Useful as the tripartite structure of detection, interpretation, and understanding is in appreciating the extended structure of discovery, we are far from showing that this is the only structure of discovery in astronomy, much less in science. In fact, we should be skeptical of such a simple view that a single structure is universal to all types of discovery. After all, we have seen that some classes (gas giant planets, giant and dwarf stars) have been inferred over very long periods rather than detected, while others (dwarf planets) have been suddenly declared. And as we saw in the last chapter, the discovery of new classes of astronomical objects is only one of many categories of discovery that historians and philosophers of science, ranging from Norwood Russell Hanson to Theodore Arabatzis, have distinguished.

Scientists themselves have cautioned against simplifying what they as practitioners recognize as a complex process. Referring to Thomas Kuhn’s attempts to find structure in the strikingly numerous discoveries from 1895 to 1905, including the contributions of Roentgen, Bohr, Planck, Einstein, Becquerel, Thomson, and Rutherford, the theoretical physicist Abraham Pais wrote that “Such facile and sweeping generalizations only serve to create serious pitfalls of simplicity. I do believe that one can discern general themes in the history of discovery in science and I shall even venture to mention some, but mainly for the purpose of emphasizing variety over uniformity. After decades spent in the midst of the fray I am more than ever convinced, however, that a search for all-embracing principles of discovery makes about as much sense as looking for the crystal structure of muddied waters.” In short, Pais confessed to “very strong reservations concerning the usefulness, let alone the necessity, of a search for general patterns or laws of history, specifically the history of discovery.”

With the exception of Tycho Brahe’s proof in 1577 that comets were celestial phenomena based on their parallax and thus distance, and his inference that the stella nova of 1572 was celestial based on its lack of parallax, the problem of the discovery and interpretation of new classes of astronomical objects begins substantially with Galileo and the telescopic era 400 years ago. Galileo’s telescopic observations revealed what we would today recognize as two new classes of astronomical objects: moons around Jupiter and rings around Saturn. And while Galileo and his contemporaries soon realized the nature of the moons of Jupiter by analogy to our own Moon, the story of Saturn’s rings is much more complicated.

As we shall see, both Jupiter’s moons and Saturn’s rings stand as early examples of what would become commonplace in astronomy: that “seeing” isn’t always “knowing,” that “detection” does not constitute “discovery,” that recognizing a new class of astronomical objects can be a difficult and multifaceted endeavor. In this chapter we begin to dissect the process of discovery in astronomy, in particular as it applies to the discovery of new classes of objects. We shall find it to be a complex and extended series of events consisting most often of at least three components: detection, interpretation, and understanding. This is particularly true of new classes of objects when the observer may have no idea of the true nature of the object. Again and again astronomers ran up against the unexpected in their reconnaissance of the heavens. Their struggle to move beyond mere detection, to enter the difficult realm of interpretation, and to seek physical understanding – often long after the original detection – is a story that has only been told piecemeal, but that deserves systematic treatment because it represents the core of astronomy and the natural history of the heavens.

The title of this volume will bring to mind for some readers Martin Harwit’s book Cosmic Discovery: The Search, Scope, and Heritage of Astronomy, published more than three decades ago. As an astronomer, Harwit’s practical aims in that book were twofold: first, to determine what fields of astronomy have the largest number of potential discoveries and, second, to explore what fields of astronomy might promise the most immediate advances and striking returns. By analyzing just enough history to answer these questions, Harwit hoped to provide answers of use both to students entering the field facing the daunting question of what part of astronomy to study and to policymakers who had only limited resources to dole out for the advancement of astronomy. In this effort at applied history he had only mixed success, but by taking the concept of discovery as a serious object of analysis, Harwit did what only very few astronomers, historians, or philosophers have done before or since. While it will become clear that I differ with Harwit on many issues, not least the crucial question of what constitutes a new “class” of object, I record here my debt to him, and others, for their boldness in attacking a concept so broad as “discovery.”

