Scribd Logo
Upload

Welcome to Scribd!

  • 110 views

    Elementary Proof of Zeeman's Theorem PDF

    Uploaded by

    Lexi
    This document provides an elementary proof of Zeeman's theorem that global causality implies the Lorentz group. It first proves that nonaffine conformal transformations necessarily violate global causality. It then shows that the group of causal automorphisms is contained within the proper orthochronous Poincaré group, establishing Zeeman's theorem with a group-theoretic proof using properties of null lines and the behavior of transformations under inversion.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Elementary Proof of Zeeman's Theorem PDF

    Uploaded by

    Lexi
    110 views5 pages
    This document provides an elementary proof of Zeeman's theorem that global causality implies the Lorentz group. It first proves that nonaffine conformal transformations necessarily violate global causality. It then shows that the group of causal automorphisms is contained within the proper orthochronous Poincaré group, establishing Zeeman's theorem with a group-theoretic proof using properties of null lines and the behavior of transformations under inversion.

    Original Title

    Elementary proof of zeeman's theorem.pdf

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document provides an elementary proof of Zeeman's theorem that global causality implies the Lorentz group. It first proves that nonaffine conformal transformations necessarily violate global causality. It then shows that the group of causal automorphisms is contained within the proper orthochronous Poincaré group, establishing Zeeman's theorem with a group-theoretic proof using properties of null lines and the behavior of transformations under inversion.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    110 views5 pages

    Elementary Proof of Zeeman's Theorem PDF

    Uploaded by

    Lexi
    This document provides an elementary proof of Zeeman's theorem that global causality implies the Lorentz group. It first proves that nonaffine conformal transformations necessarily violate global causality. It then shows that the group of causal automorphisms is contained within the proper orthochronous Poincaré group, establishing Zeeman's theorem with a group-theoretic proof using properties of null lines and the behavior of transformations under inversion.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 5

    International Journal of Theoretical Physics, Vol. 19, No.

    12, 1980

    Elementary Proof of Zeeman's Theorem


    A. J. Briginshaw

    Department of Mathematics, The City University, London EC1V OHB

    Received May 15, 1980

    We prove that the nonaffine conformal transformations of four-dimensional


    Minkowski space are necessarily global causality violators, and use this fact to
    obtain an elementary group-theoretic proof of the theorem that global causality
    implies the Lorentz group.

    1. INTRODUCTION

    The proof of the theorem that causality implies the Lorentz group was
    first given by Zeeman (1964). In a previous paper (Briginshaw, 1980), the
    author has explained how the class of causal automorphisms is contained
    in the conformal group and pointed out how the presence of the inversion
    map among the generators of the conformal group shows itself by its
    violation of global causality. It seems natural, therefore, to use this fact to
    fabricate an elementary group-theoretic proof of the causality theorem;
    that is exactly the concern of th,e present paper.

    N o t a t i o n . Let M 4 denote the vector space of quadruples of real num-


    bers of the type (x o, x l, x 2, x3) with the (indefinite) inner p r o d u c t , given
    by
    x.y = xoY 0 - xlyl - x2Y2 - x3Y3

    We define two relations in M4, as does Zeeman (1964): (i) x < y iff
    (x-y).(x-y)>O and Xo <Yo (this is a partial order); (ii) x < . y iff ( x -
    y).(x-y)--O and x o <Yo (this is not a partial order). We denote the pair
    ( M 4, <:) by/~t4 and the pair (M4, < - ) by M 4.

    899
    0020-7748/80/1200-0899503.00/0 ,~ 1980Plenum PublishingCorporation
    900 Bri~n#l~w

    Causal Automorphisms
    ^ p, ^

    A map f: M4---~M4 is said to be an automorphism of M 4 if f is a


    bijection of M 4 such that both f , f - 1 preserve < , i.e., x <yc~fx <fy.
    9 .^ .~

    A map f: M 4--->M4 is similarly an automorphism of M 4 if f is bijective


    and such that x < .yc~fx < .fy.
    Each collection of automorphisms is a group^under composition, and
    we shall refer to the.group of automorphisms of M 4 as G and the group of
    automorphisms of M 4 as G.
    Lemma 1

    This is the same result as Lemma 1 of Zeeman (1964).

