Oxford university press how the br ain evolved language

Figure 2.1. 
DNA. 
tracted to it. A bilayer of such molecules forms the cell’s membrane, a molec­
ular, water-repellent balloon enclosing the complex and vulnerable self-
replicating DNA molecules and protecting them against chemicals that would 
do them harm. But just as the primordial puddle’s rocky barrier had to be 
slightly permeable to let the first intrepid self-replicating molecule out into the 
world, the growing and surviving cell design also needed a semipermeable mem­
brane, so that nutrients could enter and waste products could be expelled. And 
this posed a significant evolutionary dilemma: how was the membrane to let 
food molecules in but keep toxic molecules out? 
Nature needed three basic solutions to this problem. The first was the 
communal solution, and it is still visible to every third grader who examines a 
drop of pond water under the microscope. Like little walled cities, the eukary-
otic cell (figure 2.2) housed within its membrane mitochondria, ribosomes, and 
various organelles: cooks, builders, and housekeepers that process incoming 
food, build new cells and cell parts, and keep things clean and healthy. 
It seems that early cells simply ingested one another, sometimes being the 
feeder and sometimes being the feed. But on a few very special occasions, sym­
biotic relationships formed. Mitochondria, ribosomes, and other organelles all 
learned to live together in self-similar domestic harmony: membranes within 
membranes, and within each membrane, molecular economies importing raw 
materials and exporting products, communicating with each other in chemi­
cal codes. Primus inter pares, at the heart of every eukaryotic cell, is the cell 

26  • 
HOW  THE  BRAIN  EVOLVED  LANGUAGE 
Figure 2.2. 
The eukaryotic cell. 
nucleus. The cell nucleus is an organelle, but it is a very special organelle. It is 
the organelle that contains DNA, the plans by which the biological city rebuilds 
itself. Now, toxins and other bad guys who scaled the cell’s first, outer cell wall 
could kill it, but in order to kill the eukaryotic cell’s progeny, toxins also had to 
penetrate the second, nucleic membrane. 
The second solution, in important respects self-similar to the first but es­
pecially important to the eventual evolution of nerve cells, was the evolution 
of the bilayer membrane itself: one layer keeps the outside out while the other 
keeps the inside in. The outside membrane has “gates” and “latches” that are 
highly nutrient-specific and only open when a nutrient is identified. The in­
side membrane has gates that open when wastes are present, allowing them to 
be expelled. 
One can imagine that the first cell membrane was selective but passive, 
allowing molecules to penetrate it only by osmosis. Like the shoreline of the 
primordial pond, which admits some runoff from puddles on higher ground 
and trickles some of its contents to ponds on lower ground, such an arrange­
ment would allow select molecules at high concentration to pass to regions of 
lower concentration. This passive membrane could recognize nutrient and 
waste molecules by their shape so as to ingest only the former and excrete only 
the latter, but Precambrian life was not so civilized that a protozoan epicure 
could float around until a molecular morsel was served with perfect presenta­
tion. The price of a too-passive and too-refined selectivity could too easily be 
starvation. The surviving and thriving protozoan needed to be a very picky eater 
only until it found something it liked. Then it needed to pig out, which leads 
us to the third solution to the problem of selective permeability. 
Fortunately, molecules become ionized in solution: they assume an elec­
trical charge. In the primordial saltwater soup, sodium chloride dissolved into 
positively charged sodium ions (Na
+
) and negatively charged chloride ions (Cl
–
). 
As a result, when an early protozoan encountered a field of nutrients, it could 
do more than passively wait for them to seep through its membrane. So long 
as it kept an internal negative charge, when it opened one of its mouths (i.e., 

