2017-18 Mitra FAMILY GRANT Recipient Song as Speech Therapy: The Trials of Pronouncing a New Language Alan Jiang, Class of 2018
Running head: SONG AS SPEECH THERAPY
Song as Speech Therapy: The Trials of Pronouncing a New Language Weilun Jiang 2018 Mitra Family Scholar Mentors: Ms. Kelly Horan and Ms. Susan Smith
Author Note This research was supported by a grant from The Mitra Family Scholar Grant Program.
SONG AS SPEECH THERAPY
2
Abstract This paper explores various published articles that examined the relationship between music and language. The relationship between music and language suggests how it might be implemented into the current education curriculum as an effective method of teaching. Choi, Tong, and Singh (2017) define the prosodic transfer process between languages and demonstrates how lexical stress sensitivity can predict lexical tone sensitivity. Higgins (2012) establishes a connection between music and language and offers insight into how music might break down the barriers of tonal and non-tonal languages. Both Christiner and Reiterer (2013) and Kleber (2010) explain how singing affects speech imitation ability and might potentially improve one's pronunciation of a foreign language. Maley (2000) connects the different ideas and gives insight into how poetry and song might be applied in a classroom setting. The paper aims to explain the connection between the various research articles and suggests how singing should be adopted into the educational curriculum as a method of teaching foreign languages.
Keywords: singing, music, pronunciation, foreign language acquisition, lexical stress sensitivity, lexical tone sensitivity, tonal and non-tonal language
SONG AS SPEECH THERAPY
3
Song as Speech Therapy: The Trials of Pronouncing a New Language A Lingua Franca is a common language that is adopted between speakers whose native languages are different. Since English is so widely used in social media, international advertising campaigns, and political negotiation between countries on the world stage, it is easy to see how citizens from Myanmar are familiar enough with English to the extent that it can be used as a bridging language with people from other countries who do not speak Burmese. The acquisition of English is increasingly important for Myanmar teenagers as the gradual opening of Myanmar to the western society has allowed them to voice their opinions and participate in democratic activities openly. The difficulty is that Burmese is a tonal language, while English is a non-tonal language. The differences between the two languages contribute to various learning obstacles that might hinder how the foreign countries communicate with countries such as Myanmar. The paper proposes that music and singing may serve as an effective way of teaching and bridge the two languages.
Literature Review Soe (2015) explains that in order to better participate economically and politically, both the Myanmar government and its citizens have a strong inclination to learn English to integrate themselves into international businesses. This is particularly important as better English will grant the country more opportunities in the global community of scientific researchers and intellectual communities who regularly publish articles in English for Western and European consumption. The study of Choi et al. (2017) suggests that the gap between tonal and non-tonal languages may be bridged as lexical tone sensitivity helps predict lexical stress sensitivity. Lexical tone sensitivity is defined as one's auditory sensitivity towards a tonal
SONG AS SPEECH THERAPY
4
language while lexical stress sensitivity is defined as one's auditory sensitivity towards a nontonal language. Moreover, music and singing contribute to language acquisition in various ways. For example, as suggested by Higgins (2012), music shares fundamental similarities with language such as the natural tendency for rhythm and discrete pitch scales. Humans even have the natural tendency to apply the gestalt principle of similarity to both language and music. Additionally, singing helps improve the learner's motor vocal control and therefore improves their ability to better imitate a foreign language.
Challenges of Learning Non-Tonal Languages To understand the challenges of acquiring English as a foreign language, it is important to first understand differences of tonal and non-tonal languages and how these differences cause obstacles for language learners. A tonal language is a language in which the tone of a word will determine the overall meaning. This means that even if two words sound identical, the variation in tones will still change the meaning. On the other hand, a non-tonal language is a language in which word meanings are not affected by pitch variation. There are biological factors that differentiate the two categories of languages. Ge et al. (2015) shows that Chinese (tonal language) speakers and English (non-tonal language) speakers use their brain's left hemisphere differently when speaking. When processing intelligible speech, both Chinese and English speakers utilize the regions of their left hemisphere such as the inferior frontal gyrus (Broca's area), posterior temporal gyrus (Wernicke's area), and the anterior temporal gyrus. However, the experimental results of Ge et al. (2015) suggests that what differentiates tonal and non-tonal language speakers is that tonal language speakers activate an additional region of their brain. For example, Ge et al. (2015) shows that tonal language speakers had an additional activation in the right anterior temporal pole, the anterior extremity of the temporal lobe. The article further
SONG AS SPEECH THERAPY
5
explains that both the right anterior superior temporal gyrus and left posterior superior temporal gyrus are involved in the processing of the tonal-language network, which is not activated when it comes to non-tonal languages. The activation in the right anterior superior temporal gyrus all the way towards the left posterior superior temporal gyrus is a result of the suprasegmental features such as pitch variation of a tonal language. In other words, tonal language speakers activate the right hemisphere when processing intelligible speech whereas non-tonal language speakers do not. Although speech is usually considered to be left-hemisphere dominant for most of the population, Ge et al. (2015) argues that the right hemisphere, especially the right anterior temporal cortex, is crucial for processing tonal languages due to its functions in tonal processing. The additional activation in the right hemisphere when processing tonal languages demonstrates, on a biological level, one of the key differences between tonal and non-tonal languages. The examined studies suggest that the differences might be a reason why it is challenging for tonal language speakers, in this case, native Myanmar speakers, to process and learn non-tonal languages. A prediction is made that it might be a challenge for native tonal language speakers with no prior exposure to non-tonal languages to learn a non-tonal language as the language processing inside their brains is fundamentally different. Ge et al. (2015) hints at a possibility that tonal language speakers might focus on processing the tonality of the received information which does not exist in non-tonal languages. Due to limited exposure towards English and nontonal languages in general, the processing mechanism of Myanmar teenagers might be fundamentally different when compared to that of native English speakers. It can therefore be inferred that tonal language speakers might encounter various pronunciation and voicing challenges when trying to speak a non-tonal language.
