2010
No ratings yet
Vocabulary: TE 301 Spring 2010
16 pages
Binder (2020) - Vocabulary Accessibility and Acquisition Do You Get More From A Financestor
No ratings yet
Binder (2020) - Vocabulary Accessibility and Acquisition Do You Get More From A Financestor
22 pages
Week 3 - Applied Ling. - Vocabulary
No ratings yet
Week 3 - Applied Ling. - Vocabulary
22 pages
Carol Fraser 1999
No ratings yet
Carol Fraser 1999
17 pages
Word Detective Strategy Related Study
No ratings yet
Word Detective Strategy Related Study
77 pages
m5 Vocabulary Skills and Knowledge
No ratings yet
m5 Vocabulary Skills and Knowledge
44 pages
The Mnemonic Value of Orthography For Vocabulary Learning
No ratings yet
The Mnemonic Value of Orthography For Vocabulary Learning
17 pages
Vocabulary Tips for Teachers
No ratings yet
Vocabulary Tips for Teachers
11 pages
Vocabulary Teaching in English Language Teaching
No ratings yet
Vocabulary Teaching in English Language Teaching
4 pages
(Furtado Et Al 2018) A - Serious - Game - For - Improving - Inferencing in The Presence of Foreign Language Unknown Words
No ratings yet
(Furtado Et Al 2018) A - Serious - Game - For - Improving - Inferencing in The Presence of Foreign Language Unknown Words
8 pages
Vocabulary Teaching for Educators
No ratings yet
Vocabulary Teaching for Educators
5 pages
Best Practices in Vocabulary
No ratings yet
Best Practices in Vocabulary
15 pages
Using Context Clues: To Understand The Meaning of Unknown Words
No ratings yet
Using Context Clues: To Understand The Meaning of Unknown Words
14 pages
SECTION C: Task 4: Forum Questions
No ratings yet
SECTION C: Task 4: Forum Questions
5 pages
Deriving Word Meanings From Context: Does Explanation Facilitate Contextual Analysis?
No ratings yet
Deriving Word Meanings From Context: Does Explanation Facilitate Contextual Analysis?
13 pages
Vocaulary Building
No ratings yet
Vocaulary Building
13 pages
AGG32 - Study Material Final
No ratings yet
AGG32 - Study Material Final
279 pages
The Influence of Grammar On English Learners Writ
No ratings yet
The Influence of Grammar On English Learners Writ
16 pages
Common Errors in Written English Essays of Form One
No ratings yet
Common Errors in Written English Essays of Form One
12 pages
Project Omni - Labelbox Evaluation Labeling Instructions
No ratings yet
Project Omni - Labelbox Evaluation Labeling Instructions
30 pages
Language Culture and Society Final Syllabus
No ratings yet
Language Culture and Society Final Syllabus
15 pages
Mtb-Mle 1
No ratings yet
Mtb-Mle 1
13 pages
Cestero Mancera, Ana M. - Nonverbal Communication in L2 Spanish Teaching (2018 Routledge)
No ratings yet
Cestero Mancera, Ana M. - Nonverbal Communication in L2 Spanish Teaching (2018 Routledge)
18 pages
Elt
50% (2)
Elt
255 pages
Testing and Assessment of Listening
No ratings yet
Testing and Assessment of Listening
19 pages
Year 6 CEFR English Lesson Plan
No ratings yet
Year 6 CEFR English Lesson Plan
8 pages
MEd Secondary Education
No ratings yet
MEd Secondary Education
2 pages
(Ebook PDF) Human Diversity in Education 8th Edition by Kenneth H. Cushnerpdf Download
100% (6)
(Ebook PDF) Human Diversity in Education 8th Edition by Kenneth H. Cushnerpdf Download
46 pages
Language Guidance Resource - Updated June 2020
No ratings yet
Language Guidance Resource - Updated June 2020
196 pages
MA Thesis in OU
No ratings yet
MA Thesis in OU
131 pages
Attitude & Motivation Scale Survey From AMTB
100% (4)
Attitude & Motivation Scale Survey From AMTB
4 pages
Language Awareness: To Cite This Article: Donglan Zhang & Christine C.M. Goh (2006) : Strategy Knowledge and Perceived
No ratings yet
Language Awareness: To Cite This Article: Donglan Zhang & Christine C.M. Goh (2006) : Strategy Knowledge and Perceived
22 pages
Serious Game Design Model For Language Learning in The Cultural Context
No ratings yet
Serious Game Design Model For Language Learning in The Cultural Context
39 pages
Social Strategy 2
No ratings yet
Social Strategy 2
10 pages
Errors Vs Mistakes
No ratings yet
Errors Vs Mistakes
2 pages
Research On English Language Teaching and Learning in Argentina 2007-2013
No ratings yet
Research On English Language Teaching and Learning in Argentina 2007-2013
34 pages
100 Questions for Kids Course
100% (4)
100 Questions for Kids Course
38 pages
A Closer Look at The Relationship of Cognitive and Metacognitive Strategy Use To EFL Reading Achievement Test Performance
No ratings yet
A Closer Look at The Relationship of Cognitive and Metacognitive Strategy Use To EFL Reading Achievement Test Performance
32 pages
Audio-Lingual Method Overview
No ratings yet
Audio-Lingual Method Overview
26 pages
Cognitive Styles
No ratings yet
Cognitive Styles
3 pages
Using Pragmatics in Elt Focus Paper
No ratings yet
Using Pragmatics in Elt Focus Paper
5 pages
1998 Brindley Assessing Listening
No ratings yet
1998 Brindley Assessing Listening
22 pages
Developing Students' Ability
No ratings yet
Developing Students' Ability
10 pages
ISHTIAQ Et Al., 2017 PDF
No ratings yet
ISHTIAQ Et Al., 2017 PDF
20 pages
Nlcs
No ratings yet
Nlcs
43 pages
Study 0
No ratings yet
Study 0
10 pages

