Class Notes (1,000,000)
CA (620,000)
Brock U (10,000)
LING (500)
katem (9)
Lecture 8

# LING 3P51 Lecture Notes - Lecture 8: Speech Recognition, Pragmatics, Praat

Department
Linguistics
Course Code
LING 3P51
Professor
katem
Lecture
8

This preview shows pages 1-2. to view the full 6 pages of the document.
Lecture 9: Issues in L2 Pronunciation Instruction (cont.)
Functional load and the teaching of pronunciation
-examines factors which should be considered in calculating the functional load
between a pair of phonemes
-uses pairs of phonemes (in received pronunciation) that are often con"ated as a
framework for statistical analysis of these factors
Brown, 1988
-issues with the use of RP? (ex. As opposed to the use of GAE-general american?)
-less of an issue when comparing consonant sounds
-but there are more vowels in received pronunciation (more diphthongs) (they
contrast
between BALM and bawm)
FL Contributing factors
1) Cumulative Frequency
-calculated by adding the individual frequencies of each phoneme in the pair
-ex. In RP the cumulative freq for the pair /e, ae/ is 11.05% (/e/=1.16% + /ae/ =
3.89%)
-pairs of phonemes with higher cumulative freqs might be more likely to result in a
loss of intelligibility if con"ated
2) Probability of occurrence
-calculated by dividing a phonemes individual frequency by the cumulative
frequency for the pair
-ex. In the above example, individual frequency of /e/ would be 7.16/11.05=.648;
individual frequency of /ae/ would be 3.89/11.05 = .352
-the closer the probability of occurrence of each member is to .50, the greater is the
potential confusion to be caused by con"ation of the pair
-if they are both equally as likely to occur, you’d expect more confusion to be
caused
-Four extreme possibilities exist here:
i) pairs with high cumulative frequency and relatively equal probability (The
condition that you’d
expect most confusion)
ii) pairs with high CF and unequal probability
iii) pairs with low CF and relatively equal probability
iv) pairs with low CF and unequal probability
3) Occurrence and stigmatization
-there are some con"ations that listeners are accustomed to making; for these
commonly occurring con"ations, intelligibility will likely not be as impacted

Only pages 1-2 are available for preview. Some parts have been intentionally blurred.

-ex. Replacement th with /d/ in “the” (we’re not unused to hearing that so it
wouldn’t cause confusion)
4) Acoustic similarity
-when sounds are acoustically more similar, its likely that the listener would be
more accustomed to recognizing the intended sound from context
-ex. /th/ and /f/ would be less ambiguity bc acoustic between those two are more
similar than vs /th/ and /s/
-Do language teachers really have this level of acoustic knowledge?
5) Structural distribution of phonemes
-con"ation of sounds will likely not lead to confusion in environments where syllable
structure/phonological rules constrains the use of one of the sounds
-if we have a rule that constrains the use of certain phones.. if the L2 learner were
to make an error and put a phoneme where it cant exist, we are likely to recognize
that it cant be “this” so it must be “this”
-ex. “ing” doesn’t occur word initially so if they say “ing-o instead of “no
6) Lexical sets
-some phonemes are not contained in many words
-ex. /u/ in book (don’t confuse this with the fat that these same phonemes might be
found in high frequency words)
-if it’s a low freq occurring sound, its less likely to cause communication breakdown
7) Number of minimal pairs
-for some phonemic contrasts, there are relatively few minimal pairs
-its unlikely that a misunderstanding will occur between these contrasts when
con"ated
-ex. There are few minimal pairs for /u/ and /u/ in book
8) Number of minimal pairs belonging to the same part of speech
-ex. English words that start with voiced /th/ are mostly function words, so they are
unlikely to be confused in context with the corresponding /d/ words
9) Number of in"ections of minimal pairs
-ex. Should we count all in"ections of fear/fare as individual examples of minimal
pairs (ex. Feared/fared, fears/fares, fearing/faring) or just count it once?
10) Frequency of members of minimal pairs
-the frequency load of a contrast in the text depends on the existence of minimal
pairs of words that are both frequent
-ex. Would, could, should, look –are all much more frequent than the corresponding
wooed, cooed, shooed, and Luke
11) Number of common contexts in which members of minimal pairs occur
-can contexts easily be supplied in which both members of a minimal pair are