This essay reviews the changing functions, and effects, of citation systems in scholarly research as they move from a range of uses primarily oriented around knowledge networking and epistemic validation, to their use as a set of metrics oriented around evaluating and rewarding certain kinds of academic performance (e.g. "impact factors"). It is argued that this shift of emphasis has a distorting effect on the uses of citations, and that the more important these metrics become, the more distorting it is. Specifically, it is argued that rewarding certain kinds of citational practices leads to "gaming" the system: this makes the use of these metrics actually less reliable--and in some cases counterproductive--and leads to what in the author's view are self-serving and sometimes unethical behaviours.