Sensitivity to grammatical gender was investigated in 22 Russian-speaking a
phasic patients, compared with young controls. Experiment 1 used a cued sha
dowing paradigm to assess gender priming (facilitation and/or inhibition of
lexical access by a prenominal modifier with congruent, incongruent or neu
tral gender). Experiment 2 used a grammaticality judgment paradigm with sim
ilar stimuli. Normals showed significant interactions between gender and pr
iming in Experiment 1 (facilitation for feminine and neuter nouns but not f
or masculines) and Experiment 2 (larger effects of context on feminine and
neuter nouns) that we interpret as a Markedness Effect. Patients showed sig
nificant priming in Experiment 1 and above-chance accuracy in Experiment 2,
but failed to show reduced effects for the least-marked masculine gender (
the Markedness Effect) in either experiment. Context effects were not relat
ed to specific aphasic symptoms or subtypes in either experiment. However,
canonical correlation revealed differential effects of specific aphasic sym
ptoms on judgment accuracy (false alarms vs. misses). We conclude that know
ledge of grammatical gender is spared in Russian aphasics, but gender proce
ssing is deviant. A possible model to account for these differences is disc
ussed.