Innovations in Statistical Physics V1.

2
arXiv:1403.6464v2 [[Link]-ph] 20 Apr 2014

Leo P. Kadanoff∗

The James Franck Institute

The University of Chicago
Chicago, IL USA 60637

April 22, 2014

Abstract
In 1963-71, a group of people, myself included, formulated and per-
fected a new approach to physics problems, which eventually came to
be known under the names of scaling, universality, and renormaliza-
tion. This work formed the basis of a wide variety of theories ranging
from its starting point in critical phenomena, moving out to particle
physics and relativity and then into economics and biology. This work
was of transcendental beauty and of considerable intellectual impor-
tance.
This left me with a personal problem. What next? Constructing
the answer to that question would dominate the next 45 years of my
professional life. I would
• Try to help in finding and constructing new fields of science.
• Do research and give talks on science/society borderline.
• Provide constructive criticism of scientific and technical work.
• Help students and younger scientists.
• Demonstrate scientific leadership.

∗
e-mail: lkadanoff@[Link]

1
Contents
1 Introduction 3

2 My accomplishment (1966-68) 4
2.1 My insight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Next steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3 New ideas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.3.1 Example: . . . . . . . . . . . . . . . . . . . . . . . . . 11

3 Afterwards. . . 12
3.1 But, what next? . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.3 New scientific directions: toward macroscopic physics . . . . . 13
3.3.1 Granular materials, sand . . . . . . . . . . . . . . . . . 14
3.3.2 Convective turbulence . . . . . . . . . . . . . . . . . . 14
3.4 The “Intelligent Design” discussion . . . . . . . . . . . . . . . 16
3.4.1 Knowledge and ignorance . . . . . . . . . . . . . . . . 17
3.4.2 About school curricular . . . . . . . . . . . . . . . . . . 18
3.5 The American Physical Society . . . . . . . . . . . . . . . . . 18
3.6 Different kinds of service . . . . . . . . . . . . . . . . . . . . . 19
3.7 Scientific criticism . . . . . . . . . . . . . . . . . . . . . . . . . 20

2
Figure 1: Stellar explosion. Computer simulation of a sequence of events
that will bring about a huge explosion of a star. This kind of event is called
a supernova. I shall return to this simulation toward the end of the talk.

1 Introduction
This is a reconstruction of a talk that I gave at the Institute of Physics,
London, on January 13, 2011. It is reconstructed to make the original power-
point style presentation suitable for book-style publication. This talk marked
a particularly important occasion for me. The Institute was honoring me
with its Newton medal for my accomplishments in physics. The actual talk
was given on the day before my 74th birthday. Not only were there many
distinguished scientists in the audience, in addition my wife and I had invited
friends and relatives including my children and grandchildren (ages 7 to 20).
I had the difficult job of trying to say things that might catch the attention
of all.
I cheated a bit by starting off with a video of a simulation of an star
exploding and becoming a supernova. (See Figure(1).) Even the 7-year-old
looked up. In this context, the video was not entirely a cheat since I had
been a (minor) part of the team that had constructed it and had expended
considerable effort in explaining the scientific significance of the work to
audiences around the world.
I had some difficulty both in constructing the London talk and putting
together this piece. In London, my quite-varied audience wanted to know
what I had really accomplished and whether I was worthy of this high honor.
For this note, I was asked to talk about myself. So I have here stressed my
own work. However, it is clear to me, as a student of the history of science,
that every advance is the product of many minds some foreshadowing, some
constructing, some extending, some explaining. In this note, I have played
down the richness of the scientific construction process and focused on just

3
one actor, myself. A more extended view of the historical process is available
in Cyril Domb’s book on the critical point[1] and in my articles on the role
of renormalization in the history of science[2, 3, 4].

