ArticlePDF Available

Don't just repeat after me: Retrieval practice is better than imitation for foreign vocabulary learning



Second language (L2) instruction programs often ask learners to repeat aloud words spoken by a native speaker. However, recent research on retrieval practice has suggested that imitating native pronunciation might be less effective than drill instruction, wherein the learner is required to produce the L2 words from memory (and given feedback). We contrasted the effectiveness of imitation and retrieval practice drills on learning L2 spoken vocabulary. Learners viewed pictures of objects and heard their names; in the imitation condition, they heard and then repeated aloud each name, whereas in the retrieval practice condition, they tried to produce the name before hearing it. On a final test administered either immediately after training (Exp. 1) or after a 2-day delay (Exp. 2), retrieval practice produced better comprehension of the L2 words, better ability to produce the L2 words, and no loss of pronunciation quality.
Dont just repeat after me: Retrieval practice
is better than imitation for foreign vocabulary learning
Sean H. K. Kang &Tamar H. Gollan &Harold Pashler
#Psychonomic Society, Inc. 2013
Abstract Second language (L2) instruction programs often
ask learners to repeat aloud words spoken by a native
speaker. However, recent research on retrieval practice has
suggested that imitating native pronunciation might be less
effective than drill instruction, wherein the learner is re-
quired to produce the L2 words from memory (and given
feedback). We contrasted the effectiveness of imitation and
retrieval practice drills on learning L2 spoken vocabulary.
Learners viewed pictures of objects and heard their names;
in the imitation condition, they heard and then repeated
aloud each name, whereas in the retrieval practice condition,
they tried to produce the name before hearing it. On a final
test administered either immediately after training (Exp. 1)
or after a 2-day delay (Exp. 2), retrieval practice produced
better comprehension of the L2 words, better ability to
produce the L2 words, and no loss of pronunciation quality.
Keywords Human memory .Human associative learning
Advances in technology have made communication, travel,
immigration, and other forms of personal exchange across
national boundaries ever more feasible and convenient. As a
result, people are increasingly likely to be confronted with
linguistic diversity, and many aspire to be proficient in more
than one language (whether by choice or by necessity). The
proliferation of language classes (e.g., English as a Second
Language courses around the world) and various language-
learning computer software are testaments to the growing
desire by many to acquire a foreign language.
In order to acquire mastery over a second language (L2),
it is critical for learners to traverse rapidly from a concept
that they wish to express and the L2 word(s) appropriate for
expressing this concept (or vice versa). In L2 instruction
programs, this has often been accomplished by drills in
which the learner sees a concept represented either in his
or her first language or in pictorial form, hears a native
speaker say the word or phrase in the L2, and then tries to
imitate. Anyone who has studied a foreign language in the
classroom or with audiotapes or CDs will recall hearing
repeat after mein these drills. Such modeling (by the
instructor) and repetition (by the learner) has long been a
prominent feature of L2 teaching practices, in traditional
classrooms as well as in self-study with audio recordings
(Macdonald, Yule, & Powers, 1994;OMalley, Chamot,
Stewner-Manzanares, Kupper, & Russo, 1985). In more
recent L2 instruction software (e.g., Berlitz Guaranteed,
Pimsleur Method, Linguaphone), this kind of drilling is still
often used.
Contemporary language instructors and software de-
signers have attempted to improve instruction by making it
more similar to natural language acquisition. For example,
the popular Rosetta Stone software provides exercises in
which people see images of several objects and hear a
spoken L2 word or phrase and then are asked to learn via
guessing to match pictures to words. The need to learn via
guessing is meant to resemble how children learn new
words. In this spirit, because children generally do not learn
language by translating to another language, the programs
for English users go to great lengths to avoid using any
English whatsoever during training. Perhaps viewing imita-
tion as the gold standard in L2 instruction, the Rosetta Stone
software also includes a CD that instructs learners to pro-
duce L2 words by imitation in additional training sessions.
This imitation is done without any reminder or guidance
about what the words mean as the learner repeats the utter-
ances (to avoid any use of English).
S. H. K. Kang :H. Pashler
Department of Psychology, University of California, San Diego,
La Jolla, CA, USA
S. H. K. Kang (*)
Department of Education, Dartmouth College, Hanover, NH
03755, USA
T. H. Gollan
Department of Psychiatry, University of California, San Diego, La
Jolla, CA, USA
Psychon Bull Rev
DOI 10.3758/s13423-013-0450-z
Relatively little research has contrasted the specific L2
instruction techniques that are commonly in use (although
see Carpenter & Olson, 2012), and given the ubiquity of
imitation in L2 instruction, it is important to investigate its
effectiveness as a pedagogical technique. Some evidence has
suggested that imitation plays a role in vocabulary learning.
For instance, spontaneous imitation by infants of words spo-
ken by caregivers has been found to predict later vocabulary
growth (Masur, 1995). Also, an early study on foreign vocab-
ulary learning showed that saying aloud the L2 vocabulary
produced better learning than did studying the items silently
(Seibert, 1927). More recently, Ellis and Beaton (1993)found
that for English-speaking college subjects learning German
vocabulary, repeated imitation (saying aloud) of the German
words produced better performance on a later production test
(i.e., giving the German equivalent when cued with the En-
glish word) than did learning using the keyword mnemonic,
which involved visual imagery to associate the English and
German equivalents, but no overt production. These findings
are consistent with the broader idea that the phonological loop
has a critical function in supporting the learning of novel
phonological forms, and hence new vocabulary (Baddeley,
Gathercole, & Papagno, 1998).
Retrieval practice
The use of imitation, in which learners model their pronun-
ciation on a native speakers, might seem particularly suited
for promoting native-like pronunciation. The ability to pro-
duce accurate pronunciations, however, is by itself insuffi-
cient for speaking in a foreign language; one needs to
associate the L2 words with the concepts that they express.
Many different kinds of evidence have demonstrated that
people learn more robust associations when they are re-
quired to actively retrieve the association, rather than having
both of the associated elements presented together (for a
review, see Roediger & Karpicke, 2006). The benefit of
retrieval practice (often referred to as the testing effect) has
been shown in studies using a wide variety of instructional
materials (e.g., prose, word lists, or paired associates). Most
relevant for the issue of foreign language vocabulary train-
ing are studies that have compared retrieval practice with
rereading EnglishL2 word pairs. For example, Carrier and
Pashler (1992) found that producing the English word given
an Eskimo (Yupik) cue strengthened the association be-
tween the two words more powerfully than did rereading
the words paired together (see also Karpicke & Roediger,
2008). In addition, Kang (2010) observed a retrieval prac-
tice advantage when the criterial task was to recall (write
out) the Chinese logographs (characters) that had been
paired with English translations during an initial study
phase. These studies suggest that imitation might not be
the most efficient way to learn the association between a
concept and a word. However, these studies did not require
people to pronounce an unfamiliar word. It seems entirely
possible that imitation might be useful, and perhaps even
optimal, when this difficult form of response learning is
required, as it is in L2 learning. Moreover, these studies
did not address the possible benefits of imitationfor in-
stance, imitation may lead to better pronunciation than do
other forms of practice.
