Spoken Language: Pronunciation and the Development of Complexity, Accuracy, and Fluency in an Intensive English Program

This chapter will focus on some aspects of spoken language development and the influence of sound systems on spoken output in an IEP. As with previous chapters, the discussion in this chapter cannot be a comprehensive overview of speaking development in an IEP. Rather, the chapter is indicative of some questions and concerns that the Pitt IEP faculty and graduate students, as well as other researchers, have addressed as they seek to understand development and improve instruction. The goal is to provide an overview of some of the findings that have emerged that might stimulate thinking and review for IEP teachers. This chapter only briefly considers some issues in phonetics and phonological development. A complete review of phonetics and phonology for IEP teaching is obviously beyond the scope of this book, but if a review is desired, some suggestions for reading are provided.

The principal focus of the chapter is on the ability of learners to produce spoken output in the IEP viewed through the lens of hypotheses about complexity, accuracy, and fluency development. CAF research has generated a great deal of interest in the second language acquisition literature and so this chapter is an opportunity to show how these concepts have been applied to speaking in an IEP. The data on this topic are derived from recorded speaking activities described in McCormick and Vercellotti (2013). Some of these data are available free of charge at https://slabank. talkbank. org/access/English/Vercellotti.html. Additionally, we will consider research in the IEP that was specifically designed to enhance students’ fluency (de Jong & Perfetti, 2011; de Jong & Vercellotti, 2016). Finally, the chapter also considers the role of morphophonology and orthography. While not directly related to speaking, the sound system of the LI of learners influences writing systems and, as a result, can affect both pronunciation and reading. The IEP faculty, like many before them, have noticed for some learners a disconnect between high spoken/lexical proficiency and their ability to spell accurately. We review research that probes this issue using experimental techniques from psycholinguistics.

A Brief Review of Topics in Pronunciation: Phonetics and Phonology

An important concern in second language speech articulation is pronunciation, the acoustic properties of speech, and the phonological systems that influence each other in multilingual speakers.¹ The influence of the first or other languages on L2 pronunciation is often very clear even among very advanced, very fluent speakers. One obvious point about ‘accent’ that is worth repeating is that accent should not (always) be seen as a negative. Clearly, expert speakers of English themselves vary widely in pronunciation across the English-speaking world. Even within one country, the United Kingdom, for example, a large range of varieties exists, not to mention local, ‘native’ standard varieties of English in South Asia, Southeast Asia, Australia, New Zealand, and of course the Americas from Alaska, Canada, the United States, the Caribbean, and South America. Although eliminating accent completely should not be the goal of instruction (Derwing & Munro, 2005), contextually appropriate comprehensibility and intelligibility are very important (Derwing et al., 2014). As Derwing and Munro (2005) pointed out, the SLA literature had not provided a great deal of guidance about pronunciation instruction, which is yet another disconnect between theoretical research and classroom instruction. They stressed that reference to good research ‘can help teachers and learners set realistic goals’ (p. 384), given that age effects in pronunciation are almost inevitable. The point about realistic goals is important. In the Pitt IEP, instructors are encouraged to provide sensitive, corrective feedback on pronunciation, but the focus on pronunciation is made during communicative activities.

In order to provide effective feedback, teachers need to be aware of the complexity of human language sound patterns. Individual sounds may be more or less difficult depending on the LI of the learners. One issue is that languages may appear to have similar sounds but differ very slightly in articulation and thus cause problems in sounding ‘native-like’. This situation is especially true of sounds that don’t involve complete closure of the oral tract: /s/ and /.[/ (sip vs. ship, for example) can be difficult for learners from southern China whose LI dialect does not make the distinction or Japanese learners for whom the LI /S/ is a predictable variant of /s/ in many cases. Of course, there are well-known issues with tense and lax vowels - the often taught ship /Jlp/ vs. sheep /Jip/ distinction. However, importantly, some ‘mistakes’ may be more crucial than others in affecting comprehensibility. For example, [d] for /6/ in ‘the’ might be less problematic than [p] for /b/) in ‘bit’ (Derwing & Munro, 2005). Thus, teachers need to consider which segments to focus on based on how important for intelligibility sounds are.

In addition, rather than being solely a function of pronunciation of individual segments, lack of intelligibility may also reside in failure to command the supra-segmental constraints on syllable structure (Hancin-Bhatt & Bhatt, 1997; Hansen, 2004; Juffs, 1990; Li & Juffs, 2015; Young-Scholten & Archibald, 2000). For example, it is well known that speakers of Lis that limit the numbers of segments at the ends of syllables tend to delete or epenthesize oral stop segments at the ends of words in English, hence, ‘bag’ [bQg] becomes ‘bage’ [bQg].

