Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Relationships among metric structure, metric tensor, special and general relativity

  1. Nov 18, 2006 #1
    hello. i'm working on a philosophical summary of general relativity. i have difficulty understanding tensor. i made the following characterization; can any expert minds here tell me if i said it correctly?

    if anyone can take a peek to see if what i got so far is correct that'd be sooooooo appreciated too:

    http://www.theophoretos.hostmatrix.org/generalrelativity.htm [Broken]

    p.s. i also did a special relativity, but it's very long.
    http://www.theophoretos.hostmatrix.org/relativity.htm [Broken]
    Last edited by a moderator: May 2, 2017
  2. jcsd
  3. Nov 18, 2006 #2
    At least, for me, your text makes very little sense. Why do you want to mess with comples numbers, which have nothing to do with the subject (here we are handling real (psedo)manifolds). In the Minkowski metric the time component has a different sign than the spatial part in order to have the causality needed. And the spacetime invariant is ds^2=dt^2 - dr^2 (or the other way round, just convention). Minkowski metric is not whatever 4-dimensional diagonal matrix.
    Last edited by a moderator: May 2, 2017
  4. Nov 18, 2006 #3
    LB, are you sure you know about relativity at all? you seem to be speaking gibberish.

    does anyone here who knows about relativity have any comment? thanks.
  5. Nov 18, 2006 #4
    I'll comment. Having read the http://www.theophoretos.hostmatrix.org/generalrelativity.htm [Broken] you linked to, I have to say that it doesn't seem to make much sense to me either. For example, you begin with

    Laboured rhetorical flourishes aside, this is completely incorrect. Firstly, the base Minkowski line interval is [itex]ds^2 = \pm((dx^1)^2 + (dx^2)^2 + (dx^3)^2 - (dx^4)^2)[/itex], with the [itex]\pm[/itex] being simply a matter of convention. Secondly, what is a "Gaussian spacetime metric"?

    Next, you say:

    Again, this is incorrect. The Minkowski metric applies to a Minkowski spacetime, not to a Euclidean spacetime (for which the appropriate metric would be [itex]\delta_{ij}=\mathrm{diag}(++++)[/itex]). In a broader context, a general spacetime is said to be Lorentzian, i.e., locally Minkowski, but most certainly not Euclidean.

    Finally, Los Bobos' comments about your use of complex numbers in relation to the Minkowski metric are completely justified.
    Last edited by a moderator: May 2, 2017
  6. Nov 18, 2006 #5


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    Last edited by a moderator: May 2, 2017
  7. Nov 18, 2006 #6
    Yes I am sure I know about relativity.
  8. Nov 18, 2006 #7
    alright the discussion is finished. feel like in a twilight zone or some kind. how can you not even know what Gaussian metric is? and if minkowski metric doesn't apply to euclidean, then i should give you my head. lorentzian? this is so strange. from which world did your guys come from? it's almost like suddenly people start saying, "the sun has always risen from the west". bizarre.
  9. Nov 18, 2006 #8
    Moderator, Help, Come Check Out What's Going On Here
  10. Nov 18, 2006 #9


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    The Minkowski metric, denoted [tex]g_{ab}[/tex], maps a pair of vectors to a real number... or a vector to a covector.

    The Kronecker delta, denoted [tex]\delta_a{}^b[/tex], is simply the identity operator (or index substitution operator), which maps a vector to a vector... or a covector to a covector.. or a vector and a covector to a number, their transvection.

    In fact, [tex]g_{ab}g^{bc}=\delta_a{}^c[/tex]. In matrix language, [tex]G G^{-1}=I[/tex]

    Could it be that you are using nonstandard and possibly archaic terms?
    If so, then it might be a good idea to clearly define your terms for your reader. It will not suffice to make a reference to an old paper in a foreign language... primary source or not.

    "feel like in a twilight zone or some kind ...from which world did your guys come from"...
    In our defense, the world that you have wandered into uses modern differential geometry and tensor analysis. Have a look at the first review paper I referenced.
  11. Nov 19, 2006 #10

    Chris Hillman

    User Avatar
    Science Advisor

    Suggest Sklar

    Therapeuter, I didn't try to follow the link you offered, since I'm not sure what you mean by "philosophical summary" and how seriously you take this. For what it is worth, a very fine nontechnical book you might like is Lawrence Sklar, The Philosophy of Space and Time, which offers a nice introduction to many classical topics, including a chapter on how philosphers have reacted to the appearance of general relativity.

    From what you wrote here it seems that you might be trying to talk about some of the same "local versus global" issues that I discussed in connection with "the" equivalence principle in another thread earlier today, but I find it too exhausting to try to explain this rather subtle concept except to an audience comfortable with tensor fields.

    Coalquay404: the paragraph you quoted didn't make much sense to me either, possibly because it was taken out of context, but it probably helps to know that the references to "shadows" probably are intended to invoke the so-called "allegory of the cave" due to Plato, which is very well known to undergraduate students of philsophy. (Therepeuter can confirm or deny this the validity of this guess.) Ironically, since we are talking about metrics, the name "Plato" is a nickname or possibly a literary pseudonym which means more or less literally "wide"; some say this moniker probably originated as a schoolboy nickname suggesting a stocky or burly build, or, some say, a chubby one.

    Chris Hillman
  12. Nov 19, 2006 #11
    I've never heard of a "Gaussian metric", either. I checked several standard GR texts and found Gaussian normal coordinates, Gauss-Codazzi equations, Gaussian curvature and, of course, Gauss's Law, but nothing on "Gaussian metric".

    Lorentzian is a standard description for a metric with signature like (-+++) (Wald, page 23). Wald is the closest thing to a canonical text in the field, so I would recommend reading the first several chapters of it for the standard terminology.

