Table 10.

Generalizations that have emerged through the phonetic study and evidence for each of the generalizations.

	Generalization	Evidence
Boundary Effects	1) Trills at the boundary of the domain are longer than trills inside the domain. Domain-Initial trills are shorter than domain-final trills, while both intervocalic and Preconsonantal trills are nearly the same length.	Figure 3, Table 3
	2) Domain-Initial trills have a higher chance of having a large number of contacts than any of the other positions. Domain-medial trills have a higher likelihood of approximance and less likelihood for full trilling than trills at the boundary.	Figure 2 Table 2
Coarticulation Effects	3) Vocalic contrasts can be detected from beginning to end of trills, except for initial part of domain-Initial trills, followed by the end of the final trill and the end of the preconsonantal trill. The latter is an interaction between boundary and coarticulation effects.	Figures 7,8,9 Table 8,9
	4) Trills contiguous to a back vowel are slighlty longer than trills near a front vowel.	Figure 3 Table 3
Contrast Effects	5) At all points in a trill, regardless of position in word, SA has an influence on the trill.	Figures 4,5 Tables 5,6
	6) The likelihood of approximant realization is higher in palatalized trills, while the likelihood of a high number of contacts is higher in non-palatalized trills	Figure 2 Table 2
	7) Palatal vs. velar Secondary Articulation does not affect the length of the trill.	Figure 3, Table 3
Syllable Effect	8) Onset palatalized trills exhibit a rise in palatalization, whereas coda palatalized trills exhibit steady palatalization throughout	Figures 4 and 6