| Teaching Articles Home | PDF File |

Volume 39
Teachers Articles
October 20009
Article 2

Article Title
“Drill, baby, drill”:
Exploring a Neurobiological Basis
for the Teaching of Segmentals in the ESL/EFL Classroom
                                           
Author
 Anna Dina L. Joaquin
 University of California, Los Angeles

Bio Data:
Anna Dina L. Joaquin is a doctoral student in Applied Linguistics at UCLA. Her experiences with ESL teaching led her to be interested in the neurobiology of language use and language learning. Her research includes resonance and alignment in conversation and the role of caregiver-child attachment and bonding in language acquisition.

Abstract
Segmentals are the individual sounds of a language that can be broken down and focused on for instruction. Problems with segmentals can cause miscommunication, embarrassment, and affect confidence and motivation. Although teaching pronunciation and thus segmentals have been suggested to be a crucial element of second language curriculum (J. Morley, 1991; P. Robertson, 2003; T. Thompson & M. Gaddes, 2005), this component has often been neglected in curriculum that emphasizes a Communicative Approach to language teaching (S. Krashen & T. Terrell, 1983). Current research in neuroscience demonstrates that when we hear people talk, we are actually simulating their articulation and matching their pronunciation to stored templates in our brains. Furthermore, the brain may have an ability to learn or modify these templates. This may be a neural basis for the perception and acquisition of segmental speech features. Thus, the research discussed in this paper supports the development of ESL/EFL curriculum that integrates a focus on segmental features with a Communicative Approach.

Keywords:  Segmentals, Speech Perception, Brain, Mirror Neurons, Phonological Acquisition

Introduction   
Communicative competence is “that aspect of our competence that enables us to convey and interpret messages and to negotiate meanings interpersonally within specific contexts” (Brown, 1994, p. 27). Pronunciation, which includes segmental and suprasegmental features of a language, can affect communicative competence, for a speaker’s pronunciation can determine whether their spoken discourse is comprehensible to the listener. In fact, Robertson (2003) states that “intelligible pronunciation is an essential component of communicative competence” (as cited in Morley, 1991, p.4).  Thus, research has shown that problems with segmentals can cause miscommunication, embarrassment, and affect confidence and motivation. Though teaching segmentals is purportedly a crucial element of second language curriculum (Thompson and Gaddes, 1995) “many pronunciation teachers would claim that a learner’s command of segmental features is less critical to communicative competence,” (Celce-Murcia, Brinton, & Goodwin, 1996, p. 131) as there are strategies for clarification available to speakers. Techniques emphasizing segmentals have also been rejected on grounds that adults are unable to improve their pronunciation (Brown, 1994) and that such methods are incompatible with the Communicative Approach (Widdowson, 1978; Brumfit & Johnson, 1979), which has strongly influenced language curriculum.
   Contrary to the notion that segmentals should not be emphasized, current research in neuroscience may support the integration of a focus on segmental features in ESL/EFL curriculum. Brain research shows that when we hear people speak, we are actually simulating their articulation and matching their pronunciation to stored templates in our brains. Furthermore, the brain may have an ability to learn or modify these templates. This may be a neural basis for the perception and acquisition of segmental speech features. Thus, in this paper I will discuss research that may support the development of ESL/EFL curriculum that integrates a focus on segmental features with a Communicative Approach.

