2018, Cilt 8, Sayı 1, Sayfa(lar) 172-183
A Matter of Controversy: Teaching New L2 Words in Semantic Sets or Unrelated Sets
Bursa Technical University, School of Foreign Languages, Bursa, Turkey
Keywords: Interference theory, Semantic set (lexical set), Semantically unrelated set, Semantic clustering, L2 Vocabulary learning and teaching
The relevant literature reveals no consensus on whether new vocabulary items should be presented in semantic or semantically-unrelated
sets in L2 classrooms. Actually, there is an excessive amount of research evidence on the interfering effect of teaching semantically-related
words at the same time. However, the majority of these studies have been carried out under strictly-controlled laboratory conditions, so
there is still need for more classroom research on this controversial issue. The present study aims to investigate the effects of presenting
novel words in lexical sets versus semantically unrelated sets on students acquisition of these words in a real classroom setting. The
participants, 44 Turkish EFL learners, were taught 12 English target words in either semantic or unrelated clusters. The vocabulary
instruction was given through pictorial flashcards accompanied by several sentential contexts, which supplied the participants with several
meaningful encounters with the target vocabulary items. The results indicated that both types of vocabulary instruction provided EFL
learners with effective recognition and production of the target words immediately after the treatments as well as two weeks later. These
findings revealed no statistically significant difference between clustering words in semantic sets and unrelated sets. Hence, the current
study did not find out any interfering effect of teaching semantically related words simultaneously in a real classroom condition.
Words are regarded as building blocks of a language. They lay
bridges into the mysterious world of meaning in a language.
Languages are by no means meaningful without words. With
this in mind, vocabulary has a fundamental role in almost all
phases of second language (L2) acquisition. First of all, the main
motivation behind L2 learning is to communicate, and words
are indispensable for a successful verbal communication. As
suggested by Wilkins (1972), Without grammar very little can
be conveyed, without vocabulary nothing can be conveyed
(p. 111). That is the reason why tourists mostly prefer carrying
dictionaries to travelling with grammar books. Second,
vocabulary knowledge is positively correlated with various
aspects of L2 development. A broad L2 lexicon is considered to
be a reliable indicator of learners overall language proficiency
as well as a good facilitator of four main language skills. If
an analogy is made between a language and a human body,
vocabulary is the heart which pumps blood to all the other
vital organs such as reading, writing, listening and speaking
(Sarıoğlu, 2014: 1). Hence, vocabulary acquisition is of prime
importance in the mastery of effective L2 comprehension and
Although there exists a general agreement on the central
role of vocabulary instruction in L2 learning and teaching,
there seems to be no consensus on the efficiency of some
vocabulary teaching techniques. Specifically, whether novel
L2 words should be taught in semantic sets or semantically
unrelated sets is a controversial issue in L2 lexical field.
Teaching vocabulary in semantic (lexical) sets can be defined as
simultaneous grouping of new L2 words within meaningful sets
such as clothes, family members and school equipment.
Semantically clustered words inherently share some common
semantic features. As for presenting vocabulary in semantically
unrelated sets, it requires categorising new L2 words which
have no meaningful association with one another (see Table 1).
Click Here to Zoom
|Table 1: Categorising L2 Words in Semantic or Semantically Unrelated Sets
Teaching L2 vocabulary in semantic sets or unrelated sets
is a matter of controversy in the related literature. On the
one hand, some studies suggest the traditional practice of
clustering novel L2 words in semantic sets as an efficient way of
lexical instruction (e.g., Gairns & Redman, 1986; Graves, 2006;
Hashemi & Gowdasiaei, 2005; Haycraft, 1993; Hoshino, 2010;
Stahl & Naggy, 2006). With respect to this standpoint, if certain semantically-associated words are taught simultaneously
within the same lexical set, it will have a facilitative impact on
L2 learners acquisition of the given words. Semantic clustering
is argued to provide language learners with well-organised
information, which is principally easier to learn (Baddeley,
1990). Such a meaningful arrangement of words is also thought
to be in line with the organisation of semantic fields in human
brain (Aitchison, 1994). Now that vocabulary items are stored
in our brain correspondingly, we can recall semantically related
words more easily than unorganised lexical items. Bearing this
viewpoint in mind, a great number of language curricula and
course books are inclined to serve new EFL (English as a foreign
language) words in semantic sets.
On the other hand, an increasing amount of lexical research
gives support to instruction of vocabulary items in unrelated
sets (e.g., Erten & Tekin, 2008; Finkbeiner & Nicol, 2003; Nation,
2000; Tinkham, 1993, 1997; Waring, 1997). These studies argue
that presenting new words in semantically related sets will have
an interfering effect on L2 learners acquisition of these lexical
items, which is a matter of frustration in vocabulary learning.
