Sergey P.

Novikov1
1.Classical and Modern Topology.
2.Topological Phenomena in Real World Physics
According to the opinion of the Ancient Greeks, the famous real and myth-
ical founders of Mathematics and Natural Philosophy like Pythagoras, Aris-
arXiv:math-ph/0004012v1 11 Apr 2000

totle and others, in fact, borrowed them from the Egyptian and Middle East
civilizations. However, what had been told before in the hidden mysteries
Greek scientists transformed into written information acceptable for every-
body. Exactly after that the development of science in the modern sense
started and had already reached a very high level 2000 years ago. Therefore
you may say that the free exchange of information and making it clear for
people have been the most important discoveries of Greeks. I would say it
is the basis of our science now. As you will see, any violation of this fun-
damental rule does serious harm to our science and inevitably leads to its
decay.
1.Classical and Modern Topology.
Prehistory. First fifty years of Topology. The first important topo-
logical ideas were observed by famous mathematicians and physicists like
Euler, Gauss, Kelvin, Maxwell and their pupils, during the XVIIIth and
XIXth Centuries. As everybody knows, it was Poincare’ who really started
Topology as a branch of Mathematics in the late XIXth Century. Many
top class mathematicians participated in the development of Topology in the
first half of our century. A huge number of mutually connected fundamen-
tal notions were invented: degree of maps and singularities of vector fields,
homotopy and homology groups, differential forms and smooth manifolds,
the fundamental idea of transversality, the simplicial/cell(CW) and singular
complexes as tools for studying topological invariants, braids, knot invari-
ants and 3-manifolds, coverings, fibre bundles and characterisic classes and
many others. Deep connections with Qualitative Analysis, Calculus of Vari-
ations, Complex Geometry and Dynamical Systems were established in this
period. Combinatorial Group Theory and Homological Calculus started from
topological sources. A great new field of topological objects unknown to the
classical mathematics of the XIXth Century appeared finally in the 1940s.
1
Math Department and IPST, University of Maryland, College Park, MD 20742-2431,
USA and Landau Institute for Theoretical Physics, Moscow 117940, Kosygina 2; phone in
USA 301-4054836(o), fax in USA 301-3149363, e-mail [email protected]; this work is
supported by the NSF Grant DMS9704613

1
At that time this new area was known to few number of mathematicians
only. However there was a very high density of really outstanding scientists
among them.
1950s and 60s: Golden Age of Classical Topology. The fundamen-
tal set of algebraic ideas unifying all these branches of mathematics appeared
in the 40s; a new era started about 1950. Spectral sequences of fibre bun-
dles, sheafs, highly developed homological algebra of the groups, algebras
and modules, Hopf algebras and coalgebras, were invented and heavily used
for the calculation of topological invariants needed for the solution of the
fundamental problems of topology. Let me point out that, in many cases,
it was a completely new type of calculations based on the deep combination
of the very general ”categorial properties” of these quantities with very con-
crete geometric, algebraic or analytical study of a completely new type. In
the previous period, people even had no dreams as to how they could be
calculated. Regular methods were built to calculate homotopy groups, for
example. It was one of the most difficult problems of topology. A lot of
them were computed completely or partially including the homotopy groups
of spheres, Lie groups and homogeneous spaces. The topologically important
cobordism rings were computed and used in many topological investigations.
The famous signature formula for differentiable manifolds was discovered.
It has an innumerous number of applications in the topology of manifolds.
Besides that, this formula played a key role in the proof of the so-called
Riemann-Roch theorem in Algebraic Geometry and later in the study index,
the famous homotopy invariant of Fredholm operators.
The mutual influence of Topology and Algebraic Geometry during that
period led to the broad extension of the ideas of homology: the extraordi-
nary (co)homology theories like K-theory and cobordisms appeared. They
brought a new type of technic to topology with many applications. Rep-
resentation theory and complex geometry of manifolds deeply unified with
homological algebra and Hopf algebras. The technic of formal groups ap-
peared here. It has been applied in particular for the improvement of the
calculations of stable homotopy groups of spheres. As everybody knows, dur-
ing this period topology solved the most fundamental problems in the theory
of multidimensional smooth manifolds:
Nontrivial differentiable structures on multidimensional spheres were dis-
covered on the basis of the results of algebraic topology combined with a
new understanding of the geometry of manifolds and bundles. The multidi-
mensional analog of the Poincare Conjecture and H-cobordism theorem were