    I think I have seen "Minkowski metric" used for the "Euclideanized" spacetime with t -> ict, usually in old books on special relativity. It's sometimes still used in QFT, though.
  13. Nov 19, 2006 #12
    The full reference is

    Robert M. Wald, General Relativity, University Of Chicago Press, 1984 (ISBN 0226870332)
  14. Nov 19, 2006 #13
    A quick use of Google seems to suggest that it's not that you can have a Gaussian metric, but Gaussian fluctuations to a metric (or even non-Gaussian ones). The metric itself is not described as Gaussian.

    Perhaps it's just because I'm not one for hardcore philosophy but if someone asked what I thought of describing special and general relativities as caves and images and shadows I'd tell them to actually learn some of relativity because that's nothing like what I'd describe it as.

    therapeuter, if you're mixing up Euclidean and Minkowski metrics, I think you might want to slow down and get your head around relativity before trying to summarise it's philosophy to other people who aren't familiar with it. Otherwise it'll be a case of the blind leading the blind by the looks of it.
    Last edited: Nov 19, 2006
  15. Nov 19, 2006 #14
    i figure out what this twilight zone circus is about now. i know where you guys are from. you have evacuated this forum. you then try to b.s. as real as possible about general relativity, as if you knew relativity and i didn't, and hope to discredit me in front of the important people.

    it's not going to work. Gaussian, euclidean = minkowski etc. are very basic concepts. you ain't gonna fool any body but you are only gonna expose your stupidity even more and sink further into disrepute. the important people have physicists on their side to verify that i know what i'm talking about here, while you guys don't.
  16. Nov 19, 2006 #15


    User Avatar
    Science Advisor
    Homework Helper
    Gold Member

    A more useful and helpful attitude would have been to enlighten us with your definitions.

    A full-text search for the phrase "gaussian metric" at JSTOR ( http://www.jstor.org ), yielded 3 results: from 1925, 1942, and 1953. None of them seem relevant, except possibly for the 1953 one, which appears in AmJMath, but is used in a very specific way. ("gaussian spacetime" and "gaussian spacetime metric" yields zero hits. By comparison, "euclidean metric" yields 1058 hits, "riemann metric" yields 2889 hits, "riemannian metric" 3196 hits)

    As you requested, I "took a peek" at one of your webpages, I am confused by your explanation of the twin paradox:
    In http://www.theophoretos.hostmatrix.org/relativity.htm [Broken] ,

    Just before the References, you say
    "The triangle representing spacetime path of the two twins as drawn by the twin remaining on Earth. The elapsed time to the traveling twin (his proper time) is less since the length of the bent path is less than the length of the straight line. "This 'inequality of the triangle of spacetime' is the reverse of the inequality of the triangle of ordinary Euclidean space, for which the sum of the sides is longer than the [straight line]. This difference is due to the particular form of the chronogeometry of spacetime, where the Pythagorean theorem contains a negative sign for the squares of the sides of a right triangle which are directed in time."
    which sounds right... but then you say...
    "For example, while the twin at rest has travelled 0 - 8 = -8 interval of space time, the travelling twin has travelled (3 - 5) + (3 - 5) = -4 (spacetime distance = spatial distance - time difference)."
    Here is your associated diagram:
    http://www.theophoretos.hostmatrix.org/spacetimetriangletwin.jpg [Broken]
    which appears to assign Euclidean lengths (i.e. lengths that would be appropriate for a Euclidean space with its Pythagorean theorem) to the legs of the Minkowski-spacetime triangles.

    (You assign 5 to the outgoing traveller's leg (the hypotenuse) [and the incoming traveller's leg], 4 to its temporal component (the adjacent side), and 3 to its spatial component (the opposite side), where the angle is taken between the vertical-on-the-spacetime-diagram stay-at-home-twin and the outgoing-traveller leg.)

    Using those edge-lengths, the correct assignments [for Special Relativity] are 3 for the outgoing traveller's leg [and the incoming traveller's leg], 5 for its temporal component, and 4 to its spatial component. Then, for the round-trip, the stay-at-home logs 10 units, whereas the traveller logs, via the spacetime-version of the Pythagorean theorem, (sqrt(5^2-4^2)+sqrt(5^2-4^2)=) 6 units.
    Last edited by a moderator: May 2, 2017
  17. Nov 19, 2006 #16


    User Avatar
    Staff Emeritus
    Science Advisor

    I'm sure then that it will be no trouble for you to cite some examples of usage of "Gaussian metric" then.

    I should add that I frequently run into situations in which the terminology varies slightly depending on whether one is talking to a mathematician or a physicist, so some questioning about terminology is not all that uncommon.

    Meanwhile, you might want to tone down your personal attacks on some of our more respected members and the forum in general.
  18. Nov 19, 2006 #17
    What important people? And besides, some of us do know a bit about GR.
    Eh?! There's no need to just turn around and say "No, you're stupid!"

    If you are adament that Euclidean = Minkowski you're either working to a TOTALLY different definition to the rest of physics or you're the one 'exposing your stupidity even more'. The two are similar but have s critical difference. Euclidean metrics are positive definite, Minkowski isn't, it's a Lorentzian metric which means it has some positive and some negative eigenvalues, not all the same sign.

    Just check Wikipedia and you'll see that that is right. This forum doesn't exist to spread misinformation, it's here to aid people. Why did you bother to ask for advice here if you just say "What?! I'm not wrong, you are!" when someone actually tried to correct you?!
  19. Nov 19, 2006 #18


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    To prevent this from deteriorating further, I'm locking this thread until one of the physics mentors can review it. They may or may not decide to unlock it again, but things seem to be getting overheated in here to let it continue unmonitored.
  20. Nov 19, 2006 #19


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    It's garbage, and should remain locked.

    - Warren
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook