Lecture 8 Flashcards
Regression
sound gets “worse”
as they learn more words they seem to get worse at producing certain sounds
down and stone
later, new oral vowel nasal words: instead of producing the initial consontate
nasal assimilation: take properties of one sound in a word and mush it onto later part of the word - assimilation of sounds at the end: meeeans (minz) instead of beans and nance (naens) instead of dance
“took over” more accurate forms
what’s going on: as they’re learning more words they’re overdoing similar across word patterns, figuring out how much they can change words and still have them be similar to the adult for (nown (naewn) for down and known (non) for stone)
issues in acquiring your language’s sound system
- not simple adding on of perfectly-prononunced words
- sometimes pronunciations change for the better or worse (saying r’s for w’s)
- sounds can show up on some words, but not on others (rocket vs. bwead for bread)
progressive phonological idiom
some words are accurate “too early”
child doesn’t produce sound “h” except in “hi” and “hello”
can’t say horse, hose, hamster: ‘orse. ‘ose, ‘amster
explanation: maybe there is no “h” they’re just learning entire word forms
child as a problem solver: how do I sound like the people around me?
- avoidance of difficult sounds (Sedaris avoiding “s” sounds)
- exploit sounds you like
- replacement of one sound with another: simplifying the information of the word: less info that you need to control motorally (wew vs. well)
- rearrangement of sounds in words (pasketti instead of spaghetti)
- one word, or approximate whole phrase
biological constraints
much easier to physically produce stop consonants than fricatives (dah vs. zah)
perceptual feedback
can’t do much if you don’t have an acoustic model from adults and internal feedback
need an acoustic model and internal feedback (that doesn’t quiet match what i hear in my head): error signal
e.g. if you’re hearing impaired, only get external feedback
internal feedback: immediate, running into an electric fence
external feedback = temporally delayed (time in between producing the sound and then relying on other person to give you feedback), poison oak (takes a day or so to show up, but doesn’t effectively teach you to avoid the plant because it’s not immediate - you don’t necessarily know where you got this horrible blistering rash)
internal reward for sounding like those around you
age 3 progressions
vowels and many consonants
tought consonants: r, l , v, th, dth
should be intelligible by….
4
by age 8 progression
slower speech rate
mostly adultlike
describing kids’ pronunciations
- rules to translate adult form into kid form: e.g. “delete /s/ in clusters” = noman instead of snowman
- constraints: e.g. avoid consonant clusters entirely
- preferred forms as templates: consonant vowel consonant
make sense of seemingly unrelated erros
caveat to describing kids’ pronunciations
how to fix?
what we hear is filtered through our own phonological system!
maybe if we use fine grained analyses we can get around that
need for multiple measurements of the same word at a give age (kids gradually add in correct sounds)
inherent variability
cuation in mispronunciation
referential style: one word at a time
expressive style: fluent speech that is less intelligible
regional variants: define correctness by dialect being acquired: are they learning the dialect they’re exposed to?
variants in contexts: reading list vs. casual conversation
-
kids pronunciations: things that are tough for kids
consonant clusters
- closed syllables are hard
coda (end of syllables) consonants
long words
initial weak syllables - guitar is hard, they’ll say tar instead
place assimilation example
gog for dog
tat for cat
putting end consonant in front
assimilation in adult speech
lean bacon to leambacon
green peas –> greempeaz (grimpiz)
NOT an error just fluent speech
