The Perception of Northern Cities Shifted Vowels by Local Speakers 249“a” in Figure 10.1). It moves into a position occupied by /ݞ/ in the pre-shift
system, but since /ݞ/ has backed into a position closer to the pre-shift position
of /ɬ/, the moved /͑/is now perhaps closer to fronted /ľ/ in the post-shift system
than anything else in the emerging system. Table 10.5 shows, however, that
misunderstandings are indeed based on pre-shift positions.
Table 10.5 Correct and Incorrect Vowel Identi¿ cation
Item Total ͑ ľ ࣜ Ԍ ɬ ͑ other/͑/ = /æ/^14212808 000x6
ľ^143112913 000x 3
^14234010800 0 x0
ࣜ^1430111 10 4^024 x3
Ԍ^1425810 0831 x 0
ɬ^142071 30066x2
͑ ࣜ 142x21^1091254Gray shading shows correct answers; bold type shows the most frequent
mistakes—overwhelmingly “pre-shift” items. Italic type shows possible mis-
understandings if the “post-shift” effect had been at work. For example, /ľ/ is
misunderstood most frequently as /æ/, even though /æ/ has shifted. If a post-
shift interpretation were at work, then one might expect misunderstandings
of /ľ/ as /͑/ (following the “b” line of Figure 10.1), but only one such misun-
derstanding occurred.
A re-examination of Figure 10.3, the Bark plot of the Peterson and Barney
(base of arrow) and experimental (point of arrow) vowels in question, also
shows this pre-shift effect to be at work except for the misunderstandings of
/ݞ/ as /ɬ/.
What about the individual vowels? The bottom section of Table 10.2
shows the GOLDVARB results for this factor group. What are some phonetic
facts about these vowels that might account for this order? Here is a list of
assumptions:
- Recency—the oldest changes in the NCS should be best understood.
- Phonetic clue—vowels which give some distinctive clue to their iden-
tity other than formant positions should be better understood.