LCD720

1
LCD720 02/04/09

• Research on teaching and learning pronunciation

2
Announcements

• New classroom RA 202

3
Reflection

• Exercise 2, p. 32/33 Rank order the ESL learners
  based on
• Age
• Amount of exposure
• Sociocultural factors
• Affective factors, incl. motivation
• Other factors

4
A short history of L2 and pronunciation research
and theories

• L2 research (in general)
• Contrastive Analysis
• Error Analysis
• Interlanguage research
• Pronunciation research
• Markedness theory
• Language universals
• Information processing theory

5
1. Contrastive analysis

• Where L1 and L2 are the same easy to learn
• Where L1 and L2 are different difficult
• For example
• L1 has /l/ and /r/, L2 has /l/ and /r/ easy to
  learn
• L1 has /l/ but not /r/, L2 has /l/ and /r/
  difficult to learn /r/
• Negative transfer from L1 plays a role in
• general segmental features (like aspiration), and
• suprasegmental features (like intonation and
  rhythm)
• Problem CA doesnt always make the correct
  predictions (e.g., directionality)

6
2. Error analysis and avoidance

• In addition to contrastive analysis Error
  Analysis
• Which predictions are borne out? How difficult
  are these problems?
• Error analysis looks at
• Interlingual errors (L1 interference)
• Intralingual errors (e.g., overgeneralization)
• Developmental errors (similar to L1 acquisition)
• Problems
• EA focuses on what is wrong, not what is right
• EA ignores avoidance (e.g., difficult words)

7
3. Interlanguage

• Interlanguage The rule-governed and systematic
  second language knowledge of learners
• Development of interlanguage is determined by
• L1
• L2
• Universals
• Communication strategies (e.g., better
  performance in more formal situations)

8
4. Markedness theory

• In any pair of sounds, one is more basic,
  neutral, frequent, earlier acquired, etc. than
  the other
• One member of the pair is unmarked
• Example
• English allows /p, t, k/ and /b, d, g/ in
  word-final position
• But German allows only /p, t, k/ in word-final
  position
• So English is more marked than German (in this
  respect)

9
4. Markedness theory

• Markedness can predict directionality
• Word-final /b, d, g/ in English (marked) are more
  difficult to pronounce for L1 German speakers
• Word-final /p, t, k/ in German (unmarked) is
  easier to pronounce for L1 English speakers
• Contrastive analysis does not make such
  predictions about directionality

10
4. Markedness theory

• /?/ can occur in syllable-final position in
  English (sing)
• /?/ can occur in syllable-final and
  syllable-initial position in Vietnamese (ngang)
• Which language is more marked (with respect to
  /?/)?
• Which learner has more difficulty?
• English speaker learning Vietnamese, or
• Vietnamese speaker learning English

11
5. Language universals

• Language universals properties all languages
  have in common
• Principles and parameters (Chomsky)
• Implicational universals
• Examples of universals
• All languages have consonants and vowels
• Consonants contrast in manner, voicing, etc.

12
5. Language universals
Fricatives are more marked than nasals. Stops are
unmarked.

• Implicational hierarchies
• stops gt nasals gt fricatives
• Languages that have nasals also have stops
• Languages that have fricatives also have nasals
  and stops
• Predictions for acquisition
• Stops are acquired before nasals nasals are
  acquired before fricatives
• Initially, fricatives are replaced by stops

13
5. Language universals

• Determine whether and how the following
  universals may affect ESL pronunciation teaching
• Stops gt fricatives gt affricates
• If a language has voiced obstruent phonemes (/b,
  d, g/), it will also have voiceless obstruent
  phonemes (/p, t, k/).
• Front vowel phonemes are generally unrounded,
  while non-low back vowel phonemes are generally
  rounded.
• What does this imply for rounded front vowels?

14
6. Information processing theory Schemata

• Tendency to interpret new information in terms of
  existing knowledge structures schemata
• E.g., schema of going to a restaurant
• Schemata also influence processing of phonology
• A L1 Spanish speaker may hear eski instead of
  ski, because that what s/he expects to hear

15
6. Information processing theoryAutomatic vs.
controlled

• Automatic vs. controlled processing
• Controlled processing requires attention and
  awareness
• Automatic processing is not controlled or
  inhibited by other processes
• Cf. learning to drive a car with manual
  transmission
• With controlled processing you have to think a
  lot its easy to make a mistake, like switching
  the order of steps or forgetting a step (e.g.,
  use the clutch)
• With automatic processing you dont have to
  think however, its difficult to change
  automatic behavior

16
6. Information processing theoryModes of
learning

• Three modes of learning
• 1. Accretion (or accommodation) adding new
  information
• 2. Restructuring changing the system based on
  existing patterns
• 3. Tuning further modify the system
• Example L1 Portuguese, L2 English
• Portuguese has /i, e, ?, a, ?, o, u/
• Accretion add /?, æ, ?, ?/
• Restructure /i, e, u/ may shift position
• Fine-tuning approaching native pronunciation

17
New directions

• So far Focus on individual sounds (vowels,
  consonants)
• More recent research
• Intonation
• Rhythm
• Connected speech
• Voice quality
• Well discuss these in more detail later

18
Intonation

• American English has a three tone system of
  intonation contours
• Other languages, like Spanish, may have only two
• To the ear of a NAE speaker, this sounds
  uninterested and bored
• NAE speaker
• This is a book
• Japanese speaker

? ? ? ?
? ? ? ?
19
Rhythm

• Stressed syllables are longer than unstressed
  syllables
• This is a common problem for ESL learners
• It is possible to learn to produce the
  appropriate rhythm, but it may take extensive
  practice
• E.g., scaffolding from single items, to phrases,
  to longer stretches of speech

20
Connected speech

• Sounds are affected by other sounds that come
  before or after them, e.g.,
• Flapping rule (e.g., water)
• Vowel reduction in unstressed syllables
• Consonant cluster simplification (e.g., sixth)

21
Voice quality

• Pitch and loudness
• Spanish and Japanese speakers tend to speak with
  higher pitch and lower volume than Americans
• Language specific and sociocultural factors

22
Derwing Munro (2005)

• Questions?
• Issues in research on pronunciation teaching that
  interest you
• How would they affect your teaching practices?
• ?

23
Next week

• Read Chapter 3 (The consonant system)
• Exercises 2 (p. 30), 1 (p. 31)
• Bring a hand mirror