In 1814, in a book about probability, Pierre-Simon Laplace described an imaginary intelligence that knows the position and velocity of every particle in the universe at a single moment, and possesses the analytical power to process that information. For such an intelligence, he wrote, nothing would be uncertain, and the future as well as the past would be present before its eyes.
The passage is the clearest statement of physical determinism ever written, and it has been called Laplace’s demon since the nineteenth century. What is usually missed is that Laplace introduced it in order to explain why we are obliged to use probability at all.
The Physics Behind the Claim
Laplace was not speculating idly. He had spent thirty years demonstrating that the solar system behaves exactly as a deterministic machine should.
Newton’s Principia had established the law of gravitation and solved the two body problem completely. It had not settled whether the solar system is stable, and Newton suspected it was not. The mutual pull of the planets on one another, particularly the large interaction between Jupiter and Saturn, appeared to produce accumulating changes in their orbits. Newton suggested that divine intervention might be required from time to time to set things right.
Laplace, working alongside Lagrange, attacked the problem analytically and showed that the troubling variations in Jupiter and Saturn are not cumulative but periodic, with a cycle of around nine hundred years. The system corrects itself. He extended the result into a general demonstration of the stability of the planetary orbits under mutual perturbation, within the approximations available to him.
The five volumes of Mécanique Céleste, published between 1799 and 1825, restated the whole of Newtonian astronomy in the language of the differential calculus and solved most of the outstanding problems. It is the book that turned the Principia’s geometry into a working computational science, and it is where the confidence behind the demon comes from.
The famous exchange with Napoleon belongs here. Presented with a copy, Napoleon observed that the work made no mention of the Creator. Laplace is reported to have replied that he had no need of that hypothesis. The remark is usually read as atheism, and William Herschel’s account of the conversation suggests something narrower and more interesting: Laplace meant that the stability of the system no longer required periodic adjustment, because he had proved that it maintains itself.
The Passage Itself
The demon appears in the Essai philosophique sur les probabilités, the introduction Laplace wrote for his treatise on probability theory. The context matters.
His argument runs like this. Every event is connected to preceding events by the laws of nature, and nothing happens without a cause. If we knew all the causes with complete precision, we would predict everything. We do not, and never will, and probability theory is the mathematics of reasoning correctly given that ignorance.
Probability, on this view, is not a feature of the world. It is a feature of us. A coin toss is fully determined by the forces applied to the coin and the air it moves through, and we assign it a probability of one half because we cannot measure any of that. The demon is a device for making the point that randomness is a description of our position, not of nature’s.
Laplace developed the mathematics to go with this: the analysis of errors of observation, a form of what is now called Bayesian inference, the rule of succession for estimating probabilities from limited samples, and the tools that became standard in astronomy for extracting a best estimate from imperfect measurements.
The First Crack: Chaos
The demon requires exact knowledge of the initial conditions. Everyone understood that this is unobtainable in practice. The assumption, unexamined for most of the nineteenth century, was that approximate knowledge yields approximate prediction: measure a little better, predict a little further.
That assumption is false, and the person who found out was Henri Poincaré.
In 1889 he entered a competition sponsored by King Oscar II of Sweden on the stability of the solar system, essentially the three body problem. He won. While the printed version was being distributed, an editor’s query led him to find a serious error in his own work, and he paid to have the printed copies recalled and destroyed at a cost exceeding the prize money.
The corrected analysis contained something stranger than the original. In certain three body configurations, trajectories that begin arbitrarily close together diverge in a tangled way that no series expansion can capture. The system is fully deterministic and its long term behaviour is unpredictable.
This is sensitive dependence on initial conditions, rediscovered and popularised by Edward Lorenz in 1963 through his work on atmospheric convection and now known as chaos theory.
Chaos does not contradict determinism. Every equation involved is exact and the future is fixed by the present. What it destroys is the demon’s practical cousin, the idea that better measurement buys longer prediction. In a chaotic system the required precision grows exponentially with the time horizon, so predicting twice as far ahead may require squaring the accuracy of the initial data. The demon still works, but only with literally infinite precision, and nothing less will do.
The Second Crack: Thermodynamics
The mechanical laws Laplace relied on are time reversible. Run a film of two colliding billiard balls backwards and it depicts a physically possible collision.
Yet the world is full of processes that never run backwards. Heat flows from hot to cold, gases expand to fill their containers, and mixed things do not spontaneously unmix. The second law of thermodynamics gives this a direction, and Boltzmann explained it statistically: irreversible behaviour is what enormous collections of reversible particles do, because the disordered states outnumber the ordered ones overwhelmingly.
This does not refute determinism either, and Laplace’s demon, tracking every particle individually, would see no arrow of time at all. It does mean that every description of the world at a human scale is statistical, and that the exact microscopic account is not merely unavailable but useless for any purpose we have.
The Third Crack: Quantum Mechanics
The decisive blow came in the 1920s, and it is a different kind of objection.
Heisenberg’s uncertainty principle states that the position and momentum of a particle cannot both have definite values simultaneously. This is not a statement about the limitations of instruments. It is a structural feature of the theory: the quantities are related in such a way that sharpening one necessarily blurs the other.
The demon’s opening move, specifying the position and velocity of every particle, is therefore not merely impossible in practice. Under the standard interpretation of quantum mechanics, there is no such specification to be had, because particles do not possess those properties jointly. The Born rule gives probabilities for measurement outcomes, and those probabilities are not a confession of ignorance in Laplace’s sense.
Einstein rejected this for the rest of his life, holding that the theory was incomplete and that a deeper deterministic description would be found. The argument has continued through Bell’s theorem and decades of experiment, and the deterministic alternatives that survive have had to accept features, such as instantaneous non locality, that Laplace would have found no more palatable.
What Is Left of the Demon
As a description of the universe, the passage did not survive the twentieth century. As a piece of intellectual history it remains valuable for three reasons.
It states, more clearly than anything else written at the time, what classical physics implied if taken seriously. It correctly identifies probability as the mathematics of limited information, which is exactly what probability is in most practical applications today, from insurance to machine learning. And it set up the questions that the next two centuries answered: how much does approximate knowledge buy you, why does time have a direction, and are the fundamental quantities even well defined.
Laplace’s own other work has aged better than his metaphysics. The Laplace transform, Laplace’s equation, the nebular hypothesis for the formation of the solar system, and his contributions to probability are all in daily use. He also proposed in 1796, independently of John Michell, that a star sufficiently massive would have an escape velocity exceeding the speed of light and would therefore be invisible, which is a recognisable description of a black hole written eighty years before anyone had a theory that could support it.
The Universe as a Machine
The clockwork picture ran from Newton to Poincaré and shaped far more than physics. It underwrote nineteenth century confidence that every phenomenon, including biological and eventually mental ones, would yield to mechanical explanation given enough time.
What replaced it is not a return to mystery. Chaos, statistical mechanics and quantum theory are all rigorous mathematical descriptions, and they predict the behaviour of the world with extraordinary accuracy. They simply do not permit the particular kind of total foreknowledge that Laplace described, and two of the three rule it out in principle rather than in practice.
No products in the cart. 