The Science of Astronomy

Astronomy comes from the Greek words for "star" and "law". As a science, it focuses primarily on the study of space, its objects, and phenomena outside of the Earth itself.

The Development of Astronomy

Astronomy is considered one of the oldest sciences. The study of the stars and plants began more or less concurrently with the oldest of human civilizations.

At the beginning, astronomy was based on observational evidence taken solely with the naked eye. Many of the world’s oldest calendars are based on these observations and were used to regulate agricultural cycles.

Astronomy and Astrology

For most of their history, up to the 17th century, astronomy and astrology were completely intertwined: for example, astronomers devoted much time and effort to compiling tables of planetary positions, not only for navigation and time-keeping but also to aid in the casting of horoscopes.

Nowadays, almost everyone would agree that astronomy is a science but astrology is not. Why is this? What do we mean by the statement "astronomy is a science" anyway? Are there any parts of astronomy which are not science (and, conversely, any parts of astrology which are science)?

The Scientific Method

If pushed, most scientists would say that science is a process (a way of working and thinking) and not simply a body of knowledge – though the acquisition of a pre-existing body of knowledge is usually necessary in order to participate in the process of science. This is not well reflected in undergraduate courses, which tend to focus exclusively on delivering the said pre-existing body of knowledge as efficiently as possible. Most scientists only learn to do science in postgraduate courses or science-based employment. The process of doing science is commonly referred to as the scientific method.

According to this picture of science, a discipline which claims to be a science can be assessed based on these key features:

This seems like an entirely reasonable approach, and it can be successfully applied to many pseudosciences. For example, astrology makes many testable predictions, so it passes the first test. When carefully tested under controlled conditions, the predictions are repeatedly found to fail. However, practitioners of astrology continue to use the same theories despite the fact that their predictions are not successful. Therefore, astrology is not scientific, because it fails the second test.

There are also philosophical issues with this model, particularly with the "key features" listed above.

Another issue which concerns philosophers of science (but rarely worries working scientists) is the status of theory. Early philosophers of science (particularly the "logical positivists") felt strongly that the basic content of science was observational results (obtained "via the senses"), and that theories are not "real" but are patterns imposed on the data by scientists. Most scientists, in contrast, feel that theories reflect a basic structure or framework which is present in the "real world" even though it cannot be directly observed. In extreme cases, "anti-realist" philosophers deny the reality of entities (not merely theories) which are not directly observed – for example, electrons (this is the point at which most working scientists start shaking their heads and muttering).

Astronomy and Philosophy of Science

The history of astronomy is an interesting case study for philosophers of science because it includes many episodes which touch directly on issues of this kind. Indeed, one of the main schools of philosophy of science (Kuhn’s theory of scientific revolutions) has its roots in a study of the "Copernican revolution" in the late 16th-early 17th centuries.

Popper and falsification

Karl Popper (1902-94) was one of the most influential 20th-century philosophers of science. His basic ideas can be summarised as follows:

Kuhn and scientific revolutions

Thomas Kuhn (1922-96), who trained as a physicist, was a historian of science as well as a philosopher of science, and therefore was well placed to recognise that real scientists do not appear to work in the way that Popper thinks they should. His description of the process of science described in his main article The Structure of Scientific Revolutions (1962), contradicts all science concepts up to that time.

Kuhn has been enormously influential, as can be seen from the extent to which the phrase "paradigm shift" has entered common usage. His picture does have merit, in that the idea of "normal science" seems to be a better match to reality than Popper's view of good science as only attempts to falsify hypotheses. However, it seems that the overall idea of single incommensurable paradigms, succeeding each other by means of crisis and revolution, is oversimplified and misleading.

In contrast to Popper, Kuhn is not enormously helpful in distinguishing science from pseudoscience: his writings indicate that he did see science as uniquely effective at understanding the world, precisely because of the complex interplay between normal science (which makes steady but gradual progress by ignoring anomalies) and revolution (which makes discontinuous progress by replacing a failed paradigm with an improved one; in some of his writing he defines "improved" in terms of "better problem-solving ability").

Experiment and Observation

One of the main issues which divides philosophers of science (and distances them from working scientists) is the status of experimental or observational knowledge. Early philosophers of science, particularly the logical positivists, regarded experimental knowledge as "pure", unassailable, and fundamental in a way that the theories built on that knowledge were not. However, this view is not borne out by history: observations are fallible and revisable in the light of new knowledge.

Some modern philosophers of science, especially those better described as sociologists of science, leap from this dependence of interpretation on theory to the conclusion that experimental results as published are entirely dependent on the experimenter’s theoretical framework, and therefore that changes is scientific theory are to be accounted for by changes in the social background rather than the effects of experiment. All working scientists will of course recognise that this view is not tenable, not least because experiments and observations have a nasty habit of not fulfilling the experimenter’s preconceived expectations!