Present study
In two experiments, we compared the relative efficacy of
retrieval practice versus imitation for learning L2 words.
English-speaking learners were trained on the Hebrew names
for40commonobjects.Intheimitation condition, the training
consisted of hearing the Hebrew name and then repeating it
aloud immediately afterward while viewing a picture of the
corresponding object. In the retrieval practic e condition, the
picture was used as a retrieval cue, and learners attempted to
retrieve the Hebrew name before the correct pronunciation
was provided. Both training conditions, it should be noted,
required activity on the part of the learner, and neither was
identical to the passive viewing/rereading control conditions
that are commonly used in studies of retrieval practice. As
compared with silent reading, overt production has been
found to improve retention via enhanced distinctiveness of
the produced items (e.g., Gathercole & Conway, 1988;Mac-
Leod, Gopie, Hourihan, Neary, & Ozubko, 2010). For the
final test, we used two different kinds of assessment. The first
involved selecting the appropriate picture (referent) when
subjects were presented with the Hebrew word auditorily.
The second required production of the Hebrew name when
subjects were given a picture of the object. The main differ-
ences between the two experiments lay in the amount of
training andin the delay before the final test was administered:
In Experiment 1, the final test occurred immediately after
training; in Experiment 2, training was more extensive, and
the final test was given 2 days later. Experiment 2 served two
purposes: as a replication of Experiment 1 and for demon-
strating that any difference in learning efficacy between the
training conditions would persist over a longer delay.
Two groups of undergraduates from the University of Cal-
ifornia, San Diego, Psychology Subject Pool participated in
return for course credit: 41 took part in Experiment 1, and
59 in Experiment 2. All of the subjects were fluent speakers
Psychon Bull Rev
of English, and none had any significant prior exposure to
Forty Hebrew nouns served as the to-be-learned foreign
vocabulary. Following common practice in foreign language
instruction, the words were trained in four lists of themati-
cally related items. Each list contained ten nouns from one
of the following four semantic categories: body parts (e.g.,
ear,hand), eating/food (e.g., fork,bread), animals (e.g.,
dog,elephant), and household objects (e.g., clock,chair).
Across the categories, the Hebrew words were equated in
terms of average numbers of syllables (M=1.95, SD =
0.55) and phonemes (M=4.65, SD =1.10).Anaudio
recording of each word was made (spoken by a native
Hebrew speaker), and these were presented in the learning
and test phases as described below (learners were never
presented the written forms of the words, whether in trans-
literated or Hebrew script). On average, the audio recordings
were about 1 s in duration. Four different photographs
depicting the referent of each noun (i.e., 40 sets of four
photographs) were selected from the Internet; a random
three from each set were presented during training, whereas
the fourth was used in the final tests. All photographs were
standardized to a size of 400 × 400 pixels.
The Hebrew nouns were learned in one of two training
conditionsretrieval practice or imitationthat were ma-
nipulated within subjects across separate blocks and seman-
tic categories. The subjectsdegrees of learning were
assessed after training in two ways: (1) comprehension, in
which learners heard each Hebrew word and had to choose
the corresponding referent from an array of 40 pictures
(representing all of the words learned during training), and
(2) production, in which learners were cued with pictures
and had to say aloud the corresponding Hebrew words.
Learners were tested individually in sound-attenuated rooms.
They were seated in front of a computer and wore headphones
equipped with a microphone. Learners were informed at the
start that they would be learning 40 Hebrew words, divided
into sets of ten according to object category. They were also
informed about the types of tests that they would be given at
the end.
During the training phase, subjects were presented with
Hebrew words to learn in four separate blocks, with each
block featuring ten words from one of the four semantic
categories. Each block began with learners hearing each
Hebrew word once, and while each word was played over
the earphones, learners saw the appropriate picture (referent)
on the computer screen. After the initial presentation, train-
ing cycled three times (Exp. 1) or six times (Exp. 2) through
all ten words, with the order of items being randomized for
each cycle (with the constraint that the first item for each
cycle could not be the last item on the preceding cycle). In
both training conditions, learners were presented with pic-
tures and audio recordings of the corresponding Hebrew
words, with the aim of learning the association between
the meaning and pronunciation of each word.
In Experiment 1, the order of the semantic categories
across the four blocks was held constant, while the assign-
ment of blocks to training conditions was counterbalanced
across learners (either the first two blocks/semantic catego-
ries were assigned to retrieval practice and the last two
assigned to imitation, or the reverse). In Experiment 2, the
assignment of blocks to training conditions was identical to
that in Experiment 1, but the order of semantic categories
was counterbalanced using a Latin-square design.
A schematic of the sequence of events in the two training
conditions is presented in Fig. 1. For the imitation condition,
pictures and audio recordings of the corresponding Hebrew
words were both presented, and subjects were asked to
imitate what they heard. On each trial, the picture was
displayed on the computer screen for 4 s, with the auditory
presentation of the Hebrew word beginning at the same
time. A 2-s blank interval was inserted between trials. Sub-
jects were instructed to repeat aloud the pronunciation as
closely as possible immediately after hearing the Hebrew
word, and at the same time to focus on connecting the word
to its meaning. For the retrieval practice condition, the
presentation of the audio recording of the Hebrew word
did not begin until about 3 s after the presentation of the
corresponding picture, and learners were asked to try to
retrieve and produce the target Hebrew name during the
lag. Each trial consisted of the presentation of a picture
(on the computer screen) for 4 s, and the timing of the
presentation of the audio recording of the Hebrew word
was such that the end of the audio recording coincided with
the offset of the picture. A 2-s blank interval was inserted
between trials. Learners were instructed to attempt the pro-
nunciation of the corresponding Hebrew word when they
saw each picture, before the audio recording of the word
The majority of our subjects reported using at home either a non-
English language or a non-English language in combination with
English (only four and six subjects in Exps. 1 and 2, respectively,
reported having an English-only home environment). To examine
whether subjectslanguage backgrounds modulated the results, we
performed additional analyses in which three language background
variables (age of exposure to English, primary language used at home,
and speaking proficiency in ones non-English language) were includ-
ed as factors in the ANOVAs. None of these factors interacted signif-
icantly with the main factor Training Condition.
