Path Integrals pp 163-200 | Cite as

Path Integrals and The Relation Between Classical and Quantum Mechanics

  • Martin C. Gutzwiller
Part of the NATO Advanced Study Institutes Series book series (NSSB, volume 34)


Our perception of the world outside ourselves can best be described in the terms of classical physics. The phenomena on the atomic scale require the ideas of quantum mechanics for their understanding which remain rather formal and abstract in our mind. Hence the need to tie the two realms together. Many problems can be treated successfully, although only approximately, with the help of classical mechanics, e.g. transition rates in chemistry and self-bound states in quantum field theory (solitons and instantons).


Periodic Orbit Canonical Transformation Position Space Conjugate Point Classical Trajectory 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    A. Einstein, Verhandlungen der Deutschen Physikalischen Gesellschaft 19, 82–92 (1917).Google Scholar
  2. 2.
    Max Born, The Mechanics of the Atom, republished by Frederick Ungar Publishing Co., New York, 1960. Neither Born nor Sommerfeld in Atombau und Spektrallinien (first published 1919) mention Einstein’s article. It is hard to understand the extent to which this important contribution of Einstein to early quantum mechanics was ignored by his contemporaries. W. Pauli wrote two long review articles on the old quantum mechanics, “Quantentheorie” in Handbuch der Physik, H. Geiger and K. Scheel, eds., Vol. 23, Springer-Verlag 1926, pp. 1–278, and “Allgemeine Grundlagen der Quantentheorie des Atombaues”, Chap. 29 in Muller-Pouillet’s Lehrbuch, Vol. 2, part 2, F. Vieweg, Braunschweig, 1929, p.p. 1709–1842. He also wrote the article “Einstein’s Beitrag zur Quantentheorie” for “Allbert Einstein als Philosoph und Naturforscher”, P. A. Schilpp, ed., W. Kohlhammer Verlag, Stuttgart, 1955, pp. 74–83, which contains a list of Einstein’s contributions to quantum theory. Yet, the article of reference 1 is mentioned nowhere.Google Scholar
  3. 3.
    There is a wide variety of articles in this area which started with the paper by J. Hadamard, J. Math. Pure Appl. 4, 27–87 (1898).MATHGoogle Scholar
  4. 3a.
    Further classic papers are by E. Artin, Abhandl. Math. Sem. Hamburg 3, 170–175 (1924);MathSciNetCrossRefMATHGoogle Scholar
  5. 3b.
    G. A. Hediund, Bull. Am. Math. Soc. 47, 241–260 (1939);CrossRefGoogle Scholar
  6. 3c.
    D. V. Anosov, Geodesic Flows on closed Riemamian manifolds with negative curvature, Proc. Steklov Inst. Math. 97, translated by Am. Math. Soc, Providence, 1969; E. Hopf, Bull. Am. Math. Soc. 77,863–877(1971).MathSciNetCrossRefMATHGoogle Scholar
  7. 4.
    The restricted three-body problem is the main topic in a monograph by V. Szebehely, Theory of Orbits, Academic Press, 1967.Google Scholar
  8. 4a.
    A lot of numerical exploration has been made since in order to establish various families of trajectories, e.g. a series of investigations by M. Henon, Ann. Astro. 28, 499 and 992 (1965);ADSGoogle Scholar
  9. 4b.
    M. Henon, Bull. Astro., Paris, 1, fasc. 1, p. 57 and fasc. 2, p. 49 (1966);MATHGoogle Scholar
  10. 4c.
    M. Henon, Astron. and Astrophys. 1, 223 (1969)ADSMATHGoogle Scholar
  11. 4d.
    M. Henon, Astron. and Astrophys. 9, 24 (1970).ADSMATHGoogle Scholar
  12. 5.