    2. T H E INTERVAL T O P O L O G Y

    Define 0x, y = (z: x < z < y }; then the collection ((0x, y }: x, y E M 4} is a


    base of neighborhoods for a topology on M4, that we call the interval
    topology (Nanda, 1976).
    We shall denote the interval topology by 1 and the Euclidean metric
    topology by E; we now have the following two results as stated and proved
    by N a n d a (1976).
    Lemma 2

    I=E

    This is Lemma 1 of N a n d a (1976).


    Lemma 3. I f f E G t h e n f i s a homeomorphism of ( M n, E).
    This is Lemma 2 of N a n d a (1976).

    Corollary. If f ~ G then f is a homeomorphism of ( M 4, E).


    Proof.

    fEG~fEG.

    Definition: Null Line. The set Np, q= (x: x = p + ttb, I~ER, p - q = b , b.b
    = 0} is said to be the null line through p and q.
    Lemma 4. There are no null line triangles in M 4.
    ElementaryProof of Zeem~n'sTheorem 901

    Proof Without loss of generality we may prove that if 0 < . a < . fl, and
    if 0 < . fl then 0, a, fl are points of the same null line. Write a - fl =p, so that
    p . p = 0 ; now fl+p is a null vector, therefore fl.fl+2fl.p+p.p=O, i.e.,
    fl.p = 0. If fl, p are not in the same direction, then

    i lp,+ 2pz+ 3p l<(9=+/ zz +B3)


    2 1/= "(p,
    2 +pz+p )'/==l&llp01

    =B0P0
    .'.fl-p>0

    Thus 0, a, fl are collinear.


    Lemma 5. If f E G then f maps every null line onto a null line.
    Proof Let Np be an arbitrary null line through p, and let q G Np be such
    that p<.q. Suppose that p, q are fixed and that ~ENp is such that
    p <- ~ < . q. Since f ~ G it follows that fp < .f~ < .fq, and that fp, f~, fq are
    collinear.
    Definition: Conformal Map. Let f: M4--+M4 (compactified M4) be such
    that f is continuous at every nonsingular point and such that for all
    appropriate x, y and their maps x'=fx, y'---fy, we have

    a x ' . dy' a x . dy
    ( ax'.ax')'/:( (ax.ax)l/:(ay.a'y)

    Then f is said to be a conformal map of M4; under composition the


    collection of such maps is a group, called the conformaI group, C.
    Definition: The Inversion Map. The conformal map I: M4--+M4 given
    by I x = x / ( x . x ) is said to be an inversion of M 4.

    3. G E N E R A T O R S OF T H E C O N F O R M A L G R O U P

    In the first of a series of papers concerning the covariance group


    associated with Maxwell's equations, Bateman (1909) proved that the
    conformal group is that generated by dilatations, Poinear6 (inlaomoge-
    neous Lorentz) transformations, and inversions. In the next lemma we
    prove that I is a global causality violator.
    Lemma 6
    902 Bri~nvlmw

    Proof Consider the null line N = (x: x = p +/~b,/z~R, b-b--O, by'O).


    Now, if x EN, then

    Ix= P+l~b ~ + #[(p.p)b-2(p.b)p]


    p'p+21~b p'p (p.p)[p.p+2#(p.b)]

    and there is clearly one singular point on N, where the value of/~ is

    ,u* = - p "p
    2p.b

    In addition, the expression for Ix has the form

    Ix= p + aX
    p.p

    where X = ( p . p ) b - 2 ( p . b ) p is null. Thus the map of a null line is a null


