Abstract
Twenty-five years have passed since a teaching technique, called shadowing, was introduced in Japanese EFL contexts. Recently, this technique has gradually been recognized around the world. Given that many studies have been published about shadowing, the time is ripe to summarize the results of the studies and to propose how shadowing should be used. This article introduces shadowing, first referring to theoretical explanations provided so far, especially in terms of attention, cognitive process, and listening process.The article then summarizes past studies on the effects of shadowing on listening and speaking. Next, it discusses challenges in research and also introduces shadowing variations for further use of shadowing. As a classroom teaching tool, this article suggests that beginner level learners should start from shadowing for listening and proceed to shadowing for speaking.
Introduction
As explained in Native Listening (Cutler, 2012), the ways in which native and non-native listeners listen are fundamentally different. As a result, non-native (L2) learners need effective ways to compensate for the gap. However, as Wilson (2003) mentions, in L2 listening, the top-down process and listening strategies, which are regarded as superior strategies, have been emphasized, while bottom-up processes have been undervalued.
In 1992, shadowing, originally a training technique for simultaneous interpreters, was imported into Japanese EFL contexts as a teaching technique for listening (Tamai, 1992) for bottom-up process improvement. Most studies about L2 shadowing have been published in Japan, and shadowing has become gradually popular in East Asia (e.g. Hsieh et al., 2013; Lin, 2009) and recently has been recognized internationally (e.g. Hamada, 2016a, Hamada, 2016b; Foote and McDonough, 2017). Despite its increasing popularity as an effective L2 learners’ bottom-up listening technique, there is no summative review of shadowing to date.
Now that 25 years have passed since attention began to focus on shadowing, the time is ripe to summarize research on shadowing and propose future directions of further study. The ultimate purpose of this article is to bridge research and teaching by synthesizing what has been found, and proposing a future research agenda, triggering more researchers’ and teachers’ interests in shadowing. This article introduces shadowing, theoretical explanations of listening and shadowing, and past case studies to propose teaching implications; it also discusses shadowing for speaking and challenges in research; and finally suggests ideas for future development.
What is Shadowing?
The idea of shadowing dates back to Cherry’s study (1953). In his experiment on speech recognition, the participants listened to two different passages. To confirm that they were listening to only one of the passages, they were asked to repeat simultaneously what they attended to, which is shadowing. Later, to learn how to listen and speak simultaneously, shadowing was used as a technique for training simultaneous interpreters. Prior to the advanced stage of translating one language into another, shadowing has been used for beginner translators to practise listening to one language and repeating it simultaneously (Lambert, 1992). Shadowing was then used for listening practice in teaching because shadowing involves listening to and repeating the input.
Shadowing is defined as ‘a paced, auditory tracking task which involves the immediate vocalization of auditorily presented stimuli’ (Lambert, 1992. p.266). A learner repeats what he/she is listening to, just as a shadow follows someone walking. Most importantly, learners must simultaneously replicate what they hear without written scripts. The examples of shadowing and repetition are shown below. As shown, learners follow the audio stimuli as simultaneously as possible in shadowing and learners repeat chunk by chunk in repetition.
Example of Shadowing (no written script provided)
Studying English is so interesting that I will keep working hard.
Studying English is so interesting that I will keep working hard.
Example of repetition (no written script provided)
Studying English is so interesting that I…
Studying English is so interesting that I
Shadowing, Listening, Speaking, and Research Trends
The next two sections summarizes shadowing research to date, highlighting the theoretical explanations of shadowing and listening proposed in past research. Their teaching implications will be discussed along with the findings of past studies. Following that, new directions in practice and research and research on speaking will be introduced.
Theories of Shadowing for Listening
Research to date has provided theoretical explanations on shadowing mainly for listening, so this section discusses shadowing in terms of cognitive and listening processes.
The Cognitive Process of Shadowing
In shadowing, learners process aural input to reproduce what they hear simultaneously, which requires the coordination of several cognitive processes. Research describes the process of shadowing using the concepts of attention and cognitive resource. First, shadowing focusses learners’ attention on the phonological aspects of what they hear. In shadowing, because there is little time lag between when they hear each word and repeat it, it is called an on-line task (Shiki et al., 2010). Learners attend to incoming sounds rather than meanings when shadowing. In contrast, repetition in which they repeat what they hear chunk by chunk or sentence by sentence is called an off-line task (Shiki et al., 2010). Next, the concept of cognitive resource (i.e. the amount of cognitive capacity available to process information) also supports the argument above. When the learner shadows something in L1, they may be able to understand it simultaneously (Carey, 1971) because of their automatized phoneme perception and adequate cognitive resource. In contrast, the small cognitive resource of L2 learners is devoted exclusively to perceiving the incoming sounds when shadowing. Consequently, they have little cognitive resource left for higher processes such us accessing meanings and comprehending the message (Kadota, 2007). Thus, when shadowing, L2 learners exclusively attend to phonological aspects of what they listen to, so as they continue to practise they will eventually improve at perceiving what they listen to (i.e. bottom-up listening skills).