Interference theory is put forward as a rationale behind the
growing opposition to the semantic clustering of words. The
theory suggest that, if a new vocabulary item has too many
semantic similarities with the words learned at the same time,
L2 learners will have more difficulty in learning that new item
because of the interfering effects of those similar words on
one another (Tinkham, 1997). According to Schmitt (2000), L2
learners usually confuse the English words right and left
since they share the exactly same semantic features except for
direction. Thus, a great number of studies lay emphasis on
the interfering effect of teaching novel L2 words in the same
Overall, whether new vocabulary items should be presented
to L2 learners in semantic sets or unrelated sets is still open
to debate in the relevant literature. Some studies support the
common principle of semantic clustering as an effective way
of L2 vocabulary instruction. Others favour teaching novel L2
words in semantically unrelated sets due to the interfering
impact of grouping vocabulary items in lexical sets. There are
also some other studies, whose findings reveal no statistically
significant difference between categorising new vocabulary in
semantic sets and in unrelated sets (e.g., Ishii, 2013; 2015).
With this in mind, the related literature reveals no consensus on this issue, and it is still worthy of further research, particularly,
by means of more authentic classroom-based studies.
Presenting new L2 words in semantic set
There may be two motivating forces behind teaching new L2
vocabulary items in semantic clusters (Tinkham, 1997). Firstly,
serving L2 words in semantic sets appears to be in harmony
with both structural and communicative approaches of
language teaching. Secondly, presenting words in lexical sets is
considered to help L2 learners realise the semantic boundaries
among the words within the sets (Gairns & Redman, 1986).
If the novel target words are clustered with regard to their
semantic features, L2 learners will easily explore the semantic
differences and similarities among the given words. Thus,
semantic grouping of L2 vocabulary items is argued to provide
learners with more well organised information so that they can
store and then recall them easily from their memory.
Aitchison (1994) also highlights the advantages of establishing
semantic relations among new L2 vocabulary items. From his
point of view, words are semantically stored and systematised
in our brain; therefore, conceptual and semantic mapping will
facilitate L2 learners acquisition, retention and recall of these
vocabulary items. In this regard, presenting new L2 words in
semantic sets seems to be more in line with the organization of
semantic fields in human brain. In a similar vein, Haycraft (1993)
makes an analogy between serving words in semantically
unrelated sets and a tree with no trunk and branches but only
leaves (p. 44). He states further that semantic grouping of
new vocabulary items will enable L2 learners to make more
effective interconnections among these words in their minds.
In their study, Hashemi and Gowdasiaei (2005) compared the
efficiency of teaching L2 words in lexical sets versus semantically
unrelated vocabulary instruction in terms of both vocabulary
breadth and depth of 60 Iranian EFL learners. The study findings
revealed positive research evidence on facilitative impact
of semantic clustering. The participants who learned new L2
words in semantically related set achieved better test results
when compared to their counterparts who learned the target
words in unrelated set. The study pedagogically suggested that
EFL teachers should systematically categorise new L2 words
under relevant topics and teach them in meaningful contexts.
Hoshino (2010) also carried out a study to investigate the
effects of different word lists on vocabulary acquisition of
119 Japanese EFL university students with various learning
styles. In a real classroom condition, the effectiveness of five
types of word lists was compared: (a) semantic/categorical,
(b) unrelated, (c) thematic, (d) synonyms, and (d) antonyms.
The results indicated that all of the participants with divergent
learning styles memorized more target words in the semantic/
categorical list when compared to the words in the other four
lists. Thus, the study findings also emphasised the facilitative
impact of categorising new words in semantic sets on L2
learners vocabulary acquisition.
Presenting new L2 vocabulary items in unrelated set
Tinkham (1993) conducted two strictly controlled experiments
to compare the vocabulary learning rates and speeds of
English native speakers in semantic sets versus in unrelated
sets. The results of both experiments demonstrated that the
participants memorised the target pseudo-words faster and
with fewer trials in unrelated sets than those in semantic sets.
In his replication study, Waring (1997) verified Tinkhams (1993)
findings as well. The native speakers of Japanese also needed
longer time and more trials to memorise the target pseudowords
in semantic sets. Both studies suggested that clustering
new words in unrelated sets facilitates learners acquisition of
these words. They revealed some research evidence to confirm
the interfering effects of serving semantically similar words
together in the same set.
Finkbeiner and Nicol (2003) carried out an experimental study
to compare these two types of clustering words in a laboratory
setting. 47 mono-lingual English speakers were presented
32 pseudo-words in either semantic or unrelated sets. The
participants were seated in front of a computer screen. They
both listened to the target words over the headphones and saw
its corresponding picture on the screen for 500 ms. They also
repeated the target words into the microphone twice. After
training all the words in the same way, the oral L1-L2 and L2-L1
translation tests were conducted. The participants translation
latencies of were measured and compared within two groups.
The results showed that the participants translated the target
words in semantic categories more slowly than those in
unrelated clusters. The study findings also offered research
evidence on the detrimental effect of arranging vocabulary
items in semantic categories.
In their study, Erten and Tekin (2008) investigated the effects of
presenting new words in semantic or unrelated sets on Turkish
EFL learners immediate and delayed recognition of these
words as well as their test completion time. Sixty fourth-grade
students were presented 80 L2 words within four 20-word
sets: two sets were introduced in semantic categories and the
rest two included unrelated words. The immediate post-test
results indicated that EFL learners achieved more vocabulary
gains when the target words were introduced in semantic sets.
The semantically related words were also recognised better in
the first delayed post-tests. However, the second delayed posttests
produced better results in favour of the words thought in
unrelated sets. In addition, the participants test completion
time was much shorter for the target words in unrelated sets.