2 My accomplishment (1966-68)
I started out in science in times much easier than our own. Sputnik had pro-
duced a flurry of support for science so that there was a temporary shortage
of scientists. I went through Harvard very well supported by the NSF and
jobs in the missile industry, and moved to an NSF postdoc in Copenhagen
in which I was better paid than the local professors. I never had to look for
a job. Through the kind offices of David Pines and John Bardeen a good
faculty job in Urbana came to me. Despite, or perhaps because, of all this
I had by the time my story starts in 1966, a set of quite respectable (but
conventional) physics accomplishments.
Like every young scientist, I dreamt of doing something really new, ideally
in a short paper, something that would have some impact upon science. A
sabbatical in England had fallen into my lap and I used that time to get into
a new (for me) scientific area: phase transitions.
In the early 1960s, the behavior of materials near the critical points of
phase transitions was a quite fashionable subject. Science pretty well under-
stood so-called first order phase transitions, as for example the boiling of wa-
ter. This kind of transition involves a sudden change in material properties,
like the jump in density upon boiling water. But a temperature adjustment
may reduce the size of the density-jump. When the jump becomes infinites-
imally small, and the phases become almost indistinguishable the system is
said to be approaching its critical point. In this region of the phase diagram,
the material’s behavior becomes harder to understand. While approaching
criticality, thermodynamic derivatives tend toward infinity and one can see
large patches of the material deviating from average behavior. These novel
effects suggest questions about the fundamentals of statistical mechanics and
the dynamics of materials. These effects called for a theory.
Consequently, along with others,1 I was trying to understand the behavior
of condensed matter systems near critical points. In England, I had done a
1
The most important work came from Lars Onsager[5], Cyril Domb[1], Michael
Fisher[8], Benjamin Widom[7], A. Patashinski, V. Pokrovski[6], and (most crucially) Ken-
neth Wilson[9, 10, 11].

4
Figure 2: The phases of water exhibited in an iceberg (solid), floating in the
ocean (liquid), surrounded by air containing water vapor (gas). The familiar
world is filled with materials and their different phases.

lengthy and detailed calculation of fluctuations near the transition of the

two-dimensional Ising model. This model provided a simplified description
of a phase transition and a critical point. I was trying to understand the
meaning of my work. It would turn out that my calculation would provide
an important clue to the behavior of all kinds of materials near their critical
points. But, since the Cavendish Lab could not afford to send preprints from
England, and because the refereeing process involved many iterations of slow
mail over oceans, I was almost the only one who knew the contents of my
lengthy calculation.
The Ising model2 is a simplified view of a fluid[12], and provides a rough,
but qualitatively accurate, description of an actual fluid. The Ising fluid is
divided into boxes of two kinds, high density and low density. The forces
within the fluid tend to pull high density regions toward other high density
regions and thereby leave islands of low density regions. (See Figure(3).)
At lower temperatures these forces can hold the different regions together,
forming huge islands composed of a single phase, so at boiling the fluid looks
2
It is named after Ernst Ising, who first studied it.

5
 (a) Two phases touching (b) High Temperature

(c) Critical point

Figure 3: Phases of water as exhibited in the Ising model. At low temper-

atures, in plate (a), we can see large high density and low density regions
in contact with one another. At high temperatures, as in the next plate, all
densities are intermixed in a fine-grained way. At the critical point, plate
(c), unmixed regions of all sizes occur.

like what you see in Figure(3a). At high temperatures, the fluid is all shaken
up and the different regions do not hold together so that red and green
intermix as in Figure(3b). My own interest turns mostly to the situation in
between high temperatures and low. Here, islands form of red regions and
green regions. At just the right temperature, called the critical temperature,
red and green islands of all sizes form and intermix as in Figure(3c).
In what follows, it will be important to recognize the two parameters that
control this system’s behavior near criticality. I just mentioned the first, the
temperature. The second parameter is the pressure, which is the primary
determinant of the proportion of high density regions relative to the low
density ones. This is included in the analysis as p, the pressure minus the
critical pressure. Both parameters are zero at the critical point.
An important empirical fact was supported by my long calculation: At
each temperature, there is a characteristic size of the fluctuating regions, a

6
size which grows as criticality is approached. The characteristic size, ξ, grows
as a power of t, a dimensionless measure of the temperature deviation from
criticality. The equation is
ξ = at−ν (1)
where a is a length which roughly gives the range of the microscopic forces
that drive the transition. The number, ν, is an example of an important
quantity, called a critical index, that we would like to evaluate. The transition
is characterized by a large number of diverging statistical or thermodynamic
quantities, each with its own characteristic critical index.
Back to my story: Because of my calculation, and my knowledge of what
others have done, I had a reasonably extensive acquaintance with the be-
havior of the two dimensional Ising model. But, all this knowledge is in
disconnected pieces and I could not make it into a meaningful picture.
Then, in one week at Christmas time in 1965, I see how to bring all these
pieces together. . .

