Double post, following on from
this post.
What you have to remember here, is that the relativistic concept of space differs fundamentally from that of prior physical theories in an important and radical aspect.
The first point to bear in mind is that any representation of space involving a coordinate system, ultimately brings you face to face with the concept of a 'metric' for that space, which in effect, defines what is meant
precisely when we talk about the 'distance' between two points. When this is combined with the concept of a manifold (which is a mathematically precise means of defining what we mean by a 'well behaved' space, and invokes some topological concepts), then you end up with what is known as a Riemannian manifold, which is considered to be the mathematical entity that best represents real physical space. Because a Riemannian manifold has a distance metric associated with it, this implies immediately that the manifold in question is associated with another mathematical entity called a vector space, which means that the full panoply of linear algebra applies to that manifold. More importantly, that vector space, because of the existence of the distance metric, is a normed vector space and an inner product space, which are two important concepts allowing us to specify in what sense the space in question is "well behaved". The existence of a norm for the vector space means that there exists a precise, rigorous and meaningful definition of 'length' in the space, and since a norm is
always associated with an inner product of some sort, this in turn allows us to place the concept of 'right angle' (or, in higher dimensions,
orthogonality) on a rigorous footing too. In truth, the inner product is the fundamental unit in a vector space, because every definable inner product leads to a norm, and every norm in turn leads to a distance metric. However, in a geometrical setting, a distance metric is considered, if you like, the means of defining the properties that a space has, because a distance metric in that setting involves relations between the coordinates of the space, and in that setting, a standard, unified concept of 'inner product' exists. Extensions to linear algebra and vector spaces arose from the fact that translating the geometrical intuitions arising from coordinate geometry to, for example, the analysis of sets of functions, proved to be wonderfully useful and unifying, and so the
generalised concepts of inner product, norm and metric arising therefrom have their own separate branch of mathematical analysis. But I digress. At bottom, the key points to remember here, before proceeding, are:
[1] Every space, once a coordinate system is defined for it, has a distance metric defined for it;
[2] That distance metric is associated naturally with a norm and an inner product, making the space a particularly well-behaved form of a vector space;
[3] Such a space can be mapped onto a Euclidean space of equivalent dimensionality, and is therefore a manifold.
To make matters even more interesting, if the space in question is
complete (i.e., in a precise mathematical sense, there are no "points missing" from the space), then the space is also a Hilbert space, but this is of peripheral interest here.
Once you have a coordinate representation of a space, all the powerful tools of manifold theory, vector space theory and linear algebra can be brought to bear upon that space, as can vector analysis (a different branch of mathematics from vector space theory, just to confuse matters a little!) and the extension thereof in the form of tensor analysis, which covers the behaviour of objects more complex than standard vectors that can exist in that space. Incidentally, tensor analysis has its own all-encompassing definition of 'inner product' for coordinate spaces, from which a norm and a metric naturally arise, and in addition, provide the existence of a 'metric tensor', which not only defines how distance is constructed within the space, but also how curvature is determined within the space.
Now, In the past, even as the full flowering of tensor analysis was starting to get under way in the latter half of the 19th century, and the idea of coordinate metrics as a means of defining the properties of a space were being placed upon a rigorous footing, along with all the other mathematical developments I've described above, time was treated as something entirely separate. Time was treated as an entity
distinct from the metrical coordinates of a space, with an independent behaviour.
What makes relativity so radical, is that
it treats time as another metrical coordinate. It integrates time and space into a greater whole, subject to all the provisions of an appropriate 'distance metric' which now has to take
separation in time into account as well as separation in space. All of which, moreover, has to behave in a well-defined manner when one space-time is subject to Lorentz transformation onto another space-time. Because time is now treated as another metrical coordinate within the coordinate system, instead of being separate therefrom, all points within that coordinate system now have a time coordinate, which as a natural corollary means that every path through such a coordinate system involves 'time travel' of some sort, because that path will pass through points that have a time coordinate, and that time coordinate can potentially vary along that path.
But, here's the fun part.
Because time is now coupled to space coordinates within general relativity, entities cannot be thought of as having places, without also including some sensible means of defining their moments as well. In short, relativity does away with the idea that 'where' and 'when' are distinct and independent. In relativity, the two are coupled by the metric for the space-time, and moreover, coupled in such a manner that even inertial motion affects your measurement of that 'where' and 'when', to the point that, for example, there no longer exists a concept of global simultaneity within a relativistic space-time - two events that appear to be 'simultaneous' to one set of observers may appear to be separated in time by a different set of observers, courtesy of the intricacies of that distance metric, and how that distance metric is transformed from a 'stationary' space to a 'moving' space.
What makes all of this even more hilarious, of course, is that making time another metrical coordinate imposes certain symmetries between time and space that didn't exist before. Prior to relativity, you didn't have to think about this, and you could treat time in an entirely intuitive manner, without worrying about the effect of so doing upon your treatment of space. Now, those days are over, at least in precise work, and time has to be treated as another metrical coordinate, because thus far, all our measurements of various relevant phenomena point to this being a necessity. This brings with it an unfortunate natural consequence, namely, that just as you can travel forwards and backwards in space (however 'forwards' and 'backwards' are defined), making time another metrical coordinate means you can, in theory at least, do the same with time. This is problematic for numerous reasons, even
before we start to consider various paradoxes. First, our intuition (and for that matter, every clock on the planet) tells us that time is moving forwards. How can it be possible to travel back to the past?
Well, the moment you make time another metrical coordinate, this becomes a possibility, simply because you now impose certain symmetries upon time that weren't present before. The relativistic view is not that
time itself is moving, rather, it's that
the universe and its contents are moving in a given direction along a time coordinate axis, and what's more, your motion in terms of pure space coordinates has an impact upon how fast you move along this time axis. General relativity adds even more nastiness, from the standpoint of intuition, to the mix by letting gravity affect your motion along this time axis, so that if you're in the neighbourhood of a sufficiently strong gravitational field, your motion along the time axis is seriously slowed. In the relativistic world, clocks simply measure your movement along this time axis, which happens even when you're stationary in terms of pure space coordinates. By treating time as "another space coordinate", if you like, relativity permits travel into the past, provided that certain somewhat exotic conditions I described in the previous post are met, and this arises because time is now treated as another metrical coordinate.
Don't worry if this buggers your mind for a while, because it has that effect even upon world class mathematical physicists during their early student days. As quite a few of them will happily admit.