Thus, the study findings concluded that Turkish EFL learners
revealed superior performance in learning new L2 words in
semantically unrelated sets, which highlighted the interfering
effect of learning and teaching lexical items in semantic sets.
The relevant literature also comprises some lexical research,
the findings of which could not differentiate between semantic
and unrelated clustering of novel L2 vocabulary items. As an
example, two recent studies conducted by Ishii (2013, 2015)
displayed no statistically significant difference between
presenting new L2 words in semantic and unrelated sets. She stated that acquiring semantically similar words is neither
more difficult nor easier than learning the given words in
unrelated sets. Likewise, Papathanasiou (2009) verified the
interference effect of teaching words in semantic sets only for
adult beginners, but not for young L2 learners at intermediate
level of proficiency. In conclusion, the existent literature reveals
no agreement on these two types of clustering new L2 words,
which is the main focus of this study.
THE CURRENT STUDY
Justification for the Study
There are two justifications for this study. First, whether novel
words should be presented in semantic sets or unrelated sets
in L2 classrooms is worthy of further investigation since it is still
a matter of controversy in the related literature. On the one
hand, some studies still encourage L2 teachers to serve new L2
vocabulary items in semantic sets (e.g., Hashemi & Gowdasiaei,
2005; Hoshino, 2010). On the other hand, an increasing number
of studies suggest teaching new words in unrelated sets due to
interference effect of semantic categorisation (Erten & Tekin,
2008; Finkbeiner & Nicol, 2003; Tinkham, 1993, 1997; Waring,
1997; Wilcox & Medina, 2013). The existent literature also
includes some studies whose findings reveal no statistically
significant difference between these two types of grouping
words (Ishii, 2013, 2015).
As for the second justification, although there is more research
evidence on the interfering effect of semantic clustering,
the experimental circumstances in most of these studies are
not natural enough to draw conclusions about vocabulary
acquisition of learners in real L2 classrooms. In most of these
a. The participants are given a very restricted time to
memorize the target words. (e.g., Finkbeiner & Nicol, 2003;
Ishii, 2013, 2015; Tinkham, 1993, 1997; Waring, 1997;
Wilcox & Medina, 2013).
b. Pseudo-words (artificial words) are assigned as target
vocabulary items. (e.g., Finkbeiner & Nicol, 2003; Ishii,
2013, 2015; Tinkham, 1993, 1997; Waring, 1997).
c. The target words are usually taught in isolation rather than
being presented within a larger context (e.g., Erten & Tekin,
2008; Finkbeiner & Nicol, 2003; Ishii, 2013, 2015; Tinkham,
1993, 1997; Waring, 1997).
Many of these studies have been conducted under such strictly
controlled experimental conditions (e.g., Finkbeiner & Nicol,
2003; Ishii, 2013, 2015; Tinkham, 1993, 1997; Waring, 1997).
Therefore, it is not likely to generalise about the real classroom
circumstances. Such kinds of laboratory-type settings cannot
offer implications about L2 vocabulary teaching and learning.
Thus, there is still a need for further research, especially
In their recent study, Bolger and Zapata (2011) found out that
the adding story context to L2 vocabulary appears to eliminate
the interfering impact of semantic clustering. Hence, we need
to know more about how the study results would be if learners
studied new L2 words in a wider context and practiced them
within a great amount of time in a real classroom.
The Aim of the Study
The present study aims to investigate the effects of both
teaching new L2 words in semantic sets and in unrelated sets
on vocabulary acquisition of EFL learners in a real classroom
setting. More specifically, it attempts to compare these two
types of lexical clustering in terms of EFL learners recognition
and production of the target L2 words after they were provided
with a number of meaningful encounters with these words in
sentential contexts. With this in mind, this study tries to answer
these two research questions:
1. What are the effects of teaching new L2 words in semantic
set or semantically unrelated set on EFL learners
recognition and production of these vocabulary items?
2. Does teaching new L2 words in semantic set or unrelated
set differ in terms of EFL learners immediate or delayed
recognition and production of these vocabulary items?
|Participants and Setting
The participants of this study were 44 Turkish EFL learners.
They were 12th-grade students from three intact classes at a
state high school in Bursa, Turkey. They were all native speakers
of Turkish and with A2 CEFR1 level of English proficiency. 33
of them were female, and the rest 11 were male. Their ages
ranged from 16 to 17. They were all considered to have similar
educational background as they had been enrolled in the given
school through the same nation-wide proficiency exam 3 and
half years before this study was carried out.
Convenience sampling was used in the selection of the
participants and the research site. The experimental
treatments were conducted in an authentic classroom setting
in a public high school. With 10 years of teaching experience,
the researcher had been working as an EFL teacher in this
school for over 5 years, which facilitated the planning and
implementation of the study. He was also the teacher of 3
participating classes for more than 3 years, thereby ensuring
natural group dynamics. All the treatments were carried out
by the same teacher so as to rule out the variations in teaching
procedure. The participants got the vocabulary instruction as
they usually did within their courses.