    line, as expected; in addition

    OL~ ,
    2(p.p)(p.b)
    [1 pp
    p.p+21~(p.b )
    ]
    N o w p is an arbitrary point on N and we may choose it such that 0 <p, i.e.,
    p . p > 0 and po > 0; equally well, without loss of generality, we may con-
    sider 0 < - b , i.e., b.b=O and bo>0, in which c a s e p - b >0. Now, if/~1 </~2 <
    #*, or if/~* <#1 </~2, then a i < ct2, and the relation < . is preserved. On the
    other hand, if/L 1< # * </~a, then a 2 < a l , and the relation < . is violated.
    Notation. We refer to the affine group generated b y translations,
    orthochronous Lorentz transformations and dilatations as P t * -
    Remark. It is easy to prove the following catalog of results concerning
    the generators of C: if ?,, t a, A are, respectively, arbitrary dilatations,
    translations, and Lorentz transformations, then
    (i) A o I - - I o X -1
    (ii) A oI = I oA
    ( i i i ) ) ~ o A = A o)~
    (iv) Aota=tA~oA
    (V) )~Ota=tXaO)~
    There is no similar commutation relation for ta and I. We now prove the
    main theorem.
    Theorem

    Pt* =G
    Elementary Proof of Zeeman's Theorem 903

    Proof. It is clear that P~* C G. Also, i f f E G, t h e n f is continuous, f r o m


    L e m m a 3; and f preserves null lines, from L e m m a 5; from Weyl (1923) it
    follows that f ~ C. Now, if f = a 1 o a2 o o~3 o a4 . . . . . aN, where each a i is an
    element of the generating subgroups of dilatation, translation, Lorentz, and
    inversion maps, then we m a y arrange f thus:

    f=)~oAotaolOtbolOtcOl ..... tpolotq

    where ~, A are, respectively, a dilatation and a Lorentz transformation a n d


    ta, t b . . . . . tq are translations. Consider the image of an arbitrary null line
    under f; without loss of generality we may take q = 0, and consider the line
    N={x:x=~+l~rl, where ~-~=~0,~/.,/=0,~/=/=0}. Suppose at least one
    b ..... p is zero; we write K ( y ) = { x : ( x - y ) . ( x - y ) = O } and refer to it as
    the null cone at y. It is clear that, for each f E C, with b , . . . , p nonzero, there
    is a corresponding network of finitely m a n y singular surfaces. Thus, I is
    singular on K(0) (as is tv oi), I o t v o I is singular on K ( O ) u K ( - p / p . p ) (as
    is tjolotpoI); if we now write A p , j = l + 2 ( p . j ) + ( p . p ) ( j . j ) , then
    Iotjo l o t , o I is singular on K ( O ) u K ( - p / p . p ) u K ( ( - p ( j . j ) / A p , j } -
    {j/Ap,j}), and so on.
    N o w consider the point x* at which N intersects the cone K(0), and
    let/x* be the parameter associated with x* as in L e m m a 6. Let/~1,/x2,/x3 be
    close enough to/x* to ensure that x i < . x * < . x 2 < . x 3 but that if S is the
    union of all singular surfaces for f, Vu, v E S - K ( O ) , u < . x I < . x 3 <.v.
    Thus I x 2 < . I x 3 < I x l ; also, for every other m a p of the collection
    comprising the nonaffine part of f these relations are preserved.
    Case (i): Suppose at least one of b, c ..... p :/: 0 and that A is nonortho-
    chronous, then < - is violated for the pair x 2, x 3.
    Case (ii): Suppose at least one of b ..... p ~ 0 and that A is orthochro-
    nous, then < . is violated for the pair xl, x 2.
    Case (iii): Suppose that every b ..... p = 0 and that A is nonorthochro-
    nous, then < . is violated for every pair on N. We conclude that every
    b, c ..... p=O and that A is orthochronous, in which c a s e f E P l ' * .
    Remark. For n > 3, the causal automorphisms of M, are always ele-
    ments of the conformal group. However, for n--2, that is not true; there
    are global causal automorphisms of M 2 which are not conformal maps.
    Thus the theorem is not valid for n = 2.

    REFERENCES

    Bateman: H (1909). Proceedings of the London Mathematical Society, 7, 70.


    Briginshaw, A. J. (1980). International Journal of Theoretical Physics, 19, 329.
    Nanda, S. (1976). Mathematical Proceedings of the Cambridge Philosophical Society, 79, 533.
    Weyl, H. (1923). Mathematicsche Ana~,se des Rawr~roblems, Springer, Berlin, pp. 71-75.
    Zeeman~ E. C. (1964). Journal of Mathematical Physics, 5, 490.

    You might also like