Psychon Bull Rev
came online. They were encouraged to guess if they could,
but if they could not, they were to wait for the audio
recording, and then imitate it immediately after hearing it.
After the training phase, learners received two final tests,
both of which were self-paced. In the test of comprehension,
learners heard the 40 Hebrew words one at a time, and for each
of the words they had to pick the picture that corresponded to
the meaning of the word. Displayed on the computer screen
were thumbnails of the 40 pictures that had not been presented
during training (each corresponding to one of the Hebrew
words that had been trained with different pictures), and
learners responded by clicking on one of the thumbnails. In
the production test, subjects were presented one at a time with
the same 40 pictures that were used in the comprehension test,
and they tried to produce the corresponding Hebrew word.
They were encouraged to guess if they were not sure. Spoken
responses were recorded by a microphone. The order of items
in each test was randomized for each learner. In Experiment 1,
the comprehension test was given first, followed by the pro-
duction test. In Experiment 2, the order of both tests was
counterbalanced across learners,
and a 48-h delay was intro-
duced before the tests were administered. Upon completion of
both tests, learners were debriefed and thanked.
Performance was analyzed separately on the comprehension
and production tests. The αlevel for all analyses was set at
Comprehension (accessing semantics when
given phonology)
Learnersability to select the correct picture/meaning (out of
40 options) when presented with the Hebrew words auditorily
was compared across the two training conditions. The retriev-
al practice condition yielded better performance than did the
imitation condition in both experiments, as is shown in Fig. 2.
In Experiment 1, this advantage (.63 vs. .57) was marginally
significant, t(40) = 1.96, p=.057,d= 0.31. In Experiment 2,
the advantage (.57 vs. .44) was significant, t(58) = 3.99, p<
.001, d=0.52.
Production (retrieving phonology when given semantics)
A native Hebrew speaker scored the spoken responses made
by learners in their attempts to pronounce the corresponding
Hebrew words when presented with the pictures. Since our
learners had no preexperimental exposure to Hebrew and were
not expected to be able to properly articulate (or perceive) the
whole range of Hebrew phonemes (several of which are not
present in the English language), we instructed our Hebrew
rater to be lenient in the scoring. She listened to the responses
blind as to condition and was asked to score them as correct if
the pronunciation was good enough that a native Hebrew
speaker would probably be able to understand the word,even
if the pronunciation deviated from what is regarded as an ideal
Hebrew pronunciation. For responses that were scored as
correct, the rater made an additional judgment of pronuncia-
tion quality on a 10-point scale (10 = perfect).
Final test of production Retrieval practice training yielded
more correct productions than did imitation training, as is
shown in Fig. 2. In Experiment 1, the advantage of retrieval
practice over imitation was significant (.40 vs. .27), t(40) =
3.95, p< .001, d= 0.62. For the responses scored as correct,
the pronunciation quality ratings were almost exactly the
same (and not significantly different) for the retrieval prac-
tice and imitation conditions (5.6 and 5.4, respectively), t<
1. Likewise, in Experiment 2, retrieval practice produced
better performance than did imitation training (.34 vs. .19),
t(58) = 6.53, p< .001, d= 0.85, and the pronunciation
quality ratings for correct responses were again not signifi-
cantly different across the retrieval practice and imitation
conditions (8.2 and 8.0, respectively), t(48) = 1.35, p= .184.
Performance during training Although it was not the main
focus of the present study, we also report production perfor-
mance during the final cycle of training (i.e., 3rd cycle of
practice for Exp. 1, and 6th cycle for Exp. 2). As one would
expect, given that in imitation but not in retrieval practice
learners had just heard a native speaker produce the target
word, performance was better in the imitation than in the
retrieval practice condition. In Experiment 1, this difference
in proportions correct (.90 vs. .35) was significant, t(40) =
20.92, p< .001, d= 3.27, and for the responses judged as
being correct, the pronunciation quality was also higher in
the imitation than in the retrieval practice condition (6.5 vs.
5.7), t(40) = 3.70, p= .001, d= 0.58. Similarly, in Exper-
iment 2, the difference in proportions correct (.90 vs. .55)
was reliable,
t(49) = 10.51, p< .001, d= 1.49, and the
pronunciation quality of correct responses was higher in the
imitation than in the retrieval practice condition (8.2 vs.
7.6), t(49) = 4.05, p< .001, d= 0.57.
The order of types of test did not have any effect, and therefore is not
discussed further.
The final production test data from a random 28 of the subjects (13
and 15 subjects from Exps. 1 and 2, respectively) were scored for
accuracy by a second native Hebrew speaker (who was blind as to
condition), and the interrater agreement was high (Cohensκ= .82).
Due to equipment malfunction, some of the spoken responses during
training for nine subjects were lost, and hence data from those subjects
were excluded from the analysis.
Psychon Bull Rev
The experiments reported here contrasted two different
ways of learning L2 vocabulary. The first was the retrieval
practice procedure, modeled after traditional flashcard-type
drills, which required the learner to produce (say aloud) the
L2 word when cued by the concept (presented in the form of
a picture), followed by presentation of the correct answer.
The second was the imitation procedure modeled after prac-
tice drills commonly found in language instruction pro-
grams; here, the learner viewed a picture of the referent of
the word and repeated aloud a native speakers pronuncia-
tion of the word. In two experiments, the retrieval practice
procedure proved robustly superior to the imitation procedure.
Although prior research has demonstrated the advantage
of retrieval practice in learning L2 vocabulary, those studies
almost exclusively used a (passive) rereading control con-
dition and assessed learnersability to recall the English
translations (presumably due to the ease of scoring English
responses). Our study is the first to have examined the
effects of retrieval practice on learning of L2 phonological
word forms. Focusing on learning of these unfamiliar sound
sequences (as opposed to learning familiar English responses)
provides a powerful demonstration of the versatility of retriev-
al practice and has significant theoretical and practical impli-
cations (discussed below). Furthermore, one could reasonably
expect when learning difficult, unfamiliar responses that imi-
tation practice would be advantageous, especially in the early
(a) Imitation
4 seconds
2 seconds
Show picture
Play L2
Learner repeats aloud L2 word
(b) Retrieval Practice
4 seconds
2 seconds
Show picture
Play L2
Learner tries to produce L2
4 s – duration of L2 soundfile
Fig. 1 Sequences of events in a
single training trial of the (a)
imitation and (b) retrieval
practice conditions
Fig. 2 Final test performance
as a function of training
condition and test type. Error
bars indicate standard errors of
the means
Psychon Bull Rev
stages of learning (i.e., a potential boundary condition for the
testing effect), or that any benefits of retrieval practice would
come at the expense of poorer pronunciation quality. Our
findings, however, disconfirm these suggestions.