SONG AS SPEECH THERAPY
6
Another variable that contributes to the challenges of learning a non-tonal language might be the differences between suprasegmental features such as variation in pitch, tone, and stress of non-tonal languages and tonal languages. For example, pitch variation determines the meanings of the words, which results in greater numbers of homophones. Homophonic words are words that are differentiated by suprasegmental features such as pitch variation. Such words have the same sound but have different meanings due to the pitch variation. Therefore, unlike non-tonal languages, the greater number of homophones in a tonal language results in a greater number of vocabulary words (Ge et al., 2015). This difference between the number of homophones can be applied towards Chinese and English. The Chinese language consists of approximately 400 syllables, whereas, the non-tonal language of English has about 12,000 possible syllables. The difference is that Chinese consists of a greater number of homophonic words. One example would be how English emphasizes lexical stress (the stress placed on syllables within words) of the sentence, whereas languages like Chinese and Burmese focuses the pitch variation of words. Yuan and Gao (2002) use an example to demonstrate how the pitch of the Chinese language differentiates the identical sounds with reference to the sound "da". One syllable "da" differentiates the meaning of the words 'to hang over something' (搭 da1), 'to answer' (答 da2), 'to hit' (打 da3), and 'big' (大 da4), where the number after each syllable indicates the tone. Since the English language does not share the same mechanisms as a tonal language, it can be increasingly difficult for tonal language speakers to differentiate individual letters. For example, an "a" with an upward pitch and an "a" with a downward pitch is just as different and distinct as the difference between "p" and "b" due to the differences in voicing (He, 2014). The stress and intonation patterns of tonal and non-tonal languages are also inherently different which contributes to more learning obstacles. Considering the differences discussed above, it can be
SONG AS SPEECH THERAPY
7
inferred that there are various difficulties for tonal language speakers (Burmese/Mandarin) to apply their knowledge of language towards English as English focuses on applying stresses onto the sentences and words. Similarly, some of the phonemes present in English do not exist in Burmese. The findings of the studies examined suggest that this contributes to another significant obstacle for tonal language speakers trying to acquire a non-tonal language. For example, the Myanmar language does not have the bilabial sound "w", a sound formed by the near closure of the lips, and the alveolar consonant "r", a consonant sound formed by the tongue touching the ridge behind the teeth on the roof of the mouth. Since Chinese and Burmese are both tonal languages and members of the Sino-Tibetan language family, it can be inferred that both Chinese speakers and Burmese speakers might face similar challenges when learning a non-tonal language such as English. For example, native Chinese speakers find it difficult to hear and differentiate between the letters l and r, and thus may mispronounce rake and rice as lake and lice. Southern Chinese speakers similarly, also have a hard time distinguishing between l and n (Grasu, 2015). Furthermore, He (2014) demonstrates that Mandarin speakers have difficulties in producing the /l/ correctly in syllables’ final position (swell, well, fall), with a high error rate of 97%. It is suggested that in Mandarin, the /l/ sound cannot occur in the final syllable position as the language only allows nasal consonants /n/ and /ŋ/ and retroflex /r/ in coda position (syllable that ends the word). Moreover, there are also phonetic differences between Mandarin and English that contribute to the pronunciation obstacles. He (2014) states that Mandarin speakers' difficulty in pronouncing the final /l/ stems from the fact that the /l/ does not occur in the position in Mandarin syllable structure patterns, and from the manners of the phoneme /l/. The examined
SONG AS SPEECH THERAPY
8
studies offer a possible explanation as to why most tonal language speakers (including Burmese) face pronunciation challenges when attempting to learn English. He (2014) further suggests that Mandarin speakers often employ three strategies to cope with pronunciation difficulties - vocalization, deletion, and retroflexion. For vocalization, the apical gesture (tip of the tongue gesture) is lost while the dorsal gesture (body of the tongue gesture) is maintained, resulting in the /l/ being heard as a back vowel of /u/ or /o/. Vocalization is problematic as English speakers might misunderstand the meaning of the word due to mispronunciation. For deletion, both the apical and dorsal gesture is lost, leaving no /l/ sound in the coda position. Retroflexion occurs as speakers curl up the tip of their tongue but do not touch the alveolar bridge. Therefore, listeners would perceive a /r/ instead of a /l/. The differences in suprasegmental features explain the potential obstacles that Burmese speakers face when conducting experiments regarding English acquisition in the future. Students are expected to have difficulties in pronouncing such syllables. It is also predicted that students will employ techniques of deletion, vocalization, and retroflexion when facing unsure syllables such as /l/ and /r/. Both biological and suprasegmental features differentiate tonal and non-tonal languages. Such differences contribute to the various challenges in second language acquisition, especially the acquisition of non-tonal languages for native tonal language speakers. The tonal language speakers are more likely to encounter English syllables or phonemes that do not exist within their language system and hence are hindered in their ability to perfect the pronunciation of certain words. Often, as suggested by He (2014), students will then employ techniques of deletion, vocalization and retroflexion to overcome such obstacles. This potentially results in the mispronunciation of words and suggests a reason why it might be difficult for native English speakers to comprehend the information delivered by Mandarin/Burmese speakers.
