Discovery is a process not an act.

This morning somebody on Twitter tweeted that William Herschel discovered the planet Uranus on this day in 1781. A typical tweet amongst history of science fans on Twitter, who like to acknowledge and celebrate births, deaths, inventions and discoveries in what amounts to a rolling history of science calendar. On this occasion my history of science soul sisterTM, Rebekah “Becky” Higgitt, who’s quite knowledgeable about eighteenth-century astronomy, tweeted, quite correctly, that Herschel initially thought he had discovered a comet and it was Nevil Maskelyne, who first suggested that he had in fact observed a new planet and not a comet. She then asked if we should not then say that it was Nevil Maskelyne who discovered Uranus and not Herschel? Becky could be considered a bit biased having fairly recently devoted several years of her life to the study of the life and work of Maskelyne and also having edited a, highly recommended, book on the man. Herschel fans might thus feel justified in dismissing her comment and maintain their position than it was the Hanoverian musician turned amateur astronomer who discovered the first new planet to be observed since antiquity. Rather than trying to stoke the fires of a discovery priority dispute, of which there are all too many in the history of science, I think this an opportunity to look critically at what the term discovery actually means in the history of science.

For some reason we love to hang a specific date, even better the exact time, when a discovery of science was made in the history of science. In fact I have about a running metre of books within arms reach of this computer full of such information. William Herschel discovered Uranus on 13 March 1781, Galileo Galilei discovered the moons of Jupiter on 7 January 1610, Simon Marius did the same just one day later, Johannes Kepler discovered his third law of planetary motion on 8 March 1618 and so on and so forth. However this accurate pinning of scientific or technological discoveries onto the ribbon of time creates a very false impression of what discovery is and this was exactly the point that Becky was trying to make on Twitter, which in turn led to me writing this post. Discovery is not a single act by a single person for which it is possible to give a stopwatch accurate moment of discovery but is rather a process spread over a period of time, which can in fact take several years and which almost always involves quite a large number of people.

To illustrate what this means let us take a closer look at Galileo’s epoch making discovery of the four largest (actually it was only three on the first day) moons of Jupiter. On 7 January 1610 whilst observing the planet Jupiter Galileo noted three stars that roughly formed a line with the middle axis or equator of the planet. When he observed again on the following evening they were still there. You might ask so what? Stars belong to the sphere of fixed stars, which are so called because they ‘always’ remain in the same place, whereas planets are called planets (the Greek for wanderer) because they move around with reference to the fixed stars. This being the case Galileo’s three new stars that he had recorded should have changed their position relative to Jupiter, or more accurately Jupiter should have changed its position relative to the three stars. Galileo was astute enough to realise that he was on to something and continued to observe and record the now four new stars and Jupiter over the following nights. The new stars did change their positions relative to Jupiter but not in the way he would have expected if they were fixed stars plus they always stayed in the vicinity of the planet. With time and enough observations Galileo realised that the four new objects were in fact orbiting Jupiter. He had discovered Jupiter’s four largest moons, or had he?

Science requires that new discoveries can be repeated by other independent practitioners/observers and discoveries are only confirmed and thus accepted when this has taken place. Now as stated above Simon Marius in Ansbach had also first observed the moons of Jupiter just one day later on 8 January 1610 and like Galileo had continued to observe them and had also reached the conclusion that they were orbiting the planet. This would have been the necessary confirmation that Galileo required but Marius only published his observations four years later, in 1614, leading Galileo, who by this time had long been acknowledged as the discoverer to denounce Marius as a plagiarist. Back in 1610 when Galileo fist published his observations on 14 March, in his Sidereus Nuncius, people were, not surprisingly, rather sceptical about his claims.

As I have recorded on several occasions on this blog it was the Jesuit mathematician astronomers under Christoph Clavius at the Collegio Romano who provide the necessary independent confirmation of his observations but this was not a simple process. At first the Jesuits did not have a telescope powerful enough to resolve the moons of Jupiter and their initial attempts to construct one failed. However Grienberger and Lembo persevered with assistance from Galileo, from afar by post, and in the end they were able to confirm all of Galileo’s observations. Another aspect of this discovery was to prove that they were actually moons orbiting Jupiter the four new objects needed to be observed consistently and accurately in order to determine their orbits so that one could predict their positions at any given time. Both Galileo and Marius undertook this task, Marius’ results were more accurate than those of his Tuscan rival, but it was first Cassini several decades later who, with much superior telescopes at his disposal, was able to produce tables of the orbits accurate enough to truly satisfy the requirements of the astronomical community.

It would now seem that we are finished with our tale of the discovery of the four moons of Jupiter but there is another extremely important factor that needs to be addressed. New discoveries often involve new methods and/or new scientific instruments, without which the discovery would not have been possible. This was very much the case with the discovery of the moons of Jupiter, which was only made possible by the very recently invented, September 1608, telescope. Any such new methodology or instrumentation must be clearly and convincingly shown to provide objective verifiable facts based on solid scientific theory. No such demonstration of objective scientific reliability existed at this point in time for the telescope. In fact all those in 1610, who doubted the telescopes ability to deliver objective verifiable scientific facts, and who tend to get ridiculed by the cheerleaders of scientism today, were perfectly correct to do so. Galileo, who when it came to optics was a tinkerer rather than a theorist, was not in the position to deliver the very necessary scientific theory of the telescope. Enter Johannes Kepler.

