The Power of Regularity-Preserving Multi Bottom-up Tree Transducers

The expressive power of regularity-preserving multi bottomup tree transducers (mbot) is investigated. These mbot have very attractive theoretical and algorithmic properties. However, their expressive power is not well understood. It is proved that despite the restriction their power still exceeds that of composition chains of linear extended top-down tree transducers with regular look-ahead (xtop), which are a natural super-class of stsg. In particular, topicalization can be modeled by such mbot, whereas composition chains of xtop cannot implement it. However, the inverse of topicalization cannot be implemented by any mbot. An interesting, promising, and widely applicable proof technique is used to prove those statements.