Marcus, Mitchell P.Santorini, BeatriceMagerman, David2023-05-222023-05-221990-07-012007-08-23https://repository.upenn.edu/handle/20.500.14332/7510This paper reports on one of the first steps in building a very large annotated database of American English. We present and discuss the results of an experiment comparing manual part-of-speech tagging with manual verification and correction of automatic stochastic tagging. The experiment shows that correcting is superior to tagging with respect to speed, consistency and accuracy.Report