Gravitational waves have not been detected as yet, but in general relativity they move at the speed of light, and thus would be neither spacelike nor timelike, but lightlike (i.e., on the light cone). Gravitons have neither been detected nor have a good theory for their behavior but they are generally assumed to both exist and be lightlike as well (and it is hard to see how, should they exist, they could not be, assuming that gravitational waves are in fact actually lightlike).
VLBI observations of the gravitational time delay caused by Jupiter have been used to infer that gravity does propagate at the speed of light,to with an error of 20% or so, but this conclusion is controversial.
And, like any radiation, gravitational waves could certainly vary in time.
@Marshall: you say "Gravitational waves have not been detected" and "assuming that gravitational waves are in fact actually lightlike". Maybe the second sentence implies the first. Maybe scientists are looking in the wrong direction (just an image, nothing to do with a direction).
In General Relativity, gravitational radiation has two senses of polarization. This is a restriction, in that a general metric theory of gravity could support up to _six_ senses of gravitational wave polarization, and thus gravitational waves will provide a sensitive test of the theory of gravity. Generally, gravitational wave detectors would be sensitive to all of these polarizations, and (depending on the signal and detector) could thus provide stringent tests of general relativity. Likewise, gravitational waves could travel at less (or even more!) than the speed of light in other theories of gravity, and the existing and planned detectors could detect such non-lightlike waves as well. It will, however, probably take the detection of gravitational radiation from a short-duration event, such as a gamma ray burst, to begin to experimentally constrain the speed of their propagation.
The lack of detection of gravitational radiation is largely due to the weakness of the signal; it is still true that none of the known sources of gravitational radiation are likely to be detected by any of the currently operating detectors. The "Advanced LIGO" currently under construction should change that, I wouldn't worry about the status of General Relativity as regards gravitational waves unless and until Advanced LIGO fails to detect any sources.
For more on gravitational waves as a test of General Relativity, see
Light-like time-like and space-like are used in regard to geodesic intervals, which tell us only about the speed of the particles and the category (massive or massless) to which they belong.
Robert has given the correct answer. I do not know, who down-voted his post. The best I can do is up-vote. These are very basic things in Relativity, a primer's knowledge.
There should be some mechanism to keep out the useless down-voters of good answers like this one by Robert. I'm surprised to find such people on RG.
The down-voter (I assume is following this thread, and was not a stray voter) should withdraw his down-voting somehow at least in this case (e.g. by writing to the editors of RG or by offering an apology to Robert in this forum) which has no room for any personal opinion.
General Relativity => Einstein's Equivalence Principle=>Theory of gravitation
I have a naive question: are these implications bijective?"
I suggest you take a look at this on the experimental basis of Einstein's theory, in particular of Einstein's Equivalence Principle: http://relativity.livingreviews.org/Articles/lrr-2006-3/
Regarding: "I just don't find the concept of light cone logical", the light cone is a logical conclusion of the constancy of the velocity of light, independent of the state of motion of an observer. So if you understand the Michelson-Morley experiment you are led to the light cone.
Note that the article I linked above does not question special relativity at all. That's so well established experimentally that it really does not bear discussion whether there is a "light cone". Muon decay lifetime observations provided a beautifully precise verification of time dilation, for instance.
What is so illogical about the light cone ? Even if we are talking about gravitons, light cones will come in when we are out to find out what type they belong to. Had you not added the parenthetic (time-like) in your question, no one would probably have gone in that direction.
Anyway, I think if remove the paranthetic "time-like" probably, people will bother to answer on other aspects of your question.
Probably, you wanted to ask : whether the graviton is entirely a temporal existence, i.e. exists only in time, and not in space !
Please, edit and clarify a little more on the question. May be, you have a good point to make.
It's not easy to explain. @Rajat, you're right, I shouldn't have added the word 'timelike'.
The 'classification' given by the light cone is not very clear to me.
When I said purely temporal, I was thinking about basic geometry in the (x,y,z,t) space, i.e. you may write the wavefunction of the graviton only with t and without x,y,or z.
I was thinking about gravitational waves in my model: I do not think that the word 'wave' is appropriate in one temporal dimension as time cannot go backward.