abstractness & psychological reality Phonology
Download
Report
Transcript abstractness & psychological reality Phonology
CHAPTER 9:
ABSTRACTNESS &
PSYCHOLOGICAL REALITY
Main Topics
Abstract Analysis:
When Underlying Representations ≠ Surface
Forms
Valid motivations/evidence or limits for
Abstract Analysis
Empirical Data
Synchronic
Diachronic (Historical Linguistics)
Extra Linguistic Data
Language acquisition
Abstractness
Underlying Representations (UR) are usually
connected to their Surface Forms (SF): they
appear as one form of the SF.
MengMe-
Men-
Mem-
Meny-
Menj-
Meng-
Abstractness
Abstract Analysis refers to when the UR never
shows up in the SF:
X
A
B
C
Problem: How abstract can you go?
Abstractness
Motivations for limiting abstractness:
Restriction of distinctive features.
Mental reality & language acquisition:
How children acquire language & make abstract
connections between SF & UR.
Abstractness
Motivations for limiting abstractness:
Abstractness and phonemic representation: URs
usually have broader SF distributions, based on
distinctive features.
mə- / _____[l, r, w, y, m, n, ny] = + sonorant
məŋ / __ [k,h,g,a,i] = no natural class
Thus, [məŋ] has a broader distribution.
Initial Principles Limiting
Abstractness
1. The UR of a morpheme must actually be
pronounced as such in some SF containing
the morpheme.
Problem: Too restrictive in some languages.
Example: Palauan. Initial vowels unstressed -> ə
Initial Principles Limiting
Abstractness
2. The UR of a morpheme must contain only
segments actually pronounced as such in
some related word containing the morpheme.
Problem: There are some languages with examples
where this doesn’t happen.
Example of Abstract Analysis:
Kimatuumbi
Focus:
Lexical class 3 prefix /mu/. Never surfaces as /mu/
Surfaces as [m], [n], [ŋ], [mw]
Example of Abstract Analysis:
Kimatuumbi
Analysis:
2 processes for nasals. Examples
Prefix /ɲ-/ -> noun & adjectives class 9
Prefix /mu-/ -> second plural subjects
Comparison between the phonological process of
these two prefixes in comparison to lexical class 3
prefix /mu-/
Example of Abstract Analysis:
Kimatuumbi
Differences between /ɲ-/ and /mu-/
1. /ɲ-/ assimilates to place of articulation of C, while
/mu-/ deletes [u], assimilates and nasalizes C (this
is optional)
/ɲ-/
bowaana -> m-bomwaana
goloka -> ŋ-goloka
/mu -/
buundike -> m-muundike or mu-buundike
laabuke -> n-naabuke or mu-laabuke
Example of Abstract Analysis:
Kimatuumbi
2. /mu -/ + nasal consonant = deletion of [u]. Whereas
/ɲ-/ + nasal consonant = deletion
/ɲ/ -> ∅
mimina -> mimina
/mu -/
mimiine - > m-mimiine
3. /ɲ-/ assimilates and turns voiceless C into voiced,
/mu-/ only assimilates, no voicing change.
/ɲ/
tinika -> ndinika
/mu-/
teleke -> nteleke
Example of Abstract Analysis:
Kimatuumbi
4. /ɲ-/ causes following glide to become a voiced stop,
while /mu -/ changes glide to nasal.
/ɲ/
wikilya - > ŋ-gwikilya
/mu -/
yikiti -> ɲ-ɲikiti
5. /ɲ-/ has no effect on vowel length, /mu-/ become
[mw] and vowel is lengthened
/ɲ/
epeesi -> ɲ-epeesi
/mu-/
eleew -> mw-eeleew
Example of Abstract Analysis:
Kimatuumbi
So what about /mu-/ class 3 noun prefix? It
behaves the same way as nouns & adjectives
class 9 prefix /mu-/
Nasalizes voiced consonants
1.
2.
3.
4.
laabuka -> n-naabuka
[u] deletes before following nasal
mulika -> m-mulika
Does not voice following voiceless Cs.
teleka -> n-teleka
Shows up as [mw] before vowels & lengthens the
vowel.
epuka -> mw-eepuka
Thus the UR of this prefix should be /mu-/ even if
it never surfaces as [mu-].
Questionable Abstract Analysis:
English
Chomsky & Halle (1968): dipthong [ɔy]
derives from [ɶ].
Problem: does not account well for
alternation, compared to Kimatuumbi
analysis.
Alternation is based on sets of words in
which have questionable synchronic
relatedness. E.g. joint-juncture, pointpuncture, boil-bullion (pg 271).
Independent evidence: Historical
Restructuring
Paul Kiparsky (1968):
a) Absolute neutralization: distinction between
phonemes in morphemes neutralized in all
cases.
b) Contextual neutralization: distinction between
phonemes in morphemes neutralized in specific
contexts. E.g. in, im, iŋ, iɲ = in only for this
specific English prefix.
Contextual neutralization more common and
more ‘real’, absolute neutralization seen as
constructed.
Independent evidence: Historical
Restructuring
Kiparsky: Historical sound change can be
used as a test of abstract analysis.
Case of Yiddish, language of Jewish
immigrants in Germany.
Picked up rule of devoicing final stem consonant
from German
tag -> tak (day)
tagn (days)
But was later lost in Northeastern Yiddish =>
reversal of sound change
tok -> tog based on the plural togn
Independent evidence: Historical
Restructuring
Exception:
gelt (money) did not reverse to geld.
Analyis:
no plural form with the voiced consonant, unlike
tagn/togn.
The singular form subject to devoicing rule, plural
was not.
In the case of tak/tok - tagn/togn, Yiddish children
could abstract from the plural the UR of tag/tog
which was resurfaces when devoicing rule was lost.
Yiddish children never heard the UR of ‘geld’ during
language acquisition.
Conclusions
Two methods of abstract analysis has been
shown:
a. Domain-internal: Kimatuumbi example, based on
feature constraints.
b. Domain-external: Yiddish example, based on data
from historical language change and not synchronic
phonological data.
Conclusions
Abstract analysis of URs must be motivated
by:
Limitations of distinctive feature theory, in the
case of domain-internal analysis.
Principles of child language acquisition. Explains
whether the URs are plausible based on how
children learn language from spoken experience.
Abstract analysis is often abstract, in the
sense that it is often difficult or even
impossible to truly prove the psychological
reality of proposed URs.