JONES
’
S  THEORY  OF  EVOLUTION 
•
27
 
one of the pores of its membrane) to eat a big, tasty protein molecule, a flood 
of the ubiquitous Na
+
 ions from the surrounding seawater soup would be elec­
trically attracted into the cell interior as well. The resulting change in voltage 
would literally shock the surrounding membrane, causing two more mouths to 
open, admitting more Na
+
, creating a still bigger shock, opening still more 
mouths. Like screaming “Pizza!” in a crowded college dormitory, the influx of 
Na
+
 made the membrane active, setting off a feeding frenzy. As we will see in 
chapter 3, this active membrane is a hallmark of the neuron, the cell type which 
sends electric signals to other cells. The other hallmark of the neuron is its 
axon: a kind of long transmission cable along which its electric signals are sent. 
Flagellates 
It is all well and good for a cell to sit in its little pond, swallowing anything that 
floats its way with an appealing shape and turning up its pores at the rest. But 
to find more food, grow big, and eat all the other little cells, it helps if a cell 
can move around. One of the first groups of moving cells, the flagellates sub­
phylum (Mastigophora), is particularly intriguing. To move around, these 
single-celled animals evolved a long, whiplike protoplasmic projection called 
a flagellum (figure 2.3). The flagellum would also have been an excellent pro­
totype for an axon, so the flagellates would have been an excellent prototype 
for the neuron. But a neuron with an axon to send an electric nerve signal is 
still missing one thing: a cell to receive its message. Perhaps it is not wholly 
accidental that the flagellates formed colonies, possibly making the important 
evolutionary transition to Porifera (sponges) and Coelenterata (jellyfish and coral), 
the first multicelled animals. 
At this still-early stage of evolution, multicellular organization is often self-
similar to cellular organization. In primitive colonial plants like slime mold and 
colonial protozoans like Porifera, when a cell divides, the two new cells do not 
swim apart; they stick together. Mobility is sacrificed for a different advantage. 
The cells on the outside of the colony have greater access to food but also risk 
greater exposure to toxins. Inner members have less access to food but are less 
Figure 2.3. 
A flagellate. 

28  • 
HOW  THE  BRAIN  EVOLVED  LANGUAGE 
exposed to toxins, and should outer cells die, the inner cells can still live long 
enough to reproduce. 
In the case of differentiated multicellular organisms like jellyfish, the 
organism’s outer membrane differentiates first into outside (the skin, or epi-
thelium) and inside (the gut, or enteron). Then the cells of the animal differen­
tiate further into such types as muscles for movement and nerves for control 
and communication. 
The Cambrian Explosion 
Up until about 600 million years ago, life left few fossils. Presumably, most life 
up to that time was one-celled and soft—not the stuff of which fossils are made. 
But then, quite suddenly in the geological record, at the boundary between 
the Permian and Cambrian periods, fossils of multicelled animals begin to 
appear in great profusion and variety. What could have caused this Cambrian 
explosion of life? One factor certainly must have been the invention of the neu­
ron, for, by definition, the multiple cells of a multicellular organism must com­
municate among themselves in order to function as an organism. While the 
electrical communication of the neuron might not have been strictly neces­
sary (the organs of your body also communicate by hormones sent through 
the bloodstream), the race in life is between the quick and the dead, and there 
is little doubt that the quick animals of the Cambrian were electric-quick: they 
had electric neurons. Evolution was no longer just a process, it was a race—an 
evolutionary arms race (Dawkins and Krebs 1979), and the neuron was its coni­
cal bullet. 
The Formula for Life 
The reader may not be ready to agree that Miller’s creation of amino acids was 
the same thing as creating life. Indeed, recent discoveries of thermophilic life 
forms (the Archaea, which live off sulfur, deep in the sea) have made the events 
of early evolution seem bizarre indeed, and we now know that Earth’s early 
environment couldn’t have been quite the combination Miller concocted. Still, 
even if we haven’t found the specific formula for life, we can be quite sure we 
have found the general formula. It is stated in equation 2.4: 
f(x) = Ax + t 
(2.4) 
Of course, if A, x, and t in equation 2.4 are simply real numbers, then 2.4 
is nothing but a straight-line function. But if x is a form and A is a geometrical 
transform of that form, then 2.4 is the formula for an affine transformation, more 
popularly known as a fractal. In the case of figure 2.4, x is the almost-triangular 
shape of a fern leaf; A is a not-very-complex function which rescales and tilts 
the shape of x; and t simply relocates the transformed shape to a different place 