    The anisotropic Kepler problem has been investigated numerically in quantum mechanics by W. Kohn and J. M. Luttinger, Phys. Rev. 96, 1488 (1954);ADSCrossRefGoogle Scholar
  13. 5a.
    R. A. Faulkner, Phys. Rev. 184, 713 (1969).ADSCrossRefGoogle Scholar
  14. 6.
    The motion of the moon is the topic of E. W. Brown, An Introductory Treatise on Lunar Theory, Cambridge LTniversity Press 1896, republished by Dover Publications 1960; a less detailed account of lunar theory aud a very readable introduction to celestial mechanics is D. Brouwer and G. M. Clemence, Methods of Celestial Mechanics, Academic Press, 1961.MATHGoogle Scholar
  15. 7.
    These general properties are derived in many textbooks on classical mechanics, e.g. L. A. Pars, A Treatise on Analytical Dynamics, Heinemann, London, 1965;MATHGoogle Scholar
  16. 7a.
    C. L. Siegel and J. K. Moser, Lectures on Celestial Mechanics, Springer Verlay, 1971.CrossRefMATHGoogle Scholar
  17. 8.
    This important property was discovered by the astronomer G. W. Hill, Am. J. Math. 1, 5–26, 129–147, 245–260 (1878); it was rediscovered by the mathematician Floquet, Ann. de l’Ecole norm, sup (2), XII, 43 (1883); and it entered physics through the work of F. Bloch, Zeits. f. Phys. 52, 555 (1928).CrossRefMATHGoogle Scholar
  18. 9.
    A good survey of the mathematical accomplishments in this area is given by Jurgen Moser, Stable and Random Motions in Dynamical Systems, Princeton University Press, 1973.MATHGoogle Scholar
  19. 10.
    M. C. Gutzwiller, J. Math. Phys. 14, 139–152 (1973) contains the numerical investigations;MathSciNetADSCrossRefGoogle Scholar
  20. 10a.
    while M. C. Gutzwiller, J. Math. Phys. 18, 806–823 (1977) gives the analytical arguments.MathSciNetADSCrossRefGoogle Scholar
  21. 11.
    A particularly instructive mechanical example is worked out by M. Henon and C. Heiles, Astron. J. 69, 73 (1964).MathSciNetADSCrossRefGoogle Scholar
  22. 11a.
    In contrast, a sequence of mathematical mappings to imitate the properties of a Poincare map is studied by M. Henon, Q. Appl. Math. 27, 291 (1969).MathSciNetMATHGoogle Scholar
  23. 12.
    Bernoulli sequences are discussed in ref. 9. Another very readable account from a mathematician viewpoint can be found in V. I. Arnold and A. Avez, Problemes ergodignes de la mecanique classique, Gauthier-Villars, 1967; translated into English at Benjamin, New York, 1968.Google Scholar
  24. 13.
    C. Garrod, Rev. Mod. Phys. 38, 483 (1966).MathSciNetADSCrossRefMATHGoogle Scholar
  25. 14.
    Cf. S. F. Edwards and Y. V. Gulayev, Proc. Roy. Soc. (London) 279, 229 (1964);ADSCrossRefGoogle Scholar
  26. 14a.
    A. M. Arthurs, Proc. Roy. Soc. (London) 318, 523 (1970); also paper I in ref. 15.ADSCrossRefGoogle Scholar
  27. 15.
    The developments in the remainder are mostly taken from a series of papers to be designated as I, II, III, IV by the author, M. C. Gutzwiller, J. Math. Phys. 8, 1979 (1967);ADSCrossRefGoogle Scholar
  28. 15a.
    M. C. Gutzwiller, J. Math. Phys. 10, 1004 (1969);ADSCrossRefGoogle Scholar
  29. 15b.
    M. C. Gutzwiller, J. Math. Phys. 11, 1791 (1971);ADSCrossRefGoogle Scholar
  30. 15c.
    M. C. Gutzwiller, J. Math. Phys. 12, 343 (1971).ADSCrossRefGoogle Scholar
  31. 16.
