Or a science equivalent to what money and market competition does with all human ventures. You need more and faster than your competitors to progress, else you'll fall into obscurity. "First to publish" is isomorphic to "first to market". If you treat research as a videogame, citations fit perfectly as an in-game currency.
That is to say, it's an example of a general problem of optimizing short-term metrics, which are a decent proxy of the actual goal up to a point.
Science could definitely use a spam filter. Journals unfortunately do a poor way of being one; they're thoroughly gamed, the same way Google is via SEO.