JONES
’
S  THEORY  OF  EVOLUTION 
•
29
 
in the picture. The result is that each frond of the fern in figure 2.4 is another 
whole frond, but on a different scale in a different place. Following Mandelbrot 
1982, this property of the “fractal geometry of nature” is often called self-
similarity, a term we have seen again and again, the principle that patterns in 
nature repeat themselves on different scales, albeit with slight mutations. 
I don’t believe, as some mystics might, that equation 2.4 implies that there 
is either a mathematical design or a mathematical precision behind Creation. 
Equation 2.4 only simply and succinctly captures a basic pattern of nature, x. 
And because it only produces the fern of figure 2.4 when parameters of A and 
t are procedurally varied, it also focuses our attention on essential processes of 
nature. Life is one such process, which faces similar problems over and over 
again, albeit in different places (t) and on different scales (A). Life first pros­
pered in a puddle, within a barrier. Similarly, on another scale, the life of the 
cell occurs within a barrier, the cell membrane; and on yet another scale, the 
life of the cell nucleus occurs within another membrane barrier. On yet other 
scales, each organ of your body is surrounded by a membrane barrier; your 
body itself is surrounded by a membrane barrier called skin. Cities are sur­
rounded by walls, nations by borders, and Earth by an atmosphere. You may 
find mystery in this if you like, but this is also just how things are. In all the 
preceding examples, despite variations in detail, there is a common, elemen­
tal force which imposes this design on all scales of life. We could say that this 
force is the second law of thermodynamics, that membranes are barriers against 
entropy. We could say that epithelia are barriers against predators. We could 
say that barriers establish the identity of Self versus Other. All of the above could 
be true, and more, but what is essential to our present story is that the barrier 
design works to enable self-replication. What works may not be True, but what 
works Survives. 
Figure 2.4. 
A fractal fern. (Barnsley 1988. Reprinted by permission of Academic Press.) 

30  • 
HOW  THE  BRAIN  EVOLVED  LANGUAGE 
Sex and Self-Similarity 
Amid all this self-similar evolution there is also differentiation. In colonial Pro­
tozoa and Porifera, any member of the colony can, in theory, split off and 
begin a new colony, self-similar to its previous colony. That means its DNA must 
contain the plans to the whole city. But by the time we move up the evolution­
ary ladder to Coelenterata, things get more complex. 
In jellyfish we begin to see a very clear differentiation of cells: there are 
muscle cells and stinger cells and brain cells. Given such a differentiated jelly­
fish, how is a jellyfish to make another jellyfish? In a protozoan, say a shapeless 
amoeba, it is simple in principle. Let it simply divide in half, and voila! one has 
two amoebae. But how do you divide a salamander in half? A salamander can 
grow a new tail, but a tail can’t grow a new salamander. How is even a jellyfish 
to make a new jellyfish? In multicellular organisms, cells become specialized, 
and so sperm cells and egg cells become specialized for reproduction. The basic 
idea is pretty obvious: keep some eggs around. What is really astonishing is that 
every jellyfish (not to mention every human!) must be rebuilt from scratch, 
from a single set of DNA plans contained in a fertilized egg. And there is basi­
cally only one way to do this: the same way nature did it. The one-celled egg 
cell must evolve into a multicelled organism with a skin and a gut, an inside 
and an outside, your basic coelenterate. Then the basic coelenterate, if it hopes 
to be a human, must evolve a backbone, making the evolving human embryo 
your basic fish. Then, if this basic fish is to become a human, it must develop 
limbs, making your basic fish look like your basic reptile. After all, one could 
hardly develop fingers before arms. 
The individual evolves in a fashion similar to the way his species evolved. 
In biology, this principle is captured in the phrase Ontogeny (the development 
of the individual) recapitulates phylogeny (the development of the species within 
a phylum). We will return to this principle in chapter 12, but for the present, 
it is pertinent to see that this is not only just another manifestation of self-simi-
larity, but a procedural self-similarity. Like the structural self-similarity I have been 
describing, it is simply the path of least resistance through a series of evolu­
tionary problems. How can nature keep toxins from the crystals? First build a 
membrane. How can nature build fingers and toes? First build limbs. This prin­
ciple is classically illustrated in figure 2.5, in which fowl, rabbit, and human 
embryos all begin by looking alike (and rather fishlike) but then diverge as 
they develop. 
But who needs sex? Puritans think we would all be better off without it, 
and a few plants and animals do seem to do nicely without it. The key word 
here seems to be “few.” In the evolutionary arms race, species must seek a new 
and better offense (or a new and better defense). This means species must 
change. One can accomplish this change by sitting around in the primordial 
puddle, waiting for an act of God to effect a mutation, or one can go out and 
actively swap DNA, combinatorially accelerating change in the fractal equation 
of life. While only a few species reproduce asexually, the myriad species inhab­
iting every nook and cranny of Earth’s ecosystem are the product of sex. 

