A while ago I wrote about some even more complex examples in Hitchcock's movies (Vertigo and North by Northwest), where an elaborate deception creates a dynamic in the world (the world of the movie, that is) which in effect makes it seem as if the deception has become reality.
Even in the real world there are some rare cases of an instance of unreality causing far-reaching developments. The first airing of Orson Welles' radio show War of the worlds (based on H.G. Wells' novel) was so convincing that it caused confusion and panic with some in the show's audience who mistook it for a real report.
Such effects are normally short-lived in the real world — reflection kicks in, people communicate and cross-check, and closeness to reality is restored all the faster the higher the number of people who are affected directly. On the other hand, as long as the feasibility of reality checks is kept low and the subject remains fascinating enough to hold a grip on the imagination, even foggy rumors can be sustained for quite some time. In his 1993 Norton Lectures, Umberto Eco cites the case of the Superb, a British submarine which was rumored to be deployed during the build-up of the Falklands crisis. It wasn't, but a combination of public imagination, media speculation and official secrecy quickly made it into a quasi-fact. "[T]he whole story grew out of vague gossip, through the collaboration of all parties. Everybody cooperated in the creation of the Yellow Submarine because it was a fascinating fictional character and its story was narratively exciting."
What all these examples have in common is that some instance of unreality brings about changes in the real world, gets people to act in a different way than they'd have acted without that instance of unreality. Unreality (sometimes) settles into reality: though unreality is unreality, and reality is reality, part of reality consists of sedimented unreality.
Let's look closer at this. When we say that a dream comes true, it's not that literally the dream events come to have happened. It's still only a dream. What does happen is that I start acting in a way that makes some future situation resemble the situation from the dream. (This future situation can be a desirable state, if the dream expresses some wishes or goals I have; it could be an undesirable state, if it is a nightmare and expresses some of my fears. In either case, the way I act towards the situation I experienced in the dream can be conscious, or unconscious, or both.) If that happens and my actions are successful in bringing that situation about, we have now two different situations: an unreal dream situation and a real situation, which I made happen partly because of my dream experience.
When unreality sediments into reality, that instance of unreality remains what it is (it's not, so to speak, transformed into something real). In the example, there's still that dream, and that is an instance of unreality. It becomes, however, the cause (at least, a partial cause) and reason (possibly one reason among others) why some further, real situation, happens the way it does.
To conclude, here are some random reflections about sedimentation. First, in the case of dreams or visions what the unreal situation (the one that's dreamt or envisioned) and the real situation have in common is some experience you have in it, or a description that applies to it. It's an experience or description that was originally a 'what-if' experience or description. In other cases, such as the panic following a fictional invasion from Mars, there's no experience in common, but rather the real situation contains an 'as-if' perception or even action. (People start evacuating as if there really was an invasion.) In the most intricate cases, a constellation might involve both a 'what-if' experience and an 'as-if' action, leading to a very potent confusion of reality and unreality: this is what Vertigo uses to great and disturbing effect.(Again: read more about it in my previous post about sedimentation of unreality in those Hitchcock movies.)
Second, people who act on the basis of some instance of unreality are sometimes aware of this (when they try to fulfill a dream or achieve a vision, or when they have deliberately assumed a scenario, e.g. as a working hypothesis); sometimes they're not (when they act under a deception or illusion). But since people invest by acting and forming views about the world that contains this sedimented unreality, it's unlikely to be reversed once it's found unreal. (There's only a limited possibility to revert your actions in the world anyway, in particular if they have caused more development already.) I've called this the irreversibility of sedimentation in the previous posts I've already linked.
And finally, not everything in reality is sedimented unreality. (Although many facts about the real world have some sedimented unreality somewhere in the chain of causes that lead up to them.) Neither does all unreality sediment into reality of some form. (Some instances of unreality will remain largely without effect in the real world.)
 Umberto Eco, Six walks in the fictional woods. The Charles Eliot Norton Lectures 1993, Cambridge: Harvard University Press 1994, 97–99, the quote is from 99.
 Technical side note: I'm using a sketchy notion of cause and effect here that leaves many unclarities and is not (yet) connected to the contemporary philosophical discussion of causation. But note that the idea of sedimented unreality as a cause is no more problematic than any notion of mental causation of events in the world. Sedimentation can unproblematically be analyzed in terms of people's views and actions. It singles out views and actions that are based on an exercise of the imagination, but that's a phenomenon that every account of human action has to deal with anyway.