this paper studies the difficulty to evaluate the validity of conditional arguments modus ponens (mp), modus tollens (mt), denying the antecedent (na) and affirming the consequent (ac) of symbolic content when explicit negations are introduced into the antecedent and/or the consequent of the main premise. no significant differences were found among the acceptance percentages in mp and ac according to the presence or absence of negations. in mt, the argument with antecedent and consequent denied obtained a significantly lower acceptance percentage than arguments with only one component denied. in na, the argument with both clauses denied has a lower acceptance percentage than the rest. the theoretical implications of these results are discussed on the basis of their comparison with the patterns proposed in schroyens, schaeken & d'ydewalle (2001).