SONG AS SPEECH THERAPY
9
Lexical tone sensitivity helps bridge Burmese and English Although there are various suprasegmental differences between tonal and non-tonal languages, on the contrary, there are also various similarities between tonal and non-tonal languages that help tonal language speakers acquire English. The study of Nguyen et al. (as cited in Choi et al., 2017) investigates how the lexical tone sensitivity of Vietnamese contributes to the lexical stress sensitivity of English in native Vietnamese speaking children. Nguyen et al. together with Choi et al. (2017), find that there is a process of lexical prosodic transfer between languages. Lexical prosodic transfer, according to Choi et al. (2017), is defined as a “process through which adults and children who learn English as second language (ESL) capitalize on similarities in the structure of lexical tones and lexical stress in a way that allows them to harness their perceptual sensitivity to L1 (first language) lexical tones in the service of L2 (second language) English lexical stress perception�. Fry's study (as cited in Choi et al., 2017) states that English stress consists primarily of variation in vocal pitch, duration, and intensity. Similarly, Chao's study (as cited in Choi et al., 2017) explains that Cantonese lexical tone also involves in the use of vocal pitch (as well as amplitude, duration and other spectral factors) to distinguish lexical items at the syllable level and a tone is assigned to every syllable. Both studies establish the idea of how tonal and nontonal languages rely on vocal pitch and duration. Choi et al. (2017) further explains that tone and stress are similar in structure as both are being driven by a similar set of acoustics. This suggests that the similarities between lexical tone and lexical stress might be able to outshine the differences. The shared similarities help Cantonese students predict lexical stress sensitivity. Specifically, Luke et al. (as cited in Choi et al., 2017) proposes that links between stress and tone are strengthened due to the "tonalization" of English lexical stress. The tonalization of English
SONG AS SPEECH THERAPY
10
lexical stress is defined as how English lexical stress is perceived as high tones by Cantonese ESL learners. This suggests that process of tonalization of English lexical stress might serve as an aid in helping lexical tone speakers learn a lexical stress language. Through predicting English lexical stress based on the tonality of their native tongue, Cantonese speakers could possibly assimilate English lexical stress into their native tonal system, directly predicting English lexical stress sensitivity through Cantonese lexical tone sensitivity (Choi et al., 2017). The establishment of the relationship between stress and tone demonstrates that it is possible to bridge the gap between tonal and non-tonal languages. It is, therefore, feasible to infer that improving lexical tone sensitivity of tonal language speakers might increase the sensitivity to lexical stress on account of evidence of transfer from the present study. Since Choi et al. (2017) establishes how lexical tone sensitivity in tonal languages could predict lexical stress sensitivity, this suggests that the mechanism of transfer might also be applied to other tonal languages such as Burmese. Choi et al. (2017) supports the idea that there is a minimized difference between how a non-tonal language speaker and tonal language speaker acquires a non-tonal language. This is particularly important to the acquisition of a foreign language as the differences between lexical tone and lexical stress languages contribute to various learning obstacles. Choi et al. (2017) suggest that the prosodic transfer between the two languages might help bridge the obstacles and help tonal language speakers predict lexical stress sensitivity and hence learn the language. Since lexical tone sensitivity helps speakers predict lexical stress sensitivity, this opens up a possibility that focusing on improving one’s sensitivity towards “tones” might actually help one learn a lexical stress language more effectively. Though this may be a matter of sound production, which is an underlying biological component to learning
SONG AS SPEECH THERAPY
11
languages, there might be another answer through the "melodic contour" theory whereby tonality and pitch variation of tonal language are linked (Higgins, 2012).
Music Shares Similarities with Language The stronger the musicality of a person, the better one might be able to apply it towards language learning. Due to the various similarities between music and language, a better music sense can help one in the field of language acquisition. One of the similarities between music and language is the natural tendency for humans to apply the Gestalt principle towards the two. Higgins (2012) draws support from the study of Lerdahl, explaining that the Gestalt Principle of psychology should be applied to both music and speech. The Gestalt Principle refers to humans' natural tendency to categorize objects in proximity into same groups. Lerdahl explains that both poetry and music involve a hierarchy of beats where such beats would be further subdivided into two or three beats (as cited in Higgins, 2012). Lerdahl’s study further explains that although different cultures have different influences on poetic and musical meters the possible combinations of two and three beats across various languages and cultures are very limited (as cited in Higgins, 2012). The limited combinations of two and three beats between languages suggests that the differences between tonal and non-tonal languages could be minimized when the languages are being divided and analyzed from a perspective of beats and meter. Lerdahl's study concludes that the difference is further minimized as humans have the tendency to construct both musical and linguistic rhythms based on a two or three beat pattern (as cited in Higgins, 2012). Higgins uses various studies and theories to support the point that music and speech are inseparable when analyzed from a micro-perspective. Higgins (2012)’s theory is particularly important as it explains how music and different languages are connected through a similar pattern of beat and meters. Although there are differences in tone and stress patterns
SONG AS SPEECH THERAPY
12
between tonal and non-tonal languages, music might be a possible solution to bridging the two as both the languages are connected through a limited combination of meter and beats. Music might help to break down the barrier of tonal and non-tonal languages Looking from another perspective, music affects speech through their shared similarities. In short, Higgin's book "The Music between US - Is Music a Universal Language" explains how music is linked to speech through four common characteristics. The first being how linguistic signals are organized in terms of melodic contour. Melodic contour is important in linguistic communication because it defines how the linguistic information is being perceived. Higgins (2012) raises an example of how non-standard ways of contouring speech, as with a foreign accent, would increase the difficulty of a native speaker's perception of the linguistic information conveyed. As mentioned earlier, the tone of a word, especially in tonal languages, can evoke various emotional responses due to the differences in the meaning of the word. Similarly, Higgins (2012) also states that melody in music also appears to express emotion in both language and music. Juslin and Laukka's experiment concludes that rising pitch contours correlate with "active" emotions such as happiness, anger, and fear (as cited in Higgins, 2012). On the other hand, falling melodic contours are often correlated with sadness and tenderness, which are regarded as less "active" emotions. An example would be how "the tone of the voice," a form of vocal expression (seen in both tonal and non-tonal languages), is involved with melodic contouring. How emotions are conveyed in a similar manner is especially evident in music, where themes such as love, death, sadness and happiness are expressed in similar ways across different cultures. The similarities demonstrate a relationship between speech and music in that similar pitch variations would evoke similar emotional responses. This suggests a possibility of how music could aid language learning due to its similarities in provoking similar emotional
SONG AS SPEECH THERAPY
13
responses as speech. Through provoking similar emotional responses as speech, music might help language learners better understand the context of the materials and hence lead to a more effective way of learning. The similarities between music and language further hint at the possibility that, when compared to normal teaching methods, singing can help language learners better understand the thematic content behind the words and hence acts as a more effective tool for students to understand and learn English. Not only do music and speech bear resemblance in provoking similar emotional responses, categorical perception can also be applied to both language and music (Higgins, 2012). Higgins explains that humans tend to categorically perceive both pitches and durations in music; similarly, humans also perceive the phonemes of language categorically. An experiment conducted by Warren and Warren (1970) concludes that through the insertion of sounds, one can normalize irregularities in both language and music, so that acoustic signal would be easier to recognize (Higgins, 2012). Warren and Warren (1970) tested this through conducting an experiment in which they told participants that a certain phoneme would be missing from a recorded sentence. The sentence selected contains the word "legislatures" in which the first letter /s/ in the word is removed and replaced with a noise. Surprisingly, the subjects did not register nor identify the missing phoneme but heard the exact word - "legislatures." Warren and Warren (1970) demonstrates in both speech and music, that despite the absence of a tone/phoneme, listeners would have the innate tendency to insert a sound to clarify the harmonic situation in the phenomenon of the missing fundamental component. The Warren experiment offers insight into how humans have a natural tendency to fill in the gaps of irregularities, no matter language or music. This further suggests that there is a relationship between music and language because most humans have the innate tendency to fill in the gaps where irregularities exist.