At first, there had been 47 participants. However, 3 of them
were excluded from the analysis in view of their pre-test
results since they had already known two target words. Thus,
the statistics from 44 participants who got 0 in the pre-test
were included into the analysis.
12 real English words were assigned as target vocabulary items.
6 of them were semantically related whereas the rest 6 were
semantically unrelated words. As seen in Table 2, the target
words in each set were homogenous in terms of their size, type
and length. They were all concrete nouns as a part of speech.
The target items in semantic set also had equal number of
syllables and letters with those in unrelated set. These words
were selected from 5,000-word level and above, which were
not likely to be known by the participants.
The target words in each set were also homogenous in terms of
their frequency bands. First, the reading texts in their nationwide
coursebook (Perşembe, Buluğ, & Eroğlu-Canmetin, 2014)
were entered into Cobbs vocabulary profiler (Cobb, n.d.) so
as to determine the suitable frequency band. The scores
indicated that the students read the texts with the 1,000 and
2,000-word level. Thus, the words from 5,000-word level and
above were regarded to be prospective candidates. Then, the
frequency band of each target word was checked from Nations
vocabulary levels (Appendix 3 in Nation, 2001: 416424).
Lastly, the final list of 12 target words was decided in the light
of the expert opinion from four EFL teachers, two of whom
were working in the given school and the rest two were PhD
students at a university.
The current study made use of one group quasi-experimental
research model in pre-post test design with repeated
measures. To start with, verbal informed consent was received
from all of the participants as well as the school principal. Two
weeks before the treatments, the pre-test was conducted to
measure the participants prior knowledge of the target words.
In one of two treatments, all the participants were taught
each of six target English words in semantic set. In the second
treatment, the other six target words were taught in unrelated
set. Both treatments were exactly the same except for the
target words. In treatments, each target concrete noun was
presented through pictorial flashcards together with sentential
contexts, in which two sample sentences were served to the
participants. The immediate post-tests were administered
in two modalities: one for word recognition and the other
for word production. The delayed post-tests were employed
two weeks later in order to measure the participants delayed
recognition and production of the target items in each set.
The orders of two treatments were also counterbalanced. The
half of the participants were first taught the semantic set of the
target words and then continued with those in unrelated sets.
However, the other half were initially presented the target
items in the unrelated sets and continued with those words in
the semantic set.
Materials and Instruments
The study employed two kinds of materials: (1) instructional
materials to teach the target words, and (2) testing instruments
to collect data. All the materials and instruments were
produced by the researcher and reviewed by four EFL teachers.
Cronbachs alpha analysis was run to measure the internal
reliability of the testing instruments. The reliability co-efficient
was 0.740 for eight items, which revealed an acceptable
internal consistency (α>0.700).
The lexical instruction was given by means of a slide show,
which comprised pictorial flashcards so that the participants
could learn and practise the target words in sentential contexts.
Three types of pictorial flashcards were developed to teach
each target item (see Figure 1). The first type had only the
picture of the related word so as to establish a context to lead
the students to the meaning of that word. Here L2 spelling of
the target word was not available so that the participants can
guess its meaning themselves in the pre-teaching and practice
stages. Used for teaching, the second type of flashcards
included the picture of the target item together with its English
label and part of speech underneath. The last type presented
the learners with the example sentences embedded in the
corresponding visuals so that they could practise each target
word in two different sentential contexts.
Click Here to Zoom
|Figure 1: Sample pictorial flashcards for teaching the target word ladle.
Testing Instruments for Data Collection
Three types of testing instruments were prepared for data
collection: (1) the pre-test, (2) the immediate post-tests, and
(3) the delayed post-tests. The pre-test was conducted to
measure the participants recognition of the target words. It
was in six-option multiple choice format, the learners were
asked to pick out L2 equivalents of target words in the light of
the pictures given as clues above these options (see Figure 2).
I dont know was also added as 7th option to preclude them
from exaggerating their test scores by guessing, as in Nation
and Beglars (2007) vocabulary size test, which is attainable at http://my.vocabularysize.com. In the pre-test, six semanticallyassociated
target words were shuffled with 6 unrelated ones.
The pre-test practice was embedded within another routineclassroom
activity so as to prevent any memory effects on the
The immediate and delayed post-tests were prepared in two
modalities in order to assess both EFL learners recognition
(L2-L1) and production (L1-L2) of 12 target words. The wordrecognition
tests were in six-option multiple-choice format,
and the corresponding pictures were provided as clues above
the options (see Figure 3). In the word-production tests, the
students were given just the related pictures, and they were
required to write (produce) L2 equivalents of the target words
in the blanks below the pictures (see Figure 4).
Totally, four immediate post-tests were employed without
prior notice. Namely, two kinds of immediate-post-tests
(one for the semantic and the other for unrelated set) were
conducted in two modalities (one for recognition and the other
for production). Each post-test included six items. Two weeks
later, these immediate post-tests were administered again as
delayed-post tests. The formats of delayed post-tests were not
completely the same as those of the immediate post-tests. The
order of test items and arrangement of options were different
so as to keep the learners away from recalling the correct
answers from their pictorial memory.
Two weeks before the treatments, the participants prior
knowledge of the target words were measured through a pretest.
