An interesting paper by James Hartle: http://arxiv.org/abs/quant-ph/0610131
He argues that an understanding of fundamental physics can be hindered by the "excess baggage" of language.
Some of the main points he makes are:
Language is best adapted for coarse-grained descriptions of everyday experience. When it is applied to fundamental physics, clarity is lost.
The surest route to clarity is to express the constructions of human languages in the language of fundamental physical theory, not the other way around.
II. SPACETIME
There is neither a unique notion of space nor time. Instead, from each point in spacetime, there is a family of timelike directions and 3x as many spacelike directions.
Questions like "What is happening on Titan now?" are ambiguous, as relativistic effects are not negligible at such distances.
A way to remove ambiguity is to rephrase "happening" questions in terms of events: "Is event B in my future light cone at event A?", "Is B spacelike separated from me at A?", and "Is B in my past light cone at A?"
Agreement on what is happening now can be reached by checking reports of our observations against those of others, if
a) the relative velocity of others is small compared to the speed of light.
b) The time it takes a signal to travel between people is small compared to the time it takes to perceive information.
c) The reliability of information changes on a time scale much larger than the time scale of perceiving information.
For astronauts on Titan communicating with astronauts on earth, b) would be violated.
Quantum gravity introduces further ambiguities. Statements like "A happened in my past light cone." become meaningless without qualification.
III. THE QUANTUM MECHANICS OF CLOSED SYSTEMS
The inputs of a quantum mechanical model universe are the initial quantum state |Ψ>, and an operator H that tells us how |Ψ> evolves.
Outputs are probabilities of alternative coarse-grained histories. Each set of alternative coarse-grained histories is called a realm. There is more than one realm, and not all realms can be combined into larger realms.
There is no non-trivial realm of fine-grained histories.
Quasi-classical realms exhibit the approximate regularities of classical physics.
Quantum theory does not privilege any realm. But observers may employ a realm for its utility.
(Appendix A goes into greater detail about graining, histories, and realms)
IV PROBABILITIES
Probabilities are measures of ignorance in classical physics, but they are fundamental in quantum physics.
Both probabilities, and conditional probabilities can be constructed by observers in the universe. Observers can ask questions like "Given that I observe a tree here today, what is the probability that there was a tree here yesterday?"
In a probabilistic theory, when we say something happened, we mean there is a probability close to unity that it happened.
Questions, answers, retrodictions, etc. are all in the context of a realm.
Questions like "Given that I observe documents dated 55 BC, what is the probability that Romans existed in 55 BC?" Or "Given that I observe fossil records, what is the probability that dinosaurs existed?" are all questions in the context of a quasiclassical realm.
Different realms, incompatible with the quasiclassical realm, will have different but equally valid descriptions of the past.
E.g. Consider a quantum particle and three boxes A,B, and C. Given some present data, we can ask what the probability is that the particle was in a particular box at an earlier time. In one realm, the answer is "It was definitely in A". In another, the answer is "definitely in B".
The usual use of "happened" has the context of one realm. Since QM does not privilege any particular realm, "happened" must be reformed. This seems crazy. When someone asks "What happened yesterday?" you do not normally respond "In what realm?"
V. REALITY
How do we reconcile the notion of the everyday, physical reality of tables chairs and stars with a quantum theory about incompatible realms with no privileged realm?
Our choice of answering questions in the context of a classical realm is an artefact of schemata we use.
Robots or aliens could use different schemata for the same reality, and so would employ a different realm.
VI. DISPENSABLE WORDS
Dispensable words can be removed from the exposition of a theory without affecting its utility.
A person will understand "The probability of rain tomorrow." and "The probability of rain to happen tomorrow." to be the same thing.
In physics "happen" can be confusing because it implies "happen" is a physical process.
All histories are "equally real" if "real" is understood in a theoretical framework to mean no history is privileged in quantum mechanics, beyond their probabilities.
Phrases like "equally real" should be replaced accordingly.