SONG AS SPEECH THERAPY
14
The third similarity between music and language is that they both utilize discrete scale pitches. Not only does this similarity bridge the gap between music and language, but also the gap between tonal and non-tonal languages. Higgins proposes that pitch plays a crucial role in acting as semantic devices in both tonal and non-tonal languages. To support this, Higgins explains the autosegmental theory. The autosegmental theory according to Steven Brown (as cited in Higgins, 2012) claims that “phonological events should be modeled as sequential movements between discrete pitch levels, often only two levels, High and Low, and that all movements between them should be reduced to the status of transitions, rather than primary phonological events of importance.� This means that levels of pitch in a language mainly serve as targets that bear little significance. The movement between pitches is a merely a transition of getting from one pitch to another. Higgins further explains that although tones help differentiate the meaning of words in tonal languages, the use of tones is not a unique feature to tonal languages. Even though non-tonal languages do not have tones that differentiate the meaning of the word, people often employ non-level tones (Higgins, 2012). Higgins, again, draws support from Brown and suggests that all variation in pitch is nothing but transitional steps from one tone to the next. Subsequently, Higgins quotes from Steven Brown's book The "Musilanguage" Model of Music Evolution that "speech (no matter tonal or non-tonal) and music are similar as they not only are based on discrete pitch scales, but also utilizes them." Higgins (2012) therefore concludes that discrete scale pitches serve as a feature that bridges tonal, non-tonal languages, and music together. Because both music and speech rely on discrete pitch scales, it can be reasonably inferred that such fundamental similarities could be emphasized in a classroom setting. This partially explains how music can help break the barrier between tonal and non-tonal languages. Since all tonal, non-tonal languages, and music relies on discrete scale pitches, it
SONG AS SPEECH THERAPY
15
might be able to help tonal language speakers understand how to apply tones to learning nontonal languages. Tonal language speakers can apply the concept of how tones are just transitional steps from one another towards the non-tonal languages. Music and discrete scale pitches might allow tonal languages to understand how non-tonal languages are similar to their language and in that, the only difference is how the scales are utilized. As such, students might find it easier to learn languages that are different from their native tongue. Teaching English through music might be a possible method in bridging the differences between tonal and non-tonal languages through shared similarities such as melodic contours, the innate tendency to fill in the gaps of irregularities, and discrete scale pitches. Singing and Music Helps with English Learning Both music and language heavily rely on not only rhythmicality but also tonality. One of the major similarities that categorizes language and music is how the linguistic signals are organized in terms of melodic contour (Higgins, 2012). This suggests the possibility of how the tonality and pitch variation of tonal languages might be linked to the melodic contour of music, hence the field of linguistic signaling. It therefore suggests that the tone sensitivity of native tonal language speakers, together with the reinforcement of music and singing, might help such L1 speakers acquire L2 languages such as English. Tonality and pitch perception play a major role in the acquisition of lexical-stress-based language such as English. Tones are inseparable elements of music, and melodic contour influences the organization of linguistic signals. This suggests that singing is an effective way to help Burmese speakers acquire English. This brings us to the investigation of how singing contributes to the acquisition of a foreign language. Christiner and Reiterer (2013) examine how singing contributes to the acquisition of a foreign language, specifically how singing ability links to the speech imitation ability. Christiner
SONG AS SPEECH THERAPY
16
and Reiterer (2013) conducted an experiment to determine whether singing ability and musicality would influence one's speech imitation ability. Forty-one participants of varying singing ability were selected for this study. All participants were native German speakers and had no prior exposure to Hindi. The study tested two factors, the first being the participants' ability to spontaneously read and repeat unknown words for English and unintelligible utterances for Hindi. The second factor is the participants' abilities in singing and the imitation of pronunciation of a foreign language and Hindi. The study demonstrates that the three major factors that influence speech imitation ability are working memory, singing ability, and educational background. Since the pre-educational influence is controlled for, this suggests that singing ability and working memory are the two biggest factors that influence second language learning. Singing can be divided into two sub-categories - rhythm and the quality of voice. It demonstrates that singing has a stronger correlation to speech imitation than other musical factors. This is explained by the fact that singing and speaking are underpinned by the same speech generation process. The experimental results of Christiner and Reiterer (2013) suggests that the excellent speech imitation ability of a good singer stems from their vocal flexibility which might be the result of their physical training. Singers are more aware of the sound production process and have a vocal apparatus with good fine-tuning ability of the tongue and the lips as well as the larynx. This again exemplifies the relationship between singing and speaking as both share the same sensory network and vocal apparatus. Research by Christiner and Reiterer (2013) further shows that, on a biological level, singers showed greater complexity in certain white matter tracts, a result from their extensive vocal-motor training. A study by Halwani et al. explains that this leads to improvements in both sound perception and sound production (as cited in Christiner and Reiterer, 2013). Moreover,
SONG AS SPEECH THERAPY
17
increased activity in the primary somatosensory cortex, part of the brain that is responsible for processing sensations related to touch, has also been observed in the brains of classically trained singers (Kleber et al., 2010). Kleber et al. (2010) further explains that experienced classical singers have been extensively trained to focus on timbre, vowel quality, intonation accuracy, and the sound intensity in the context of increased emotional expression. This requires the singers to have intense control over laryngeal motor activity in conjunction with subglottal air pressure, whereby vocal folds oscillate rapidly between 50 and 1500 Hz. Singers have to have the aforementioned vocal motor control to produce the changes in tension and positioning of vocal folds. This offers a possible explanation for why singers have increased activity in the primary somatosensory cortex - singers have higher vocal sensitivity due to the better vocal motor control. Kleber et al. (2010) draws support from the study of Wyke et al. and explains that vocal motor control during the process of sound production heavily relies upon kinesthetic feedback. According to the study of Wyke et al., kinesthetic feedback is defined as “information about position, movement, and receptors in muscle spindles and tendons as well as join receptors from laryngeal mechanoreceptorsâ€? (Kleber et al., 2010). Kleber et al. (2010) draws support from MĂźrbe and continues to explain that several studies have previously explained the significance of auditory feedback in singing, but the kinesthetic feedback circuit seems to be particularly improved after years of training in classical singing. This could correspond with the development of more accurate feedforward commands of vocal tract adjustments based on kinesthetic expectations (or "kinesthetic motor memory"). Kleber et al. (2010) study proposes that excellence in classical singing might be accompanied by increased involvement of the primary somatosensory cortex, possibly reflecting enhanced kinesthetic motor control of the vocal tract for sound production (Kleber et al., 2010). The higher level of kinesthetic motor control of the