They were allowed 12 minutes to complete the pre-test,
but most students completed it much earlier since they did not
know the target words. The participants were given no further
instruction about target words in two-week period up to the
Within the scope of the treatments, one set of six target
words was presented in semantic sets and the other set of
six target items was served in unrelated sets. By means of a
slide show, all the participants were taught target words in
each set through pictorial flashcards. They also practiced the
meaning of each word in two sentential contexts embedded
in the corresponding visuals. The same amount of instruction
was given on both sets of target words by the same teacher
through the same technology and similar teaching materials.
The instructional procedure for each target word was carried
on as follows: Initially, the students were shown the unlabelled
pictorial flashcard to establish a meaningful context for learning
(see Figure 5). Here the teacher tried to elicit the word meaning
from the students. When they properly guessed the meaning of
the target word in L1 (Turkish), the teacher pronounced it three
times: sledge, sledge, and sledge. Next, the students were displayed the labelled pictorial flashcard (see Figure 6).
Seeing the spelling of the word on the flashcard, they repeated
the L2 pronunciation of sledge three times after the teacher.
Then, the students were presented two different sample
sentences accompanied with related visuals (see Figure 7 and
8) to practise the target word in a larger context. Meanwhile,
the teacher read the examples to the students, and checked
their understanding of the sentences.
After all of 6 target items in each set were taught in a similar
fashion, all the words and their corresponding pictures were
shown to the students for the last time (see Figure 9). While
looking at the screen, they pronounced each target word
three times after the teacher once again. Lastly, all the target
vocabulary items in the set were practised as a whole class through a simple activity. Here the students were asked to
guess and produce the target words after they were shown
the unlabelled pictorial flashcards of each item only once. The
vocabulary instruction and practice for each set lasted about
Two kinds of immediate post-tests were conducted after
five-minute distraction activity. First, word production tests
were administered to measure the participants active recall
(productive knowledge) of the target words. Then, their
immediate recognition of these words was measured by
means of another post-test in the multiple-choice format. Five
minutes were given to the students for each immediate posttest,
but they completed both tests in shorter time. The whole
instructional procedure lasted about 40 minutes for one set of the words. The other set was taught and tested in another
successive 40-minute session.
During the treatments, the participants were asked not to
take notes about the target vocabulary, and no assignment
was given. Two weeks later, four post-tests were conducted
again to measure the learners delayed recognition and
production of the target words in each set. They were also
allowed five minutes to complete each delayed post-test. No
prior notice was given about these tests so as to avoid the
regular attempts to study the taught words. In this two-week
period, the participants also did not take any English course
because of the exam-weeks which were arranged by the school
Piloting and Scoring
The pilot study of the treatments and tests were employed
with eight 10th grade students on a voluntary basis. Moreover,
the instructional materials and testing instruments were
revised by four EFL teachers, two of whom were EFL colleagues
of the researcher and the rest two were PhD students at a
university. While scoring the word-production tests, one point
was assigned to each correctly produced target word, and 0.5
point was given for the answers with one spelling mistake. The
responses with more mistakes were not regarded as true.
Data Collection and Analysis
Data were collected in 2015/2016 academic year. The study
made use of the pre-test, immediate post-tests and delayed
post-tests as data collection tools. The pre-test was mostly
employed to exclude the participants who had already known
some of the target words instead of testing their recognition of
these words before the treatments. The statistical data from
three participants were eliminated from the analyses in view
of their pre-test scores.
The data were analysed through a statistical software program.
The quantitative data gathered from the pre-test and posttests
were entered into the software. First, descriptive statistics
of the mean scores and standard deviations were computed
individually for each test. Second, one-way repeated measures
ANOVA analyses were run to investigate the effects of teaching
new L2 words in semantic sets and unrelated sets on the
participants acquisition of these words. So as to determine
the significance levels of differences, pair-wise comparisons
with Bonferroni adjustment were calculated across the mean
scores of pre-tests, post-tests, and delayed post-tests within
each treatment. Finally, the independent samples t-tests were
applied to explore whether teaching words in semantic and
unrelated set differs in terms of EFL learners immediate or
delayed recognition and production of these vocabulary items.
The pre-test results showed that the participants had no
previous knowledge of the target words in both semantic and
unrelated set. Due to elimination of three participants knowing
two of 12 target words, the pre-test mean scores were accepted
to be 0 before each treatment. For this reason, the reliability co-efficient was also not calculated for the pre-test. Briefly, the
pre-experimental measures indicated no statistically significant
differences between two sets of target words in terms of EFL
learners prior lexical knowledge of these words.
RQ 1: What are the effects of teaching new L2 words in
semantic set or semantically unrelated set on EFL learners
recognition and production of these vocabulary items?
One-way ANOVA with repeated measures was employed
to investigate the effects of presenting new L2 words in
semantic and unrelated sets to EFL learners. The results
showed statistically significant differences between the pretest
and post-test mean scores of the participants within each
type of clustering. Therefore, the pair-wise comparisons with
Bonferroni adjustment were computed among the pre-test,
immediate post-tests and delayed post-tests so as to determine
where the significant difference existed.