2
proved. Counterexamples to the so-called ”Hauptvermunung der Topologie”
were found. A classification theory for the multidimensional smooth (and for
PL-manifolds as well) was completely constructed. The role of the fundamen-
tal group in this theory led to the development of a new branch of algebra:
the algebraic K-theory. Topological invariance of the most fundamental char-
acteristic classes was finally proved. The so-called ”Annulus Conjecture” was
proved. No matter how elementary these results can be formulated, nobody
has succeeded to avoid the use of a whole bunch of results and tools of al-
gebraic and differential topology in the proof. The classification theory for
the immersions of manifolds was constructed. The theory of multidimen-
sional knots was constructed. Several classical problems of the theory of
3-manifolds also were solved during that period: the so-called Dehn’s pro-
gram was finished after a 50 year break; the algorithm for recognizing the
trivial knot in three-space has been theoretically constructed as a part of
the deep understanding of the structure of 3-manifolds and the surfaces in
them. As a by-product of topology, the fundamental breakthrough in the
topological understanding of generic dynamical systems was reached. A new
great period started in this area. Qualitative theory of foliations has been
constructed with especially deep results for 3-manifolds.
As a summary, I would like to add one more very important characteristic
of the topological community in the golden age of classical topology:
All important works have been carefully checked. If some theo-
rem had not been proved, it immediately became known to every-
body.
So you can find a full set of proofs in the literature. Unfortunately, a
full set of textbooks covering all these developments (1950-1970) has not
been written yet. Many modern textbooks are written in a very absract way.
Even if they cover some pieces formally, it is more difficult to read them than
the original papers. Let me recommend to you the Encyclopedia article [1]
written exactly for the exposition of these ideas.
1970s: Period of decay. In my opinion, the period of the 1970s can
be characterized as a period of decay for classical topology. There are many
indications for that. Several leading scientists left topology for the new ar-
eas like algebra and number theory, riemannian and symplectic geometry,
dynamical systems and complexity theory, functional analysis and represen-
tations, PDEs, and different branches of mathematical/theoretical physics....
It is certainly a good characterization of the community if it could gener-
ate such a flux of scientists in many different areas and bringing to them

3
completely new ideas. Anyway, this community dispersed.
What can we say about the topological community after that?
First of all, some important new ideas appeared in the 70s (like localiza-
tion technic in homotopy topology, the nicely organized theory of the rational
homotopy type, hyperbolic topology of 3-manifolds). However, a huge in-
formational mess was created in the 1970s. Let me point out that a
series of fundamental results of that period was not written, with full proof,
until now. Let me give you a list:
Sullivan’s Haupvermutung theorem was announced first in early 1967.
After the careful analysis made by Bill Browder and myself in Princeton of
the first version in May 1967 (before publication), his theorem was corrected:
a necessary restriction on the 2-torsion of the group H3 (M, Z) was missing.
This gap was found and restriction was added. Full proof of this theory has
never been written and published. Indeed, nobody knows whether it has been
finished or not. Who knows whether it is complete or not? This question
is not clarified properly in the literature. Many pieces of this theory were
developed by other topologists later (they used sometimes different ideas).
Nobody has unified them until now. Indeed, these results were used by
many others later. In particular, the final Kirby-Siebenmann classification
of topological multidimensional manifolds therefore is not proved yet in the
literature.
The second story is the theory of Lipshitz structures on the manifolds. In
the mid-seventies Sullivan distributed a preprint containing the idea how to
prove existence and uniqueness of such structures on the manifold N n , n 6=
4. This idea obviously included (for the uniqueness) the direct use of the
Annulus Conjecture (and therefore of all ideas and technic needed in the
proof of topological invariance of the rational Pontryagin Classes inside).
Proof of the Lipshitz Theory has never been published. Indeed, many years
later, already in the 1990s, some brilliant younger scientists developed a very
nice theory of Fredholm (elliptic) operators on Lipshitz manifolds. As a
corollary, they claimed that a new proof of topological invariance of rational
Pontryagin classes has been obtained from Analysis (it was a problem posed
by Singer in the 60s). Young scientists made a ”logical circle” believing in the
classical results. Nobody told them that corresponding theorems have never
been proved. How could it happen? This funny story shows the modern
state of information in the topological community.
Another informational mess has been created in 3D Hyperbolic topology.
This beautiful area was started by Thurston in the mid-70s. For many years

4
people could not find out what was proved here. In this area the situation has
been finally resolved: it has been aknowledged that these methods lead to the
proof of the original claim (the so-called Geometrization Conjecture) only for
the special class of Haken manifolds. The Geometrization Conjecture means
more or less that (in the case of closed 3-manifolds) the fundamental group
can be realized as a discrete subgroup acting in the 3D Hyperbolic space if
trivial necessary conditions are satisfied: all its abelian subgroups are cyclic
and π2 = 0. However, it is difficult to find out who actually proved this
theorem? It seems for me that the younger mathematicians who managed
to finish this program did not receive proper credit.
I would like to mention that this kind of informational mess has happened
since 1970 not only in topology. For example, the famous results of KAM
in the three-body problem known since the early 60s were found recently
unproved. It was announced for the first time at the Berlin Congress last
year. In this case, some works supposedly containing full proof were published
in the first half of the 60s. Does this mean that nobody accually read them
for at least 30 years?
Do you think that algebra is better? Let me tell you as a curious remark
that all works of the Steklov Institute (i.e. Shafarevich’s) school in algebraic
number theory, algebraic geometry and theory of finite p-groups awarded
by the highest (Lenin and State) prizes in the former Soviet Union since
1959, did not contain full proof. The gaps in the proofs were found many
years later. Not all these gaps were really deep. However, some of these
authors knew their mistakes many years before they became publicly known
and could not correct them. They managed to fulfill gaps after many years ,
using much later technical achievements made by other people. Does it mean
that in the corresponding time, despite many public presentations, nobody
in fact read these great works? Can we say that all proofs are known now in
all these cases?
There are much worst cases in modern algebra indeed. How many of you
know that the so-called classification of simple finite groups did not exist as
a mathematical theorem until now? In this case we can even say that in fact
(as a few number of real experts have known since 1980) no one work existed
claiming that this problem was finished in this work. All public opinion has
been based only on the ”New York Times Theorem” for the past 20 years.
1980s and 90s: Period of recovery. The role of Quantum Field
Theory.
It became clear already in the late 70s that modern quantum field the-

