About seven years ago Mary and I drove two hundred miles down to Columbus and paid our admission fee at Origins for the sole purpose of demoing one of my euro-style games for an interested publisher. This was probably the best of my games in that style, an economic engine-builder about private equity firms eating up small companies, building them up, and then selling them for a profit. The development process on this one was lengthy, and it was the first one I was really confident about, the game that I thought would at last break through the wall of rejections and catapult me into the ranks of mid-weight euro designers.
I was about forty seconds into my demo, explaining the activation mechanism that drove the game - a sort of worker placement twist - when the publisher stopped me.
"I have a theme question. Why do I need these workers? Why does my sawmill stop running because I have a guy over in the plastics factory?"
I fumbled. "I don't exactly have a theme answer," I admitted.
"Then we have a big problem," said the publisher. "I'm not saying you have to shut up, but I'm not going to buy this."
Which made the subsequent two hundred mile drive back to Michigan fairly disheartening. On the way home of course I came up with that theme answer: the workers represent the amount of bandwidth you have as a CEO, and the amount of oversight you can give to the businesses in your portfolio. Not that that explanation would have persuaded him; I think his modus operandi was to find reasons to shoot down pitches as quickly and efficiently as possible. (Quick aside: the mechanism from that game has been repurposed, with added limitations to reflect internal factionalism, for Ensi, my solo civ-building game set in Mesopotamia.)
One of the complaints often levied against Richard Borg's Commands and Colors series - heck, it's even been said in this household - is that you'll have situations where you have a bunch of troops on your left, but you only have cards that activate your center or your right. And no general in history has ever said, well, those men are doomed; I didn't draw the cards I wanted. In Combat Commander, how many squads have found themselves exposed without any Fire cards in hand, and how many players take potshots to churn through their deck to speed up the game? Those experiences aren't reflective of the reality in any way, shape, or form. And why does the sawmill stop running because you have your pawn over at the plastics factory?
In all these cases, and in many more, the mechanism is a metaphor. It's an artificial thing that allows you to get at the essential truth of the situation - the lack of total control for a battlefield commander, the chaos of a real engagement, or the decision whether to emphasize your lumber or your plastics - but that, in and of itself, perhaps doesn't represent anything literal or tangible at all. Playing a Bad Weather card on your opponent doesn't literally mean that you are controlling or predicting the weather; the card is a metaphor for other things which allows the game to create an experience with some degree of verisimilitude.
Wargame designers sometimes call this "design for effect", and euro designers, being rather agnostic about theme, don't bother to call it anything at all. I prefer to think of it as a metaphor because it both highlights its purpose - to express something true through an elegant lie - and its pitfalls: sometimes there's nothing there but pretty, empty nonsense. While I don't think that every mechanism needs a "theme answer", they do need to connect and contribute to the experience in a way that coaxes players in rather than pushing them away.