Role of Mirror Neurons in Speech Perception
At the Haskins Laboratory at Yale University, Alvin Liberman and his colleagues attempted to create a device that could transform text into spoken words so that persons who had lost their eyesight would be able to “read” books, magazines, and newspapers. When Liberman tested his device among war veterans, he found that the veteran’s perception of the device’s speech output was incredibly slow – much slower than even the perception of distorted human speech. This observation led to a theory of speech which posits that the way our brain perceives speech is by simulating the talk ourselves.  According to this motor theory of speech (Liberman & Mattingly, 1985), speech perception involves the transformation of an acoustic signal to an articulatory representation through the coordination of more than a hundred muscles in the mouth.  Thus, the relationship between acoustic and articulatory forms raises the possibility that the motor system might play a role not only in producing speech but also in perceiving. Theorists suggest that when a person speaks to us, we are not only receiving an acoustic signal, but are in fact also “articulating” the phonemic sounds within our brains. This perspective has been supported with the discovery of mirror neurons in the brain.
   In the 1990s, a new class of neurons was discovered in the ventral premotor cortex (vPMC) or F5 region of the macaque monkey brain.  These neurons are observed to discharge not only when the monkey executed actions, but also when observing similar actions executed by another. Therefore, they are suggested to be the mechanism that subserves action-recognition.  They are called mirror neurons and have also been found in humans (Cochin, Barthelemy, Roux, &  Martineau, 1998; 1999). Some researchers have also suggested that this mirroring mechanism has an essential role in speech perception.
   Several studies have already demonstrated that there is neural activity in the brain, particularly in the premotor cortex (PMC), during passive speech perception. Among the first, was an functional magnetic resonance imaging (fMRI) study conducted by Buccino et al. (2001) in which participants observed a number of videotaped actions such as biting and chewing. The results of the study confirmed the speculation that mirror neurons in the PMC coded mouth actions. The activation of these areas in viewing speech has been confirmed by another study that found the same area used specifically during lip-reading, and more importantly not being activated during observations of other movements (Santi, Servos, Vatikiotis-Bateson, Kuratate, & Munhall, 2003). In another study carried out by Ferrari, Gallese, Rizzolati, and Fogassi (2003), researchers found that one-third of mouth motor neurons in the PMC become active during the execution and observation of communicative mouth gestures.
   Watkins, Strafella, and Paus (2002) by using transcranial magnetic stimulation (TMS) techniques, recorded motor evoked potentials (MEPs) from a specific lip and hand muscle. Subjects were exposed to four stimuli: continuous prose, nonverbal sounds, speech-related lip movements, and eye and brow movements. Compared to control conditions, listening to speech enhanced the MEPs recorded from the specific lip muscles. Furthermore, MEPs did not increase from the hand muscles when the subjects listened to speech. Similarly, Sundara, Namasivayam, and Chen (2001) found that visual observation of speech movement enhanced the MEP amplitude in muscles involved in the production of the observed speech. 
   In addition to evidence of neurons in the PMC activating while listening to passive speech, additional studies also suggest that neurons in this region are at work when an individual listens to specific phonological material. For example, in one study, Fadiga, Buccino, and Rizzolatti (2002), recorded the MEPs from tongue muscles in normal participants who were instructed to listen carefully to verbal and nonverbal stimuli.  The stimuli were words, pseudowords, and bitonal sounds. Either a double /f/ or a double /r/ was embedded in the middle of words and pseudowords (i.e. baffo, terra). /r/ a linguo-palatal fricative, in contrast to /f/, a labio-dental fricative, requires more tongue muscle movement. During the experiment, the participants’ left motor cortices were stimulated. Interestingly, the results showed that compared to the stimuli with /f/, listening to words and pseudowords containing double /r/ created a significant increase of MEPs recorded from the tongue muscles.
   In another fMRI study, subjects listened passively to meaningless monosyllables (i.e. /pa/ and /gi/) and produced the same speech sounds to examine whether motor areas involved in producing speech would be activated (Wilson, Saygun, Sereno, & Iacoboni, 2004). The research found that listening to speech bilaterally activated a superior portion of the vPMC. Another study confirmed the essential role of the PMC in speech perception (Meister, Wilson, Deblieck, & Wu, 2007).  In this study, the researchers used fMRI to observe the PMC during three different perceptual tasks. The first was a speech perception task which involved discriminating between voiceless stop consonants in single syllables (i.e. pa, ta, ga) at baseline. The second task was a color discrimination task. The final task was a tone perception task, which involved recognition of pitch changes. At baseline, the average percentage of correct responses was 78.9% for the speech perception task, 76.6% for the color perception task, and 85.5% for the tone discrimination task. The researchers then disrupted the activity of the premotor cortex with TMS. When this area was disrupted the correct responses for the speech task fell to 70.6%. However, participants’ color perception abilities were not disrupted and remained at 76.5%, and their tone discrimination abilities also did not decrease significantly. The results suggest that there is a strong trend toward a decrease in performance after TMS is applied to the PMC for the speech condition. Thus, in addition to the auditory system, mirror neurons in the left PMC are crucially involved in speech perception. Taken together, these studies indicate that when we listen to others, our motor speech brain areas are activated as if we were talking.

