This review orders the likely components of the pathogenesis of diabetic neuropathy into vertical (temporal) and horizontal dimensions. It is argued that the effects of hyperglycaemia are transduced to neuronal dysfunction via at least three secondary biochemical disturbances--the sorbitol (polyol) pathway, non-enzymatic glycation of proteins and oxidative stress--and that there are clear interactions between them. Because of these interactions, interference with one of these biochemical transducers could either worsen or attenuate the effects of the others. Examples of these alternatives are given. It is suggested that the prime goal for pharmacological intervention should be a combined attack on all three sources of disturbance. Interventions further on in the sequence of pathogenesis are also considered, and the arguments for the use of neurotrophic factors are persuasive because of their selectivity for different neuronal phenotypes, even though side-effects may be inevitable. Finally, a novel conjugate of gamma-linolenic acid and alpha-lipoic acid is considered as an agent with the potential to correct effects arising from more than one pathway of disorder in experimental diabetic neuropathy. The preliminary results with this agent have been encouraging.