Wiener Ch 1 — Newtonian vs Bergsonian Time
Weisst wie viele Sterne stehen / In dem blauen Himmelszelt? / Weisst, wie viele Wolken gehen / Weit hinüber alle Welt?
Knowest thou how many stars stand in the blue tent of heaven? Knowest thou how many clouds pass far over the whole world?
A German children’s hymn opens the chapter by accident — it puts two sciences side by side that share only a sky.
Astronomy vs meteorology
You can count the stars. A star is a definite object; the Durchmusterung — the catalogue — only stops short because human telescopes do. Ask a meteorologist for an analogous catalogue of the clouds and he’ll laugh. In all the language of meteorology there is no such thing as a cloud with quasi-permanent identity. A topologist might define one as a connected region of space where the water content in solid or liquid state exceeds a threshold, but the definition is useless: that region is gone in minutes. What concerns the meteorologist is a statistical statement — Boston, January 17, 1950. Sky 0.38 overcast. Cirrocumulus.
Astronomy is the oldest science. Meteorology is among the youngest. Eclipses can be predicted centuries out; tomorrow’s weather is hard and often wrong. There is cosmic meteorology too — galaxies, nebulae, star-clusters — but that’s a young branch outside the classical tradition. The tradition is the solar system: Copernicus, Kepler, Galileo, Newton. The wet-nurse of modern physics.
It’s an ideally simple science. The Babylonians already knew eclipses ran in cycles extending backwards and forwards through time. Time itself was best measured by the motion of the stars. The pattern for every event was a wheel — Ptolemaic epicycle or Copernican orbit — and in any such theory the future repeats the past. When Newton reduced it all to a closed mechanics, the fundamental laws were unaltered when you replaced t with −t.
The music of the spheres is a palindrome. The book of astronomy reads the same backwards as forwards.
Run a film of the planets in reverse and it’s still a possible planetary motion. Run a film of a thunderhead in reverse and everything looks wrong — down-drafts where up-drafts should be, turbulence growing coarser, lightning preceding the cloud it usually follows.
Why the difference
Meteorology involves a vast number of approximately equal particles, many of them closely coupled. The solar system has few particles, wildly different in size, loosely enough coupled that second-order effects don’t change the picture and higher-order effects vanish. The planets move under conditions more favorable to isolating a limited set of forces than any laboratory experiment we can set up. Compared to the distances between them, even the sun is nearly a point. Compared to the deformations they suffer, the planets are nearly rigid. The space they move through is almost empty. Their masses are almost constant. The departure from the inverse-square law is minute. Their positions, velocities, and masses are well-known. Computing their past and future positions is, in principle, easy.
Meteorology is the opposite. The number of particles is so enormous that an accurate record of their initial positions and velocities is impossible — and if you had it, you’d have an impenetrable mass of figures. Cloud, temperature, turbulence — these refer not to one physical situation but to a distribution of possible situations, of which only one is realized. If every meteorological station on Earth reported simultaneously, the readings wouldn’t give a billionth of the data needed to characterize the atmosphere from a Newtonian point of view. They give constants consistent with infinitely many atmospheres. With a priori assumptions, you get a probability distribution over the set of possible atmospheres. Using Newton’s laws, or any other causal laws, the most you can predict is a probability distribution of the constants — and that prediction fades with time.
The observer is directed in time
Even in a perfectly reversible Newtonian system, questions of probability and prediction give asymmetric answers, because the questions themselves are asymmetric. If I set up an experiment, I bring the system from the past into the present, fixing some quantities and assuming statistical distributions for others. I then observe what comes out after a given time. I cannot reverse that. To reverse it I’d need to find a fair distribution of systems that, without intervention, end up within tight statistical limits — and then ask what they were doing a given time ago. For a system from an unknown initial state to land in any tight statistical range is so rare it would be a miracle, and you can’t build an experimental technique on counting miracles.