Semantic Clustering: Serving new L2 words in semantic
sets was found to have statistically significant effect on the
participants immediate and delayed recognition of these
words [F (2000, 42000)=7022.709, p<0.001] (see Table 4).
The pair-wise comparisons verified the significant difference
between the pre-test (M=0, SD=0) and the immediate post-test
mean scores (M=5.93, SD=0.33) as well as between the pre-test
and the delayed post-test means (M=5.14, SD=1.30) at p<0.001
level. Semantic clustering also yielded superior vocabulary gain
scores in word production tests [F (2000, 42000)= 958.375,
p<0.001]. The post hoc tests revealed statistically significant
differences not only between the mean scores of the pre-test
(M=0, SD=0) and immediate post-test (M=5.53, SD=0.83) but
also between those of the pre-test and the delayed post-tests
(M=1.74, SD=1.41) at p<0.001 level. These research findings
indicated that teaching semantically related words together
had a facilitative impact on EFL learners immediate and
delayed recognition and production of these target words.
Click Here to Zoom
|Table 4: The Effect of Semantic Clustering on EFL Learners Vocabulary Learning
Unrelated Clustering: The participating students also got a
high-level of vocabulary recognition scores when L2 target
words were taught in unrelated sets [F(2000,42000)= 2642.870,
p<0.001] (see Table 5). According to the pair-wise comparisons,
there is a statistically significant difference between the
participants pre-test (M=0, SD=0) and the immediate posttest
mean scores (M=5.89, SD=0.54) as well as between their
pre-test and delayed post-test means (M=5.39, SD=1.22) at
p<0.001 level. The statistics also released the significant effect
of unrelated clustering on L2 learners production of target
words [F (2000, 42000)=556.826, p<0.001]. The post hoc
analyses displayed statistically significant difference not only
between the pre-test (M=0, SD=0) and immediate post-test
(M=5.34, SD=1.05) but also between the pre-test and delayed
post-test scores (M=1.99, SD=1.53) p<0.001 level. Hence, it
can be concluded that teaching new L2 words in unrelated sets
equipped EFL learners with superior vocabulary gains in terms
of their recognition and production of these words.
Click Here to Zoom
|Table 5: The Effect of Unrelated Clustering on EFL Learners Vocabulary Learning
Given these findings, semantic and unrelated grouping of
new L2 words were both found to be significantly effective in
providing EFL learners with higher vocabulary recognition and production scores not only immediately after the instructions
but also two weeks later. Thus, the study revealed positive
research evidence about the beneficial effects of presenting
novel L2 words in semantic and unrelated sets on EFL learners
immediate and delayed recognition and recall of these words.
RQ 2. Does teaching new L2 words in semantic set or
unrelated set differ in terms of EFL learners immediate or
delayed recognition and production of these words?
Although teaching new L2 words in semantic set and unrelated
set were both found to have favourable impact on EFL learners
acquisition of these words, independent samples t-tests were
also employed to statistically compare these two kinds of
The immediate recognition and production: Shortly after
the treatments, the participating students recognised about
99% of the semantically-associated target words properly
(M=5.93, SD=0.33) while the figure was 98% for the unrelated
vocabulary items (M=5.89, SD=0.54). As for immediate word
production tests, the mean scores were almost equal for each
type of vocabulary instruction. The participants produced 89% of target words correctly in both semantic (M=5.33, SD=0.83)
and unrelated set (M=5.34, SD=1.05). All in all, both types of
clustering provided the participants with superior vocabulary
gains immediately after treatments. However, independent
samples t-tests revealed no statistically significant difference
between teaching new L2 words in semantic and unrelated
sets in terms of EFL learners immediate recognition [t(86) =
0.476, p=0.635] and immediate production of these target
words [t(86) = -0,056, p=0.955].
To sum up, teaching novel L2 words in semantic and unrelated
sets both gave rise to very high rates of vocabulary gains.
EFL learners achieved equally well through both types of
vocabulary instruction. The achievement rates of 99% and
88% were really favourable results for immediate recognition
and production of target words, respectively. Consequently,
the research findings found no interference effect of both
semantic and unrelated clustering of new L2 words on EFL
learners immediate acquisition of these words.
The delayed recognition and production: The delayed posttest
results demonstrated that the high achievement scores of
the participants in word recognition tests were stable even two weeks later (see Table 7). Although they did not recycle and
review the target words, they were able to recognise about
86% of semantically-related words (M=5.14, SD=1.30) and 90%
of the target items in unrelated set (M=5.39, SD=1.22). On the
other hand, the word production rate was 29% for the words
taught in semantic set (M=1.74, SD=1.41) whereas it was 33%
for the target vocabulary presented in unrelated set (M=1.99,
SD=1.53) two weeks after the treatments. Overall, teaching
new L2 words in unrelated sets led to slightly higher receptive
and productive vocabulary gains than semantic clustering.
However, these research results once again revealed no
statistically significant difference between two methods
of vocabulary instruction in terms of EFL learners delayed
recognition [t(86)= -0,927, p=0.357] and production of the
target vocabulary items [t(86)= -0,796, p=0.428].