Vocabulary Learning from Context

Uploaded by

Vocabulary Learning from Context

Uploaded by

8 Learning words from context

Incidental learning from context is the most important of all the

Which is best: intentional or incidental learning?

Context sentences and phrases are valuable aids in intentional,

In the meaning-inferred method, where guesses were veriﬁed or cor-

How are reading and vocabulary growth related to

In a later paper, Perfetti (2010) elaborated on what were good vocabu-

Learners with good skills at inferring from context develop larger

Decoding skills Skill at accessing

Vocabulary size Comprehension

Inferring from Amount of input

The circle is also relevant to L2 learners. Pulido and Hambrick

What proportion of unknown words can be guessed

They found no signiﬁcant difference between learners who had context

There is useful partial information available from this context. First,

Fifth, in discussions of learning from context, it is important to distin-

How much vocabulary is learned from context?

a certain time may be an important factor in learning from context.

per year, then considerable vocabulary learning is possible. Nagy

There are several implications to be drawn from the ﬁndings on the

What can be learned from context?

An example of aberration would be:

What clues does a context provide and how effective

Number of items in % correct in Rankin and

Words in series sonnets and plays of William Shakespeare 31 69

(physical properties), functional, value (worth or desirability),

What are the causes of poor guessing?

gradually has its words replaced by L2 words until eventually it turns

Do different learners approach guessing in the same way?

does not discuss this, it may be that there is a causative connection

proﬁcient learners were distracted by the form of the unknown words

How can learners be trained to guess from context?

Does drawing attention to words help learning from context?

(1999), in a meta-analysis of 20 studies of learning from context,

1. Drawing attention to the word

Do glossing and dictionary use help vocabulary learning?

What formats should be used for testing or

missing item is a blank, to unaltered texts where learners guess words

reader. If the density of unknown words is too great then learners do

Table 8.2 Features of formats for

Dunmore (1989) reviewed exercise types in ﬁve different coursebooks

What are the steps in the guessing-from-context strategy?

Table 8.3 Conjunction relationships and their meaning

How should we plan the training of learners in the strategy

You might also like