We are directed in time. Our relation to the future is not our relation to the past. Every question we can ask is conditioned by this asymmetry, and every answer is too.
So why does the directional thermodynamics we learn on Earth still work for astrophysics, where the experiment is a single observation of a remote heavenly body? Because we observe via incoming light. We dark-condition the eye to avoid after-images; we wrap plates in black paper to prevent halation. We see stars radiating toward us. If somewhere there were a star whose evolution ran the other way, it would attract radiation from the whole heavens, and we’d never know — its existence would be invisible by construction. The very fact that we see a star means its thermodynamics is like our own.
Imagine a being whose time runs the opposite of ours. Any signal he sends arrives as the logical consequent (from his side) of a chain of antecedents already in our experience. If he drew us a square, we’d see the remains of his figure as its precursors, and the square’s appearance would look like a curious crystallization — explainable by natural laws, looking as fortuitous as faces in cliffs. Drawing the square would, to us, be a catastrophe in which a square ceased to exist. His view of us would be the same.
Within any world we can communicate with, the direction of time is uniform.
Every science slides toward meteorology
Most sciences sit between Newton and the weather, and most sit closer to the weather. Astronomy itself contains tidal evolution: the moon raises two water hills on the Earth; the wave gets tangled in coastal shallows and lags; friction drags the moon backward in its orbit and accelerates the Earth’s rotation forward. Across deep time, the day and the month converge — the moon already shows the Earth one face, the result of an ancient tidal evolution when the moon could deform. Even gravitational astronomy involves frictional processes that run down. No science conforms precisely to the strict Newtonian pattern.
Biology runs one way. Birth is not the reverse of death; anabolism is not the reverse of catabolism; cell division and fertilization are time-asymmetric. The individual is an arrow through time, and so is the race. The paleontological record bends from simple to complex. By mid-19th century this trend was obvious to any honest scientist, and it’s no accident that Darwin and Wallace arrived at the mechanism near-simultaneously: fortuitous variation, carved by differential viability into a near-directional progress. Mendel sharpened it; de Vries added mutation; the chromosome and gene followed; Haldane made the statistics work.
We’ve already seen the tidal evolution of George Darwin — Charles Darwin’s son. The name “evolution” is not chosen at random. In tidal evolution as in the origin of species, fortuitous variability is converted by a dynamical process into a pattern that reads in one direction. Tidal evolution is an astronomical application of the elder Darwin.
The third Darwin, Sir Charles, is an authority on quantum mechanics — fortuitous in genealogy, not in significance. The succession Maxwell–Boltzmann–Gibbs is a progressive statistical mechanics: Newtonian mechanics applied not to one system but to a statistical ensemble, where conclusions concern overwhelming majorities, not individual cases. By 1900 it was clear something was wrong with thermodynamics, particularly around radiation: the ether absorbed high-frequency radiation much less than existing theory predicted (Planck’s law). Planck gave a quasi-atomic radiation theory that fit the data but quarreled with the rest of physics. Bohr added a similarly ad hoc atom. Newton and Planck are the thesis and antithesis of a Hegelian triad; the synthesis is Heisenberg’s statistical theory of 1925, in which even Newtonian dynamics has become a picture of average behavior — an account of an evolutionary process.
Bergson and the moved wall
This transition — from Newtonian reversible time to Gibbsian irreversible time — has philosophical echoes. Bergson distinguished the reversible time of physics, in which nothing new happens, from the irreversible time of biology, in which there is always something new. The realization that Newton was the wrong frame for biology was the central point of the old vitalism-mechanism controversy, complicated by the desire to preserve some shadow of soul or God against materialism.
The vitalist proved too much. Instead of building a wall between life and physics, the wall got pushed outward to enclose both. The matter of the newer physics is not Newton’s matter, but it’s just as remote from the vitalist’s wishes. The chance of the quantum-theoretician is not the ethical freedom of the Augustinian, and Tyche is as relentless a mistress as Ananke.