2.1 My insight
As I said, at the critical point the Ising system contains regions of all sizes.
How can we analyze this situation? Maybe we should use different descrip-
tions based upon viewing the Ising system at different magnifications. We
know that the basic atomic constituents of water are very small, so there is
plenty of room for magnification between their size and our own. But how
to change the magnification of our view. Well that is the new insight which,
forty-five years later, brought me to London.
Look at Figure(4), please. The first panel shows a small region of the
Ising model at its critical point with its mix of high density and low density
squares. In the second panel, Figure(4b), we have divided this fragment of
fluid into three by three boxes. Nine boxes cover the whole panel. These
bigger boxes each have a color. If the majority of the small covered squares
are red we give the covering box a red color. Otherwise, it gets a green color.
The right hand panel shows just the covering boxes. That panel looks like a
more coarse version of the critical Ising model.
I then made the hypothesis, now called a universality hypothesis, that
at criticality very much the same description will tell you about both the
new box problem and the original Ising model. This hypothesis is based
in the experimental observation that different critical phenomena problems

7
 (a) Ising criticality (b) Box construction (c) A new magnification

Figure 4: A renormalization calculation in the Ising model. The first panel

shows a configuration of the model with high density boxes in red. In the
second panel, we have covered the system with bigger boxes, each with a
side three times larger than in the first panel. These bigger boxes are colored
according to the “majority rule” in which a bigger box is colored to agree with
the majority of boxes beneath it. The final panel shows the renormalized
model, with bigger colored boxes.

have similar observable behavior. I went further and assumed that they were
really the same so that exactly the same description could work for problems
on different length scales. In this form, my assumption was one of scaling,
i.e. that all feature of criticality would remain unchanged as one changed the
length scale from that of the Ising model to that of the box.
Instead of solving either the original Ising model or the box problem, I
talked about how to relate the two problems. The usual way of describing
liquid-gas criticality is in terms of the two already-mentioned measures of how
far the actual system is from the critical point, t, a dimensionless measure of
the temperature deviation from criticality, and p, a similarly dimensionless
measure of the pressure deviation3 .
In this way, I could describe the same physical system in two different
ways. The old description was of N squares, with size a, and thermodynamic
variables t and p. The new, box-based, description is just the same except
that the parameters have different values. These might be written as tnew
and pnew . These values have changed because we have increased the size, a,
3
For example we could define t = 1 − Tc /T , where T is the absolute temperature and
Tc is the critical temperature.

8
of the basic box by a factor of l. (In the figure, l = 3.) Thus, the box size is

anew = la

while the number of boxes is

Nnew = N/l2

since the system is in two dimensions. The important new hypothesis is

that in this situation the two measures of the deviation from criticality vary
according to
tnew = q t
pnew = r p
where q and r depend upon the length change, l. Because physical quantities,
like ξ, tend to vary as powers of t and p, I could guess that I should express
q and r as powers of l. As we shall see in a moment, this choice enabled the
elimination of l from physical results. Thus, I defined

q = ly

r = lz
In 1965, y and z function were unknown numbers.
This transformation provided a description of the effect of scale changes
upon the Ising model in terms of the numbers y and z. All the qualitative
properties of the phase transition can be expressed in terms of these two
numbers. For example, from what we have so far, the typical fluctuation
size, called the coherence length has the value in the “new” picture.

ξnew = anew (tnew )−ν = a l (tly )−ν = at−ν l1−νy . (2)

But ξ is a physical quantity which does not change with our representation
of the system. Therefore, all the factors of l must cancel out of the right
hand side of Eq.(2) so that
ν = 1/y.
The physically significant index, ν is thus expressible in terms of our model
quantity, y. In fact all the qualitative properties of the transition can thus
be expressed in terms of y and z.
If we could but find y and z. . .

9
2.2 Next steps
I gave seminars about this work at Urbana, Cambridge, Paris, Moscow, and
many other places. Almost everyone loved the work, and encouraged me.
However, somewhat incongruously, for six years nobody could put the finish-
ing touches upon it. So what could I do? Like everyone else I was stymied.
However the Ising model was the site of my big achievement. Again and
again I tried to find y and z, but instead went in circles, making no progress
at all.
Then, in 1971, Kenneth G. Wilson showed how to marry this box analysis
with the earlier work of Murray Gell-Mann and Francis Low[18] and thereby
produce a complete theory. Wilson called the outcome, using the name from
the earlier work, the renormalization group theory.
Given the fullness of time, and 1971 was forty-three years ago, we can
appreciate the power and impact of this work and of the ideas that came out
of it.

