Gnu.org: Studies Find Reward Often No Motivator
This is an old article. It's cited in the context of the Free Software movement, as a refutation of the arch-capitalist strawman, "but if people don't do things for money, there will be UTTER CHAOS!!!". Interesting enough in that sociological sense. I of course applied it to game design, and it was something of a mini-epiphany.
The phrase that jumped out to me was "work that involves creativity". Most people probably do not think of videogame playing as "creative". This is partly because few games really recognize or foster creativity, and players who do play creatively in such games are often going well beyond the demands of the game's rules (see: speedrunning, trickjumping, etc)  . However, there is more chicken and egg to this than you might think. I think there is a very compelling argument for saying that games to date have not fostered creative/stylish/interesting play precisely because they have tended towards extrinsic rewards, which in turn springs from their desire to "addict" players in consistent and reliable ways, as per the fetishization of "pacing" Aubrey identified recently.
GTA3 is an example of a game I would say offers many intrinsic rewards. Doing basic stuff in the world is just fun: driving, doing crazy stunts, getting into and escaping from trouble. Indeed, the extrinsic rewards of many actions, eg the money you get from killing people and completing missions, is pretty much an afterthought. It's not the reason you were doing X, X was just fun to do on its own.
One of the reasons Aubrey and I dislike the trend towards making "unlockable" things a major feature of games is that it turns many things that were previously about intrinsic rewards into extrinsics. You play 60 hours to unlock Waluigi not because the game is fun for 60 hours, but because you get Waluigi.
Probably the biggest reason MMOs turn me off is that they have such a brazen, rigidly codified extrinsic reward structure, and for most people that is the only reason to play the game. The carrot becomes your entire reason for living, and it's tied into a business model that is designed to keep you pressing on that Skinner Box lever while the cash continues to flow out of your wallet.
Interestingly, when you hit the level cap in many of these games, the biggest reasons to stick around are the social connections you have built with other players - intrinsic rewards (unless you're a sociopath who views friends as instruments).
It's almost startling in retrospect how much of my personal design philosophy can be explained by this principle. I am very passionate about expanding game design beyond the (now very established) realm of extrinsic rewards. I think the result will be games that people feel are more entertaining and even enriching, rather than just a way to kill time. To put it another way, it's hard to imagine humankind's enjoyment of Art as anything but a big, profound intrinsic reward. Games will be further down the road to becoming Art when they lose the obsession with being Disneyland rides and graphically intensive slot machines.
 Players who are creative enough to find their own forms of expression within a game must implicitly go beyond merely understanding the game mechanics. That is a base requirement. Expression in games (finding one's own "way" - one's own "style", like Jeet Kun Do) seems only to occur in the emergent strategies thrown up by the core rules. In Deux Ex, placing a Mine is not creative - it has been coded strictly. Equally, jumping is not creative - it has merely been allowed by the developers. But in combination, players have been able to climb sheer surfaces with these two simple actions. That's creativity - not because it was unanticipated (which it happened to be) but because it was at a level of expression higher than the simple use of an existing verb. For many games, this stratosphere of creativity is only met after autistic levels of engagement with the game - typically, video games are quite complex when compared to folk, and board games. There are such a breadth of verbs to learn that the entry level for actual strategic creativity often lies well after total appropriation of a tombs' worth of button combinations and context sensitive verbs. And often, this breadth simply exists to hide the lack of depth present (I do not believe that breath is necessarily inversely proportional to depth - sadly, it often seems the case).
 We are not well versed, on the whole, on the art of creating creative tools. They are difficult to conceptualize, design, and produce, and it is easy (too easy) to dismiss the undesireable emergent strategies as un-forseeable... a product of the magic of game design. I wish that we did focus on this side more - the management without contrivance of our own possibility spaces. I have my own scrappy ideas for methodologies which seem to get me by. They basically come down to this: systems that are complex enough for its results to become percievably random not only to its own designer, but to its player, are pointless. Obviously this is rather subjective, but the point is, while we should not sit on our laurels and make shallow game play out of fear of emergence, we should equally recognize that too much depth can turn dull, too. After all, it could be the overwhelming amount of depth that hinders us from being intentionally creative within our play - if we cannot percieve the full result of our actions easily, we may be frightened to make certain expressions.
 I had posted something earlier on the subject, but took it down for a re-write (still pending).
 I've felt this quite often, and seen friends succum to it too. The technique of well-paced-unlockables doesn't make a game implicitly bad, but it has certainly seen me play more than a few games more than they probably deserved (Need for Speed: Underground, I look in your direction). I'm certainly averse to the idea that this technique "saves" otherwise uncompelling games. They say that you can't polish a turd... but this pavlovian approach makes me think otherwise.
 It's also at this point that the exploration of the breadth of the game (the constant opening of new verbs as you level through your different disceplines) dries up, at which point, you've been exposed to the idea of exploring the depth thrown up by your endless verbs. I witnessed four great friends from my childhood, this weekend, all sit around a table during a wedding celebration, trying to figure out the most efficient barrage of spells to take out a single, incredibly tough enemy. So, my point in  stands. With enough saturation in the problem area, you establish an accurate mental model, and start to really be able to riff creatively with the game-qua-musical-instrument.
 I agree (of course!) but can also sympathise with why such games have been successful - or atleast, why mankind has allowed them to be successful. It's the Need to Want. As Jesper Juul pointed out of Sid Meier's fleeting description of games: "Games are a set of interesting and uninteresting choices". Yes, a player's voice has been sadly unheard in most of the games we play, but often, we like to follow paths. We sometimes like to press the correct button. We sometimes like to be told "you did the right thing", rather than "any approach is as good as the next". We also just like extrinsic rewards (at least until their novelty wears off due to too much repetition). But what this article clears up so elegantly is why, even though intrinsic rewards are so much more fulfulling, why so many people put up with extrinsically biased games. Things like Guitar Hero, where we follow the same track every single time, or Half Life, where we try to follow the "Hero's Journey" - both cases punish unsanctioned creativity. But these games seem to work because they are the ones where the extrinsic end is the intrinsic desire. The need to want to sound like Jimmy Hendrix. The need to want to be Gordon Freeman. The Need to Want is built into all of us through evolution. The need to suffer for even an unfulfilling reward still fills us with a sense of worth. It's the reason we build pyramids and climb mountains and fuck each other over for the top spot. It's evolutionary. It's a bit sad, really.
Up creatively empowering games, I say.