Unlike the realms of the planets and stars, long distinguished by the “wanderers” moving among the fixed stars, the realm of the galaxies had to be discovered. Curiously, however, some of its members had been detected long before they were known to be outside our stellar system. As we saw in Chapter 3, a few objects such as Andromeda and the Magellanic Clouds had long been seen with the naked eye, and in the eighteenth century Charles Messier and William Herschel detected numerous fuzzy objects catalogued as “nebulae.” But these nebulae were largely believed to be in the realm of the stars, as indeed many of them were. How it was determined that some comprised separate systems of stars far beyond our own is a storied part of the history of astronomy, approached in this chapter through the lens of discovery and its complexities. Despite prescient early guesses and some more scientific inferences, it was only Edwin Hubble in the early twentieth century who provided definitive proof that “extragalactic” objects existed beyond our own Milky Way Galaxy.

In contrast to the discovery of new laws, processes, or properties, one of the hallmarks of the discovery of localizable natural objects such as we have been discussing in this volume is an almost irresistible temptation to classify them. A new object will barely have been discovered before the human mind tries to determine where it “fits” in the order of things already known. This is undoubtedly an ancient instinct, and one of the basic characteristics of a developing and even a mature science, as basic to astronomy as it is to biology, geology, chemistry, and physics. To put it another way, the recognition of new classes and their classification is an integral part of natural history, and a particularly challenging activity when the discovery involves a new class of objects. The origin of the categories in which we think is a question rarely raised except in the rarefied reaches of philosophy, but it goes to the core of discovery. In discovery it may flare briefly and sporadically, as in the case of Pluto, or at more sustained length as in the case of stellar luminosity types. But once routine, the origin of the categories is often forgotten, even as they determine the way we think of things.

Throughout this volume we have witnessed many narratives of discovery, each groundbreaking in its own way. But one of the greatest achievements of the twentieth century was synthesizing these discoveries into one great master narrative, itself constituting a great discovery. That master narrative is cosmic evolution, the story of the universe from its beginning with the Big Bang 13.7 billion years ago stretching down to the present moment when evolution still continues. The discovery of cosmic evolution was foreshadowed by the luxuriant gardens of William Herschel, who recognized already in the late eighteenth century that individual celestial objects at varying stages of development allow us to “extend the range of our experience to an immense duration.”

As historian David DeVorkin has written, by the early twentieth century, American astrophysicists were becoming world leaders in astronomical natural history thanks to the unparalleled power of astrophysics to reveal the nature of celestial bodies. “Akin to the naturalist, the typical American professional astronomer was collector and classifier. Instead of museum shelves and cases, astronomers stored their systematic observations in plate vaults and letterpress log books, and displayed them in catalogues sponsored by universities and observatories.” In particular, as more and more stellar spectra were gathered, they hinted at numerous variations in the nature of the stars. Once stellar physics was understood later in the twentieth century the reasons became clear: stellar structure depended on mass, temperature, and luminosity, and the range of all of these physical variables was enormous. Moreover, stars existed in a variety of different ages, and (it turned out) in a variety of stages. Stars were born, lived, and died; once the concept of stellar evolution was accepted, the problem was determining which stars were in which stages and how the physics worked under varying conditions, among the greatest puzzles in the history of science. In short, William Herschel’s gardens, first explored with his work on the nebulae, were luxuriant beyond his wildest dreams when it came to the stars themselves.

The Natural History of Discovery

Our natural history of discovery in this volume, analogous to the natural history of the heavens that astronomers have undertaken over the last 400 years, has uncovered some surprising conclusions. Focusing on new classes of astronomical objects – defined in our study by those eighty-two objects listed in Appendices 1 and 2 and as discussed in Chapter 8 – we have found that the discovery of virtually every class of object displays an extended structure. Thomas Kuhn and others foreshadowed such a structure already fifty years ago, but unlike Kuhn’s anatomy of discovery, which involved anomaly, paradigms, a distinction between normal and revolutionary science, and a subordination of discovery to the concept of revolutions, we place discovery at center stage in science. We find a macrostructure consisting most often of detection, interpretation, and understanding, and a microstructure involving not only the nuanced problems of those three stages and their conceptual content, but also social and technical elements.