Listening Processes and Shadowing
When listening, people use both top-down and bottom-up processing (Vandergrift and Goh, 2012). Shadowing improves bottom-up processing in listening because it helps learners perceive and recognize words (Hamada, 2016a). Low-listening proficiency learners depend on top-down processing to compensate for their weaker bottom-up processing (Rost, 2011). The reliance on top-down processing will prevent bottom-up processing from improving, leaving it weak. Shadowing helps learners attend to perceiving what they listen to (i.e. bottom-up processing) by blocking top-down processing, so through practice learners’ word recognition skills, especially phoneme perception skills, improve. This means they will be able to process more information in bottom-up processing and free up the cognitive resource they would originally have devoted to word recognition for higher processing, which leads to better listening comprehension.
Connecting Research and Teaching for Listening
Case studies on shadowing have been conducted to show its effectiveness for listening comprehension skills, and most studies have been conducted in Japan. Kato (2009) conducted shadowing-based lessons with 40 Japanese university students for five months (for approximately 15 minutes once a week) using EFL textbooks; Tamai (1997) gave 90-minute shadowing-based listening lessons to 25 Japanese university students for five days; Lin (2009) gave shadowing-based lessons to 25 Taiwanese junior high school students using EFL textbooks for five weeks (for 50 minutes three times a week). In all these studies, the participants showed significant improvements in listening comprehension. Additionally, to examine the assumption that learners will improve their phoneme perception skills, Hamada (2016a) conducted shadowing-based lessons with 44 Japanese university students for a month (for 20 minutes twice a week) using an EFL textbook. The participants statistically improved on the 20-item partial-dictation pre-post tests, confirming that their phoneme perception skills improved. Taken together, the classroom research supports the assumption that shadowing improves learners’ bottom-up listening processing, which helps develop overall listening comprehension.
Research also shows who benefits from shadowing and when and how it should be used. First, past research has reported that shadowing is effective mainly for low-proficiency listeners (Hamada, 2016a; Kato, 2009; Tamai, 1997). In these studies, learners were divided into two proficiency levels and the lower-proficiency listeners were shown to have improved their listening comprehension skills the most. This indicates that shadowing helps low-listening proficiency learners’ bottom-up listening skills, which advanced learners already possess.
Second, Hamada (2014) investigated whether learners should practise shadowing before or after learning the content by giving eight shadowing-based lessons to 56 Japanese university freshmen using an EFL textbook. Thirty-two students practised shadowing before learning the contents, while 24 did so after. Only those who practised shadowing after learning the content showed statistically significant improvement in the listening comprehension test. Therefore, in daily English lessons, learners should study the contents first.
Third, to compensate for the disadvantages of shadowing, studies recommend that learners record their shadowing performance for reflection. Despite the importance of monitoring performance (Vandergrift and Goh, 2012), L2 learners cannot monitor their shadowing performance because most of their attention is on listening to input and reproducing it orally. In Nakayama and Suzuki (2012), 35 university students were placed in three groups; the self-monitoring group shadowed and recorded themselves on an IC recorder, checking their performance using a written script afterwards. The pair-monitoring group checked their performances with each other using the written script. The control group shadowed and checked their performance, recalling their performance, using the written script. Each group repeated the shadowing activity three times. The results indicate that the self-monitoring group outperformed the pair-monitoring and control groups in the shadowing reproduction rate. Hamada (2015) applied the results in classroom teaching; 30 university sophomores worked on pair-shadowing and 31 worked on self-monitoring shadowing for 30 minutes twice a week for a month using a low-level graded-reader book. The results show that the low proficiency listeners in the self-monitoring group alone improved on listening comprehension questions. Thus, the use of IC-recorders is advisable especially for low-proficiency listeners to compensate for the theoretical weakness of shadowing.
Connecting Research and Teaching for Speaking
There is much less research on shadowing for speaking than for listening, but by applying the concepts of attention and cognitive resource, we can theorize that once learners’ bottom-up listening skills improve, they can allocate more attention and cognitive resource to speaking when shadowing.