SONG AS SPEECH THERAPY
18
vocal tract of opera singers offers a partial explanation to why opera singers might acquire a language faster than non-singers. The optimization of the opera singers' vocal motor system might be the reason to their speech imitation ability when compared to non-singers. As we consider this in the context of the majority of language learners, it can be inferred that better speech imitation ability might help language learners master the pronunciation of a foreign language. Although playing an instrument and singing both involve music, the study by Lotze (as cited in Kleber et al., 2010) explains that instrumentalists demonstrate an increase in activity in the primary motor cortex (as cited in Kleber et al., 2010) rather than the primary somatosensory cortex of the singers. The difference demonstrates that although both instrumentalists and singers are commonly exposed to musical elements, singers have different processing mechanisms that are more closely related to speech. This difference suggests that, even though playing instruments and singing both involve music, singing might still be a better method of teaching English. This comparison subdivides musicality into two subgroups - singers and instrumentalists. Not only is singing similar to speech, GarcĂa-LĂłpez and GavilĂĄn Bouzas' study also establishes the argument that, according to singing teachers, singing and speaking are nearly identical. Both share principles such as body posture, emission, resonance, and articulation. An exception is breathing, which is more active during exhalation in song than in speech (as cited in Christiner and Reiterer, 2013). Christiner and Reiterer (2013) demonstrate not only the close resemblance between song and speech, but also illustrate how singing aids speech imitation ability. The findings of the papers examined propose that through singing, students will exercise greater vocal motor control, generating more accurate feedforward commands of vocal tract
SONG AS SPEECH THERAPY
19
adjustments based on kinesthetic expectations, therefore improving their ability to imitate a new language. Furthermore, learning a language through singing might also help the students achieve better second language pronunciation because of the enhanced auditory working memory and vocal flexibility. Music and Song in a Classroom Setting Maley (2000) examined how song and poetry serve as effective language learning activities. The author suggests that both poetry and song are effective due to memorability, rhythmicality, performance/recitability, universality, reactional language, motivation, and interaction. From the memorability perspective, Maley claims that fragments of poems and songs remain in the mind for an extended period even after communicative competence has faded away. Song and poetry phrases are often particularly poignant and striking; furthermore, such phrases also seem to go on repeating themselves in the inner ear without conscious will. The second factor in how singing might help Burmese students learn English faster is related to the rhythmicality of music in songs. Brown's study (as cited in Maley, 2000) claims that the structure of information in spoken message is built upon rhythmicality. This partially explains why singing could contribute to better learning retention as, during singing, the patterns of sound and stress are repeated in regular sequences. Better learning retention might contribute to the learning process and further improve the pronunciation of the foreign language. Maley (2000) also suggests that simple repetition and similar drills are often lackluster and boring. On the other hand, Maley claims that written words fit naturally into singing, especially in choral form. This suggests that singing in groups, specifically in a classroom setting, naturalizes the process of learning which helps mask individual error and contributes to building up self-confidence among the students. By building up stronger self-confidence, Maley's claims suggest that singing could
SONG AS SPEECH THERAPY
20
naturalize the process of learning English for Burmese students and hence make the process easier when compared to the non-singing group. Brown and Yule's study (as cited in Maley, 2000) explains that the uses of spoken language are often divided into “interactional” and “transactional.” Interactional language means that people are mainly concerned with social lubrication. It is an act of making speakers feel comfortable with each other. “Transactional” language, on the other hand, has a goal of communicating a utilitarian message. An example of a transactional language would be giving instructions, stating opinions, and describing objects. Maley claims that both poetry and song give speakers a third type of spoken language use called “reactional” language. The purpose of “reactional” language is to "make people react personally to another person's verbal sensibility" (Maley, 2000). This proposes the idea that singing might not only help Burmese speakers effectively learn English but also help them better in communicating with other English speakers. This provides a partial explanation as to how, due to singing, Burmese speakers might be able to have better verbal sensibility, which could be beneficial to English-speaking Burmese people as they can better communicate with English-speaking foreigners, providing them with a wider range of economic opportunities. Maley (2000) reasons that as English is the language of popular music, this provides obvious motivation for language learners to learn English. This motivation stems from the learners' realization that even though one might be relatively inexpert in a language, one can still appreciate, to a certain degree, what is considered to be a "difficult use of language" (Maley, 2000). As such, Maley claims that it would be a morale booster for language learners to learn the language, should they have this realization. This contributes to the fact that when conducting the proposed experiment, the singing group of Burmese students might have higher levels of motivation and morale when compared to the non-singing group in learning
SONG AS SPEECH THERAPY
21
the selected word phrases. The higher level of motivation and morale suggests that the singing group might be more engaged and focused when learning the selected word phrases, yielding better results. Last but not least, Maley (2000) also suggested that not only do song and poetry help language learners acquire a foreign language, it is, overall, a more effective way of teaching as it offers a variety of opportunities for both student-student and student-teacher interactions. Poetry and song, when compared to normal teaching methods such as repetition, give both the teachers and students opportunities to participate in activities. Since the majority of such in class singing activities involve more than just the individual, it follows that there is also more communication between the language learners. Language is meant to be used with other people. This increased level of interaction between students stimulates a real-world situation where students have to communicate with others. This is particularly important as better spoken English will connect Myanmar students to a wider variety of job opportunities. Discussion Maley (2000) and Higgins (2012) suggest that singing and music are forms of language that are universal among humans. In most cultures, higher rising pitches in music are associated with happiness, while lower pitches are associated with sadness. Music connects different cultures through this shared similarity. This might also be a reason why music can be helpful in language acquisition. By adding musical elements into a classroom setting, students might be able to better understand the context of the materials and hence better learn the language. A better understanding of the materials helps the student learn and absorb the materials more effectively. Similarly, music and singing might make the classes more interesting, causing