Engineering follows the philosophy
The thought of every age is reflected in its technique.
Ancient civil engineers were surveyors, astronomers, navigators. In the 17th and early 18th centuries they were clockmakers and lens-grinders. The craftsmen made tools in the image of the heavens. A watch is a pocket orrery, moving by necessity as do the celestial spheres; friction and energy dissipation are nuisances to be minimized so the hands move as periodically as the planets. Huygens and Newton produced the engineering of the great age of navigation, which converted ocean commerce from chance to business. The engineering of the mercantilists.
To the merchant succeeded the manufacturer; to the chronometer, the steam engine. From Newcomen onward the central field of engineering has been the study of prime movers. Heat is converted into rotation and translation; Newton is supplemented by Rumford, Carnot, Joule. Thermodynamics arrives — a science in which time is eminently irreversible. The conservation of energy and the statistical explanation of the Carnot Principle fuse thermodynamics and Newtonian dynamics into the statistical and non-statistical aspects of one science.
If the 17th and early 18th centuries are the age of clocks, and the late 18th and 19th are the age of steam engines, the present age is the age of communication and control. Electrical engineering splits into the technique of strong currents and the technique of weak currents — what we call power and communication engineering. Communication engineering can handle currents of any size, including the ones that swing massive gun turrets; what distinguishes it from power engineering is that its main interest is not the economy of energy but the accurate reproduction of a signal. Tap of key to tap of receiver; sound at one end of a telephone to sound at the other; turn of a ship’s wheel to the angular position of the rudder.
Communication engineering started with Gauss, Wheatstone, and the first telegraphers; got its first scientific treatment from Lord Kelvin after the first transatlantic cable failed; was brought to modern shape from the 1880s by Heaviside. Radar and anti-aircraft fire control in the Second World War brought in a large cohort of well-trained mathematicians. The automatic computing machine belongs to the same realm of ideas.
The automaton through the ages
At every stage since Daedalus, the ability to make a working simulacrum of a living organism has fascinated. The form of the automaton tracks the technique of the age.
- The age of magic: the Golem. A figure of clay into which the Rabbi of Prague breathed life with the blasphemy of the Ineffable Name.
- The age of Newton: the clockwork music box, little effigies pirouetting stiffly on top.
- The 19th century: a glorified heat engine, burning fuel instead of glycogen.
- The present: opens doors with photocells, points guns at radar tracks, computes solutions to differential equations.
The clockwork automaton played a serious role in modern philosophy that we tend to forget. Descartes treats the lower animals as automata — partly to dodge whether they have souls. How they function he doesn’t say, but the related question of how the human soul couples to its material environment he does discuss, unsatisfactorily, placing the coupling in the pineal gland. Whether mind acts directly on matter, he’s unclear; the validity of human experience he attributes to the goodness of God.
God’s role here is unstable. Either He’s entirely passive (in which case Descartes hasn’t explained anything), or He’s an active participant (in which case His honesty is itself a participation in sensation). The causal chain of matter is paralleled by a causal chain starting from God’s act. The Occasionalists — Geulincx, Malebranche — take this path. Spinoza, the continuator, treats mind and matter as two self-contained attributes of one God, but isn’t dynamically minded enough to address the mechanism.
Leibniz is. He replaces the pair (mind, matter) with a continuum of analogous elements — monads — each living in its own closed universe, with its own perfectly determined causal chain from creation onward, but coordinated with the others through the pre-established harmony of God. Leibniz compares them to clocks wound up at the creation to keep time together for eternity. Unlike human clocks, they never drift, by the workmanship of the Creator. The monads have no real influence on the outside world and aren’t influenced by it. They have no windows.
The monad is a Newtonian solar system writ small.
