3.3.2. Perceptual representation: duration structure

Jegyzet elhelyezéséhez, kérjük, lépj be.!

Figure 8 plots the mean duration that could be computed for the synthesized vowel stimuli that were identified with each of the 11 response categories, separately for the L2 (left panel) and the L1 (right panel) participants. It can be seen that the two groups of EFL learners associate approximately the same durations with the vowel types. The correlation between the vowel durations assigned to the 11 vowel types by the two learner groups is r = .904 (p < .001). Therefore, we plot the mean of the two groups (large markers, connected by lines) but also indicate the data points for monolinguals and bilinguals separately. The red markers represent the seven long/tense AE vowels, while the four short/lax vowels are indicated by green markers in both panels.
 

Jegyzet elhelyezéséhez, kérjük, lépj be.!

Figure 8. Mean duration (ms) of synthesized vowel tokens identified as one of seven long/tense or one of four short/lax vowels of American English by monolingual Persian, bilingual Azerbaijani/Persian (left panel) or American native (right panel) listeners.
 

Jegyzet elhelyezéséhez, kérjük, lépj be.!

The vowel durations obtained for the L2 groups differ substantially from the durations selected by the native listeners. The correlation in assigned vowel durations is much poorer but just significant for the bilingual learners and the American native listeners, r = .542 (p = .043, one-tailed), while the correlation between the monolinguals and the controls fails to reach significance, r = .308 (p = .178, ins.). Nevertheless, all three groups associate short durations with the lax vowels. Divergence between the EFL learners and the native controls is seen prominently for a number of tense vowels. Typically, the duration of /i/ and /u/ is overestimated in the perceptual representation of the EFL learners, possibly because they are intent on signaling a difference between lax /ɪ, ʊ/ (which indeed have the shortest durations selected) and tense vowels /i, u/, so that the duration contrast is exaggerated, to compensate for a lack of contrast in vowel quality. On the other hand, the EFL learners underestimate the duration that is appropriate for the vowels /e, æ, ɑ/, which should be long in American English.

Jegyzet elhelyezéséhez, kérjük, lépj be.!

To illustrate the different weight of the duration parameter in the choice between the long/tense versus short/lax counterparts in the high vowel pairs /i-ɪ/ and /u-ʊ/, Figure 9 shows a count of these vowels as responded by our three groups of listeners to synthesized vowel durations of 200 and 300 ms.
 

Jegyzet elhelyezéséhez, kérjük, lépj be.!

Figure 9. Number of /i/ vs. /ɪ/ responses (left panel) and /u/ vs. /ʊ/ responses (right panel) by three groups of listeners for short (200 ms) and long (300 ms) synthesized vowels. Chi-square and p-values are indicated per subpanel.
 

Jegyzet elhelyezéséhez, kérjük, lépj be.!

It is obvious from the chi-square values computed on the skew in the proportions of long/tense vs. short/lax responses to the 200 and 300-ms vowel stimuli that the stimulus duration has a large effect on the responses by the Iranian groups of participants (20.4 < χ2 < 58.5, p < .001 in all cases), while the effect of stimulus duration is much smaller for the native listeners in the case of /i-ɪ/ (χ2 = 5.6, p = .019) and even insignificant for /u-ʊ/ (χ2 = .8, p = .378, ins.).
 
Tartalomjegyzék navigate_next
Keresés a kiadványban navigate_next

A kereséshez, kérjük, lépj be!
Könyvjelzőim navigate_next
A könyvjelzők használatához
be kell jelentkezned.
Jegyzeteim navigate_next
Jegyzetek létrehozásához
be kell jelentkezned.
    Kiemeléseim navigate_next
    Mutasd a szövegben:
    Szűrés:

    Kiemelések létrehozásához
    MeRSZ+ előfizetés szükséges.
      Útmutató elindítása
      delete
      Kivonat
      fullscreenclose
      printsave