JONES
’
S  THEORY  OF  EVOLUTION 
•
31
 
Figure 2.5. 
Ontogeny recapitulates phylogeny. (von Baer 1828. Reprinted by 
permission of the New York Public Library.) 
Some Basic Brains Evolve: The One-Celled Brain 
Just as the evolutionary criterion of survival dictated how simple organisms had 
to be structured and how they had to reproduce, there seem to exist only a few 
paths along which the modern neuron could evolve and survive. One path is 
exemplified by arthropods like the crayfish. In the crayfish and other lower phyla, 
the nervous system takes on the appearance of a serial anatomy (figure 2.6).
3
 In 

32  • 
HOW  THE  BRAIN  EVOLVED  LANGUAGE 
Figure 2.6. 
An avalanche anatomy. 
this brain design, sometimes called an “avalanche,” a single signal, originating 
in the crayfish brain, is sent down the tail. Each swimmeret of the tail is then 
successively innervated by an axon collateral, and the crayfish is propelled along, 
swimmeret by swimmeret. This branching of the axon into multiple axon col­
laterals is an important and ubiquitous feature of axons. The neuron doesn’t 
really use its own energy to send its electrical signal; it uses the local Na
+
, which 
surrounds its axon everywhere. As a result, it hardly takes any more energy to 
send a signal along a hundred or a thousand axon branches than it does to send 
a signal along one axon. As a result, even the lowly crayfish can use a single, central 
nervous system to coordinate a multitude of distal swimmerets. 
Figure 2.6 is not intended to insult the intelligence of your average lobster; 
arthropods actually have many more than one cell in their brains. Rather, figure 
2.6 is a minimal anatomy. It uses just one diagrammed cell to self-similarly repre­
sent a population of many cells. (Technically, we should perhaps call each “cell” 
or “neuron” of such a minimal anatomy by a more abstract name like “site” or 
“neurode” or “population.” But concrete terms like “neuron” are more readable, 
and now that I have made my point, belaboring the reader with self-similar ex­
ample upon self-similar example of self-similarity, I will strive for readability.) 
A one-celled brain can be simple-minded in more than one way. Consider, 
by contrast with the crayfish, how the common jellyfish moves. To move, its 
brain gives a single command which contracts the enteron, forcing water out 
the rear and propelling the jellyfish forward. In order for this to work, all the 
muscles around the periphery must contract and relax together. This is accom­
plished with a radially branching axon, as in figure 2.7a. This minimal anatomy 
is only a slight evolutionary variation on figure 2.6: figures 2.6 and 2.7a are both 
“one-celled brains” with axon collaterals. But they are also significantly differ­

