The Computation of Prosody
W. J. Idsardi, 1992
This thesis presents a new theory of metrical representations and computations. This theory emphasizes that the metrical grid is a separate module of the phonology, devoted to the calculation of partitionings of phonological elements. The metrical grid consists of parallel tiers composed of three kinds of elements: grid marks and left and right boundaries. A single boundary serves to define a metrical constituent: a left boundary creates a grouping of the elements to its right, a right boundary creates a grouping of the elements to its left.
The calculation of the metrical grid is accomplished through the use of both rules and constraints. This division of labor accounts for observed properties of stress systems in a succinct manner. Metrical rules apply successively in a derivation, thus modelling the functional character of metrical structure assignment. The constraints prevent the application of metrical rules that would generate universal or language-particular disfavored configurations.
The interface to the metrical grid module is controlled by two parameters of projection, which provide the initial grid marks and boundaries. Further parameterized rules of Edge marking, Iterative constituent construction and Headedness complete the construction of the grid.
This theory allows the derivation of Extrametricality effects through the interaction between Edge marking and Iterative constituent construction. Constraints against particular configurations yield both clash effects and a ternary parsing ability.
Further, the Edge marking parameter provides the requisite formal power to deal with stress introduced by specific morphemes in such languages as Turkish, Macedonian, Polish, Russian, Cayuvava, Shuswap (Salish) and Moses-Columbian (Salish).
Thesis Supervisor: Morris Halle
Title: Institute Professor
Table of Contents
1 Basic mechanisms for constructing metrical grids 1
1.1 Introduction 1
1.2 Koya 2
1.3 Warao and Weri 6
1.4 Tubatulabal 8
1.5 Macedonian I 9
1.6 Selkup and Khalkha Mongolian 10
1.7 Turkish 11
1.8 Polish 14
1.9 Macedonian II 16
2 Constraints on the construction of metrical grids 20
2.1 Clash avoidance 20
2.1.1 Malayalam and Wolof 20
2.1.2 Garawa 22
2.2 Edge avoidance 24
2.2.1 Latin 24
2.2.2 Western Aranda 25
2.3 Ternary constituents 26
2.3.1 Cayuvava 27
2.3.2 Chugach Alutiiq 30
2.4 Parenthesis deletion 32
2.4.1 Diyari 32
2.4.2 Winnebago 34
3 Conflation and multiple metrical parameters 38
3.1 Introduction 38
3.2 Bidirectional stress rules 38
3.2.1 Cahuilla 38
3.3 Conflation and circumsciption 39
3.3.1 Lenakel 43
3.3.2 Auca 46
3.3.3 Klamath 48
3.3.4 Seneca 50
4 Case studies in lexical stress 51
4.1 Introduction 51
4.2 Russian 51
4.2.1 Noun inflections 52
4.2.2 Jers and stress shift 54
4.2.3 Stress retraction 56
4.2.4 Verbal inflections 58
4.3 Stress in Shuswap and Moses-Columbian 64
4.4 Shuswap glottalized sonorants 70
5 Rules and representations in metrical theory 74
5.1 Introduction 74
5.2 Representations of bracketed grids 74
5.2.1 Boundary representations 74
5.2.1.1 Left and right parantheses 75
5.2.1.2 Foot separators 76
5.2.2 Grid marks as feature matrices 77
5.2.3 Autosegmental representations 77
5.2.3.1 Autosegmental edges 77
5.2.3.2 Autosegmental feet 78
5.2.4 Summary 80
5.3 Constraints versus rules 80
5.3.1 Iterative paranthesis insertion 80
5.3.2 Constraint-only metrical theory 83
6 Conlcusions 90