Theoretical implications
One potential explanation for the observed superiority of
retrieval practice over imitation is that the former engages
the same sorts of mental operations that are required on the
final production test (in both cases, the learner is cued with a
picture and has to retrieve and produce the appropriate L2
word). This line of reasoning is consistent with the transfer-
appropriate processing framework, which proposes that per-
formance is optimized when the processes required at test
overlap with those recruited during learning (Morris,
Bransford, & Franks, 1977). It should be noted, however,
that the advantage of retrieval practice was not limited to the
criterial test that most resembled the retrieval practice train-
ing procedure. The benefits extended to a reception test that
assessed learnerslistening comprehension of the L2 words.
The fact that retrieval practice can enhance learning even
when the criterial test differs from the testing procedure used
in training has been observed in past studies of learning
from prose passages (McDaniel, Anderson, Derbish, &
Morrisette, 2007), maps (Rohrer, Taylor, & Sholar, 2010),
and paired-associate learning (Carpenter, Pashler, & Vul,
The present results are also consistent with the neural-
network model of test-enhanced learning proposed by
Mozer, Howe, and Pashler (2004). According to this model,
learning entails a comparison between a desired output and
the actual output, upon which the connections between input
and output units are adjusted so as to reduce the discrepancy
between the desired and actual outputs. When the cue and
target are presented together (imitation condition), the error
correction mechanism is short-circuited, reducing the effi-
ciency of learning. But when the network is allowed to
produce a response to a cue and then receives feedback
(retrieval practice condition), error correction is facilitated
and the learning system reaches the desired state more
Some recent accounts of the testing effect have empha-
sized the role of mediators in promoting later retrieval of the
target. For instance, when one is presented with a cue (e.g.,
donor?) and attempts retrieval, there is a greater tendency
for information related to the cue to become activated (e.g.,
blood) than when one merely rereads the cuetarget pair
(donorheart), and the activated information then serves as
an effective mediator for subsequent retrieval (Carpenter,
2011). When subjects are explicitly instructed to generate
mediators (between cues and targets), evidence also shows
that retrieval failures during practice encourage a shift to
more effective mediators (Pyc & Rawson, 2012), thus im-
proving later retrieval. However, these studies involved
target responses that were English words; it is unclear how
mediators could be generated (whether spontaneously or
deliberately) to support the retrieval or production of unfa-
miliar Hebrew words.
Practical considerations
We found numerically larger effect sizes in the criterial tests
of production (cf. comprehension). Given that productive
abilities in language learning generally lag behind (or are
more difficult/slower to acquire than) receptive abilities, it is
noteworthy that retrieval practice produced especially robust
gains in the more difficult task (i.e., from meaning to L2
phonology; e.g., Kroll & Stewart, 1994; Schneider, Healy,
& Bourne, 2002). When considering various procedures for
use in training, presumably an ideal candidate would be one
that enhances learning in more challenging (or harder-to-
learn) domains.
Why, then, is imitation so popular a learning strategy
in L2 instruction programs? One possibility is that de-
signers of training programs often assume that condi-
tions that optimize performance during training are the
conditions that best promote the goals of trainingthat is,
long-term posttraining performance (Bjork, 1994). Indeed, if
performance during training were the determining criterion,
imitation would seem to be an ideal training procedure, as
evidenced by the near-ceiling production accuracy and quality
(during imitation training) found in our experiments. Howev-
er, procedures that yield quick acquisition and high perfor-
mance during training often produce illusions of competence,
and may not support durable learning and performance in the
long term. A more effective way to encourage long-lasting
learning is to incorporate desirable difficulties during training
(Bjork, 1999), of which retrieval practice is one example.
Of course, our results do not imply that imitation
practice is a completely futile strategy for L2 learning
(see Ellis & Beaton, 1993), nor that L2 instruction
programs rely on imitation as their sole pedagogical
strategy (in fact, many programs include some time
spent on some form of testing/retrieval practice). How-
ever, given that the present findings unambiguously
demonstrate a more effective method of training than
imitationone that requires only a subtle modification
in training procedure, without any increase in training
timewe contend that for L2 vocabulary learning, im-
itation should not be automatically assumed to be the
tried and true instructional procedure. On the basis of
our data, it would appear that time spent imitating a
native speakers utterance could be much better spent
engaging in retrieval practice with corrective feedback
on pronunciation.
Psychon Bull Rev
Author note S.H.K.K. is now at the Department of Education,
Dartmouth College. This work was supported by the Institute of
Education Sciences (US Department of Education, Grant No.
R305B070537 to H.P.), the National Science Foundation (Grant No.
BCS-0720375, H.P., PI; and Grant No. SBE-0542013, G. W. Cottrell,
PI), a collaborative activity award from the J. S. McDonnell Founda-
tion, and the National Institutes of Health (Grant Nos. NIDCD 011492
and NICHD 050287, T.H.G., PI). We acknowledge the contributions of
the following individuals: Jeff Estacio programmed the experiment;
Noriko Coburn assisted with data collection; and Shira Cabir, Efrat
Golan, and Ronit Snyder scored our subjectsoral pronunciations.
Baddeley, A., Gathercole, S., & Papagno, C. (1998). The phonological
loop as a language learning device. Psychological Review, 105,
158173. doi:10.1037/0033-295X.105.1.158
Bjork, R. A. (1994). Memory and metamemory considerations in the
training of human beings. In J. Metcalfe & A. Shimamura (Eds.),
Metacognition: Knowing about knowing (pp. 185205). Cam-
bridge, MA: MIT Press.
Bjork, R. A. (1999). Assessing our own competence: Heuristics
and illusions. In D. Gopher & A. Koriat (Eds.), Attention
and performance XVII: Cognitive regulation of performance:
Interaction of theory and application (pp. 435459). Cam-
bridge, MA: MIT Press.
Carpenter, S. K. (2011). Semantic information activated during retriev-
al contributes to later retention: Support for the mediator effec-
tiveness hypothesis of the testing effect. Journal of Experimental
Psychology: Learning, Memory, and Cognition, 37, 15471552.
