Coleman(1995: 376) and Palmada (1995: 308–310) have pointed out that there is a cross-linguistic tendency for fronted velar stops to affricatize in the palatal region, and have both proposed a [palatal] → [continuant] implication to account for this dependency.