Phonology Class Notes – Week 5 (Autosegmental
Phonology, Tone Spreading, and Vowel Harmony)
1. Introduction to Non-Linear Phonology
So far: we've worked with linear models – phonemes and features in sequence.
But some phenomena cannot be represented well in a single linear string.
Autosegmental phonology (Goldsmith, 1976) introduces multiple tiers of representation.
2. Autosegmental Representations
Sounds/features are placed on separate tiers and associated by lines.
Useful for representing:
- Tone
- Vowel harmony
- Nasal harmony
Association lines can show one-to-many or many-to-one relationships.
3. Tone Spreading
Tone often spreads from one syllable to another.
Example: High tone spreading in some Bantu languages.
Representation:
H tone linked to multiple vowels:
H
/\
VV
4. Tone Stability
When a vowel is deleted, its tone can remain and attach to a neighboring vowel.
Shows independence of tone from segmental tier.
5. Vowel Harmony
A process where vowels within a word agree in certain features (e.g., [±ATR], [±back]).
Common in Turkic, Uralic, and Niger-Congo languages.
Example (Turkish):
- ev-ler 'houses' (front vowel harmony)
- kap■-lar 'doors' (back vowel harmony)
6. Feature Geometry
Autosegmental approach can represent features as hierarchically organized.
Allows spreading of single features without affecting others.
E.g., [nasal] can spread from one segment to another while place and manner remain unchanged.
7. Association Conventions
Well-formedness condition: every segment must be associated with all necessary features.
No crossing association lines.
Rules can add, delete, or spread associations.
8. Benefits of Autosegmental Analysis
Captures long-distance effects.
Separates independent layers of phonological information.
Explains phenomena like floating tones and partial feature spreading.