    M. Morse, The Calculus of Variations in the Large, Am. Math. Soc, Providence, Rhode Island, 1935. A more easily accessible account is given by N. Seifert and W. Threlfall, Variationsrechnung im Grossen (Theorie von Marston Morse) B. G. Teubner, 1928; republished by Chelsea, New York, 1951; J. Milnor, Morse Theory, Princeton University Press, 1962.Google Scholar
  32. 17.
    The simplicity of Kepler’s problem in momentum space has been widely recognized in quantum mechanics since the treatment of V. Fock, Z. Phys. 98, 145 (1935).ADSCrossRefGoogle Scholar
  33. 17a.
    It rests on the connection with the group of rotations 0(4) which was first used by W. Pauli, Z. Phys. 36, 336 (1926).ADSCrossRefMATHGoogle Scholar
  34. 18.
    This result was first derived by the author in I, but independently, by A. Norcliffe and I. C. Percival, J. Phys. B. Ser. 2, 1, 774, 784 (1968).ADSGoogle Scholar
  35. 19.
    The classical quantization of non-separable mechanical systems with invariant tori has become a very active area of research, and only some early representative references can be given. R. A. Marcus, Faraday Discussions of the Chemical Society, 55, 34 (1973);CrossRefGoogle Scholar
  36. 19a.
    R. A. Marcus, J. Chem. Phys. 61, 4301 (1974)ADSCrossRefGoogle Scholar
  37. 19b.
    R. A. Marcus, J. Chem. Phys. 62, 2119 (1975);ADSCrossRefGoogle Scholar
  38. 19c.
    M. V. Berry and M. Tabor, Proc. Roy. Soc. London A 349, 101–123 (1976);MathSciNetADSCrossRefGoogle Scholar
  39. 19d.
    K. S. Sorbie, Molecular Physics, 32, 1577 and 1327 (1976);ADSCrossRefGoogle Scholar
  40. 19e.
    K. S. J. Nordholm and S. A. Rice, J. Chem. Phys. 61, 203 and 768 (1974);ADSCrossRefGoogle Scholar
  41. 19f.
    K. S. J. Nordholm and S. A. Rice, J. Chem. Phys. 62, 157 (1975).ADSCrossRefGoogle Scholar
  42. 20.
    This argument has been put in doubt by K. F. Freed, Faraday Discussions of the Chemical Society 55, 68 (1973). He reasons that q’ and q” need only be inside the volume d q of integration which is assumed to be of linear extent δq, and the momenta may differ by an amount δp. The variation of S(q”q’E) is, therefore, of the order δp δq, and destructive interference does not occur as long as δp δq<h. Trajectories might be important when they close approximately, even if they cannot do it exactly. Our skill in treating mechanical systems, ergodic or not, does not seem sufficient at present to resolve this issue. For a start we shall assume the extreme position of excluding all except perfectly closed trajectories.Google Scholar
  43. 21.
    The authors’ results have been misconstrued and misrepresented almost consistently in this respect. Cf. D. W. Noid and R. A. Marcus, J. Chem. Phys. 62, 2119 (1975);ADSCrossRefGoogle Scholar
  44. 21a.
    M. V. Berry and M. Tabor, Proc. Roy. Soc. Lond. A. 349, 101 (1976); and others in ref. 19 and 22. On the contrary, the first applications of (84) in III show that all periodic orbits are needed to give rise to resonance, not only a single primitive one. In IV, however, an example was investigated where only one periodic orbit was known at the time, and it was used to construct approximate eigenvalues for the lack of more information about the system, the anisotropic Kepler motion in two dimensions. With the many more periodic orbits which have been discovered since (cf. ref. 10), the approximate treatment of IV will have to be reexamined.MathSciNetADSCrossRefGoogle Scholar
  45. 22.