5
ory started to generate new ideas in topology. It gave several new alterna-
tive ways to construct topological invariants: Path integral for the metric-
independent actions on manifolds was used for the first time. The famous
self-duality equation appeared first in the works of physicists. It was applied
in the 80s for the solution of fundamental topological problems in the theory
of 4-manifolds. Quantum string theory brought in the early 80s new deep
results in the theory of the classical Fuchsian groups and moduli spaces. At
first physicists (like t’Hooft and Polyakov) were not interested very much in
such by-products of their activity. They always said that they were doing
physics of the real world, not pure mathematics. However the next wave
of brilliant physicists (like Witten, Wafa and others) started to solve prob-
lems of pure mathematics. Such purely topological subjects like the Morse
theory and cobordism theory associated with action of compact groups on
manifolds, were developed in the 80s from the completely new point of view.
Symplectic Topology reached a very high level in the late 80s. We are facing
now impressive development of Contact Topology.
Certainly Quantum Theory brought new beautiful ideas. Besides that,
the fundamental new invariants of knots were discovered in the 80s by the
topologists who came from functional analysis and theory of C ∗ algebras.
These invariants also received quantum treatment in the late 80s. The beau-
tiful connection of the specific Feinmann diagrams with surfaces was bor-
rowed from physics literature. It became a very effective tool for the solution
of several topological problems. Unfortunately, only a few number of math-
ematicians learned this technic and started to apply it in topology. I know
only Singer, Konzevich and a very small number of others. Even if you will
add here the names of pure mathematicians who learned this with the in-
tention to do real physics, this list will increase inessentially. I do not count
here people who were trained originally in the physics community. A large
number of them moved into pure mathematics with the intention to prove
rigorous theorems about the models serving (in their opinion) as an ideal-
ization of theoretical physics. They call this area Mathematical Physics, but
not everybody agrees with such a definition of mathematical physics. This
community does not do topology.
I would like to make a remark here concerning a beautiful work of Konze-
vich calculating certain Chern numbers on the punctured moduli spaces of
Riemann Surfaces through the special solution to the KdV hierarchy. This
folmula has been known as a Witten Conjecture. You have to specify for this
some compactification of the moduli spaces of punctured Riemann surfaces,

6
otherwise it makes no sense. Konzevich accually proved this formula for one
specific (”Strobel-Penner”) compactification in 1991. What about the stan-
dard Deligne-Mumford compactification? Konzevich claimed in 1992 in his
work in Inventiones that it is true. However, no proof has been presented
until now. So this problem is open. There was a mistakable statement about
this at the Berlin Congress.
Let me point out that the physics community did not create any infor-
mational mess in topology. According to their training tradition, theoretical
work produces Conjectures which should be proved only by some kind of
experiment. Starting to do beautiful nonrigorous mathematics, they do not
claim that they ”proved” something. They are saying that they ”predicted
this fact”. In the case of pure mathematics, the final proof done by pure
mathematicians these people may treat as an ”experimental confirmation”.
In the past ten years several deep results have been obtained in the 4D topol-
ogy. We cannot say this about 3D topology: quantum invariants here created
some sort of ”invariantology”: a lot of people are constructing topological
invariants but no one new topological result has been obtained for almost 10
years. Indeed, these ideas look beautiful in some cases. In my opinion, new
deep results will appear after better understanding of the relationship of new
invariants with classical topology.