If discovery in the realm of the planets was difficult, in the realm of the stars the problems were only multiplied. While Pluto, representing the outer edge of the solar system as it was known in 1930, was about forty times the distance of the Earth from the Sun (40 astronomical units, or 4 billion miles), the nearest star was 25 trillion miles, a factor of nearly 7000 times more distant. Put another way, rather than light minutes or light hours for the planets, the nearest stellar distances were measured in light-years, and the more distant ones in hundreds or thousands of light-years. Whereas a light beam from Pluto would take a few hours to travel to Earth, the nearest starlight (other than our own Sun!) would take 4.2 years. In terms of distance, astronomers studying the stars clearly entered a new realm in discovery and interpretation when it came to techniques of observation and methods of inference.

Astronomers at the time, however, knew little about the distances they were dealing with. The fixed stars of time immemorial notwithstanding, so-called sidereal astronomy did not become an important part of astronomy until the eighteenth century, when observers such as the French astronomers Nicholas-Louis de la Caille and Charles Messier focused on objects in the realm of the stars, and especially when William Herschel began using his newly constructed large telescopes in England. The Copernican theory implied vast stellar distances, but those distances remained unknown until 1838 with the first determinations of stellar parallax, and even then only the distances to the nearest stars could be measured. A few years earlier Auguste Comte famously highlighted the composition of the stars as the very prototype of a problem that could never be solved. But the development of spectroscopy in the 1850s and 1860s enabled just that, a stunning feat in the annals of human thought that unlocked the secrets of the sidereal realm – but only slowly and grudgingly.

On August 24, 2006, the International Astronomical Union (IAU) – the only institution that counts when it comes to official designations of astronomical bodies – declared that Pluto was not a planet. More specifically, astronomers demoted Pluto from a planet to a dwarf planet, and (to the chagrin of many scientists and the confusion of the general public) declared that a dwarf planet was not a planet all, thus reducing the number of classical planets in the solar system to eight for the first time since 1930 when Pluto was discovered. Pluto’s demotion not only meant a rewriting of the textbooks, but also set off a surprisingly intense scientific and public outcry – an interesting cultural phenomenon indicating not only the importance of classification to scientists, but also a deeper investment in astronomy among the general public than one might have thought.

As a longtime member of the IAU, I was among those voting on that fateful day in Prague. Although I had attended every triennial IAU General Assembly since 1988, many of them as an officer in its History of Astronomy Commission, I had never seen the meeting dominated by a single issue as it was on this occasion. Though literally hundreds of sessions were held over the two weeks of the meeting, discussing a broad panoply of astronomical subjects, and though numerous other resolutions were considered and passed at this General Assembly, the resolutions involving Pluto were the center of attention, the subject of numerous sessions, and the topic of the buzz in the hallways. As was tradition, the resolutions were voted on during the last day of the General Assembly, after much discussion the previous two weeks, leaving only 424 delegates to vote out of the thousands who had attended.

As explained in Section 8.2, Appendix 1 constitutes an exercise in constructing a comprehensive classification system for astronomy. The “Three Kingdom” system begins with the three Kingdoms of planets, stars, and galaxies – the three canonical divisions adopted in textbooks for almost a century, since it became clear that galaxies were indeed a separate “realm of the nebulae,” as Hubble put it. For each Kingdom six astronomical Families are then delineated, based on the object’s origin (Proto-), location (Circum- and Inter-), subsidiary status (Sub-), and tendency to form systems (Systems), in addition to the “Central” Family (planet, star, or galaxy) with respect to which the other Families are defined. These considerations give rise to astronomy’s eighteen Families, and the symmetry of the six Families of each Kingdom reflects their physical basis in gravity’s action in all three Kingdoms. The system then distinguishes eighty-two classes of objects, a large subset of which are the subject of this volume.

Like biology, the Three Kingdom (3K) system is hierarchical, extending from Kingdom to Family to Class, with the possible extension to further categories lower in the hierarchy such as Type and Subtype. As in biological classification it occasionally adds an intermediate Subfamily level wherever useful. With the benefit of hindsight (and with utility in mind), the system incorporates some classes as they have historically been defined, and adds others as they might be defined in a coherent and consistent system. The criterion adopted for class status is, wherever possible, the physical nature of the object, rather than orbital, dynamical, temperature, morphological, spectral, or any other characteristics. Some of the principles of classification are discussed in Section 8.2.