Carpenter, S. K., & Olson, K. M. (2012). Are pictures good for learning
new vocabulary in a foreign language? Only if you think they are
not. Journal of Experimental Psychology: Learning, Memory, and
Cognition, 38, 92101. doi:10.1037/a0024828
Carpenter, S. K., Pashler, H., & Vul, E. (2006). What types of learning
are enhanced by a cued recall test? Psychonomic Bulletin and
Review, 13, 826830. doi:10.3758/BF03194004
Carrier, M., & Pashler, H. (1992). The influence of retrieval on reten-
tion. Memory and Cognition, 20, 633642. doi:10.3758/
Ellis, N., & Beaton, A. (1993). Factors affecting the learning of foreign
language vocabulary: Imagery keyword mediators and phonolog-
ical short-term memory. Quarterly Journal of Experimental Psy-
chology, 46A, 533558. doi:10.1080/14640749308401062
Gathercole, S. E., & Conway, M. A. (1988). Exploring long-term
modality effects: Vocalization leads to best retention. Memory
and Cognition, 16, 110119. doi:10.3758/BF03213478
Kang, S. H. K. (2010). Enhancing visuospatial learning: The benefit of
retrieval practice. Memory and Cognition, 38, 10091017.
Karpicke, J. D., & Roediger, H. L., III. (2008). The critical importance
of retrieval for learning. Science, 319, 966968. doi:10.1126/
Kroll, J. F., & Stewart, E. (1994). Category interference in trans-
lation and picture naming: Evidence for asymmetric connec-
tion between bilingual memory representations. Journal of
Memory and Language, 33, 149174. doi:10.1006/
Macdonald, D., Yule, G., & Powers, M. (1994). Attempts to improve
English L2 pronunciation: The variable effects of different types
of instruction. Language Learning, 44, 75100. doi:10.1111/
MacLeod, C. M., Gopie, N., Hourihan, K. L., Neary, K. R., & Ozubko,
J. D. (2010). The production effect: Delineation of a phenomenon.
Journal of Experimental Psychology: Learning, Memory, and
Cognition, 36, 671685. doi:10.1037/a0018785
Masur, E. F. (1995). Infantsearly verbal imitation and their later
lexical development. Merrill-Palmer Quarterly, 41, 286306.
McDaniel, M. A., Anderson, J. L., Derbish, M. H., & Morrisette, N.
(2007). Testing the testing effect in the classroom. European
Journal of Cognitive Psychology, 19, 494513. doi:10.1080/
Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of
processing versus transfer appropriate processing. Journal of Ver-
bal Learning and Verbal Behavior, 16, 519533. doi:10.1016/
Mozer, M. C., Howe, M., & Pashler, H. (2004). Using testing to
enhance learning: A comparison of two hypotheses. In K. Forbus,
D. Gentner, & T. Regier (Eds.), Proceedings of the Twenty-Sixth
Annual Conference of the Cognitive Science Society (pp. 975
980). Mahwah, NJ: Erlbaum.
OMalley, J. M., Chamot, A. U., StewnerManzanares, G., Kupper, L.,
& Russo, R. P. (1985). Learning strategies used by beginning and
intermediate ESL students. Language Learning, 35, 2146.
practice beneficial for memory? An evaluation of the media-
tor shift hypothesis. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 38, 737746. doi:10.1037/
Roediger, H. L., III, & Karpicke, J. D. (2006). The power of testing
memory: Basic research and implications for educational practice.
Perspectives on Psychological Science, 1, 181210. doi:10.1111/
Rohrer, D., Taylor, K., & Sholar, B. (2010). Tests enhance the transfer
of learning. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 36, 233239. doi:10.1037/a0017678
Schneider, V. I., Healy, A. F., & Bourne, L. E., Jr. (2002). What is
learned under difficult conditions is hard to forget: Contextual
interference effects in foreign vocabulary acquisition, retention,
and transfer. Journal of Memory and Language, 46, 419440.
Seibert, L. C. (1927). An experiment in learning French vocabulary.
Journal of Educational Psychology, 18, 294309.
Psychon Bull Rev
... Exercises have also frequently been provided in app-assisted vocabulary learning. Having been extensively investigated in previous research on vocabulary education (Goossens et al., 2014;Kang, Gollan, & Pashler, 2013), exercises are reported to be overall more effective in enhancing students' knowledge than restudying of the target knowledge or no re-exposure to the knowledge, termed "the testing effect" (Kang, Gollan, & Pashler, 2013). This study identified that all apps which provided exercises also presented learners with feedback on their performance, possibly reducing the risk of students' mistaking the wrong information as correct and enhancing vocabulary learning efficiency (Butler, Karpicke, & Roediger III, 2008). ...
... Exercises have also frequently been provided in app-assisted vocabulary learning. Having been extensively investigated in previous research on vocabulary education (Goossens et al., 2014;Kang, Gollan, & Pashler, 2013), exercises are reported to be overall more effective in enhancing students' knowledge than restudying of the target knowledge or no re-exposure to the knowledge, termed "the testing effect" (Kang, Gollan, & Pashler, 2013). This study identified that all apps which provided exercises also presented learners with feedback on their performance, possibly reducing the risk of students' mistaking the wrong information as correct and enhancing vocabulary learning efficiency (Butler, Karpicke, & Roediger III, 2008). ...
... The majority of vocabulary learning apps applied interactivity/feedback to enhance students' efficiency in exercises. Despite the overall positive effects of the exercise (Goossens et al., 2014;Kang, Gollan, & Pashler, 2013), students might not achieve satisfying learning outcomes using this type of material if they are not given any performance feedback (Roediger & Butler, 2011). Students may fail to retrieve the correct information from their memory, mistake the wrong information as correct, and end the test with learning errors if they receive no feedback about their problems, termed "the negative suggestion effect" (Butler, Karpicke, & Roediger III, 2008). ...
In recent years, mobile applications (apps) have been increasingly used and investigated as a vocabulary learning approach. Despite the extensive use of commercial English as a Foreign Language (EFL) vocabulary learning apps in China, there is a lack of a review of these apps for a systematic Knowledge Management & E-Learning, 13(3), 250-272 251 understanding of the components and usefulness of app-assisted vocabulary learning. To fill this knowledge gap, this study presents a systematic review of 15 EFL vocabulary learning apps that were most downloaded in China, focusing on how these apps help students develop word knowledge. The results of this study showed that most apps enabled students to access word knowledge through translating words into their native language. Notably, word knowledge was usually presented through text-plus-image and text-plus-image-plus-audio. Most of these mobile apps provided sentence examples as vocabulary learning materials. Many of these apps were integrated with game elements, especially in interactivity or feedback systems and reward systems. Based on the review results, we have provided three recommendations to vocabulary learning app developers concerning the use of video for the input of word knowledge, the efficiency of vocabulary learning, and the integration of more game elements.
