This paper presents empirical evidence for the spread of the feature [sonorant], based on data from Bilaala (Nilo-Saharan, Chad). The analysis assumes that this feature is a dependent of the root node rather than part of the root node (as previously assumed). An alternative analysis, involving the spread of the feature [nasal], is shown to be inferior to one in which [sonorant] spreads.
Olson, Kenneth S. and Schultz, Paul H.
"Can [sonorant] spread?,"
Work Papers of the Summer Institute of Linguistics, University of North Dakota Session: Vol. 46, Article 1.
Available at: https://commons.und.edu/sil-work-papers/vol46/iss1/1