SONG AS SPEECH THERAPY
22
students to be more engaged while learning. This might enhance the effectiveness of teaching and hence help students learn foreign languages faster. Studies of Christiner and Reiterer (2013) and Kleber (2010) hint at the possibility of how singing training can be applied to teaching a foreign language, especially second language pronunciation due to a singer's enhanced auditory working memory and vocal flexibility. This is made possible because good singers retain perceptual plasticity and are open to new and unusual sound combinations throughout adulthood. Singers are also more aware of the sound production process and have a vocal apparatus with good fine-tuning ability of the tongue and the lips as well as the larynx. The benefits of singing towards vocal control suggest how it could be applied towards a classroom setting. Since singing improves vocal control and speech imitation ability, it might also be applied towards to learning a foreign language. Teaching through singing might be a more effective way of teaching as it might also help students better imitate the language that they are trying to learn. Singing might improve both vocal motor control and speech imitation abilities of a student, hence also improving their pronunciation of the language that they are learning. The above factors suggest that when conducting experiments, singing can be considered as an experimental variable. It is expected that the singing group of students will demonstrate not only a better recall ability but also better pronunciation of English words when compared to the non-singing group. The more interesting classes will help with the memory of participants and might also improve their concentration. Better concentration yields better results in learning as the students will be able to absorb more information. Moreover, singing helps with one's speech imitation ability, hence helping students imitate the language better than the non-singing group.
SONG AS SPEECH THERAPY
23
Proposed Methods of Experiment Singing will help non-native English speakers achieve a greater individual rate of improvement in spoken English pronunciation compared to mere repetition. Christiner and Reiterer (2013) suggest that singing and regular vocal practice are mechanical factors that lead to better English pronunciation, specifically vocal imitation ability. There are myriad aspects of how singing might influence the acquisition of a language - this paper will examine how the memorability, rhythmicality, and even the universality of song can contribute to acquiring a foreign language. In this proposed experiment, the independent variable is identified as the teaching method of each group. The experimental group will be taught through playing the songs from the respective phrases and repeatedly singing the song with emphasis and repetition on the specifically selected phrase. The control group will be taught through simple speech repetition of the selected phrases. The dependent variable is to measure the effects of such teaching methods on improvements of pronunciation of these selected phrases. The proposed statistical analysis will be a test regarding the comparison between the proportion of successes, in this case, the proportion of accurate pronunciation for each individual within the singing and non-singing groups. The proposed experiment shall be conducted within a period of 14 days. The aim of this experiment is to test the hypothesis of whether singing helps non-native English speakers achieve a greater individual rate of improvement in spoken English as opposed to simple repetition in a short period of time. The time period is particularly important as it gives researchers brief insight into the effectiveness of implementing singing in a classroom setting. It is suggested that each group include at least 50 participants (singing and non-singing). The 50 participants ensure that the data collected will be robust and is not subjected to chance or other
SONG AS SPEECH THERAPY
24
confounding variables. The experimental group is expected to learn English through singing, whereas the other group, the control group, is expected to learn English through simple methods such as vocal repetition. Simple vocal method is being defined as repeating the English materials vocally without any musical aid. During the 14 days of instruction, the experimental group will be taught by both a singing coach and an English teacher. The singing coach will teach basic singing techniques. Basic singing techniques not only prevent students from damaging their voices, but also align with the finding of how singing might be able to improve vocal motor control. The control group will be taught through simple vocal methods, such as repeating after an English teacher. A survey will be conducted to eliminate possible confounding variables. The possible confounding variables include the students' previous exposure towards English and the family's household language. If the household language is English, the student might have an unfair advantage compared to other students. Another factor is whether the students attend tutoring or singing classes outside of school. Attending extra classes might give them an unfair advantage and affect the experimental results. The family's socioeconomic background is also an important factor as it might give certain participants unfair advantages. Each of these variables might be determinants in affecting a student's process of learning. Since each individual student's language background is different, it is important to note that the experiment will focus on the individual rate of improvement for each individual rather than focusing on the discrepancy between the experimental and control group. The two groups must be taught for exactly one hour for exactly 14 days, including weekends. All the participants are required to attend class throughout the extended 14 days. It is strongly advised to minimize discussion within groups. To prevent interaction between two groups, the experimental group and the control group should be drawn from different schools
SONG AS SPEECH THERAPY
25
where interaction and communication are less likely. Information leaked to the other group might cause students to change their way of learning and hence affect the results of the experiment. Since the experiment is proposed to be conducted in Myanmar, it is also suggested to provide 5000 kyats (local currency) as an incentive for participation. An incentive should be provided no matter where the experiment is being conducted. An incentive serves to motivate the participants to abide to the rules of the experiment. Informed consent should be obtained. Each participant should know that the experiment is completely voluntary, and one will not suffer from any consequences or any loss if one wishes not to participate. The students who wish not to participate in the experiment can still participate in the normal classes that will be taught before conducting the experiment. On the other hand, a briefing session will be conducted before the start of the experiment. Every student should understand that they would be participating in a study with different teaching methods. The briefing session should also include the possible risks and discomforts to the participants. The possible risks and discomforts include the damage to the voice due to unhealthy singing techniques and the anxiety that one might face when singing in front of the class. Each participant should also understand every procedure of the experiment, the potential benefits to the individual, the expected length of time of the experiment, the right to confidentiality, and the right to withdraw at any time without suffering any consequences. There will also be a debriefing session shortly after the end of the experiment. The debriefing session should discuss areas including the procedures and the findings of the research. Specifically, the Myanmar students should understand their specific role within the experiment. This code of ethics will be followed strictly.