2.3 New ideas

• Scaling. In physics we are often concerned with connecting prob-
lems at different length scales. For example, we start from atomic
forces, based upon phenomena at 10−10 meters and use these to pre-
dict the properties of solids, including scales up to many meters. The
renormalization method enables one to connect the physics at different
scales, thereby understanding the relationship among different domains
of physics.

• Universality. As one traverses the different length scales, one only

retains a few characteristics of the original problem. Thus many dif-
ferent microscopic problems have the same macroscopic manifestation,
depending mostly upon the symmetry of the original problem. For ex-
ample, the Ising model, a magnetic solid, and a liquid-gas phase tran-
sition all have the same behavior near their respective critical poinst.

• Universality Classes. Problems then fall into a few different cate-

gories depending upon the nature of their solutions. These categories
are called universality classes. It is now commonplace to say that criti-
cal behaviors fall in a few universality classes. Moreover, scientists now
use this phrase to describe many other kinds of things beyond phase

10
 transitions. One of these days I expect to hear about a description of
pop music in terms of universality classes.

• Renormalization. Wilson described the effect of many successive

renormalizations in which the length scale was changed again and again,
as the resulting motion of the interactions toward a fixed point, a state
of constant and unchanging interaction. In doing this, he provided a
calculational framework from which one can derive scaling, universality,
and the values of the critical indices. In this way, he tied all aspects of
the theory into a coherent whole.

• New Calculational Paradigm. Previously one started with a “prob-

lem,” i.e. a description of the interactions among the particles in the
system. The job of the theorist was mostly to calculate an “answer”,
that is a detailed description of the the behavior of the particles.
In the new era, one does renormalization calculations in which one
starts with one problem (i.e. a set of interactions) and constructs an-
other, equivalent, problem by changing the length scale. A partial, but
often sufficient, description of the behavior is encoded in the relation
between the different scales.

2.3.1 Example:
Old era: One said that quantum electrodynamics was characterized by an
interaction strength called α, having the value 1/137.036.
New era: One says that at smaller distance scales, α gets larger and
eventually this electromagnetic interaction becomes very strong. In informal
language, one says that the interaction strength runs.
The practice of physics has changed and moved away from the mode of
calculation of Newton, Boltzmann, Einstein, and Dirac, going from solving
problems to discussing the relationship among problems. Of course, this
change is not all my doing. In addition to the aforementioned people (See
footnote number 1.) who pushed in this direction, all the workers in the field
determine the subject’s contents.

11
3 Afterwards. . .
So here I am in 1971. Wilson had just put the beautiful finishing touches
upon the theory that I partially developed. My initial reaction:
I am disappointed and angry. How could someone else finish up the
description of my beautiful world!
But that view, if held too hard and strong, would have paralyzed me and
prevented any future work. I concentrated on the fact that Wilson had quite
impressive skills and knowledge.
He could apply his thought processes in ways inspired by computer pro-
gramming.
He knew about previous work in particle physics and dynamics of which
I was ignorant.
He was, I could acknowledge, a very great physicist.
I was gradually able to put aside my childish chagrin, and take pleasure in
the fact that, with Wilson’s additions, an edifice of transcendentally beauty
and importance had been constructed.

3.1 But, what next?

It was evident to me that I had been wonderfully creative, but also wonder-
fully lucky. It should have been, and perhaps was, evident that I would never
do anything that approached my previous level of creation.
History suggests that theoretical physicists hit their top level of creation
early in their professional lives and then never again approach the same
pinnacle. (See Table(1).)

3.2 Perspective
The historian of science, Thomas Kuhn[13] says that most scientific work is
“normal science,” that is constructing a small examples or extensions of a
subject in which the basic approach is already accepted.
Occasionally the basic approach comes under attack and something really
new is done. This is “revolutionary science.”

12
 person age accomplishment year
Isaac Newton 25 mechanics 1667
Albert Einstein 26 best year 1905
P.A.M. Dirac 26 Dirac equation 1928
Hans Bethe 30 energy from sun 1936
Brian Josephson 25 Josephson effect 1965
Leo Kadanoff 29 scaling, universality 1966
K.G. Wilson 35 renormalization 1971
Alan Guth 32 inflationary universe 1979
exceptions
Galileo Galilei 58 relativity 1632
Neville Mott 44 Mott transition 1949
Edward Witten passim many

Table 1: Table showing the ages at which theoretical physicists constructed

their most creative accomplishment, along with the accomplishment and the
year in which it was achieved. Aside from myself, the people in this table
were great physicists who far outdistanced me in the quality of their work.
But their experience was nonetheless probably relevant to my work life.