7
 Topological Phenomena in Real World Physics
Topological ideas in physics in the period of the early 80s. I
spent about 10 years learning different parts of Modern Theoretical Physics
in the 60s and 70s. After joining the physics community (i.e. Landau school)
in the early 70s I found out that most physicists did not know at all the new
areas of mathematics like topology, dynamical systems and algebraic geome-
try, including analysis on Riemann surfaces. The quantum people knew some
extracts from the group theory and representations because they needed it
in Solid State Physics as well as in Elementary Particles Theory since the
1960s. A lot of them knew something about Riemannian Geometry because
of the Einsteinian General Relativity. However, these people had already
heard something about the new mathematics of the XXth century and badly
wanted to find its realization in physics. You have to take into account that
between them there was a great number of extremely talented people at that
time with very good training in practical mathematics. In some cases I was
able to help physicists (like Polyakov, Volovic and some others) to learn and
to use topology in the 70s. I worked this period in General Relativity (Homo-
geneous Cosmological Models) and Periodic KdV Theory with my pupils and
collaborators. We found completely nonstandard applications of Dynamical
Systems and Algebraic Geometry in these areas. However, until the late 70s
I did not produce any new topological ideas. My very first topological work
in physics was made in 1980 (see [2]). I started to use in the spectral theory
of the Schrodinger operators in periodic lattice and magnetic field the idea of
transversality applied to the families of Hermitian matrices or elliptic oper-
ators on the torus. This idea led to the discovery of the series of topological
invariants, Chern Numbers of Dispersion Relations. They are well-defined
for the generic operators only. The classical Spectral Theory in mathemat-
ics never considered such quantities because they are not defined for every
operator with prescribed analytical properties of coefficients. The ideology
of transversality is important here. This work was not understood by my
colleagues-physicists at that time (the vice-editor of JETP did not want to
publish it as ”nonphysical”, so I published it in the math literature). Peo-
ple thought that the important integer-valued observable quantities in Solid
State Physics may come from symmetry groups only. Indeed, the Integral
Quantum Hall phenomenon was discovered soon. Some famous theoretical
physicists rediscovered my mathematical idea after that. It is certainly a
sum of the Chern classes of dispersion relations below the Fermi level.

8
 My next topological discovery was made in the joint work with student
I.Schmelzer in 1981, dedicated to the very special problem of classical me-
chanics and hydrogynamics (see [3]). I immediately realized its value for
modern theoretical physics, as well as for mathematics, and developed this
idea in several directions in the same year ([4]). The series of work in the
Theory of Normal Metals which I am going to discuss today, is also one of
by-products of that discovery. Doing the Hamiltonian factorization proce-
dure for the top systems on the phase spaces like T ∗ (SO3 ) by the action of
S 1 , you are coming to the systems mathematically equivalent to the motion
of the charge particle on the 2-sphere. This sphere is equipped by some
nontrivial Riemannian metric. What is important and has been missed by
the good experts in analytical mechanics like Kozlov and Kharlamov is that
the effective magnetic field like Dirac monopole appears here for the nonzero
values of the ”area integral” associated with S 1 -action. It means precisely
that the magnetic flux along the sphere is nonzero. The reason for this is
that the symplectic (Poisson) structure after factorization is topologically
nontrivial. In terms of modern symplectic geometry, the magnetic field is
equivalent to the correction of the symplectic structure. This fact is not
widely known in the geometric community even now. The appearance of the
topologically nontrivial symplectis structures after S 1 -factorization of sym-
plectic manifolds was independently discovered and formulated in geometric,
nonphysical terminology in 1982 in the beautiful work [6] for different goals
(calculating of integrals).
It has been realized in [3, 4] that the action functional for such systems
is in fact a closed 1-form on the spaces of loops. These functionals have been
immediately generalized for higher dimensions, to the spaces of mappings F
of q-manifolds in some target space M where a closed q + 1-form is given
instead of magnetic field. We are coming finally to the action functional
well-defined as a closed 1-form on the mapping spaces F . The topological
quantization condition for such actions was formulated in 1981 [4, 5] as a
condition that this closed 1-form should define an integral cohomology class
in H 1 (F, Z). It is necessary and sufficient for the Feinmann amplitude to be
well-defined as a circle-map

exp{iS/h} : F → S 1
For the case q = 1 the original Dirac requirement was based on a different
idea: the magnetic field should be a Chern class for the line bundle whose

9
space of sections should serve as a Hilbert space of states for our Quantum
Mechanics. Therefore it should be integral in H 2 (M, Z).
In pure topology and in the Calculus of Variations these ideas led to the
construction of the Morse-type theory for the closed 1-forms on the finite-
and infinite-dimensional manifolds. Let me refer to the last publication of
the present author (with P.Grinevich) in this direction [7] where the survey
of results and problems is discussed. I would like to point out that for
the compact symplectic manifolds the action functional for any nontrivial
Hamiltonian system is multivalued. The cohomological class of symplectic
form cannot be trivial here. I do not know of such cases in real physics
where the symplectic manifold is compact. However, even in the community
of symplectic geometers nobody paid attention to such properties of action
functional until the 90s.
After that I started to think about different aspects of the Hamiltonian
Theory where the class of one-valued functions naturally can be extended to
the class of all closed 1-forms. For every symplectic (Poisson) manifold M
with H 1(M) 6= 0 we may consider Hamiltonian Systems generated by the
closed 1-form dH where the function H is multivalued. Instead of energy
levels H = const we have to consider nontrivial codimension 1 foliation
dH = 0 with Morse (or Morse-Bott) singularities. We are coming to the
topological problems of studying such foliations. It has been posed in [5].
Several participants of my seminar (A.Zorich, Le Tu Thang, L.Alania) have
made very important contribution to the study of this subject. Interesting
quasiperiodic structure appears here. It is not revealed fully in my opinion
(see references and discussions in the article [9]).
Multivalued Hamiltonians in real physics. I started to look around
in 1982 asking the following question: can you find such systems in real
physics where Hamiltonian or some other important integral of motion is
multivalued (i.e. dH is well-defined as a closed 1-form)? Much later people
realized that in the theory of the so-called Landau-Lifshitz equation (which is
a well-known physical integrable system with zero-curvature representation
elliptic in the spectral parameter) the momentum is a multivalued functional.
At that time (1982) I found only one such system describing motion of
the quantum (”Bloch”) electron in the single crystal D-dimensional normal
metal (D=1,2,3) under the influence of the homogeneous magnetic field B.
We are working here with one-particle approximation for the system of Fermi
particles whose temperature is low enough. For the zero temperature our
electrons fill in all one-particle quantum Bloch states ψp below the so-called