SONG AS SPEECH THERAPY
26
Data Collection At least three recordings for each individual should be made throughout the experiment to monitor the individual rate of improvement in English pronunciation. Before the initial recording, each participant's English ability will be recorded and assessed. The purpose of the survey is to determine the previous amount of exposure to English each student had before the experiment. Since individual students have varying English levels, the survey will serve as a baseline in helping to measure the individual rate of improvement in spoken English. The initial recording will be recorded on at least two recording devices with good quality. Backup files should always be made in case of accidents. Recordings should be made individually without the interference of other classmates. Students should be required to read the list of phrases selected from songs without skipping any words even if they do not know the correct pronunciation; however, chances of self-correction should be given as it is part of the learning process. The dependent variable that will be compared and analyzed is the student's individual ability to pronounce the words with the correct pronunciation. The accuracy of pronunciation will be measured by an application software developed by SpeechAce, a Seattle based startup company that specializes in pronunciation analysis. SpeechAce has developed an application that specializes in speech recognition. The software can score a speech learner's accuracy in pronunciation through their patented technology and specify individual syllable and phoneme level mistakes in pronunciation. Each analysis will provide feedback on stress, intonation, and speaking rate. Therefore, SpeechAce will help analyze the accuracy of the pronunciation of the Myanmar students. The results from SpeechAce will be the data used to determine whether singing outperforms regular repetition as a more effective way of teaching English. The collected data will be analyzed through the program provided by Speechace. The data will be in the format
SONG AS SPEECH THERAPY
27
of a raw score ranging from 0 to 100 in determining the pronunciation accuracy of Burmese students when vocalizing the selected word phrases. The raw score data will be collected, and a statistical analysis shall be conducted. After the initial recording is made, every participant will receive a word list which will be given out before the start of every class and collected back before the class ended to prevent students from going home and studying it on their own. Both groups will receive the identical word list. It is strongly advised that no electronic devices/pen/notebooks should be allowed throughout the entire period of the class to prevent students from recording the list of words for purposes outside of the classroom. After a period of 7 days (mid-point), a recording will be made for both the singing and non-singing groups to determine the improvement in the pronunciation of the selected words and phrases. The purpose of the mid-point recording is to evaluate the time period required to see an observable difference between the experimental and the control group. The mid-point recording also serves as a way to eliminate the confounding variables of time. This is particularly important as the participant in the control group might be able to catch up to the progress made by the experimental group due to the long exposure towards the material. Both the experimental group and the control group after a 14-day period might achieve similar results due to the excess length of the experiment. Both groups might be accustomed to the selected words and phrases and hence will be difficult to determine the differences in results between the two groups. Therefore, the mid-point recording serves to eliminate the lurking variable of time, and the data will also be analyzed. This will help future experiments as researchers will get a better sense of the optimal time period when conducting similar experiments or for determining learning language through these different modalities.
SONG AS SPEECH THERAPY
28
After 14 days, a final recording will be made for both singing and non-singing groups to determine each individual's improvement in the pronunciation of the selected word phrases. The goal of the proposed experiment is to measure whether singing is a more effective method of helping language leaners' pronunciation than mere repetition. The analysis of the data should be focused on the combined differences of individual growth between groups to determine whether singing is a more effective method of teaching that helps Burmese individuals improve English pronunciation when compared to simple vocal repetition. Conclusion The data from the proposed experiment bears paramount importance as it wil help determine whether singing is an effective teaching method that can be applied to a classroom setting. If the results show that singing is more effective teaching method over simple vocal repetition. This suggests a variety of possibilities of how it can be implemented to the foreign English learning curriculum. With a more effective teaching method at hand, Burmese students will have a better method to learn English, improving their ability to communicate with the international world. This would signify that by learning English through singing, Burmese students will have an increased level of access to new resources that will allow them to further pursue various opportunities both economically and academically. The possible findings of this paper can also be applied to other tonal language speaking countries. Singing might bridge the differences between tonal and non-tonal languages and help speakers learn a new language more effectively.
SONG AS SPEECH THERAPY
29
References Choi, W., Tong, X., & Singh, L. (2017). From Lexical Tone to Lexical Stress: A CrossLanguage Mediation Model for Cantonese Children Learning English as a Second Language. Frontiers in Psychology. Retrieved from https://www.frontiersin.org/articles/10.3389/fpsyg.2017.00492/full This research gives insight into the acquisition of a non-tonal language (English) for tonal language speakers (Cantonese). The journal aimed to understand how Cantonese lexical tone sensitivity contributes to English lexical stress sensitivity. This journal proposes that there could be a direct correlation between Cantonese lexical tone perception and English lexical stress perception. The results of the experiment explains how lexical tone perception helps predict lexical stress perception. The results of the experiment not only demonstrate the interdependency between lexical stress and lexical tone sensitivity, but also suggests a possibility of overcoming the gaps between tonal and non-tonal languages. Christiner, M., & Reiterer, S. (2013). Song and speech: Examining the link between singing talent and speech imitation ability. Frontiers in Psychology. Retrieved from https://www.frontiersin.org/articles/10.3389/fpsyg.2013.00874/full This source demonstrates a direct correlation between singing ability and the ability to imitate speech. Both musicality and the ability to sing aid the process of language acquisition by improving pronunciation skills in foreign languages. As auditory signals are the fundamentals of day-to-day communication, the ability to sing greatly improves the ability to perceive complex signals, improving foreign language acquisition. This journal article gives background information on the links between musicality and
SONG AS SPEECH THERAPY
30
