Slide Set 10-Tone Representation and Derivation-F2024
Tone: Representation and Derivation
1. Introduction to Features and Segments
Two views regarding segments:
Bundled Features View: Segments composed of features bundled in a matrix (e.g., +F1 and -F1)
Autonomous Features View: Segments defined by distinct, independent features
2. Implications of Feature Autonomy
Autonomy means shared features affect all segments with that feature:
Example: A rule for a feature impacts all segments sharing it.
Features can be maintained even if the associated segment is lost.
Rules may require identical feature specifications across sequences of segments.
Rules may prohibit sequences from sharing identical labels.
Evidence from Tone: Tone reinforces the autonomous nature of features against the view of independent feature matrices.
3. Understanding Tone Languages
Definition: A tone language uses pitch contrastively to differentiate meaning between morphemes.
Types of Tone Languages:
Register Tone Languages:
Utilize level tones (e.g., high, mid, low)
May also feature contour tones derived from level tones.
Example: Nupe (Volta-Niger):
Level:
High: ba + H (bitter)
Mid: ba + M (cut)
Low: ba + L (count)
Contour:
Rising: ba + LH (indeed)
Falling: ba + HL (defamation)
Contour Tone Languages:
Have both contour and level tones but contour tones are not based on level tones.
Example: Mandarin:
ma + high level (55) 'mother'
ma + rising (35) 'hemp'
ma + falling-rising (214) 'horse'
ma + falling (51) 'scold'
Focus Area: This lecture emphasizes register tone languages.
4. Tone Identification in Register Tone Languages
Tone Features:
Tones distinguished by three features: High (H), Mid (M), Low (L).
Contour tones can represent pitch variation (e.g., HL, LH).
Tone Transcription Conventions:
Level tones:
H: bá (high)
L: bà (low)
Contour tones:
HL (falling): bâ
LH (rising): bǎ
M: bā or ba (mid)
Tone-Bearing Units (TBUs):
Typically vowels bear tones, but some consonants can as well.
Example: Efik (Benue-Congo):
èdèp ‘you (sg) buy’
édèp ‘s/he buys’
ńdèp ‘I buy’
5. Autosegmental Representation of Tones
Autosegmental Phonology:
Tones not represented in a linear manner; they occupy a separate tier from other features.
Association Lines:
Show connections between tonal features and segments, allowing for complex mappings.
6. Examining Relations in Mende
Examples of tone relationships:
One-to-One: [ŋ͡ɡílà] 'dog'
One-to-Many: [pɛ́lɛ́] 'house' (single H)
Many-to-One: [m͡bû] 'owl' & [m͡bǎ] 'rice'
7. Underlying Specifications of Tones
Underlying Representations: Each morpheme identified with a tone or toneless.
Floating Tones: Not directly linked to segments, e.g., /ma, L/
Prespecified Tones: Indicate tone tied to a specific segment, e.g., /ma-L/
8. Universal Association Convention (UAC)
Key Principles:
Tones associated with TBUs one-to-one, left-to-right or right-to-left.
Association lines must not cross (No Crossing Constraint).
Parameters leading to tone spreading (sharing) or tone docking (creating contour tones).
Each TBU must link with at least one tone (or more).
9. OCP and Derivations in Tone Languages
Obligatory Contour Principle (OCP): Identical adjacent tones prohibited within a morpheme or word,
OCP Outcomes: Represents various conflicts in derived forms needing resolution.
10. Repairing the OCP
Options for OCP Repairs:
Substitution: Adjusting tones to prevent adjacency violations.
Deletion: Removing tones to align with OCP requirements.
Example Cases:
Mende tone patterns highlight strategies for managing adjacent identical tones, demonstrating rules and derivations clear through structure.
Conclusion
Tones in register tone languages demonstrate the complex navigation between representation, rules, and phonological behaviors that reveal deeper structural functionalities of language.