MNS and Segmental Acquisition in SLA
The studies mentioned are applicable to situations in which the participants have the same phonemic inventory.  What happens in situations in which speakers have different phonemic inventories? Researchers in another fMRI study examined neural responses to non-native (non-English) phonemes, speculating that activity in brain areas involved in transforming the acoustic signal to a phonetic code would differ for native and non-native phonemes (Wilson & Iacoboni, 2006). The researchers chose 30 phonemes for the study: 25 non-English phonemes (i.e. postalveolar click, alveolar click, uvular ejective stop, dental click) and 5 English phonemes (i.e. voiced palatal fricative, voiced bilabial stop, voiced alveolar fricative). As expected, when native English speakers were asked to produce the phoneme sounds they were able to easily produce the English-like phonemes. The results also showed, not surprisingly, that motor areas were activated during speech production activities. However, activity in motor areas differed for native versus non-native phonemes. Bilateral superior temporal auditory regions were more active when participants attempted to produce the more difficult phonemes. The results also showed that motor areas were functionally connected to the superior temporal cortex. The finding that motor areas distinguish between native and non-native phonemes suggests that these regions are sensitive to whether or not phonemes are part of the speaker’s inventory and that motor areas are actively involved in the speech perception process. 
   However, Wilson and Iacoboni (2006) suggest that our motor mouth neurons do not stop at distinguishing. When non-natives speak to natives, though their phonemes may not always be accurate, natives may still be able to comprehend what was said. Wilson and Iacoboni propose that when perceiving non-native phonemes, the PMC generates forward models of native phonemes to the superior temporal cortex, which matches auditory input to stored templates (Hickok & Poeppel, 2007). Therefore, when the speech is “not quite native” or a different dialect, the PMC provides top-down information and can facilitate the perception of less intelligible speech if there is a stored template that is similar (Wilson & Iacoboni, 2006, p. 322). This model may explain how it is possible that though a non-native’s pronunciation may be “imperfect,” natives are still able to comprehend the non-native’s speech when their pronunciation is “close enough.” Such research supports pedagogical notions that “pronunciation should be taught to a level of intelligibility rather than accuracy” (Celce-Murcia, Brinton, & Goodwin, 1996 p. 16).

   Wilson and Iacoboni’s  (2006) study also raises the question of whether mirror neurons can develop mirroring properties or whether mirroring properties are fossilized. A study by Ferrari (Iacoboni, 2008) may provide some evidence that mirror neurons may be able to learn. Ferrari and his colleagues recorded the activation of neurons in the macaque monkey as the monkey observed the actions of experimenters. Though these monkeys did not use tools themselves, the researchers found that 20% of all the recorded cells responded to actions performed with the hands and mouth and also to actions performed with tools though much more weakly. This was the first evidence of mirror neurons being activated during the observation of actions supposedly absent in the motor repertoire of the monkeys since they are not tool users. The researchers and Iacoboni (2008) propose that “it is likely that these 20 percent of mirror neurons…are the result of repeated exposure of the animals to the sight of human experimenters using tools…[which] suggests that mirror neurons can acquire new properties…” (p. 42). Though this study involved monkeys and was related to actions of tool use, it has possible implications for SLA notions related to phonological acquisition. Research has generally shown that while adults are superior learners to children in vocabulary, syntax, and literacy (Scovel, 1969 cited in Brown, 2000), when it comes to pronunciation, adults are not expected to achieve native-like abilities (Brown, 2000). However, if mirror neurons involved in speech and perception can be modified through repeated interactions, then acquiring the phonology of a second language may be more attainable than previously suggested.
   Furthermore, taken together with Wilson and Iacoboni’s study (ibid), Ferrari’s findings support notions that native language transfer can play a role in a learner’s acquisition. It is generally believed that first language transfer plays a role in the acquisition of the sounds of the second language (Tarone, 1987). If a non-native’s first language is phonologically closer to the target language, then the learner may actually have more difficulty in developing the motor template for a sound because the premotor cortex may mitigate the perception of the speech and thus interfere with learning and accuracy. This may explain, for example, why students will convert unclear input into a similar sound in their own language (Dalton, 1997). It is also possible that because the non-native’s first language may be phonologically similar to the target language, repetition, practice, and exposure to the target language may facilitate learning. Further studies of mirror neurons such as those done by Wilson and Iacoboni can hopefully provide more insight into the role of the first language in second language acquisition.

Conclusion
Success in communication does not simply rely on perceiving the segmental aspects of language; the prior discourse, gesture1, context, and suprasegmental features of language are valuable resources that participants can depend on to facilitate communication. Furthermore, when those resources fail, there is a system of repair available to speakers  (Schegloff, Jefferson, & Sacks, 1977; Schegloff, 1979). Thus, some researchers suggest that teaching segmentals is not as important as teaching other aspects of communicative competence. However, segmental features are still an aspect of communicative competence, and problems in this area can cause miscommunication, embarrassment, which can affect confidence and motivation.
   The studies presented in this paper, may provide a neurobiological basis for speech perception at the segmental level. The studies suggest that when we hear people talk, we are actually simulating their articulation in our brains through mirror neurons. We are matching their pronunciation to stored templates in our brains, and if their non-native or dialectically different   pronunciation does not perfectly match with ours, our brains “work” to find a match that facilitates comprehension. This provides some explanation as to why speakers are intelligible despite differences in dialect or problems with phonemes in their speech. Such findings also support the notion that teaching segmentals, at least to a level of intelligibility, to a non-native speaker can improve a learner’s communicative competence. In addition, research demonstrates that mirror neurons can learn. Thus, if research also demonstrates that mirror neurons involved in speech perception can be acquired, then such a finding supports the value of including a focus on segmentals (i.e. drilling) in a Communicative Language Teaching (CLT) framework (Hammond, 1995).