In the 1970s, I had read Kuhn and could recognize the revolutionary
aspect of my past work, but also that I was likely to have only normal science
in my future. So be it. But within this limitation I could do worthwhile
things, including
• Try to exemplify a standard of good work..

• Lead my colleagues into new areas of research.

• Offer worthwhile commentary about the relation between science and

society.

• Offer careful criticism of scientific work.

3.3 New scientific directions: toward macroscopic physics

For decades, physics and chemistry departments had been studying atomic
and subatomic phenomena, and mostly ignoring the observable world of wind,
water, waves, and sand. Partially because the existence of computers made

13
possible new modes of investigation, in the 1970s major opportunities arose
in the scientific study of study of things that happen in the observable world
around us. Commonplace things like the flow of water or the piling up of
sand.
Albert Libchaber, Sidney Nagel and I introduced4 the study of these
things at the University of Chicago, and through Chicago into physics de-
partments all around. For example, I did a theoretical study of the pinching
off of the neck that connected a drop to a larger body of liquid. The mathe-
matical content of this work included the study of singularities, infinities in
fluid flow. This led to experimental work, most notably by Sidney Nagel5
and additional theoretical work at Chicago on the shape of drops of fluid.
For example, the first panel in Figure(6) shows a drop of fluid, connected to
a larger mass by a thin neck, just about to separate and fall off.
At the same time, P. de Gennes was broadening perspectives of scien-
tists in France, as was Sam Edwards in England. Like these other people,
Chicago’s workers on these topics had to overcome some conservative bias,
but we persevered, and physicists at other institutions gradually began to
work on macroscopic behavior.

3.3.1 Granular materials, sand

Along the same lines, I began a series of calculations to understand what
would happen in real collisions among small pieces of material like gains
of sand or of sugar[22]. This too led to experimental work at Chicago by
Heinrich Jaeger and Nagel. I show only one example of very many. The
second panel in Figure(6) shows a combination of work on drops and granular
materials. In the this panel, a granular material is mixed into a fluid. Then
the mixture is allowed to form a drop and fall off. Some steps in the process
are shown[15].

3.3.2 Convective turbulence

Another thread of work is in “convective turbulence.”
That sounds formidable but it is exemplified by a most common situation:
heating water to boil an egg. You heat the water from below and cool it from
4
Not quite introduced. The earlier work of Chandrashekhar and Eugene Parker pre-
pared the University to accept work in classical physics.
5
For a recent example of Nagel and coworkers see[21].

14
 (a) Cartoon Cell (b) Laboratory Study (c) Computer Study

Figure 5: Convective turbulence. Three pictures of a box containing a fluid

heated from below and cooled from above showing temperature patterns.
The left-hand panel is my cartoon view, partly improved from the picture I
drew before the system was actually observed in the lab. Here, and on the
right, red indicates warm fluid, blue cold fluid. The view in the center is
from Penger Tong’s lab in Hong Kong. Here, shadows outline regions where
the temperature changes. The right hand panel is the result of a computer
simulation by Richard Stevens, R. Verzicco, and D. Lohse. In all panels,
hotter regions are buoyant and rise, cold regions sink. . As you can see from
all three views, this simple machine, driven by only a heater, has developed a
complex pattern with mushroom-like objects called “plumes” in the interior
and wall regions at top, bottom, left, and right with a very different flow
pattern from the center.

above. There is hotter water toward the bottom of the pot and cooler water
toward the top. The heated water tends to expand so the water at the bottom
is less dense. The resulting disparity in gravitational force pushes the bottom
fluid upward and the top fluid downward. A flow begins.
The resulting motion was being investigated in the laboratory of my col-
league, Albert Libchaber. Before I saw the fluid in motion, I drew a picture
of what I thought was happening. To see my picture and the real thing, look
at Figure(5). This shows three different views of the motion. A cartoon pic-
ture of the convective cell shown in the first panel of the figure. This cartoon
showed different regions of the cell: jets at the side walls, thin boundary layers
top and bottom, thicker mixing zones above the bottom boundary layer and
below the bottom one, and a large central region containing “plumes.” These
are mushroom-shaped objects formed in the mixing zones that can group

15
together in the jets and move freely through the central region. My work
then described in a qualitative fashion how all these parts worked together
to make an efficient “machine-like” fluid flow that moved heat from bottom
to top in a quite efficient fashion. I also made quantitative predictions, now
outmoded; but the qualitative picture of the flow’s structure remains valid.

