We discuss the import, validity and implications of the identity thesis: th
e idea that the P600 component of the scalp recorded event-related brain po
tential is identical with the P3b, a domain-general component elicited by i
mprobable task-relevant events. We point to data reported in Coulson, King
and Kutas (1998), as well as more recently published findings, which sugges
t that the P600 component is sensitive to domain-general factors of probabi
lity, salience and task relevance. Exploring the epistemological complexiti
es of the issue, we suggest what the identity thesis does and does not impl
y about the existence of a modular parser.