As a guest user you are not logged in or recognized by your IP address. You have
access to the Front Matter, Abstracts, Author Index, Subject Index and the full
text of Open Access publications.
For incomplete sentences, different structural descriptions with different levels of prediction are possible. Two existing dependency parsers have been modified to generate sequences of output hypotheses in an incremental manner. The parsing results can be characterized with respect to different criteria like the amount of predicted information, its quality, monotonicity, delay, inclusiveness and connectedness. We propose an evaluation scheme able to capture these properties and apply it to the parsers in different configurations.