Recent evidence suggests that glutamate-induced neuronal damage may contribute importantly to neuronal death in several neurological diseases, including cerebral hypoxia-ischemia. This review outlines a range of measures that might be used to protect neurons from such excitotoxic damage. The organizing thesis is a speculative consideration of glutamate neurotoxicity as a sequential three-stage process--induction, amplification, and expression--each perhaps specifically amenable to therapeutic interference. Overstimulation of glutamate receptors likely induces the intracellular accumulation of several substances, including Ca2+, Na+, inositol-1,4,5-trisphosphate, and diacylglycerol. Blockade of this induction might be accomplished most easily by antagonizing postsynaptic glutamate receptors, but also might be accomplished by reducing glutamate release from presynaptic terminals, or improving glutamate clearance from synaptic clefts. Following induction, several steps may importantly amplify the resultant rise in intracellular free Ca2+, and promote the spread of excessive excitation to other circuit neurons. Protective strategies operative at this level might include blockade of additional Ca2+ influx, blockade of Ca2+ release from intracellular stores, and interference with the mechanisms coupling glutamate receptor stimulation to lasting enhancements of excitatory synaptic efficacy. Following amplification, toxic levels of intracellular free Ca2+ might trigger destructive cascades bearing direct responsibility for resultant neuronal degeneration--the expression of excitotoxicity. The most important cascades to block may be those related to the activation of catabolic enzymes, and the generation of free radicals. Broad consideration of possible methods for antagonizing glutamate neurotoxicity may be needed to develop therapies with the greatest efficacy, and least adverse consequences for brain function.