3.4 The “Intelligent Design” discussion

The convective turbulence result formed a major part of a series of lectures
that I gave dozens of times to a very wide variety of audiences including
philosophy departments, theology schools, high schools and various physics
groups. This talk included some other examples of how fluids “naturally”
developed complex motions. An important part of the argument is a demon-
stration that the laws of fluid motion are not mysterious, but follow directly
from a molecular picture of a fluid.
I gave this talk as a professor of physics and also in my role as president of
the American Physical Society. The point of the talk was not only that fluid
flow is an interesting subject. As a scholar I felt I had a duty to participate
in the public discussion on the issues of the day. The basic idea is that
fluid-motion, while it can be vastly complex, can be understood by direct
application the laws of physics. But complexity is a fraught word in US
science and culture. That is because there is a large body of people who
refuse to believe that humankind is a result of an evolutionary process, but
instead argue that we are a product of “intelligent design” by a creator or more
likely “Creator.” Many of them would like this view to become a regular part
of school curricula. Among the people who have contributed to this line of
thought are Reverend William Paley (1743-1805), Michael Behe (a chemist),
Bill Dembski (in mathematics and philosophy)6 . The best argument for
intelligent design comes from the Reverend Paley who argued that if someone
found a watch on a beach that person would immediately conclude, in view
of the careful design and workmanship, that the watch had been designed for
a purpose. Thus, in this line of argument, living things that are even more
intricate and involved than the timepiece were probably also created by an
intelligent being to serve some purpose.
In our era, Professor Behe wrote that living cells are amazingly complex
machines (true).
6
My former student

16
 (a) Drip of Fluid (b) Fluid containing granular material

Figure 6: Drips. The first panel, due to Itai Cohen and Nagel, shows a thin
neck just about to break. The neck is formed by a drop of fluid coming off
of a larger mass that has moved through a pipette. The second panel due to
Heinrich Jaeger and coworkers[15] shows a drop composed of a fluid loaded
with a granular material. It too form a drop and a neck, but the detailed
shapes in the two panels are entirely different. These flows and shapes fall
into different universality classes.

One cannot imagine how they could have arisen through any natural
process. (He cannot; I cannot.)
Behe concluded that the only reasonable thing is to assume they were
put in place by an intelligent force.
My reply: Fluids are basically very simple, they can be described as
equivalent to billiard balls in motion, occasionally bump into one another.
Nonetheless, if you apply heat, they naturally and spontaneously form a
convective cell, an object resembling a simple machine. Thus we can see a
natural start toward the possible formation of a complex machine. Perhaps
a cascade of such processes might have produced a living cell. Perhaps.

3.4.1 Knowledge and ignorance

We don’t have much knowledge about how complexity arises, in the present
biological world or for the earliest biological forms. Dembski and Behe do us
some good by pointing out that ignorance, but neither they nor the Intelligent
Design people generally seem to have much positive to add to the discussion.
My own view is closer to that of Saint Augustine of Hippo (354-430)

The Universe was brought into being in a less than fully

formed state, but was gifted with the capacity to transform itself

17
 from unformed matter into a truly marvelous array of structure
and life forms.

translation by Howard J. van Till.

3.4.2 About school curricular

So, my complexity talk ends with

In the last 1600 years, we have learned some specifics to help

flesh out Augustine’s good overview. We should not let the partic-
ular view of the universe provided by Intelligent Design’s followers
replace science in the school curricula. Instead, we should notice
that evolutionary biology, paleontology, and cosmology are not
speculations. They are root parts of science and human knowl-
edge. They belong in schools as part of the basics of our cur-
riculum and of our understanding of the material world. Other
subjects, unrooted in experiment or scientific observation, should
be avoided in school science classrooms.

3.5 The American Physical Society

I was proud to serve as President of the American Physical Society (APS),
an organization composed of almost 50,000 physicists from the US and the
rest of the world. However, in the three years I was an officer of that group,
it seemed to be burdened by a huge inertia. I wanted the organization to
move in directions different from its past history in which it devoted itself to
managing its stellar journals, running huge scientific meetings, and lobbying
for more government spending on research. I argued that, in addition, we
should have a major effort aimed at the improving scientific education for
all US students. Only slow progress was made in this direction. I also felt
that the APS should be a voice for physicists worrying about tenure, job
conditions, and remuneration. Instead, the APS seemed to be rooted in its
past when it spoke for the employers of physicists.
In the years since my APS presidency, the organization has indeed deep-
ened and broadened its educational effort. It has spoken out against the
assassination of scientists in Iran. So there has been some movement in the
right direction.