Kepler had already ready written extensively on theoretical optics including one of the earliest scientific analysis of how lenses functions. He was also an unabashed cheerleader for Galileo’s telescopic discoveries, sight unseen, writing the first positive, rather gushing in fact, review of Sidereus Nuncius, which Galileo used for his own propaganda purposes. Kepler realised at once that in order to confirm those discoveries a theoretical description of how the telescope functions was necessary and he sat down and wrote one. His Dioptrice, which explains the science of single lenses, the convex/concave two lens Dutch telescope used by Galileo, the convex/convex two lens astronomical or Keplerian telescope, the three lens terrestrial telescope and even the telephoto lens, was published in 1611. Galileo, arrogant and egoistical as ever, dismissed it as unreadable but it successfully silenced those who doubted the scientific objectivity of the telescope.

All of the factors that I have described above played an important and indispensible part in the discovery of the four largest moons of Jupiter. What we have here is not the act of one person at a specific point in time, in this case Galileo’s first observation of those three stars, but a chain of intertwined events or a process spread over a period of several years. There is nothing exceptional in the discovery of the moons of Jupiter but all scientific and technological discoveries involve a similar complex process carried out by a group of people over a period of time. Discovery is not the single act of a single person but a process involving several and sometimes many people spread over a period of time. The anniversaries that we like to celebrate are mostly just the starting point to that process.



Filed under History of Astronomy, History of science

16 responses to “Discovery is a process not an act.

  1. Rebekah Higgitt

    Useful post, Thony, thanks, both on the idea of ‘discovery’ and the specifics of Jupiter’s moons. Just to correct any apparent Maskelyne bias, my original tweet ( asked whether they should be given joint credit for the discovery, rather than demanding that Maskelyne get it alone. It should probably also be shared with others who made confirmatory observations.

  2. Well said!

    Kuhn argued this point over and over again in Structures, but I like the way he put it in the Afterword to Black-Body Theory, where he gives an answer to your implied question, “For some reason we love to hang a specific date, even better the exact time, when a discovery of science was made in the history of science.”:

    In the socialization and professionalization of scientists the concept of the unit discovery plays a profound and generally functional role. … unit discoveries are the bricks from which, in a familiar image, the edifice of science is piecemeal built. When the student later enters the profession, it is with the understanding that success is measured by the size and number of bricks the individual is able to put in place. That is why, since at least the seventeenth century, attempts to establish priority in discovery have played so large a part in scientific development. … Among historians, however, it is now a truism that that concept will not do. Discoveries are extended processes…

    Of course, Kuhn was not the first to make this point…

    It is

    • Of course there were also the pre-discovery observations of Uranus (and Neptune too), which proved invaluable in calculating the planets’ orbits once they had been identified correctly. In that sense, we can say that the process extended from the first observation (John Flamsteed, 1690, for Uranus) until the correct identification of Uranus as a planet.

  3. Another great example of discovery as an extended process is the inverse square law. Grosseteste (c.1168-1253) discusses the expanding sphere of light in On light, and points out that this accounts (qualitatively) for decreasing intensity with distance. The theme is taken up by Nicolaus Cusanus (1401-1464), cited by Kepler, the one who explicitly states the inverse-square law for light.

    Kepler, however, proposes an inverse first-power law for his “sweeping” force (not gravity) that carries the planets in their orbits; this is justified by the particulars of Kepler’s physics.(*) Boulliau rejects the very idea of explaining planetary orbits with forces, but does throw off the comment that if there was such a force, it would be inverse-square — apparently ignoring, misunderstanding, or rejecting Kepler’s analysis.

    Bouilliau’s off-hand remark makes it way to England where it becomes part of the conversation. Hooke, in his Micrographia, explicitly states an inverse square law for gravity. Later of course he and Newton have a famous priority dust-up over this. Newton complains to Halley that just suggesting the ISL shouldn’t count for much, compared to deriving the elliptical orbits as a consequence; if just saying “inverse square!” is all that important, why not give the credit to Bouilliau instead of Hooke? However, Newton declines to credit Bouilliau for the ISL in the Principia.

    (*) Koyré gave a nice understandable summary. According to Kepler, spokes of “image” radiate out from the sun and are carried around by the sun’s rotation; these serve to sweep the planets around. (We’re in an Aristotelian context, where speed is proportional to applied force.) The density of spokes decreases according to the inverse-square law, but in the ecliptic plane, this is partially compensated by the increased speed with which the spokes “hit” the planets. That speed increases directly with the distance from the sun. Net result, an inverse first-power law.

    Of course, you’ve already told some of this story before: The man who inverted and squared gravity.

    • laura

      I think Koyre is being pretty generous to Kepler in that passage! But his is one of my very favourite books on 17th century science — I especially enjoyed the Borelli chapter on how Borelli tried to Galileify Kepler’s physics… and ended up with a big mess. It really drives home just how difficult the concepts we take for granted in physics were for pre-Newton thinkers.