... Age of acquisition (e.g., Hartshorne et al., 2018;Johnson & Newport, 1989) and linguistic context (Bice & Kroll, 2019;Linck et al., 2009) influence the relative amount of exposure to the L1 and L2 (see Gollan et al., 2008 andWhitford & for a discussion of the weaker links hypothesis). Intuitively, more opportunities to practice retrieving the L2 have been linked to higher L2 proficiency outcomes in both L2 immersion (Serrano et al., 2011;) and classroom environments (Barcroft, 2007;Kang et al., 2013). L1-L2 distance contributes to the relative difficulty of L2 acquisition. ...
... Younger age of L2 acquisition, higher L2 exposure and use, and L2 immersion improved performance on all measures of L2 proficiency. These results are consistent with previous reports that early (Hartshorne et al., 2018;Johnson & Newport, 1989) and frequent (Serrano et al., 2011;) L2 exposure leads to higher L2 proficiency outcomes by providing learners more opportunities to practice retrieving the L2 (Barcroft, 2007;Kang et al., 2013). Higher L2 exposure also predicted higher L1 semantic fluency and faster L1 word reading, consistent with the idea that L2 use may boost L1 activation (Higby et al., 2020) as a consequence of L1 mediation (Kroll & Stewart, 1994). ...
... We report that more L2 exposure and use improves L2 fluency and may train cognitive control mechanisms. The pedagogical implication of these findings is that learners may benefit from opportunities for retrieval practice (Barcroft, 2007;Kang et al., 2013). Whether naming objects, describing pictures, telling stories, or engaging in conversation, more opportunities for L2 retrieval help learners link the L2 word form directly to the concept, gradually moving away from L1 mediation (Kroll & Stewart, 1994). ...
We evaluated external and internal sources of variation in second language (L2) and native language (L1) proficiency among college students. One hundred and twelve native-English L2 learners completed measures of L1 and L2 speaking proficiency, working memory and cognitive control and provided self-ratings of language exposure and use. When considering learner-external variation, we found that more frequent L2 exposure predicted higher L2 and L1 proficiency, while earlier L2 exposure predicted higher L2 proficiency, but poorer L1 maintenance. L1-L2 distance limited cross-linguistic transfer of print-to-sound mappings. When considering learner-internal variation, we found that L1 and L2 proficiency were highly correlated and that better working memory, but not cognitive control accounted for additional variance in L2 and L1 proficiency. More frequent L2 exposure was associated with better cognitive control.
... Second and subsequent learning cyclesretrieval attempt In these cycles, participants were presented with the Hebrew words and were requested to attempt to produce their Arabic translations. Retrieval-based learning (Karpicke, 2012) was incorporated because practice testing has been shown to be a more efficient learning strategy in comparison to rehearsal or imitation (Kang, Gollan & Pashler, 2013;Rice & Tokowicz, 2020;Tokowicz & Degani, 2015). On each trial, following the 1000 ms fixation cross and the 500 ms blank screen, participants were visually presented with a Hebrew word for 1000 ms, followed by a question mark that appeared on the screen until the participant's vocal response triggered the voice key. ...
... On each trial, following the 1000 ms fixation cross and the 500 ms blank screen, participants were visually presented with a Hebrew word for 1000 ms, followed by a question mark that appeared on the screen until the participant's vocal response triggered the voice key. Following this retrieval attempt, a 1000 ms blank screen appeared followed by aural presentation of the correct Arabic translation, regardless of the participant's answer (Kang et al., 2013), to keep the number of presentations of the Arabic word equal for all participants. An additional 1500 ms blank screen then appeared before the next trial. ...
The study examined whether false-cognates, overlapping in form but not meaning across languages, are easier to learn due to form overlap, or more difficult to learn due to meaning competition, compared to unambiguous control and cognate words. Fifty-four native Hebrew speakers learned 14 cognates, 14 false-cognates, and 28 control Arabic words in one session. Cognates were learned better than control items. There was no overall difference in learning false-cognates relative to controls, but individuals with higher phonological short-term memory, or with lower L1 verbal fluency, did exhibit a false-cognate learning-advantage. For these individuals, form overlap was more influential than meaning competition. Lexical decisions to Hebrew words following Arabic learning were slower for false-cognates than controls, indicative of backward influences. The findings reveal the influence of prior knowledge on learning and processing, and highlight the importance of jointly considering item-based and learner-based characteristics during the initial stages of vocabulary learning.
... In contrast, little attention has been directed toward the productive knowledge of spoken form or the knowledge of word pronunciation. A few studies elicited the spoken form of L2 words (via L1-to-L2 translation or picture naming tasks) and measured word pronunciation accuracy by counting the number of mispronounced phonemes ( Barcroft & Sommers, 2005 ), identifying misplacement of word stress ( Bürki, 2010 ), or deriving a listener's judgement of how nativelike the pronounced words sounded ( Kang et al., 2013 ). What appears missing in the L2 vocabulary literature is an in-depth discussion and justification for the choice of pronunciation measures from a listener's perspective. ...
Full-text available
The current study aims to explore the validity of measuring comprehensibility versus accentedness of L2 words as a construct of word pronunciation knowledge. Two research questions were addressed by investigating (a) the interrelationships among four listener-based measures (comprehensibility, accentedness, intelligibility, processing time) and (b) the relative contribution of linguistic features of L2 speech (segmental, word stress, rhythm, fluency) to comprehensibility and accentedness ratings. Nineteen native speakers of English rated L1 Japanese speakers’ productions of 37 English words elicited through a picture naming task for comprehensibility and accentedness. Two expert raters were recruited to complete a timed dictation task from which measures of intelligibility (orthographic transcription of L2 words) and processing time (how fast raters can initiate word transcription) were derived. The analyses of rating responses and relationships among listener-based measures showed that the current results were consistent with previous L2 speech studies measuring comprehensibility and accentedness at the sentence or paragraph level (Derwing & Munro, 2009). Three linguistic measures (segmental, word stress, rhythm) were significantly related to comprehensibility and accentedness ratings. The length of words (number of syllables) significantly predicted comprehensibility but not accentedness, indicating that longer words were easier to understand than shorter words. These findings provide initial evidence supporting the partial independence of comprehensibility and accentedness when L2 speech is measured at the word level. This study provides methodological implications for L2 vocabulary research and suggests using a word-level comprehensibility measure as an additional tool to gauge the employability of L2 words in real-life spoken communication.