18
3.6 Different kinds of service
Another one of my “service” roles was to serve two terms as director of the
University of Chicago’s materials research laboratories. These NSF funded
laboratories have as their goal the fostering of studies of condensed matter
by diverse teams of scientists. I found my role immensely satisfying. I could
help my university move in new scientific directions without getting much
involved in bureaucratic nonsense. I could foster work on education and help
my colleagues make contact with local schools and science museums, thereby
building our educational outreach.
In the same vein, for about a decade, I managed a program aimed at
bringing graduate students into contact with science museums. During five
years, we brought about a dozen students each year, half in the sciences and
half in the social sciences, to classrooms where they learned about science
museums and to the museums themselves where they helped design exhibits.
We found scientists and museum professionals to meet with the students and
spend one afternoon a week teaching them about museums. The aims of the
program included
• To teach the students about museums and prepare them for museum
outreach and museum careers in their future professional life.
• To bring additional up-to-date science to the museums.
• To teach the graduate students scientific communication skills.
• To make university people and museum people more aware of the pos-
sibilities of working together.
The program worked quite well for about five years. But, after a time, NSF
funding for the program disappeared and support for science museums across
the country weakened. At that point, my attention was turning to the APS
and with reluctance I gave up that particular museum activity.
I have been happy to see continued museum outreach within and around
our present materials lab. I still serve this lab, now as “director” of outreach
activities. I see my materials lab colleagues and their work on granular
materials projected 25 feet high on a wall in Chicago’s Museum of Science and
Industry. By the opposite wall I see exhibits for which the early design came
from my museum graduate students. A few of the grad students from the
program have begun museum careers. Others continue to work on presenting
science to the public.

19
3.7 Scientific criticism
From a very early point in my professional career, I have worked on a critical
assessment of scientific activity. Many students of the sciences have taken
the point of view that science is uniformly wonderful, that all is progress, and
that great scientists are always great. A few others have taken the opposite
tack. I have tried to have a more balanced view.
Among the most recent work is a set of papers about the history of
science. In addition to some laudatory assessments, I have noted that in
his first paper on the relevant Fourier series J. Willard Gibbs[16] failed to
notice the phenomenon now called the “Gibbs overshoot.” In a more serious
vein, an assessment of Landau’s and BCS’s work on the superfluidity[17]
indicates that these scholars’ important advances in our understanding of
microscopic behavior of superfluids and superconductors was matched by an
almost complete lack of understanding of the macroscopic wave functions
that drive superfluidity.
My most serious criticism, however, has been devoted to large-scale com-
puter models. As early as my Urbana days, I started with an assessment
of Jay Forrester’s Urban Dynamics[14] model. This was a set of difference
equations which modeled the economic and social growth of an urban area.
He applied a range of different public policy initiatives within the model, and
examined their various consequences, saying some initiatives were better and
some were worse. Finding his conclusions in disagreement with my political
prejudices, I formed a group that showed how one could reach the opposite
“conclusions” by changes in the focus of the model and in evaluation criteria.
In the same vein I studied the use of very large models to simulate the ef-
fects of rather turbulent hydrodynamic behavior, not only in convective cells,
but also in studies of sonoluminescence and supernova explosions. (See Fig.
(1).) The results were delivered in papers[20] and in a frequently delivered
seminar on “Excellence in Scientific Computing.” These models were only
partially calibrated, and only included a portion of the physics of the phe-
nomenon at hand. I concluded with the statement that partially calibrated
models were a form of argumentation, and then
How does the power of argumentation provided by exploratory
simulations compare to that of rhetorical or order-of-magnitude
discussions? Since the simulations must include everything to
make a star go boom, they provide an internal check of consis-
tency and completeness not available through words. On the

20
 other hand, some intermediate steps in the argument may have
their weaknesses hidden in unexamined computer processes. Words
may be better than computer output for showing up weak ar-
guments. Computer arguments often force us to rely upon the
integrity and care of the investigators. So computers provide a
useful but dangerous tool for the exploration of complex systems.
Many people including my students, my postdocs and myself have done
some very well controlled simulations on smaller computers and thereby ex-
posed the qualitative properties of well-defined systems. When one can guar-
antee that the results of a computer simulation accurately reflect what will
happen in a specific sharply defined situation, they can be quite meaningful.
If, in addition, they answer a well-posed scientific or technical question they
are of some value. Otherwise they provide nothing more than a scenario for
some imagined event.
As I see it, very few of the largest and most expensive simulation models
answer well-posed questions. In addition, most of these models give results
that should be taken with a grain of salt.