The 19th-century body as heat engine
In the 19th century, automata — humanly built ones and the natural ones, the animals — get studied from a very different angle. Conservation and degradation of energy are the ruling principles. The living organism is above all a heat engine, burning glucose or starch into carbon dioxide, water, and urea. The metabolic balance is the center of attention. The low working temperatures of animal muscle, compared to a heat engine of similar efficiency, get glibly explained by contrasting chemical energy with thermal energy. The fundamental notion is potential. The engineering of the body is a branch of power engineering. This is still the view of the conservatively-minded physiologist, and Rashevsky’s biophysics carries it forward.
The reversal: the body is not a conservative system
We’re coming to realize the body is far from a conservative system, and that its components work in an environment where available power is much less limited than we’d assumed. The electronic tube showed that a system with an outside energy source, almost all of which is wasted, can be highly effective if it operates at a low energy level. Neurons — the atoms of the nervous complex — work under much the same conditions as vacuum tubes, with relatively small power supplied from outside by the circulation. The bookkeeping that matters for their function is not one of energy.
In short: the newer study of automata, in metal or in flesh, is a branch of communication engineering. Its cardinal notions: message, noise (a term taken from telephone engineers), quantity of information, coding technique.
What the modern automaton is
The modern automaton is coupled to the external world by a flow of impressions, of incoming messages, and of the actions of outgoing messages — not merely by an energy flow. Its parts:
- Receptors (sense-organs): photo-electric cells, radar receivers, hydrogen-ion-potential recorders that taste, thermometers, pressure-gauges, microphones.
- Effectors: motors, solenoids, heating coils.
- A central control between receptor and effector that recombines incoming impressions into the form needed to produce a desired response.
- Proprioception: the information fed back to the central control often concerns the functioning of the effectors themselves — the analogue of the kinaesthetic organs.
- Memory: information need not be used at once; it may be delayed or stored for later.
- Learning: while the automaton is running, its rules of operation are susceptible to change on the basis of data that has passed through its receptors.
These are not the dream of the sensationalist. They already exist as thermostats, gyrocompass ship-steering, self-propelled missiles that seek their target, anti-aircraft fire control, automatically controlled oil-cracking stills, ultra-rapid computing machines. The steam-engine governor is an old member of the family. The present age is as truly the age of servo-mechanisms as the 19th was the age of the steam engine or the 18th the age of the clock.
The synthesis
The relation of these mechanisms to time needs care. Input-output is consecutive — a definite past-future order. What’s less obvious is that the theory of the sensitive automaton is a statistical one. We’re not interested in performance for a single input; the machine must give a satisfactory performance over a class of inputs, which means a statistically satisfactory performance for the class it’s statistically expected to receive. Its theory belongs to Gibbsian statistical mechanics, not to classical Newtonian mechanics.
The modern automaton exists in the same Bergsonian time as the living organism. There is no reason in Bergson’s considerations why the essential mode of functioning of the living organism should not be the same as that of the automaton of this type.
Vitalism has won to the extent that even mechanisms correspond to the time-structure of vitalism — but this victory is a complete defeat. From every point of view with the slightest relation to morality or religion, the new mechanics is fully as mechanistic as the old. Materialism has come to be little more than a loose synonym for mechanism. The whole mechanist-vitalist controversy has been relegated to the limbo of badly posed questions.
Structure choice: contrast. The chapter opens with two sciences (astronomy: palindromic Newtonian time; meteorology: directional Bergsonian time), shows every science slides toward the meteorological pole, then re-examines the modern machine and finds it lives in Bergsonian time too. The opposition is set up to be dissolved — not by picking a winner but by moving the wall out to enclose both sides.
What I cut: the long passage on Sir Charles Darwin III’s quantum biography (kept the punchline); most of Descartes’s pineal-gland metaphysics (kept the structural move); the bulk of Leibniz on monads (kept the clockwork-eternity image and “no windows”); the Boston cirrocumulus example trimmed; the second cosmic-meteorology aside; numbered subsections (Wiener has none in this chapter but the chapter has many dependent clauses I straightened).
What I added: the load-bearing one-liners are bolded — those are scannable structural markers I introduced, not Wiener’s italics.