Alzheimer's disease (AD) is characterized clinically by dementia and pathologically by two hallmark lesions, senile plaques and neurofibrillary tangles. About a quarter century ago these hallmark lesions were purified and their protein constituents identified, precipitating an avalanche of molecular studies as well as substantial optimism about successful therapeutic intervention. In 2009, we now have copious knowledge on the biochemical cascades that produce these proteins, the different modifications and forms in which these proteins exist, and the ability to selectively target these proteins for therapeutic intervention on an experimental basis. At the same time, there has been no discernible alteration in the natural course of AD in humans. While it may be that the complexity of AD will exceed our capacity to make significant treatment progress for decades or more, a paradigm shift from the reductionism that defines amyloid-beta and tau hypotheses, to one that more accurately reflects the meaning of neuropathological changes, may be warranted. We and others have demonstrated that AD pathology is a manifestation of cellular adaptation, specifically as a defense against oxidative injury. As such, AD pathology is therefore a host response rather than a manifestation of cytotoxic protein injury, and is unlikely to be a fruitful target for therapeutic intervention. An "expansionist" view of the disease, we believe, with oxidative stress as a pleiotropic and upstream process, more aptly describes the relationship between various and numerous molecular alterations and clinical disease.