10
”Fermi Level” ǫ ≤ ǫF . Its value depends on the number of electrons in the
system. It is the intrinsic characteristic of our metal. The index p here may
be considered finally as a point in the torus T D defined by the reciprocal
lattice dual to the crystallographic one

p ∈ T D , T D = R3 /Γ∗
There is a Morse function ǫ(p) : T D → R (dispersion relation) such that
the domain ǫ ≤ ǫF in the torus T D is filled in by Bloch electrons. Its boundary
ǫ = ǫF is a closed surface MF ⊂ T D for D = 3. We call it Fermi Surface.
It is homologous to zero in the group H2 (T 3 , Z. For finite but very small
temperature all essential events are happening nearby the Fermi Surface.
Add now a homogeneous magnetic field to our system (i.e. put metal
in the magnetic field B). Nobody succeeded in constructing a suitable well-
founded theory for the exact description of electrons in the magnetic field and
lattice. Irrational phenomena appear in the spectral theory of Schrodinger
operators and destroy all geometric picture. However, since the late 50s
physicists have used some sort of adiabatic approximation which they call
”semiclassical”. Let me warn you that this approximation has nothing to do
with the standard understanding of semiclassical approximation. We take
dispersion relation ǫ(p) as a function on the torus T 3 extracted from the
exact solution of the one-particle Schrodinger operator in the lattice with-
out magnetic field. We consider a phase space T 3 × R3 with coordinates
pi , xj , i, j = 1, 2, 3 and Poisson bracket of the form:

{xj , xk } = 0, {xj , pk } = δkj , {pj , pk } = Bjk

where Bjk (x) are components of the magnetic field B treated as a 2-form.
Our space R3 is Euclidean, so we can treat magnetic field as a vector B with
components B j .
Take now the function ǫ(p) as a Hamiltonian. It generates through the
Poisson structure above a Hamiltonian system in the phase space T 3 ×R3 . For
the homogeneous (i.e., constant) magnetic field we can see that our phase
space projects on the torus T 3 with Poisson bracket {pj , pk } = Bjk . This
Poisson bracket has a Casimir (Annihilator) CB (p) = ǫijk pi Bjk = B i pi . This
P
Casimir is multivalued: it is defined by the closed 1-form ωB = i B i dpi
on the torus. As you will see, this is the main reason for the appearence of
nontrivial topological phenomena in this problem.

11
 Our Hamiltonian H = ǫ(p) depends on the variable p only. Therefore
all important information can be extracted from the Hamiltonian system on
the 3-torus with Poisson bracket defined by the magnetic field. The electron
trajectories for the low temperature can be described as a curves in this torus
such that

ǫ(p) = ǫF , CB (p) = const

However, the levels of the Casimir on Fermi surface are in fact leaves of
foliation given by the closed 1-form restricted on the Fermi surface
X
ωB | MF = B i dpi |MF = 0
i

Some people in ergodic theory studied in fact the most generic ergodic
properties of ”foliations with transversal measure” on the Riemann surfaces.
In a sense, our situation is a partial case of that. However, our picture in
3-torus is nongeneric in that sense. We cannot apply any results of that
theory. We have to work with foliations obtained in the 3-torus by this
special procedure only. Our use of word ”generic” here is resticted by that
reqiurement. As we shall see, ergodicity is a nongeneric property within this
physically realizable subclass of foliations 2-surfaces given by the closed 1-
form. What is interesting is that ergodic examples exist in our picture but
they occupy a measure zero subset on the sphere of directions of the magnetic
fields (if generic Fermi surface is fixed).
As I realized in 1982 (see [5]), this picture leads to nontrivial 3-dimensional
topology, and I posed it as a purely topological problem to my students.
The first beautiful topological observation was made by A.Zorich [8] for the
magnetic fields closed to the rational one. After new discussion and recon-
sidering all conjectures (see [9]), I.Dynnikov made a decisive breakthrough
in the topological understanding of this problem for the generic directions of
magnetic fields (see [10]). S.Tsarev constructed in 1992 the first nontrivial
ergodic examples, later improved by Dynnikov see [14]).
However, several years passed before some physical results were obtained
(see the first remark about the possibility of that in my article [11]). We made
a series of joint works with A.Maltsev (see [12, 13]) dedicated to physical
applications. Essentially, we borrowed topological results from the works
of Zorich and Dynnikov. However, the needs of applications required that
we not apply their theorems directly, but extract the key points from the

