Consistent histories and the measurement problem