    The idea that the stability angle α includes the effect of the conjugate points was clearly stated in IV. Yet, Miller, J. Chem. Phys. 63, 996 (1975) states the contrary.CrossRefGoogle Scholar
  46. 22a.
    Even more unfortunate is his assertion in J. Chem. Phys. 64, 502 (1976) that the author’s method based on (89) is just “an approximate version of Keller’s formalism”. The reference is to J. B. Keller, Ann. Phys. (N.Y.) 4, 180 (1958);ADSCrossRefMATHGoogle Scholar
  47. 22b.
    J. B. Keller and S. I. Rubinow, Ann. Phys. (N.Y.) 9, 24 (1960) whose work is based entirely on the construction of wave functions for a classical mechanical system with invariant tori in phase space. Ergodic systems are not considered there.ADSCrossRefMATHGoogle Scholar
  48. 23.
    R. Balian and C. Bloch, Am. Phys. (N.Y.) 60, 401 (1970);MathSciNetADSMATHGoogle Scholar
  49. 23a.
    R. Balian and C. Bloch, Am. Phys. (N.Y.) 63, 592 (1971);ADSGoogle Scholar
  50. 23b.
    R. Balian and C. Bloch, Am. Phys. (N.Y.) 64, 271 (1971);MathSciNetADSMATHGoogle Scholar
  51. 23c.
    R. Balian and C. Bloch, Am. Phys. (N.Y.) 69, 76 (1972);MathSciNetADSMATHGoogle Scholar
  52. 23e.
    A. Voros, Am. Inst. H. Poincare 24, 31 (1976);MathSciNetGoogle Scholar
  53. 23d.
    L. Hormander, Acta Math. 127, 79 (1971);MathSciNetCrossRefGoogle Scholar
  54. 23f.
    J. J. Duistermaat, Comm. Pure Appl. Math. 27, 207 (1974);MathSciNetCrossRefMATHGoogle Scholar
  55. 23g.
    J. Chazarain, Inventiones Math. 24, 65 (1974);MathSciNetADSCrossRefMATHGoogle Scholar
  56. 23h.
    Y. Colin de Verdiere, Compositio math. 27, 83 and 159 (1973);MathSciNetMATHGoogle Scholar
  57. 23i.
    J. J. Duistermaat and V. W. Guillemin, Inventiones Math. 29, 39 (1975);MathSciNetADSCrossRefMATHGoogle Scholar
  58. 23k.
    A. Weinstein, in Fourier Integral Operators, Lecture Notes in Mathematics, Springer, 459, 341 (1975).CrossRefGoogle Scholar
  59. 24.
    I. C. Percival, J. Phys. B 6 L, 229 (1973).ADSCrossRefGoogle Scholar
  60. 25.
    N. Pomphrey, J. Phys. B 7, 1909 (1974).ADSCrossRefGoogle Scholar
  61. 26.
    W. H. Miller, J. Chem. Phys. 64, 2880 (1976).MathSciNetADSCrossRefGoogle Scholar
  62. 27.
    Cf. ref. 22. Also W. H. Miller, J. Chem. Phys. 56, 38 (1972) and, in particular,ADSCrossRefGoogle Scholar
  63. 27a.
    W. H. Miller, J. Chem. Phys. 62, 1899 (1975) are of interest in this connection. Path integral arguments are used in order to compute transition rates in quantum statistics. Periodic orbits now occur in the inverted potential and the stability exponents are indeed the frequencies for the transverse degrees of freedom. The expression (6) gives their occupation probability.ADSCrossRefGoogle Scholar
  64. 28.
    This fact has been known for some time, in particular since the geodesics on a surface of negative curvature can be constructed quite explicitly, cf. ref. 3. A detailed numerical study in a system where no such complete description is available, was started by G. H. Lunsford and J. Ford, J. Math. Phys. 13, 700 (1972). They try to give a mathematical formulation (based on their empirical data) to the way in which periodic orbits occur in an ergodic system.MathSciNetADSCrossRefMATHGoogle Scholar

Copyright information

© Springer Science+Business Media New York 1978

Authors and Affiliations

  • Martin C. Gutzwiller
    • 1
  1. 1.IBM Thomas J. Watson Research CenterYorktown HeightsUSA

Personalised recommendations