language acquisition. The author proofs that the ability to sing is highly correlated with the ability to imitate speech. Singing, therefore, should not be categorized differently from instrumental music. It can therefore be concluded that singing is an effective means that can be used and applied in teaching and acquiring a foreign language as it enhances auditory working memory and vocal flexibility. This suggests that through training, singers could achieve better control of their voices to help them imitate a speaking language. This signifies singing would be an effect way helping students learn a new language as it would not only help them in improving their vocal flexibility, but also their ability to imitate the language (better pronunciation). Ge, J., Peng, G., Lyu, B., Wang, Y., Zhuo, Y., Niu, Z., . . . Gao, J.-H. (2015). Cross-language differences in the brain network subserving intelligible speech. Proceedings of the National Academy of Sciences of the United States of America PNAS, 112(10). Retrieved from http://www.pnas.org/content/112/10/2972.full Ge’s study suggests a new idea that although intelligible speech processing is often regarded as a left-hemisphere dominant responsibility, his experiments show otherwise. Ge conducted an experiment to examine the biological differences of the brains of tonal and non-tonal language speakers. Ge indicates that although tonal and non-tonal languages share similarities within the brain such as both would activate the the inferior frontal gyrus, the posterior temporal gyrus, and the anterior temporal gyrus; however, the interactions between these areas within the brain are different. Additionally, tonal language speakers activate and additional part of their brain - an additional activation of the right anterior temporal pole. The biological differences demonstrate the differences of how tonal and non-tonal language speakers process intelligible speech. This is
SONG AS SPEECH THERAPY
31
important as such differences might contribute to various learning obstacles for tonal language speakers who are interested in learning non-tonal languages. Grasu, D. (2015, June 24). Tonal vs. Non-Tonal Languages: Chinese vs. English. In Lexington. Retrieved January 25, 2018, from http://www.lexington.ro/en/blog/item/29-tonal-vs-nontonal-languages-chinese-vs-english.html Grasu’s article explains the pronunciation difficulties that tonal language speakers such as Chinese face when trying to pronounce the syllables /r/ and /l/. This article relates to the experiment conducted by He (2014), demonstrating why such pronunciation obstacles occur. He, Y. (2014). Production of English Syllable Final /l/ by Mandarin Chinese Speakers. Journal of Language Teaching and Research, 5(4). Retrieved from http://www.academypublication.com/issues/past/jltr/vol05/04/03.pdf He’s experiment explains the various pronunciation obstacles faced by tonal language speakers when acquiring English. He specifically analyzed the reasons why the syllable /l/ causes a great amount of errors in tonal language speakers. He concludes that it is because of the phonological differences between tonal and non-tonal languages that causes the rise of such an obstacle. He also demonstrates that tonal language speakers often employs the techniques of deletion, vocalization, and retroflexion when facing pronunciation obstacles. This experiment partially explains why Burmese or Chinese speakers make various pronunciation errors through the perspective of phonology. Higgins, K. M. (2012). The Music Between Us: Is Music a Universal Language. Chicago, IL: The University of Chicago Press.
SONG AS SPEECH THERAPY
32
Higgins claims that music and language shares many fundamental similarities. She states that music and language are similar in terms of melodic contour, categorical perception, discrete scale pitches, durations of syllables, rhythm, and gestalt principles. Kleber, B., Birbaumer, N., Gruzelier, J., & Lotze, M. (2010). The Brain of Opera Singers: Experience-Dependent Changes in Functional ActivationO. Oxford Academic, 20(5). Retrieved from https://academic.oup.com/cercor/article/20/5/1144/334526 This study by Kleber et. al (2009) states how opera singers demonstrate stronger vocal control of the vocal folds due to their training which made them more sensitive to timbre, vowel quality, and intonation accuracy. Therefore, the study states that that excellence in classical singing might be accompanied by increased involvement of primary somatosensory cortex, possibly reflecting enhanced kinesthetic motor control of the vocal tract for sound production. This hints at how singing might help improve vocal control and might lead to a better ability to imitate a language. (Christiner and Reiterer). Maley, A. (2000). Interactive Language Teaching (10th ed.) (W. M. Rivers, Ed.). Cambridge: Cambridge University Press. Maley's [Poetry and Song as effective language-learning activities] suggests that both poetry and song serves as effective ways of teaching a foreign language due to memoriability, rhythmicality, performance/recitability, universality, reactional language, motivation, and interaction. Soe, T. (2015). A study of contemporary trends and challenges of English language teaching in Myanmar. In Burma library. Retrieved from http://www.burmalibrary.org/docs21/Language/Thandar-Soe-
SONG AS SPEECH THERAPY
33
2015A_Study_of_Contemporary_Trends_and_Challenges_of_English_Language_Teachi ng_in_Myanmar-en.pdf Soe summarizes and proposes the various reasons why it is important for Burmese students to acquire English. She points out that the reasons for Burmese students to acquire foreign languages such as English is because of the opportunities that it provides. English bears paramount importance as it helps bridge the gap between Burmese citizens not just economically, but also academically, as most western research is published in English. Through acquiring English, Soe states, it would connect Myanmar to the globalized world. Soe’s study also states the various methods to improve English teaching proficiency in a Burmese classroom setting and proposes future research possibilities to explore the elements that make English an important language in Myanmar. Warren, R. M., & Warren, R. P. (1970). Auditory Illusions and Confusions. Mark Wexler. Retrieved from http://wexler.free.fr/library/files/warren%20(1970)%20auditory%20illusions%20and%20 confusions.pdf The Warren and Warren experiment states a natural tendency for human to bridge irregularities of language and music through the method of normalization. Humans innately normalize irregularities in languages to achieve easier acoustic signals. Experiments demonstrate that when it comes filling in the gaps of irregularities in music and language, humans do not differentiate the two and naturally fill in such gaps. This hints at the similarities between music and language. When facing irregularities in both
SONG AS SPEECH THERAPY
34
music and language, humans have the innate tendency to react through similar means to fix them. This experiment establishes a close correlation between music and language. Yuan, H., & Gao, S. (2002, August 22). Tones. Retrieved from Western Kentucky University website: http://people.wku.edu/shizhen.gao/Chinese101/pinyin/tones.htm Yuan and Gao explain, through relative pitch, how the Chinese language applies different tones towards the same sounds. This explains the suprasegmental differences between tonal and non-tonal languages. The authors summarize that even if the words have the same sounds, the pitch of the word can drastically change the meaning. This difference suggests another obstacle for tonal language speakers when learning non-tonal languages: changes in pitch would be as significant of a change as due ( high pitch “a” and low pitch “a”) as a change in the consonants of a word ( “p” and “b”). This maybe be an obstacle for tonal language speakers as they often have to overcome their old habits of assuming different pitches convey different meanings in-order to learn English.
The Harker School | 500 Saratoga Ave., San Jose CA 95129 OofC: 3/26/18 (RM)