In response to this post, Julian asked if in the process of making a twin-cylinder as opposed to a single, "Wouldn't just the engine cost double? Not the whole bike?"
Well, the way it was explained to me was this. In pure material terms, obviously, the top the engine - valves etc, double. The middle of the engine, obviously doubles (pistons, bores, mounting studs etc). The crankshaft becomes a lot more complicated and nearly doubles in cost, if not in material. But more importantly, now you have to upgrade the chassis to handle the extra performance. This means that the tyre specification and size, rims, all the suspension, most of the frame has to be upgraded. Again, most of the time, this results in a nearly ground-up reworking rather than a simple make all tubes thicker kind of engineering. This, of course, presumes that the original motorcycle was not a platform. If it were a platform, then the ability to upgrade the frame would be built in - and cost less over the life of the platform. That last bit is a severely complex calculation that is beyond my ability to explain any further.
The upshot is that the twin cylinder engine might appear to be the simple addition of another bore-piston-valve set to the single. But that's the illusion. By the time you've finished re-engineering the product, you are looking at close to double the cost. Now consider the sale price. This will proportionally be higher. But remember that the single - since it is cheaper - will logically sell bigger numbers. This in turn allows smaller margins to justify the products since the volume will compensate and bring your the profit you need. But the twin is more expensive so it will naturally sell lower volumes. So the manufacturer, then, has to command bigger margins on that bike to make similar profit on that model. What results, I am given to understand, is nearly twice the price.
I must admit though, that it never occurred to me to ask if going to twin to triple or four will cause similar price rises but empirical evidence suggests that the big leap in price and complexity is from single to multi as opposed to twin to triple/four and hence the an inline-four is usually not twice the price of the single. Again, this is my conclusion and I could be wrong.
If there's any R&D engineers who are reading this, I would appreciate a clarification on that last bit.