12
proofs and reformulate them. So the modern topological formulations of
these results are by-products of these works with applications (see the most
modern survey in [14]). Let me formulate here our main physical results and
after that explain the topological background and generalizations.
This picture has been extensively used in solid state physics since the
late 50s. The leading theoretical school in that area has been the Kharkov-
Moscow school of I.Lifshitz and his pupils, like M.Azbel, M.Kaganov, V.Peschanski,
A.Sludskin and others. You may find all proper quotations to physics lit-
erature in the survey article [13]. The following fundamental Geometric
Strong Magnetic Field Limit was formulated by that school (and fully
accepted later by the physics community):
All essential phenomena in the conductivity of normal metals in strong
magnetic field should follow from the geometry of the dynamical system
described above.
How to understand this principle? You have to take into account that
this picture certainly will be destroyed by the ”very strong” magnetic field
where quantum phenomena (of the magnetic origin) are important. It should
happen for such magnetic fields that magnetic flux through the elementary
lattice cell is comparable with quantum unit. However, the lattice cell in
solid state physics is so small that you need for that magnetic field the order
of magnitude B ∼ 108 Gauss or B ∼ 104 t where 1t is equal to 104 Gauss.
Therefore we are coming to the conclusion that even for the ”real strong”
magnetic fields like 102 t this picture still works well.
For our goal we need to consider such metals that Fermi surface is topo-
logically nontrivial. It means precisely that the imbedding homomorphism
of fundamental groups

π1 (MF ) → π1 (T 3 ) = Z 3
is onto. As people have known already for many years, the noble metals like
copper, gold, platinum and others satisfy to this requirement. Probably the
very first time this property was found was by Pippard in 1956 for copper.
Many other materials with really complicated Fermi surfaces are known now.
By definition, the electron orbit is compact if it is periodic and homo-
topic to zero in T 3 . Therefore it remains compact on the covering surface in
R3 , where R3 is a universal covering space over the torus T 3 . All other types
of trajectories will be called noncompact.
Normally all pictures in physics literature are drawn in R3 , but everybody

13
knows that quasimomentum vectors p1 , p2 ∈ R3 , such that p1 − p2 belongs to
the reciprocal lattice Γ∗ , are physically identical.
The Lifshitz group started to study this dynamical system about 1960
and made the first important progress. For example, Lifshitz and Peschanski
found some nontrivial examples of noncompact orbits stable under the varia-
tion of the direction of magnetic field. It looks like nobody could understand
them properly in the physics community at that time. It was several decades
before this community started to understand the geometry of dynamical sys-
tems. The Lifshitz group was ahead of its time. They made some mistakes
leading to wrong conclusions and investigations were stopped. You may find
the detailed discussion in our survey article [13]. Their mistakes have been
found only now because they contradicted our final results describing the
conductivity tensor.
Our main results:
Consider projection of the conductivity tensor on the direction orthogonal
to magnetic field. This is a 2 × 2 tensor σB . Applying any weak electric field
E orthogonal to B, we get current j. Its projection σB (E) orthogonal to B
is only what is interesting for us now. We claim that for the strong magnetic
field |B| → ∞ of the generic direction in S 2 only two types of asymptotics
are possible:
Topologically Trivial Type:

σB → 0, |B| → ∞

More exactly, we have σB = O(|B|−1) for the topologically trivial type.

All directions with trivial type occupy a set U0 of measure equal to µ0 on the
two-sphere U0 ⊂ S 2 .
Topologically Nontrivial Type:

σB → σB0 + O(|B|−1

Here 2×2 tensor σB0 is a nontrivial limit for the conductivity tensor. We claim
that it has only one nonzero eigenvalue on the plane orthogonal to B. Let us
describe the topological properties of this limiting conductivity tensor. It has
exactly one eigen-direction η = ηB with eigenvalue equal to zero. Consider
any small variation B ′ of the magnetic field B. For the new field B ′ we have
an analogous picture if perturbation is small enough. We have a new 2 × 2
tensor σB′ with one zero eigen-direction ηB′ = η ′ . Our statement is that the
plane aη + bη ′ , a, b ∈ R, generated by this pair of directions, is locally stable

14
under the variations of magnetic field. This plane is integral (i.e. generated
by two reciprocal lattice vectors). It contains zero eigen-directions ηB′′ for
all small variations of the magnetic field B. It can be characterized by 3
relatively prime integer numbers m = (m1 , m2 m3 ). This triple of integer
numbers is a measurable topological invariant of the conductivity tensor. An
open set of directions Um ⊂ S 2 with measure µm corresponds to this type.
The total measure of all these types is full:
X
µ0 + µm = 4π
m∈Z 3