Language learners whose first languages are Chinese, Korean, and Japanese are particularly noted for changing syllable-final structure in English when speaking. All of those languages restrict syllable final consonants compared to English. English is unusual in that it allows up to four consonants in the coda, for example, ‘sixths’ [slksOs].² It is important to realize that native English speakers also reduce the number in fluent speech, but epenthesis tends not to be a strategy that LI English-speaking children use - they more often delete consonants. Li and Juffs (2015) pointed out, however, that the source of LI influence may in fact be different and that the patterns of modification vary based on the first language. While teaching in the IEP, Li had noticed that the students from Japan, China, and Korean seemed to be pronouncing words differently even though they were predicted to make the same changes based on a simple comparison of the LI and L2 English possible syllable structures. Li therefore looked at the RSA data, which were in .wav format, using acoustic analysis and coding using PRAAT (speech analyzing freeware) (www.fon.hum.uva.nl/praat/).

Tables 6.1 and 6.2 summarize our first analysis of the types of changes each language group makes.

Table 6.1 shows the Japanese speakers, while making a similar number of total errors, actually preserved more consonant codas than the Chinese and Korean speakers, who tend to delete more. In addition, Table 6.2

Table 6.1 Results of the Number of Preserved and Deleted Consonant Codas

Table 6.2 Results of the Number of Errors From Preserved Consonant Codas

LI Epenthesis

Aspiration

Devoicing

Japanese 134 (21%)

Chinese 154 (22%)

Korean 95 (18%)

• 117(19%)
• 33 (5%)
• 27 (5%)
• 269 (43%) 186(27%) 187(35%)

Spoken Language 173 shows that even when they preserve a consonant in syllable final position, Japanese speakers devoice that segment or aspirate that segment more often than the speakers of the other two languages. Note that both aspiration and devoicing are laryngeal features, whereas epenthesis involves adding a whole extra vowel segment.

Li and Juffs (2015) suggested that this pattern could be the result of two factors: first, Japanese is a ‘mora’ timed language. Briefly, this means that the timing units of Japanese at the level of syllable and mora are different. The following words in 1 exemplify how the mora and syllable are perceived differently by LI Japanese. The word for Tokyo ‘too-kyo’ has two syllables, Amazon ‘a-ma-zon’ has three syllables, and America ‘a-me-ri-ka’ has four syllables; yet all three words for Japanese speakers have four morae (adapted from Kubozono, 2002), with the ‘long’ vowels in ‘Tokyo’ and the nasal in Amazon each having the status of a whole mora.

1. a. To-o-kyo-o b. a-ma-zo-n c. a-me-ri-ka

Thus, the timing units of Japanese could be influencing their retention of consonants because the Japanese speakers are using a different unit of speech organization in their L2 English. Second, devoicing of vowels in word final position is common in Japanese inflectional morphology -thus a word final devoicing ‘strategy’ is already available to Japanese speakers, who may utilize this in indirect transfer to make syllable final consonants easier to pronounce. The pedagogical implication of these findings is that instruction on pronunciation should take account of LI influence, but for all learners, the role of syllable structure and intonation are important in addressing pronunciation challenges at the level of the segment. Thus, interventions that only address segments may be less effective than interventions that address the whole ‘melody’ of the language.

In this regard, it is important to recognize that intonation patterns in sentences may rely on different acoustic cues across languages: for example, falling pitch is used in English to indicate clause boundaries, while rising pitch is often used to indicate a question in a yes/no question: compare Mary played tennis and John played chess, where the pitch on tennis and chess falls; in contrast, Did John play chess? the pitch rises on chess. In both cases, chess means the same thing. In many tonal languages, pitch movement changes lexical meaning. For instance, in Chinese ‘ma4’ with falling tone means ‘scold’, whereas 'ma3’ with falling-rising tone means ‘horse’. Tone also plays a role in many African languages and interacts with voicing of segments in complex patterns, including grammatical inflection. All of these phenomena rely on control of vocal fold vibration in the larynx; thus, languages deploy physiological aspects of speech production differently for creating meaning (e.g., lexical differences

[Chinese] vs. sentential meaning [questions in English]). This physiological control of the larynx is subconscious, very finely tuned, and therefore difficult to change once it has been deployed as a key component of the sound system.

These are all just specific examples; we mention them in this amount of detail because it is important for language professionals to realize that the sound systems of human languages are complicated and can have interesting effects on the comprehensibility of students’ speech. It is very important for curriculum supervisors and teachers to build in appropriate activities to address comprehensibility, but the design of activities or ‘on-the-spot’ focus on form feedback can only be achieved efficiently if the instructor has enough background knowledge in both basic phonetics (e.g., a sound knowledge of the International Phonetic Alphabet) and the effectiveness of interventions that can improve intelligibility throughout the sound system.