Experimental results are therefore influenced by both the external world (with whose laws they must obviously be consistent) and the experimenter's theoretical framework (which will determine how the experiment is designed, which variables it is sensitive to and to what precision, and how both positive and negative results are interpreted). Although scientists instinctively feel that carefully controlled and well designed experiments are scientifically preferable to observations of naturally occurring phenomena, it can be argued that the latter are actually less subject to bias introduced by the prevailing theoretical framework: experiments are specifically designed not to be sensitive to effects other that the particular one being investigated, and it is therefore difficult to spot the truly unexpected.

The history of physics is littered with people who failed to make important discoveries which, in hindsight, are clearly visible in their experimental results. Observational campaigns, on the other hand, are often designed around what one can observe, rather than any more theoretical motivation, and any unexpected results are more likely to be noticed. This does not mean that observations are entirely independent of the theoretical framework; the framework will still influence which instruments are built, which objects are singled out for observation, and how the results are interpreted.

Astronomy as a Science

Astronomy has provided philosophers of science with two key case studies: the Copernican revolution of the late 16th and early 17th centuries, and the replacement of Newtonian by Einsteinian mechanics in the early 20th century. These have been extensively used as test-beds for the ideas of various philosophical schools, with varying degrees of success. Philosophers of science therefore have no doubt that astronomy is a science – otherwise it would not make sense to use its history in this way. However, some aspects of astronomy, particularly the more modern variants of astrophysics and cosmology, do present problems.

The difference between astronomy and the "ideal" definition of a science – perhaps physics – is that astronomy is very rarely an experimental science. Most of our knowledge about astronomical objects is obtained indirectly, through techniques such as spectroscopy, analysis of binary star orbits, redshift surveys, etc. This raises the question of realism and anti-realism that we mentioned above: to what extent are these indirectly determined quantities "real"? Working astronomers of course are entirely confident that they are, but in some cases this confidence requires justification. Even if we accept them as real, they are clearly "theory-laden" – i.e. theoretical assumptions are required in order to convert the observations into physically relevant quantities.

Even if we accept them as real, they are clearly "theory-laden" – i.e. theoretical assumptions are required in order to convert the observations into physically relevant quantities. A current example of this is the interpretation of galactic rotation curves: everyone agrees on the data (curves of velocity against distance from the centre of the galaxy), but, while the majority of astronomers interpret these curves as demonstrating the presence of large quantities of dark matter, a minority prefer to see them as demonstrating a deviation from Newtonian gravity at small gravitational accelerations (Modified Newtonian Dynamics, or MOND). The two camps therefore deduce entirely different physical properties from the same empirical data. In physics, this sort of ambiguity can usually be resolved by designing an appropriate experiment; in astronomy, however, it is not in general possible to do this (the threshold gravitational acceleration relevant to MOND is far smaller than anything accessible in the solar system).

How do working scientists interpret the scientific method as applied to an observational science such as astronomy? (Other observational sciences include palaeontology and geology).

The key feature of a scientific theory, according to both the naïve scientific method and Popper, is that it should make testable predictions. The definition of a "prediction" is obviously that it refers to a measurement which has not yet been made – explanations of existing unexplained effects are less highly regarded, because the framer(s) of the theory knew about these beforehand and may have "tuned" the theory to explain them. This does not, of course, mean that explanations of existing results have no value at all: if they are quantitatively precise and come "naturally" out of the theory with no obvious evidence of fine-tuning, they can be very persuasive.

Historically, observation has often outstripped theory in astronomy, and therefore the "natural explanation" class of support has more influence in astronomy than it does in most experimental sciences. Occasionally theories come first, and wait for a long time before they are experimentally tested. The classic picture of new theory => prediction => test is actually fairly uncommon in astronomy (Eddington's test of Einstein's prediction of gravitational bending of starlight is perhaps the most widely known example). This is a clear difference between physics, for example, and astronomy, and accounts for a good deal of the notoriously inconsistent and illogical nomenclature of astronomy.

Despite these differences in methodology imposed by the nature of an observational science, astronomy clearly does satisfy both Popper's and the "scientific method" criteria for science: it consists of a body of observational data interpreted and explained by a framework of theoretical ideas, all of which have been or are intended to be tested by observation - either by providing a natural, quantitative explanation of a previously unexplained body of data or by making predictions about the results of new observations. Theories which make predictions which are not borne out by observation are abandoned, e.g. the Steady State cosmology – albeit, as Kuhn would expect though Popper would not, rather more slowly and hesitantly than seems appropriate in hindsight. New theories generally have greater explanatory power than those that they replace: in Popper's terms, they are bolder; in Kuhn's, they have greater problem-solving potential. Both of these features are characteristic of science, and both help us to differentiate between astronomy and astrology: astrologers do not discard (even reluctantly!) theories which have been shown to make inaccurate predictions, and consequently astrological theory has not made the progress that we see in astronomy.

Philosophy of science is a large and complex field.