In the light of the findings from the delayed post-tests, it can
be concluded that the participants perfectly recognised a great
deal of the target words, on average of 88%, even though they
did not recycle and revise these vocabulary items in two-week
period after the treatments. Naturally, there was a decrease in
their word production gains in each type of clustering after two
weeks interval. However, this amount of decrease was fairly
reasonable without further practice and review of the target
words in the given time. Nevertheless, EFL learners actively
recalled and produced more than one third (31%) of target
lexical items even two weeks later, which was regarded as very
satisfying outcome in relevant literature.
The present study concluded that teaching novel L2 words in
semantic and unrelated sets produced superior vocabulary
gains in terms of EFL learners immediate or delayed recognition
and production of these words. In a real classroom setting, the
real L2 target words were presented to 44 Turkish EFL students
by means of pictorial flashcards. They also experienced several
meaningful encounters with the target items in sentential
contexts embedded in corresponding visual images. Under
such circumstances, both types of clustering equipped
EFL learners with high rates of receptive and productive
vocabulary knowledge both immediately after the treatments
and two weeks later. The results did not show any statistically
significant difference between presenting words in semantic
and unrelated sets. Thus, this study revealed no interference
effects of semantic or unrelated clustering on students
acquisition of the target words in an EFL classroom, especially
when they practised newly-learned words in a broader context
and within a great deal of time.
The study findings add a new dimension to the relevant
literature in relation to the controversial issue of presenting
new L2 vocabulary in semantic or unrelated sets. It proposes
that the way of clustering new words is not as important as
how many words to teach in per class period and how to teach
these lexical items. On the one hand, teachers should set more
realistic goals on the number of new words to be presented to L2 learners. Naturally, the size of vocabulary to be taught
at a time relies on many factors such as the difficulty of the
words, their similarity to L1 and the levels, needs, interests of
the learners. In this regard, Schmitt (2000) suggests serving an
average of 10 novel vocabulary items in a 60-minute lesson.
Gairns and Redman (1986), consider ideal vocabulary load as
eight to twelve productive items per lesson. That is the reason
why the current study has attempted to teach six vocabulary
items at a time. On the other hand, the quality of teaching
and practising new L2 words is also more important than
categorising these words. Obviously, there is no best way of
teaching L2 vocabulary which suits all circumstances. However,
the relevant literature reveals several practical guidelines
which should be born in mind by L2 teachers and researchers.
This study takes a few of guidelines into consideration: using
dual coding, exemplification of the concept the word refers
to and providing a number of encounters with a word. First,
dual coding requires using both linguistic and visual elements
together so as to convey the meaning of target words (see
Clark & Paivio, 1991; Paivio, 1991 for further information about
the dual coding theory). Likewise, the current study makes use
of both pictorial flashcards and verbal linguistic elements to
present the target words to EFL learners. Second, meaningful
examples about the concept of a word also facilitate L2
learners mastery of the given word. According to Nation
(2001: 215), examples help bring a message alive. With this
in mind, this study supplies EFL students with several sample
sentences so that they can easily conceptualise the meanings
of target vocabulary items in their minds. Third, learning a
word involves knowing many different aspects of that word
such as spelling, pronunciation, meaning and use. Therefore,
limited amount of exposure to a word is not sufficient for L2
learners even to acquire only one meaning sense of that word.
They should repeat, recycle and practise the newly-learned
words through various exercises, tasks and activities. Similarly,
this study, to a certain extent, attempts to provide EFL students
with several encounters with the target words in different
In the relevant literature, many studies try to compare the
effects of grouping new L2 words in different clusters on
vocabulary acquisition of L2 learners. However, most of these
studies are conducted in strictly-controlled experimental
conditions. For instance, the study by Wilcox and Medina
(2013) allows the participants only two seconds to learn each
target word. Likewise, Ishii (2015) gives EFL learners totally
45 seconds to memorise 6 new words. Such studies can be
claimed to deal with memorization, rather than vocabulary
learning. Thus, it is not convenient to make generalisations
about vocabulary learning and teaching in real L2 classrooms
in view of the research findings from such studies conducted
in laboratory-like settings. These studies also select artificial
words as the target items rather than using real L2 words, and
they present these words in isolation. However, the present
study anticipates that practising new words in a larger context
and within a great deal of time may provide L2 learners with
better learning or less confusion.
In view of these limitations in the related literature, the current
study attempts to compare two ways of grouping novel L2
words in natural classroom setting, and it offers EFL learners
more opportunities to practice the real target words in various
meaningful contexts which were also enriched with visual
materials. The research findings suggest that both semantic
and unrelated clustering facilitate EFL learners acquisition
of new L2 words provided that they are provided with more
meaningful learning environment. In this respect, a study by
Bolger and Zapata (2011) reveals that adding a story context
to the target words is likely to eliminate the interfering effect
of presenting semantically-associated words in semantic sets.
The present study also concludes that there is no significant
difference between teaching new L2 words in semantic
sets and in unrelated sets, especially when these words are
sufficiently practised through visual materials and meaningful
sample sentences in a real L2 classroom setting.
