In some propositional reasoning tasks people exhibit a ''matching bias
'', i.e., a tendency to select information which matches the lexical c
ontent of the premises, irrespective of the presence of negations. Acc
ording to Houde and Moutier (1996), the matching bias is due to the in
sufficient ''inhibition'' of misleading reasoning schemes. In the pres
ent paper, I argue that Houde and Moutier's results do not support the
ir interpretation, which is both inadequate (it makes erroneous predic
tions) and indeterminate (it cannot predict the matching phenomena in
a definite way). I conclude that the reported results do not provide a
ny new justifications for the presumption of rationality of human thou
ght.