... In this task, raters first listened to each of the speech samples and pressed an "f" key for correct and a "j" key for incorrect word pronunciation. Pronunciation was considered correct if it was sufficiently intelligible with minor errors or foreign accents present (Kang et al., 2013). Raters were first presented with 40 target words produced by the model speaker as a reference point and asked to evaluate whether L2 speech samples were intelligible to an average speaker of their variety of English. ...
Full-text available
This study examined how mode of input affects the learning of pronunciation and form-meaning connection of second language (L2) words. Seventy-five Japanese learners of English were randomly assigned to one of 3 conditions (reading-while-listening, reading-only, listening-only), studied 40 low-frequency words while viewing their corresponding pictures, and completed a picture-naming test 3 times (before, immediately, and about 6 days after treatment). The elicited speech samples were assessed for form-meaning connection (spoken form recall) and pronunciation accuracy (accentedness, comprehensibility). Results showed that the reading-while-listening group recalled a significantly greater number of spoken word forms than did the listening-only group. Learners in the reading-while-listening and listening-only modes were judged to be less accented and more comprehensible compared to learners in the reading-only mode. However, only learners receiving spoken input without orthographic support retained more targetlike (less accented) pronunciation compared to learners receiving only written input. Furthermore, sound-spelling consistency of words significantly moderated the degree to which different learning modes impacted pronunciation learning. Taken together, the findings suggest that simultaneous presentation of written and spoken forms is optimal for the development of form-meaning connection and comprehensibility of novel words, but provision of only spoken input may be beneficial for the attainment of targetlike accent.
... We are aware of a handful of studies that have examined the effect of retrieval practice on L2 vocabulary learning and retention. In a laboratory study by Kang, Gollan and Pashler (2013), advantages for retrieval practice (recalling the word before hearing it) as compared to imitation (simply repeating the word) were reported for both the comprehension and production of Hebrew words in English native speakers, both immediately subsequent to training and after a two-day retention period. In another laboratory study, native English speakers learning Spanish were trained on picture-word pairs (with unknown Spanish words) under either retrieval practice or restudy conditions (Barcroft, 2007). ...
Full-text available
We investigated whether learning and retaining vocabulary in a second language (L2) can be improved by leveraging a combination of memory enhancement techniques. Specifically, we tested whether combining retrieval practice, spacing, and related manipulations in a ‘multidomain’ pedagogical approach enhances vocabulary acquisition as compared to a typical learning approach. In a classroom-laboratory design, 48 Turkish university students studying L2 English were trained on 64 English words over 17 days. They were assigned to either a ‘typical’ study regimen of (re)studying the words on the first day (initial study) and last day (cramming) of training, or an ‘optimized’ regimen of retrieval practice (retrieving the words), moreover with feedback, spaced throughout the period, moreover with expanding gaps. The target words were tested before training (pre-test) and one and 11 days afterwards (post-tests). Mixed-effects modeling revealed a training-group by test-session interaction, due to greater improvements from optimized training (a striking 18 percentage-point accuracy increase from pre-test to both post-tests) than typical training (an 8 percentage-point increase). Further analyses showed that the optimized training advantages were mainly driven by high (rather than low) frequency words. Overall, the results suggest that a multidomain approach of combining different memory enhancement techniques can lead to substantial gains in both the learning and retention of L2 words, as compared to a typical study pattern. The findings have implications for L2 learning and pedagogy.
... Following this retrieval attempt, participants heard the correct form of the German word. Such retrieval attempts have been shown to strengthen learning (Kang, Gollan, & Pashler, 2013;see Tokowicz & Degani, 2015, for a review). In this type of training cycle, each word was presented twice in a random order for a total of 110 training trials. ...
When learning novel vocabulary in a third language (L3) through translations in the first language (L1), bilinguals may have more available cognitive resources and more accumulated experience in language regulation compared to when learning through translations in the second language (L2). In a study designed to test language of instruction (LOI) effects, 59 Hebrew–English bilinguals auditorily learned over two sessions 55 words in German, including three word types: cognates, overlapping in form and meaning between English and German; false cognates, overlapping in form but not meaning; and controls. Critically, half of the participants learned through their (dominant) L1 Hebrew, and half through their L2 English (which is also more similar to German). Results showed a significant LOI effect, with better learning through the (less similar) L1, especially for control items. Cognates were learned better in both LOIs, but false cognates were learned better relative to controls to a greater extent when the LOI was English. Together, results highlight the importance of LOI and item‐based language similarity during multilingual novel word‐learning.
Syntax famously consists of abstract hierarchical representations. Less famously, most theories of syntax also assume a higher level of abstract representation: one that abstracts over the hierarchical representations. The existence of such representations would imply that, under certain circumstances, speakers should be able to produce syntactic structures they have never been exposed to. We test this prediction directly. In particular, different types of relative clauses have different surface word orders. These may be represented in two ways: with many individual representations, or with one general representation. If the latter, then learning one type of relative clause amounts to learning them all. We teach participants a novel grammar for only some relative clause types (e.g. just subject relative clauses) and test their knowledge of other types (e.g. object relative clauses). Across experiments, participants consistently produced untrained types, providing the first experimental evidence for this higher level of abstract syntactic knowledge.
Full-text available
examine 2 . . . contributors to nonoptimal training: (1) the learner's own misreading of his or her progress and current state of knowledge during training, and (2) nonoptimal relationships between the conditions of training and the conditions that can be expected to prevail in the posttraining real-world environment / [explore memory and metamemory considerations in training] (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Full-text available
Students learning facts such as foreign language vocabu- lary often rely on a self-testing procedure in which they cue themselves with the English word and try to recall the foreign language target, instead of simply memorizing cue-target pairs. The value of this strategy has been empirically verified by a long history of research, yet existing computational models of human learning do not address the enhancing-learning-through-testing phenom- enon. Using a simple, well studied model—a feedforward neural net with no hidden units—we propose two differ- ent hypotheses for characterizing the phenomenon. Hypothesis 1 is that self-testing generates a target which is used for additional training. Hypothesis 2 is that self- testing produces a more reliable error signal for training than rote memorization. Through simulation studies, we find that hypothesis 2 readily explains the phenomenon whereas hypothesis 1 does not. Further, hypothesis 2 makes predictions worthy of further empirical study, and can be viewed as a natural consequence of temporal dif- ference learning.