There are two noteworthy empirical studies of this type of shadowing. Hamada (2016b) conducted shadowing-based training with 29 Japanese university sophomores of intermediate English proficiency (for-contrast group; repetition-based training for 29) for approximately 20 minutes twice a week for a month using an EFL textbook. Six native speakers of English rated the participants’ pre/post recordings of a sentence-read task on a 1–9 Likert scale, but little improvement was found in either group in comprehensibility or accentedness. In Foote and McDonough (2017), 16 high proficiency L2 speakers engaged in shadowing for eight weeks (at least four times and for 10 minutes a week), using sitcoms such as Friends and The Big Bang Theory, with an iPod to record performance. The raters, 22 L1 English listeners, rated the participants’ performance on shadowing and picture-narrative pre-and post-tests using MATLAB, a proprietary programming language developed by MathWorks. The participants statistically improved their imitation skills, comprehensibility, and fluency but not accentedness. The primary difference between Foote and McDonough (2017) and Hamada (2016b) is that they were conducted with advanced and average students, respectively. As discussed above, shadowing focussed on accurate replication of speech, especially TV shows, requires a high level of English proficiency. The comparison of these two studies suggests that shadowing is effective for pronunciation development but requires a degree of high proficiency.
An additional finding of Foote and McDonough (2017) is that the participants’ comprehensibility in extemporaneous speech improved, meaning that the participants were able to transfer skills in imitating a speech model to daily performance. Comprehensibility is one of the most important pronunciation features (e.g. Grant et al., 2014), and thus improvement of comprehensibility through shadowing is a valuable finding for classroom practice and further research.
In summary, advanced learners are displayed improved comprehensibility from practising shadowing. To use shadowing for pronunciation development, learners need to attend to their pronunciation when shadowing.
Challenges of Shadowing and Shadowing Variations
In this final section, two challenges that arise from the nature of shadowing studies are discussed. Ideas for more shadowing activities are introduced to expand the possibilities of shadowing in terms of practice and research.
In terms of listening, firstly, the depth of learners’ attention when shadowing (e.g. levels of perception, semantic, and syntactic) needs further exploration. It is assumed that beginner level learners pay most attention to speech perception when shadowing, and as their listening skills improve, they are able to attend to semantic and syntactic processing as well. A more extensive examination of this issue will contribute to expanded usage and research on shadowing. Additionally, most studies were conducted for a month, and the optimal amount of shadowing practice has not been identified. By using delayed post-tests, investigation into the length of the effect can be conducted. Lastly, most case studies have been conducted in Japan, so in order to strengthen the findings and explore further, more research in various contexts is required.
Regarding speaking, a theoretical framework that explains speaking development through shadowing needs to be established. The research on shadowing and speaking is still undeveloped, and problems such as a lack of robust research design, careless and questionable data analysis, use of biased data, and over reliance on students’ self-reports are often observed. Next, acoustic analysis (Mori, 2011) may show some possibilities for pronunciation development, but does not reveal if it is noticeably improved among people. Thus, such studies as Foote and McDonough (2017) and Hamada (2016b), in which human raters were actually involved, are recommended. Additionally, to examine the transferability of shadowing practice to speaking performance, researchers should examine pre-post speech performance, and not only pre-post shadowing.
For the sake of further exploration on shadowing, research has proposed two variations on standard shadowing. First, in selective shadowing (Murphey, 2001), learners shadow only certain words and phrases (e.g. only content words, prepositions, or keywords). Because in selective shadowing not only phoneme perception but word recognition needs to occur, depth of attention may change. In standard shadowing, semantic and syntactic analyses of what is heard are less likely to occur, but in selective shadowing, these analyses must occur for learners to perform selective shadowing successfully.
Second, in text-presented shadowing (Kuramoto et al., 2007), learners shadow together with a written script of the target passage. Because scripts are involved in this process, the learners’ cognitive process changes as their attention is split between sound information, meanings and letters. Because of this change in cognitive process, a combination of shadowing and text-presented shadowing may improve reading skills as Kuramoto et al. (2007) shows, and possibly pronunciation.
Conclusion
The accumulated research provides implications for using shadowing in the classroom. A key implication is that beginner level learners should start from standard shadowing for listening. Once learners achieve the upper-intermediate or advanced level, they are ready for shadowing for speaking. The review of the research reveals what has been found and what still remains to be investigated. Learners’ attention when shadowing should especially be further examined, and the foundation of a theoretical model for speaking skill development by shadowing needs to be developed. To strengthen the findings that past studies have found and to expand the idea of shadowing for the sake of language learners, more research in different contexts and are needed. Research on shadowing should be directed towards examining more variations and exploring ways to develop speaking skills.
Footnotes
Acknowledgements
I would like to thank Dr. Michael Rost, Dr. Choonkyong Kim, Dr. Tomoko Nemoto, Dr. Yuichi Suzuki, Dr. Wataru Suzuki, Dr. Yasuhiro Fujiwara, and the RELC Journal reviewers for their helpful comments.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was partly supported by a Grant-in-Aid for Scientific Research (JP15K16788).