The causes of regeneration failure in the mammalian central nervous system (CNS), and how to circumvent them, have long been subjects of controversy. The more general issue of postinjury CNS recovery, of which regeneration is one distinct aspect, has undergone conceptual changes over the last two decades as a result of certain key discoveries. First, it was recognized that at least some CNS nerves do have the capacity to regenerate, but fail to accomplish this under normal circumstances because the cellular milieu of the regrowing axons is apparently unable to acquire growth-supportive properties and/or to overcome the influence of growth inhibitors. Another landmark was the recognition that in addition to the need for regeneration, the injured nerve would benefit from neuroprotection, as a means of halting the progression of secondary degeneration, i.e. injury-induced spread of damage to neurons that escaped the primary injury. These two aspects of nerve recovery appear to have distinctive and even conflicting requirements. Over the years, the various approaches aimed at promoting regeneration have met with varying degrees of success. Common to these approaches was an attempt to overcome the growth hostility exhibited by the axon’s immediate environment following injury. Our own attempts to identify rate-limiting or other key steps that might account for regeneration failure led us to suggest that the failure may be linked, at least in part, to the postinjury inflammatory response, which differs in the CNS from that in regenerative tissues, including the peripheral nervous system. The inflammatory reaction in any injured tissue includes phagocytic activity, cytrotoxicity and production of cytokines and growth factors, all needed in order to clear the damaged tissue of cell and other debris and to rebuild new tissue. In the CNS, however, the inflammatory processes occur only to a limited extent, possibly because of a default resulting from the fact that the CNS has evolved as an immune-privileged site and consequently its communication with the immune cells is restricted. In this review we discuss the multidisciplinary studies that have gradually thrown light on the possible role of macrophages in regeneration, the apparently conflicting inflammatory needs of neuroprotection and neuroregeneration, and a possible explanation, in evolutionary terms, for regeneration failure.