Full-text available
A powerful way of improving one's memory for material is to be tested on that material. Tests enhance later retention more than additional study of the material, even when tests are given without feedback. This surpris- ing phenomenon is called the testing effect, and although it has been studied by cognitive psychologists sporadically over the years, today there is a renewed effort to learn why testing is effective and to apply testing in educational settings. In this article, we selectively review laboratory studies that reveal the power of testing in improving re- tention and then turn to studies that demonstrate the basic effects in educational settings. We also consider the related concepts of dynamic testing and formative assess- ment as other means of using tests to improve learning. Finally, we consider some negative consequences of testing that may occur in certain circumstances, though these negative effects are often small and do not cancel out the large positive effects of testing. Frequent testing in the classroom may boost educational achievement at all levels of education. In contemporary educational circles, the concept of testing has a dubious reputation, and many educators believe that testing is overemphasized in today's schools. By ''testing,'' most com- mentators mean using standardized tests to assess students. During the 20th century, the educational testing movement produced numerous assessment devices used throughout edu- cation systems in most countries, from prekindergarten through graduate school. However, in this review, we discuss primarily the kind of testing that occurs in classrooms or that students engage in while studying (self-testing). Some educators argue
Three experiments are reported in which picture naming and bilingual translation were performed in the context of semantically categorized or randomized lists. In Experiments 1 and 3 picture naming and bilingual translation were slower in the categorized than randomized conditions. In Experiment 2 this category interference effect in picture naming was eliminated when picture naming alternated with word naming. Taken together, the results of the three experiments suggest that in both picture naming and bilingual translation a conceptual representation of the word or picture is used to retrieve a lexical entry in one of the speaker's languages. When conceptual activity is sufficiently great to activate a multiple set of corresponding lexical representations, interference is produced in the process of retrieving a single best lexical candidate as the name or translation. The results of Experiment 3 showed further that category interference in bilingual translation occurred only when translation was performed from the first language to the second language, suggesting that the two directions of translation engage different interlanguage connections. A model to account for the asymmetric mappings of words to concepts in bilingual memory is described. (C) 1994 Academic Press, Inc.
Laboratory studies show that taking a test on studied material promotes subsequent learning and retention of that material on a final test (termed the testing effect). Educational research has virtually ignored testing as a technique to improve classroom learning. We investigated the testing effect in a college course. Students took weekly quizzes followed by multiple choice criterial tests (unit tests and a cumulative final). Weekly quizzes included multiple choice or short answer questions, after which feedback was provided. As an exposure control, in some weeks students were presented target material for additional reading. Quizzing, but not additional reading, improved performance on the criterial tests relative to material not targeted by quizzes. Further, short answer quizzes produced more robust benefits than multiple choice quizzes. This pattern converges with laboratory findings showing that recall tests are more beneficial than recognition tests for subsequent memory performance. We conclude that in the classroom testing can be used to promote learning, not just to evaluate learning.
Three methods of studying French vocabulary—(a) studying silently, (b) studying aloud, (c) studying aloud with written recall—were tried out upon 81 students who constituted three sections of a French class. The vocabulary, as far as possible, was selected so that the words were unknown to the subjects, who had already studied French for one year. A series of words was learned by each method named above, then relearned after an interval of two days, ten days, and forty-two days. For the first learning, method C was slightly superior to the others. Methods B and C are decidedly superior for relearning. For accuracy, the learning aloud is best, and method B is better than C. Students who do well in one method do well in the others ( r about .60). The correlation between time and accuracy is negative. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Limitation and lexical development of infants ( n = 20) were examined longitudinally at ages 10, 13, 17, and 21 mo. Infant lexicons were determined from maternal interviews. Words, vocalizations, and actions either outside or within infants' current repertoires were presented at ages 10 and 13 mo. More verbally advanced 13-mo-olds imitated more words inside their repertoires, but only reproduction of words outside their current lexicons predicted their future lexical development. When early vocabularies were statistically controlled, Ss who imitated more words outside their repertoires at 13 mo acquired larger noun lexicons by 17 mo and larger non-noun lexicons by 21 mo. Imitation of other behaviors was not related to vocabularies. Infants' early imitation of words not in their repertoires predicts and may facilitate their future lexical development. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
This study was designed to (a) identify the range, type, and frequency of learning strategy use by beginning and intermediate level ESL students and (b) determine the types of language tasks with which the strategies tend to be associated. Students at beginning and intermediate levels in English proficiency were interviewed in small groups to determine the strategies used to assist in learning each of a number of language tasks: pronunciation, grammar, vocabulary, following directions, listening, making a brief presentation in class, social communication, and functional communication (e.g., applying for a job). In addition, ESL and other teachers of limited English proficient students were interviewed to detect their familiarity with student use of strategies, and to determine whether or not they introduced strategies to their students during instruction. Findings indicated that (a) strategies could be classified into three broad categories—metacognitive, cognitive, and social mediating strategies, (b) students tended to use strategies most often with less complex language tasks, (c) strategies students used most often tended to require little cognitive processing of the learning materials, and (d) teachers were generally unaware of students' strategies and rarely introduced strategies while teaching.
This study compared the pronunciation of targeted vocabulary items in spontaneous speech by 23 adult Chinese L1 learners of L2 English grouped into four different conditions reflecting current pedagogical practices: (a) traditional drilling activities, (b) self-study with tape recordings, (c) interactive activities, and (d) a no-intervention control condition. One hundred and twenty native-speaking listeners judged whether there was improvement or deterioration in pronunciation before and at two separate times subsequent to each of the four conditions. Because none of the results appeared to overwhelmingly favor one teaching technique, we included a discussion of the range of patterns of change brought about by the four input types. We also present arguments for a more serious consideration of the complex effects potentially involved when setting out to modify a learner's L2 pronunciation.
Levels of processing were manipulated as a function of acquisition task and type of recognition test in three experiments. Experiment 1 showed that semantic acquisition was superior to rhyme acquisition given a standard recognition test, whereas rhyme acquisition was superior to semantic acquisition given a rhyming recognition test. The former finding supports, while the latter finding contradicts, the levels of processing claim that depth of processing leads to stronger memory traces. Experiment 2 replicated these findings using both immediate and delayed recognition tests. Experiment 3 indicated that these effects were not dependent upon the number of times a rhyme sound was presented during acquisition. Results are interpreted in terms of an alternate framework involving transfer appropriate processing.