[Absctract]: We introduce a method to reduce constituent parsing to sequence labeling. For each word wt, it generates a label that encodes: (1) the number of ancestors in the tree that the words wt and wt+1 have in common, ...