      • Yeah, I liked the Borelli chapter a lot too. But then, I have a soft spot for short-lived theories that nonetheless have at least one follower/developer. (I think the Sarpi-Galileo theory of tides falls into this category too, since, if memory serves. John Wallis liked it, and tried to account for the moon’s influence in this framework.)

        While Koyré gave a more lucid explanation than Kepler did, I think Kepler’s analysis stands up to scrutiny. (Of course, if we grant his assumptions.) Stephenson’s discussion is also good.

  4. Why not split the issue into: Herschel discovered something, but formed a wrong theory about its nature. Maskelyne then corrected that theory and explained the observation correctly.

    A detailed narrative can have a discoverer, an explainer or understander etc. etc.

    It’s not very useful, IMHO, to reduce everything to discovery and then try to expand that concept, to make it into a process. The word discovery seems to suggest a discrete moment rather than a process.

    • I guess it’s a question of granularity. The traditional potted history of science features a number of major “discoveries”, which are (as Thony puts it) pinned onto the ribbon of time. Like, say, the discovery of oxygen. Most students probably emerge from high-school chemistry with the notion that oxygen was discovered on some definite date by one person, maybe Carl Wilhelm Scheele in 1772, or perhaps Priestley in 1774, or perhaps Lavoisier in either 1775 or 1777.

      Kuhn gives an extensive discussion of just this case in Structure. A couple of highlights:

      Priestley’s sample was not pure, and if holding impure oxygen in one’s hand is to discover it, that had been done by everyone who had ever bottled atmospheric air.

      … in 1777 and to the end of his life Lavoisier insisted that oxygen was an atomic “principle of acidity” and that oxygen gas was formed only when that “principle” united with caloric, the matter of heat. … the principle of acidity was not banished from chemistry until after 1810, and caloric lingered until the 1860s. Oxygen had become a standard chemical substance before either of those dates.

      “The discovery of oxygen” really alludes to a whole bunch of insights and observations, all glommed together into one big Concept. Science has to do this kind of chunking, otherwise it would be impossible to learn and discuss. Historians like to zoom in on the details.

      On the other hand, zoom in far enough and you do reach individual thoughts and sense perceptions of specific individuals — or you would, if the historical record was that comprehensive. Concepts are not literally “in the air”, they develop in discrete jumps in the minds of particular people.

      • I don’t see why historians of science should not provide more chunks or tags than ‘discovery’ and possibly ‘justificstion.’ Potted histories would work as fine with something like:

        Date a: person b discovered object c, but screwed on explaining it. Date d: person e invented f (device, theory or model), but failed to apply it to object c. Date: g: person h applied invention i to object c and thus explained it, but had no impact. Date j: person k did same as h just better, had enormous impact and gained almost all the credit.

      • Lucy M

        That’s a fabulous ‘levels of detail’ explication. Does anyone know if thony has written anything about this? If thony himself sees this and has time to answer, all the better!

      • I totally agree with Joachim and have propagated this point of view in all my work as a historian.

  5. Will Thomas

    It’s worth pointing out that the problematization of “discovery” was a big deal circa 1980.

    Michael mentions Kuhn above, but the key reference is his “The Historical Structure of Scientific Discovery,” Science 136 (1962): 760–764, reprinted in The Essential Tension (Chicago UP, 1977), 165–177.

    S. W. Woolgar, “Writing an Intellectual History of Scientific
    Development: The Use of Discovery Accounts,” Social Studies of Science 6 (1976): 395–422.

    Augustine Brannigan, The Social Basis of Scientific Discoveries (Cambridge UP, 1981).

    Peter Galison, “The Discovery of the Muon and the Failed Revolution against Quantum Electrodynamics,” Centaurus 26 (1982): 262–316. This is his first discussion of discovery as aesthetically informed, which he would develop into the concept of discovery as an expanding “circle” or “circles of belief.”

    Simon Schaffer, “Discovery and the End of Natural Philosophy,” Social Studies of Science 16, no. 3 (1986): 387–420. Although a latercomer to this literature, Schaffer’s piece is notable here, because it actually discusses the Uranus discovery.

    • Yes, that Kuhn essay contains the marvelous line, “To make a discovery is to achieve one of the closest approximations to a property right that the scientific career affords.”

      Don’t you think though that the paper, “Energy Conservation as an Example of Simultaneous Discovery”, from the same collection, gives at least as good an exposition of complex, compound nature of the so-called “unit discovery”?

  6. Sometimes the joy of writing a blog consists of throwing a pebble into the pond and then watching as the ripples spread.

  7. The issue of what constitutes a bona fide discovery is merely nominal wordplay. And of course science is a process not an isolated act. The real question to be addressed for the period in question, is that Kepler alone had a superior methodology to all of his contemporaries in the ongoing process of hypothesis generation. Which process of discovery Kepler sought to make intelligible for posterity in his wonderful opus called Harmonices Mundi.

  8. Pingback: Whewell’s Gazette: Vol. #39 | Whewell's Ghost

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s