Implications, Limitations and Suggestions for Further
The current study has some significant implications about L2
vocabulary learning and teaching. First of all, the quality of
teaching and practising new L2 words is more important than
the categorisation of these words. Second, the use of visual
materials plays a fundamental role in almost all stages of L2
vocabulary instruction in that it facilitates learning and retention
of lexical items. Third, apart from the explicit teaching of novel
L2 words in isolation, teachers should supply L2 learners with
opportunity to practise newly-learned vocabulary items in
larger contexts in order that they can easily conceptualise
the meaning of these words in their minds. Fourth, they
should recycle, practise and revise new L2 words in a range of
meaningful contexts via several practical exercises, productive
tasks and activities. Hence, EFL teachers should benefit from
a number of methods and techniques rather than utilising a
single approach in L2 vocabulary instruction.
In the light of the limitations of the present study, some
suggestions can be made for further research. Initially, there
is still a lack of classroom research to compare the effects of
semantic or unrelated clustering on L2 learners acquisition
of new vocabulary items. Such research evidence needs to be
substantiated by similar kinds of studies conducted in real L2
classroom settings. Second, the scope of this study is restricted
in not only the size of the participants but also the number
of target words. Further studies should be conducted with
greater sample size and with different vocabulary items so as to
verify these findings. Third, it would be better to confirm these
research findings with various types of learners at different
levels of language proficiency and at different ages. Fourth,
this study just adds sentence contexts to the target words to
be presented. Further research can focus on the teaching of
the target vocabulary items in broader contexts. Finally, the
target words used in the study were concrete nouns; therefore,
further studies can assign the words from other parts of speech
as target vocabulary items.
1) Aitchison, J. (1994). Words in the mind: An introduction to the
mental lexicon. Oxford: Blackwell Publishers.
2) Baddeley, A. (1990). Human memory. London: Lawrence Erlbaum
3) Bolger, P., & Zapata, G. (2011). Semantic categories and context
in L2 vocabulary learning. Language Learning, 61(2), 614-646.
4) Clark, J. M., & Paivio, A. (1991). Dual coding theory and education.
Educational Psychology Review, 3(3), 149-210.
5) Cobb, T. (n. d.). Range for texts v. 3 [computer program]. Retrieved
6) Erten, İ. H., & Tekin, M. (2008). Effects on vocabulary acquisition
of presenting new words in semantic sets versus semantically
unrelated sets. System, 36, 407-422.
7) Finkbeiner, M., & Nicol, J. (2003). Semantic category effects in
second language word learning. Applied Psycholinguistics, 24,
8) Gairns, R., & Redman, S. (1986). Working with words. Cambridge:
Cambridge University Press.
9) Graves, M. (2006). The vocabulary book: Learning and instruction.
New York: Teachers College Press.
10) Hashemi, M. R., & Gowdasiaei, F. (2005). An attribute-treatment
interaction study: Lexical-set versus semantically-unrelated
vocabulary instruction. RELC Journal, 36, 341361.
11) Haycraft, J. (1993). An introduction to English language teaching.
12) Hoshino, Y. (2010). The categorical facilitation effects on L2
vocabulary learning in a classroom. RELC Journal, 41, 301312.
13) Ishii, T. (2013). Reexamining semantic clustering: Insight from
memory models. Vocabulary Learning and Instruction, 2(1),
14) Ishii, T. (2015) Semantic connection or visual connection:
Investigating the true source of confusion. Language Teaching
Research, 19(6), 712-722.
15) Nation, I. S. P. (2000). Learning vocabulary in lexical sets: Dangers
and guidelines. TESOL Journal, 9(1), 6-10.
16) Nation, I. S. P. (2001). Learning vocabulary in another language.
Cambridge, England: Cambridge University Press.
17) Nation, P., & Beglar, D. (2007). A vocabulary size test. The Language
Teacher, 31(1), 9-13.
18) Paivio, A. (1991). Dual coding theory: Retrospect and current
status. Canadian Journal of Psychology - Revue Canadienne de
Psychologie, 45(3), 255-287.
19) Papathanasiou, E. (2009). An investigation of two ways of
presenting vocabulary. ELT Journal, 63, 313-322.
20) Perşembe, E., Buluğ, N., & Eroğlu-Canmetin, Z. Z. (2014). Yes you
can: Students book A2.3. Ankara: MEB Yayınları.
21) Sarıoğlu, M. (2014). The use of mnemonic devices for minimizing
cross-association in teaching vocabulary to primary school
EFL learners. Unpublished Masters Dissertation. Uludağ
22) Schmitt, N. (2000). Vocabulary in language teaching. Cambridge:
Cambridge University Press.
23) Stahl, S. A., & Nagy, W. E. (2006). Teaching word meanings.
Mahwah, N. J: Erlbaum.
24) Tinkham, T. (1993). The effect of semantic clustering on the
learning of second language vocabulary. System, 21, 371-380.
25) Tinkham, T. (1997). The effects of semantic and thematic clustering
on the learning of second language vocabulary. Second
Language Research, 13, 138-163.
26) Waring, R. (1997). The negative effects of learning words in
semantic sets: A replication. System, 25, 261-274.
27) Wilcox, A., & Medina, A. (2013). Effects of semantic and
phonological clustering on L2 vocabulary acquisition among
novice learners, System, 41, 1056-1069.
28) Wilkins, D. A. (1972). Linguistics and language teaching.
Cambridge: Cambridge University Press.