We describe a syntactically based salience algorithm for pronominal an
aphora resolution and a procedure for reevaluating the decisions of th
e algorithm on the basis of statistically modeled lexical semantic/pra
gmatic preferences. We report the results of an extensive blind test o
f both systems on computer manual text. We discuss the implications of
these results for the comparative roles of syntactically defined sali
ence and statistically measured lexical preference in determining the
references of pronouns in text.