... We initialize the sampler by randomly as-signing each string in the training corpus a randomtree generated by the grammar. Then we randomlyselect a string to resample, and sample a parse of ... grammar, in which the subtrees are specified in advance, in an adaptor grammar the subtrees, aswell as their probabilities, are learnt from the train-ing data. In order to make parsing and inferencetractable ... terminal in the grammar. If Ais a terminal then GAis the distribution that puts all of its mass on the unit tree (i.e., tree consisting of asingle node) labeled A. If A is a nonterminal thenGAis...