JONES
’
S  THEORY  OF  EVOLUTION 
•
33
 
Figure 2.7. 
A radial anatomy, or “outstar.” 
ent in how they move and behave. Whereas figure 2.6 is a “serial anatomy,” 
figure 2.7a is a “parallel anatomy”: all terminal nodes are activated at once, in 
parallel. In figure 2.7b, this radial anatomy is further schematized as an “outstar.” 
Note that such outstar minimal anatomies do not violate the biological fact that 
every cell has one and only one axon leaving its cell body. Outstar minimal 
anatomies are drawn to emphasize the parallel branching of axon collaterals 
or the many axons emanating from a population of many neurons. 
The Two-Celled Brain 
The first two-celled brain may be supposed to have evolved when two neurons 
accidentally synapsed with each other. The result is the minimal anatomy of 
figure 2.8. Note first that this minimal anatomy implies a parallel organization. 
Figure 2.8. 
A two-celled brain. 

34  • 
HOW  THE  BRAIN  EVOLVED  LANGUAGE 
No matter whether cell x
1
 initially excites cell x
2
 or vice versa, in short order 
they are both excited together, at once, in parallel. But like many other evolu­
tionary accidents, the two-celled brain entailed both good news and bad news. 
The good news is the following. Let either cell start “firing.” It thereby 
activates the other cell, which thereby activates the first cell, and so on. We will 
say that the two cells resonate, and in chapter 5, adaptive resonance theory will 
show how this can be a very good thing. It is a form of memory, albeit only short-
term memory. The bad news is that this resonance is also the model of a ner­
vous system out of control, a kind of miniature model of an epileptic convulsion, 
and as in a convulsion, resonance and contraction cannot go on forever. Such 
a brain works for jellyfish: parallel activation causes the jellyfish’s radial mus­
culature to contract convulsively, and the jellyfish “swims” forward, but only 
when the convulsion stops from nervous exhaustion can the muscles relax and 
the jellyfish ready itself for another convulsive forward stroke. When coelenter­
ates evolved, it was an evolutionary marvel that they could move at all, but in 
the race between the quick and the dead, other organisms soon found a faster 
way to move. 
The Six-Celled Brains 
Several phyla evolved a six-celled brain model, most famously the phylum 
Chordata, the vertebrates, you and I. Like many other phyla, the chordates 
abandoned radial symmetry and evolved bilateral symmetry: their bodies have 
a left and a right side. 
Unlike most surviving phyla, in which the brain is connected to periph­
eral nerves and muscles along a route ventral to (beneath) the enteron, in 
Chordata this central route runs dorsal to (above) the enteron, within a noto-
chord. In higher chordates, the notochord became bony and rigid, and this 
structural backbone, in combination with the six-celled brain, gave chordates 
a particularly fast form of locomotion. 
A simple, bilateral vertebrate like a fish moves forward by successively con­
tracting first the muscles on the left side of its backbone, then those on the 
right, then the left, and so on. But to get this rhythm, the minimal anatomy of 
the chordate brain needs two two-celled brains—one for each side of its body— 
as well as two inhibitory cells like R
i
 and L
i
 in figure 2.9a. This brain, however, is 
still “convulsive” and slow like a jellyfish. Neither L
e
 nor L
i 
will stop signaling 
until it has exhausted itself. Only then will it become possible for R
e
 to activate 
and drive the fish’s contralateral (“other-side”) stroke. 
The anatomy of figure 2.9a will reappear time and again in subsequent 
chapters, where its resonant architecture will be used to drive learning. But to 
survive in the primordial soup, the first vertebrates had to be fast and efficient, 
and for this they had to give the anatomy of figure 2.9a a further twist. In fig­
ure 2.9b, the motor drive signals from L
e
 and R
e
 are twisted contralaterally.
4
 In 
chapter 4, we will use this twist to associate L
i
 and R
i
 with the modern verte­
brate cerebellum. For our ancient ancestors, however, it was enough that L
e 

JONES
’
S  THEORY  OF  EVOLUTION 
•
35