We started to look in the old experimental data obtained in the Kapitza

Institute in the 60-s by Gaidukov and others (see references in [13]). They
measured resistance for the single crystal gold samples in the magnetic field
about 2t-4t following the suggestion of Lifshitz. Confirming the ideas of the
Lifshitz group, several domains with nonisotropic behavior of conductivity
were found and many suspicious ”black” dots (maybe domains of small size)
on the sphere S 2 . It is not hard to see even now that several larger domains in
these data with nonisotropic conductivity should correspond to the simplest
stable topological types like (±1, 0, 0), (±1, ±1, 0), (±1, ±1, ±1) up to permu-
tation in the natural basis of this cubic lattice. However, for good checking
it would be nice to increase magnetic field to 20t-40t for a more decisive
conclusion. The black dots either correspond to the smaller domains with
larger values of the topological integers or to some ergodic regimes occupy-
ing measure zero set on the sphere. For the final decision these experiments
should be repeated and increased about 10 times magnetic field and smaller
temperature like 10−2 K.
Let me explain now the topological background of these results. Consider
the generic Morse function ǫ : T 3 → R and its generic nonsingular level
MF ⊂ T 3 , ǫ = ǫF in the torus and in the covering space M ′ ⊂ R3 . We call
the surface M ′ ⊂ R3 a periodic surface. Apply now generic magnetic field
B and make the following construction:
Remove all nonsingular compact trajectories (NCT) from the periodic
surface M ′ and its image MF in the torus. The remaining part is exactly
some surface with boundary if it is nonempty:
[
MF \(NCT ) = Mi
(i.e. Fermi surface minus all NCT is equal to the union of surfaces with

15
boundary). We call these surfaces Mi and their closure below the Carriers
of Open Trajectories. All boundary curves are the separatrix type tra-
jectories homotopic to zero in T 3 . They bound 2-discs in the corresponding
planes orthogonal to magnetic field B. Let us fill them by these discs in
the planes. We get closed piecewise-smooth surfaces M̄i . We denote their
homological classes by zi ∈ H2 (T 3 , Z).
We use the following extract from the proofs of the main theorems of
Zorich and Dynnikov (see [8, 10]; their theorems have not been formulated
in that way, but you may extract these key points from the proofs):
In the generic case all these homology classes are nontrivial and
equal to each other up to sign 0 6= zi = ±z ∈ H2 (T 3 , Z) where z is
some indivisible class in this group. All these closed surfaces have
a genus equal to 1.
As you may see, this statement means in fact some kind of the ”Topo-
logical Complete Integrability” of our systems on the Fermi surfaces for the
generic magnetic field.
For obtaining our final result on the conductivity tensor, we need to use
the Kinetic Equation for the quasiparticles based on Bloch waves nearby the
Fermi level. This equation has been used a lot by solid state physicists for
the past 30 years. For the small (but nonzero) temperature, strong magnetic
field and apropriate general assumptions on the impurities, the motion of
quasiparticles concentrates along the electron trajectories above. This fact
leads to our conclusions. Despite the fact that this theory is considered a
well established one already for many years in the physics community, any
attempt to prove such things as the rigorous mathematical theorems would
be a huge mess. As we see, our final conclusion is separated from all theorems
by some gap which cannot be eliminated. Let me point out that it is always
so. ”Rigorous proofs” in mathematical physics never prove anything in real
world physics.
What about nongeneric trajectories? Tsarev and Dynnikov constructed
very interesting examples where genus of carriers of the open trajectories is
larger than 1 (see[14]). We call such cases stochastic. Sometimes we call
them ergodic. There were some attempts to extract from their properties
highly nontrivial asymptotics of the conductivity tensor in the strong mag-
netic field [15]. However, these attempts need a better understanding of the
properties of such trajectories. We have to answer the following questions:
1.How many directions of the magnetic field on the sphere S 2
admit ergodic trajectories?

16
 According to my conjecture, for the generic Fermi surface, this set of di-
rections has a Hausdorf dimension not greater than some number a < 1 on
the sphere S 2 . For the special Fermi surfaces ǫ = 0 of the even functions
like cosp1 + cosp2 + cosp3 = 0, we expect to have ergodic trajectories for
the set of directions with Hausdorf dimension like 1 < a < 2. Dynnikov
started to investigate this example in his Thesis and proved several general
properties. Recently R.Deleo investigated such kinds of examples more care-
fully and performed more detailed calculations ([16]). His results confirm our
conjectures. However, the Hausdorf dimension of this set has been unknown
in this example until now.
2.Which geometric properties does ”typical” ergodic trajectory
have?
According to the conjecture of Maltsev, these trajectories are typically
the ”asymptotically self-similar” plane curves in the natural sense. His idea
(if it is true) leads to the interesting unusual properties of the asymptotic
conductivity tensor. Anyway, this problem is very interesting.
Dynnikov investigated also the dependence of these invariants on the level
ǫF of the dispersion relation (see [14]). These results are useful for the right
understanding of our conjectures.
Multidimensional Generalizations.
Consider the following problem: What can be said about topology of
the levels f (x, y) = const of the quasiperiodic functions with m periods on
the plane x, y?
For the case m = 3 this problem exactly coincides with our subject above:
By definition, quasiperiodic function on the plane is a restriction on the
plane R2 ⊂ Rm of the m-periodic function. Our space R3 was a space
of quasimomenta (more precisely, its universal covering). Our plane was
orthogonal to the magnetic field. Can this theory be generalized to the case
m > 3? According to my conjecture, it can be generalized to the case m = 4.
I think that for small perturbations of the rational directions this theory can
be generalized to any value of m. We consider now any 4-periodic function
f : R4 → T 4 → R and pair of the rational directions l10 , l20 corresponding to
some lattice Z 4 in R4 .
Let me formulate the following theorem.
Theorem. There exist two nonempty open sets U1 , U2 on the sphere S3
containing the rational directions l10 , l20 correspondingly such that:
For every plane Rl2 ⊂ R4 from the family given by 2 equations l1 =
const, l2 = const, the quasiperiodic functions fl have only the following two