Acknowledements
I want to thank Betsy Kadanoff, Michal Ditzian, and Jim Langer for helpful
advice. This work was partially supported by the University of Chicago
NSF-MRSEC under grant number DMR-0820054.

References
[1] Cyril Domb, The Critical Point, Taylor and Francis, (1996).
[2] Leo P. Kadanoff, “More is the Same; Mean Field Theory and Phase Tran-
sitions,” J. Stat Phys. 137 777-797 (2009).
[3] Leo P. Kadanoff, “Theories of Matter: Infinities and Renormalization,”
The Oxford Handbook of the Philosophy of Physics ed. Robert Batterman,
Oxford University Press, 143-185 (2013).
[4] Leo P. Kadanoff, “Relating Theories via Renormalization,” Studies in His-
tory and Philosophy of Science Part B: Studies in History and Philosophy
of Modern Physics 44 (1):22-39 (2013).

21
[5] L. Onsager, “Crystal Statistics: A two-dimensional model with an order-
disorder transition,” Phys. Rev. 65, 117-149 (1944).

[6] A.Z. Patashinskii, V.L. Pokrovsky, Soviet Phys. JETP 19, 667 (1964),
A.Z. Patashinskii, V.L. Pokrovsky, Fluctuation Theory of Phase Transi-
tions, Elsevier (1979).

[7] B. Widom, J. Chem. Phys. 43, 3892 (1965) and ibid page 3896.

[8] K.G. Wilson and M.E. Fisher, “Critical Exponents in 3.99 Dimensions,”
Phys. Rev. Lett. 28 240-243 (1972).

[9] K. G. Wilson, “Renormalization Group and Critical Phenomena. I. Renor-

malization Group and the Kadanoff Scaling Picture,” Phys. Rev. B 4
3174. (1971).

[10] K. G. Wilson, "Renormalization Group and Critical Phenomena. II.

Phase-Space Cell Analysis of Critical Behavior," Phys. Rev. B 4 3184
(1971).

[11] K. G. Wilson, “The renormalization group and critical phenomena,” Re-

views of Modern Physics 55 583 (1983).

[12] C. N. Yang and T. D. Lee, Phys. Rev 87,404 ( 1952). T. D. Lee and C.
N. Yang, Phys Rev. 87, 410 (1952).

[13] Thomas S. Kuhn, The Structure of Scientific Revolutions, Univ. of

Chicago Press, Chicago, Illinois, (1962).

[14] Jay Forrester, Urban Dynamics, Pegasus Communications (1969).

[15] Marc Z. Miskin and Heinrich M. Jaeger, “Droplet Formation and Scaling
in Dense Suspensions,” Proc. Nat’l Acad. Sci. 109, 4389-4394 (2012).

[16] Leo P. Kadanoff, J. Stat. Phys. accepted, (2014).

[17] Leo P. Kadanoff, “Slippery Wave Functions,” J. Stat. Phys. 152, 805-823
(2013).

[18] Murray Gell-Mann and F.E. Low, “Quantum Electrodynamics at Small

Distances,” Phys. Rev. 95, 1300-1312 (1954).

22
[19] Leo P. Kadanoff, “Scaling Laws for Ising Models Near Tc ,” Physics, 2
263-267 (1966).

[20] Leo P. Kadanoff, “Excellence in Computer Simulation,” Computing in

Science and Engineering, 6, 57-67 ( 2004). “Computational Physics:
Pluses and Minuses,” Phys. Today, 39 (1986).

[21] J.D. Paulsen, J.C. Burton, S.R. Nagel, S. Appathurai, M.T. Harris, and
O.A. Basaran, PNAS 109, 6859 (2012)

[22] Leo P. Kadanoff “Built Upon Sand: Theoretical Ideas Inspired by the
Flow of Granular Materials,” Rev. Mod. Phys., 71 (1) 435-444 (1999).

23