Lability and spontaneity
In my article, I address the factors which favor using a verb as labile (both transitive and intransitive, with no formal change required).
Haspelmath (1993) proposes that the key feature which conditions a way of marking (in)transitivity of verbs in the transitive / intransitive verb pair is the spontaneity parameter.
However, the statistical analysis of Haspelmath’s data shows that for labile / ambitransitive verb the main parameter is the lexical semantic class of the verb, not the degree of spontaneity. This lets us discover a more general principle: phenomena which are not purely grammatical, but rather lexico-grammatical (as lability) depend on lexical features, not on generalized grammatical or semantic parameters.