Water Was an Element

Four things

Around 450 BC Empedocles proposed that everything is made from four roots: earth, water, air, fire. Aristotle took the scheme up, gave it a physics, and it held.

It is easy to file this under ancient error, but that misreads what it was. It was a working classification, and a decent one. Each root was defined by a pair of qualities — water was the cold and the wet — and the scheme explained why ice becomes water becomes steam, why oil floats, why flame rises. It predicted things. It was wrong in the way a good model is wrong: usefully, and for a long time.

It also placed water among the fundamentals. Within that scheme water was fundamental rather than compounded from more basic materials — Aristotle allowed the four to turn into one another, but none of them was built out of anything simpler. That was precisely what late-eighteenth-century chemistry overturned.

Two years

In 1781 Henry Cavendish burned what he called inflammable air with common air and found that the residue condensed into water. He did not publish until 1784, in the Philosophical Transactions, and even then he described the result in the vocabulary of phlogiston — a theory already beginning to lose its grip.

Cavendish's assistant Charles Blagden carried the news to Paris in 1783. Antoine Lavoisier saw immediately what it meant, repeated the synthesis, and in November 1783 reported to the French Academy of Sciences that water was not an element but a compound: a combination of oxygen and what was then called inflammable air. In the reformed nomenclature that followed in 1787, that gas acquired its enduring name — hydrogen, the water-former.

It became one of the ugliest priority disputes in the history of chemistry — Cavendish, Lavoisier and James Watt each with a claim, and Watt publishing in the same volume of the same journal as Cavendish. The dispute has outlived the interest of the question. What matters is the date. Somewhere between 1783 and 1784, water stopped being an element and became a compound — a substance whose composition could be taken apart.

The proof came from the other direction in 1800, when William Nicholson and Anthony Carlisle put a Volta pile into water and watched two gases come off two wires. Not synthesis this time but decomposition: the thing that could not be taken apart, taken apart, in a glass, by anyone with a battery.

The second error

The correction of 1784 was so complete, and so satisfying, that its shorthand hardened into a synonym. H₂O came to mean water, and water came to mean H₂O, and a formula that describes composition at the molecular level was quietly promoted into a description of a glass.

The formula is not wrong. But most water we encounter — from springs and taps to bottles and the sea — is not a chemically pure sample; it is a solution. Its water molecules are H₂O, and the formula describes them exactly. It does not describe everything else present in the sample, and it says nothing about which isotopes those molecules are built from.

Three things the formula leaves out

Dissolved solids. Rain begins with relatively little dissolved mineral matter. Once it reaches rock and soil its chemistry starts to diverge: calcium, magnesium, sodium, potassium, bicarbonates, chlorides, sulfates, in proportions set by whatever it passed through and how long it took. Many of these naturally acquired ions are not contaminants; they are part of what gives one source a different chemical profile, and a different taste, from another. WHO sets no health-based guideline for total dissolved solids, and the taste-panel data it summarises describe water with very little of them as flat.

Ions. Even chemically pure water is not only intact molecules. Proton transfer between water molecules continuously creates tiny populations of hydronium and hydroxide ions, which recombine just as continuously. At 25 °C, pure neutral water contains roughly one hydronium ion per 550 million H₂O molecules, with an equal concentration of hydroxide — minute, but not zero, and the reason pure water has a pH at all.

Isotopes. These add another layer. The formula H₂O specifies elemental composition, not isotopic identity. Natural water is therefore a family of isotopologues — mostly H₂¹⁶O, alongside smaller populations carrying deuterium, oxygen-17 or oxygen-18. In the international standard for ocean water, deuterium accounts for about 156 hydrogen atoms in every million, oxygen-18 for about 2,005 oxygen atoms in every million, oxygen-17 for another 380. Put together, roughly one molecule in 372 contains at least one isotope other than hydrogen-1 or oxygen-16.

Every one of them is water, and every one of them is H₂O. But a litre of chemically pure water holds on the order of 10²³ molecules built from the less abundant isotopes — a population the molecular formula does not distinguish. Chemical purity does not imply isotopic uniformity.

A formula is a description of a molecule. It was never a description of a glass.

The same mistake, twice

Set the two errors side by side and they turn out to be one error, made in opposite directions.

The ancients had a working simplification — four roots, water among them — and mistook it for a complete account. It survived twenty-two centuries because it kept being useful, and usefulness is very easily confused with truth.

We inherited a far better simplification, H₂O, and are making the identical move. The formula is not wrong. It is exact about what it describes. It simply does not describe everything in the glass, any more than the four roots described what fire is.

Which suggests the safer habit is not to distrust simplifications — they are how anything gets understood — but to keep asking what each one leaves out, and to notice when the answer has quietly become nothing.