Bring a comfy chair to the lab in your search for inspiration

The armchair is ideal for idle fancy, while the lab bench can be the graveyard of creativity. But combine them and, argues Peter Atkins, the results can be spectacular.

Armchairs alone have never been a source of reliable discoveries concerning the fabric of reality. The laboratory bench, on its own, is also a sterile region of the intellectual world. Put them together, though, and they become gloriously potent. For thoughtful reflection, imaginative speculation and controlled, publicly accessible shared experience have jointly resulted in that most extraordinary of human achievements, science.

Aristotle must take the Oscar for armchair misconception. I do not mean to denigrate and scorn him unduly, for he is a beacon of questing intellect - searching, asking and coming up with what then passed for answers. However, his legacy, elaborated as it was by his medieval commentators, was a benighted world.

Take his mechanics, for instance. One of Aristotle's most durable explanations was that of an arrow in flight. Extrapolating from the mud-locked oxcart of the farmyard to the flight of a slender arrow through the air, he supposed that the aerial oxen that drove the arrow through the air were vortices that somehow impelled it forward. How absurd it would be for him and his contemporaries to suppose that motion was natural, for then farmers would have to harness their oxen behind the cart to stay its forward progress! Had Aristotle had a laboratory to hand, he could have tested his conjecture by shooting an arrow through a vacuum: according to him, it would have flopped to the ground. Now we know, by experiments originating with Galileo and reflected on by Newton, that exactly the opposite of Aristotle's explanation of motion is correct: objects remain in their state of motion unless acted on by a force: vortices impede, not impel.

Biology has not been free of the misleading fruits of armchair speculation any more than physics has been. Deeply puzzling to mankind has been the origin of that extraordinary property of matter - life - and the amazing variety of its curious manifestations that we call the biosphere. Aristotle had a view here, of course, as little is more worthy of the attention of fine intellects. His solution was straightforward: the various organisms had simply dropped to Earth from some celestial factory. The archdeacon of Carlisle, William Paley, took a similarly relaxed view, with his famous analogy of stumbling across a watch and being impelled to the inescapable conclusion, in his view, that there must be a watchmaker. That view is still adopted, in slightly modified form, by creationists.

Empedocles, shortly before he unwisely chose to demonstrate his own divinity by leaping into the crater of Mt Etna (but at least we should admire his experimental spirit), had taken a more secular view, and had a glimmer of an idea that a mechanism was at work. According to him, extant organisms are viable combinations of a universal kit of parts, in which unviable combinations, after a brief hobble, wallow or flutter, simply flopped over and died, disappearing from our ken. Incidentally, I have a sneaking admiration for the dismal and much-derided Lamarck, professor of worms and brief feeder of them while tenant of a rented pauper's grave, for although his view of the emergence of creatures was quite wrong, at least he sought a mechanism - the striving for perfection.

Experiment in biology accompanied the eruption of interest in examining the natural world, with the recognition that fossils were the mineralised remains of the once-living rather than, as Plato had presumed, discarded templates for erstwhile organisms. The hunt was on for scientific explanation of this amazing litter, and it is hardly surprising that it was to emerge in the minds of two observers - Charles Darwin and Alfred Russel Wallace - who had immersed themselves for years in the raw data of the natural world. I relish the picture of Darwin tucking into the flesh of the tortoises he had helped to slaughter on the Galapagos Islands, and picking through the corpses of the finches he had captured and slain, as the unconscious, undirected, serendipitous mechanism of their gradual emergence - natural selection - dawned on him. As his bulldog, Thomas Huxley, was later to say, "How very stupid not to have thought of that."

Natural selection is one of the simplest ideas of science, but its application to the biosphere is one of the subtlest problems because the biosphere reacts back on itself. There are some who whinge that natural selection cannot make predictions and is therefore an act of faith that might as well be enacted in the bench-free armchair. That is not true: once molecular biology had been established and it was realised that DNA changes over time as organisms evolve, it was possible to predict that some sequences would emerge in the same order as the organisms themselves had done. Not a single conflict has ever been observed between the empirical assessment of evolving organisms and their molecular encoding.

Chemistry also benefited enormously when its practitioners rose from the armchair, discarded the secrecy of the alchemist's workshop and stood at a laboratory's public bench. Few chemists could have had more tolerant workmates than those of Hennig Brand, who noted that urine and sand were both golden and so presumed that their colour sprang from an inner gold. He collected barrels of human urine, allowed it to ferment, distilled it to a pasty residue, added sand and distilled the mess. Instead of manufacturing the gold he sought, he produced phosphorus, the first new element to be discovered since antiquity. Brand turned this element into gold metaphorically by guarding the secret of its manufacture and hawking it as a cure for everything - no doubt killing many in the process.

His contribution of a new element to chemistry's pantheon was the beginning of a minor avalanche, for element after element started to be isolated and recognised, not least on account of Humphry Davy at the Royal Institution applying his electrodes to almost everything in sight, and - as his brother John records - dancing with delight on obtaining little globules of new elements, such as sodium and potassium. By the middle of the 19th century, enough elements had been recognised for patterns to be detected by armchair speculators. It was a pioneering writer of chemistry textbooks, Dimitri Mendeleev, who identified the pattern - the periodic table - that is one of chemistry's greatest achievements. It showed, astoundingly, that matter was not a random collection of disparate entities, but families of brothers, sisters and cousins, with their underlying relationships fully understood.

The epitome of armchair activity is mathematics - commonly regarded as a subject that holds reality in some disdain. Indeed, Bertrand Russell made the waspish comment that "pure mathematics is the subject in which we don't know what we are talking about, or whether what we are saying is true". But even mathematics has an experimental aspect. While the ancient Egyptians supplied the equivalent of Kepler's orbits, Darwin's finches and Mendeleev's elements, it was the Greeks who brought proofs, in the form of Euclid's percipient axioms about the shape of empty space.

Beware Greeks bearing proofs, however, for they were not the intellectual gifts that they seemed for the first 2,000 years. That extraordinarily subtle philosopher Immanuel Kant was seduced by the staying power of Euclid's thought to suppose that the shape of space embodied in his axioms was a synthetic a priori , and that therefore the world was incapable of being different. How wrong he was. Carl Friedrich Gauss, the greatest of almost modern mathematicians, wrote that, partly in jest, he had amused himself by wondering if Euclid's axioms could be modified and empty space have a different shape. With theodolite and ruler, he even tried to measure the angle between mountain tops to ascertain that three angles of a huge triangle did indeed add up to 180 degrees. Then Bernhard Riemann stole Gauss's timid thunder, and found that geometries other than Euclid's were logically possible. At a stroke, geometry had become an experimental subject: mathematics had to confront reality in a laboratory.

All of which goes to show that armchairs are indeed fine incubators of fantasies, but they take wings and edge us towards reliable understanding only when they confront the bench.

Peter Atkins is professor of chemistry at the University of Oxford and fellow of Lincoln College. His new book, Galileo's Finger: The Ten Great Ideas of Science , was published this week by Oxford University Press (£20.00).