17
types of connectivity components of the levels fl = const on the plane Rl2 .
1.The connectivity component of the level is a compact closed curve on the
plane. 2. The connectivity component of the level is an open curve lying in the
strip of finite width between 2 parallel straight lines with the common direction
η. This situation is stable in the following sense. After any small variations
of the directions l1 ∈ U1 , l2 ∈ U2 , of function f on T 4 or the level we still
have such open component with direction η ′ . For all possible perturbations
this set of directions η, η ′ , ... belong to some integral 3-hyperplane in R4 .
This property can be formulated in terms of the integral homology class
in the group H3 (T 4 , Z) and of the torical topology of the carriers of the open
trajectories. The idea of the proof was recently published by the author in
[17].
We may reformulate this problem in terms of Hamiltonian systems. Let
the constant Poisson Bracket Bij be given on the torus T m whose rank is equal
to 2. Any Hamiltonian f generates such systems whose trajectories are equal
to the levels of f on the planes. Our theorem means that in these cases this
Hamiltinian system is Completely Integrable in the specific topological sense
described above.

References
[1] S.Novikov.Topology-I. Encyclopedia of Mathematical Sciences, vol 12,
Springer-Verlag, Berlin-Heidelberg-New-York.

[2] S.Novikov. Bloch functions in a magnetic field and vector bundles. Typy-
cal dispersion relations and their quantum numbers. Doklady AN SSSR,
1981, v 27 n 3 pp 538-543

[3] S.Novikov, I.Schmelzer. Periodic solution of the Kirchhoff equations.

Functional Analysia Appl., 1981, v 15 n 3 pp 54-66

[4] S.Novikov.Multivalued functions and functionals, Doklady AN SSSR,

1981, v v 260 n 1 pp 31-35

[5] S.Novikov. The Hamiltonian formalism and multivalued analog of the

Morse theory, Russian Math Surveys, 1982, v 37 n 5 pp 3-49

18
 [6] J.Duistermaat, G.Heckmann.On the variation in the cohomology of the
symplectic form of the reduced phase space, Inventiones Math, 1982, v
69 pp 259-269

[7] P.Grinevich, S.Novikov.Nonselfintersecting magnetic orbits on the plane.

Proof of the Overthrowing of the cycles principle, Americam Math So-
ciety Translations, 1995, ser 2 v 170 (Topics in Topology and Math
Physics, edited by S.Novikov), pp 59-82

[8] A.Zorich.Novikov’s problem on the semiclassical electron in the homoge-

neous magnetic field. Russian Math Surveys, 1984, v 39 n 5 pp 235-236

[9] S.Novikov.Quasiperiodic structures in topology. in the proceedings

of the Conference ”Topological methods in modern mathematics-
Stonybrook,June 1991 ”, Stonybrook University, 1993

[10] I.Dynnikov.Proof of Novikov Conjecture on the semiclassical electron.

math Zametki, 1993, v 53 n 5 pp 57-68

[11] S.Novikov.The semiclassical electron in a magnetic field and lattice.

Some problems of low-dimensional ”periodic topology”. Geometric and
Functional Analysis, 1995, v 5 n 2 pp 433-444

[12] S.Novikov, A.Maltsev.Topological quantum characteristics observed of

the conductivity in normal metals, Letters of JETP, American Institute
of Physics

[13] S.Novikov, A.Maltsev.Topological Phenomena in the Normal Metals. Us-

pekhi Phys Nauk, 1998, v 168 n 3, pp 249-258, translated in English by
the American Institute of Physics

[14] I.Dynnikov.Geometry of stability zones in the Novikov’s problem on the

semiclassical motion of an electron, Russian Math Surveys, 1998 v 54 n
1 pp 21-60

[15] I.Dynnikov, A.Maltsev. JETP-Journ. Experimental and Theoretical

Physics, 1997, v 112 pp 371-376

[16] R.Deleo.Existence and measure of ergodic leaves in Novikov’s problem

on the semiclassical motion of an electron, Russian Math Surveys, 1999
v 54 n 6

19
[17] S.Novikov.The levels of Quasiperiodic functions on the plane and topol-
ogy. Russian Math Surveys, 1999 v 54 n 5

20