What are the future prospects in astrophysics


Christmas is approaching, the year is drawing to a close, and whether you like it or not you will witness a general urge to transform the past and future into a ...

Christmas is approaching, the year is drawing to a close, and whether you like it or not you will witness a general urge to put the past and the future in order. Everywhere, annual reviews are being created and the highlights of 2012 are being collected, at the same time the horoscopes in the magazines are getting longer and more categorized, resolutions and plans are being forged for 2013, and one would love to know what the new year will bring, in concrete terms.

This desire to be able to look into the future is of course not new and by no means limited to the time of the turn of the year. Even scientists who do not recognize horoscopes or other hocus-pocus as a source of knowledge have an interest in being able to foresee the development of their subject, if not out of curiosity, then at least in order to be able to plan for the long term. Now, of course, we know that such predictions are anything but easy. Historians know a thing or two about how often science has taken surprising turns in the last few centuries when scientists were particularly convinced that they knew the future. For example, the career advice given to Max Planck in Munich in 1874 is legendary, in the course of which physics professor Philipp von Jolly advised him against studying physics because there was nothing new to discover in the field of physics. As is well known, Planck nevertheless became a physicist and, through the discovery of energy quantization, impressively showed that Jolly had been completely wrong with his assessment. Philosophers like Thomas S. Kuhn have used such historical lessons as an opportunity to emphasize the discontinuous nature of science. Science can therefore not be understood as a continuous more and more knowledge (which would have a certain predictability), but one must always reckon with breaks that throw everything that was assumed to be correct up to now.

Illustration: A look into the future of the astronomical future.

Despite all these cautionary signs, there are still physicists who are ready to make prognoses about the longer-term future of their subject. For example, the astrophysicist Martin Harwit published in 1981 in his book "The Discovery of the Cosmos - History and Future of Astronomical Research" (Nature Letter 1984) an astonishingly concrete study of the question of where we are with our astronomical research, ie how much we already know , and how much is still undiscovered in the universe.

Harwit's theses are briefly: 1) We know how many astronomical observations are in principle possible, and therefore we can estimate how many of them we have realized. 2) From the history of the discovery of the cosmic phenomena known to us, we can estimate how many cosmic phenomena there are in total. 3) If we can transfer the course of the frequency of discovery from fields such as geography or zoology to astronomy, then by the year 2200 we will already know 90% of all cosmic phenomena.

Illustration: Hubble eXtreme Deep Field: How much undiscovered can there be in the cosmos? (Credit: NASA; ESA; G. Illingworth, D. Magee, and P. Oesch, University of California, Santa Cruz; R. Bouwens, Leiden University; and the HUDF09 Team)

Such daring theses naturally arouse curiosity about their scientific justification. An important starting point for Harwit's consideration is the fact that astronomy is largely an observational science. The observed system sends out information by itself without the observer being able to manipulate it directly. According to Harwit, the observer only has one fundamental decision to make: he tries to understand the signals received from heaven, or he ignores them. Astrophysics thus has an apparently simpler structure than the experimental sciences, which, in the enormous variety of possible experiments, are more complex with regard to the empirical information that is potentially available.

The information available to the observing astronomers therefore seems comparatively clear. There are a finite number of information carriers that can reach us from the universe. Harwit puts the number of their types at five: electromagnetic radiation, cosmic particle radiation, meteors and meteorites, neutrinos and antineutrinos, and gravitational waves. We can evaluate these sources of information on the basis of observations; they are our only possibility of empirically learning something about the phenomena of the universe. A concrete, elementary astronomical observation is then defined by the choice of one of these information carriers, together with a number of other parameters that are partly set by the observer and partly specified by the concrete technology. These include the wavelength or energy of the carrier, the spatial resolution of the observation instrument, the temporal resolution, i.e. the minimum distance between two pulses that can still be detected, or the spectral resolution, i.e. the minimum distance between different wavelengths that is resolved. The crucial point is that all parameters that define an observation each have only a finite range of possible values. This is due, for example, to restrictions due to Heisenberg's uncertainty principle, the nature of the information carriers or the non-reducibility of the observed complexity. In other words: according to Harwitt, the total number of all possible astronomical observations in a given period is finite. Assuming the universe is homogeneous and isotropic, i.e. uniform everywhere and in all directions, this statement also applies regardless of the choice of a given point in time and place.


Figure: The permeability of the atmosphere to electromagnetic waves: the information that can reach us from the universe is subject to a number of restrictions.

If one estimates the finite space of all possible observations in such a concrete way, Harwit can then determine that at the time of his investigation about 5 percent of all electromagnetic observation possibilities were realized. Taken together with the four other information carriers in addition to electromagnetic radiation, he estimates the status of the observation methods developed to be 1 percent of all possible methods. Harwit's next question is what else we can conclude about the completeness of our astrophysical knowledge. How many of the existing phenomena in the universe have we already discovered and how many are still waiting for us?

To do this, Harwit must first objectively define what is to be viewed as an independent phenomenon. According to its definition, two phenomena are to be distinguished if at least one of their observational features differs by a factor of 1000. For example, open star clusters comprise between 100 and 1000 stars, while globular clusters are composed of 100,000 to 1,000,000 stars. According to the definition, both represent two different phenomena. In this way of counting Harwit comes to a number of around 43 known cosmic phenomena. Some of these phenomena were not discovered just once, but in different, completely independent ways, for example on the one hand with optical wavelengths, on the other hand in the radio range.

Harwit uses this fact of the existence of independent rediscoveries for the final, statistical step of his argument. Imagine collecting football stickers. As long as the collection is still small, the present stickers differ, but from a certain size of the collection you get more and more stickers twice. If one comes across duplicates, one can conclude, according to Harwit, that the set of all printed soccer motifs is finite, otherwise the occurrence of duplicates would be improbable. At the same time, the fewer the total number of different stickers, the earlier the number of duplicates compared to the stickers that are only available once. If the stickers only cover the German national team, you will have to deal with duplications earlier than if all international players in the World Cup are shown. Harwit transfers this observation to the case of cosmic phenomena by converting it into a statistically motivated formula that determines the total number of phenomena in the universe from the number of singly discovered and the number of repeatedly discovered phenomena. Of the 43 phenomena he identified, there are 7 rediscoveries. According to his statistical analysis, Harwit deduces from this the existence of 123 cosmic phenomena, of which about a third was known in 1981. The total number could increase to around 500 if one takes into account that there could be phenomena that in principle only show up in a single information channel (in the analogy: sticker motifs that were only printed once).

The field of cosmic phenomena appears for Harwit 1981 in relation to astronomical research like the earth in relation to geography: the latter experienced the time of great discoveries in the 15th and 16th centuries, while today practically everything has been discovered. The frequency of discovery in such scientific fields with a finite number of phenomena to be discovered is described by Harwit using a bell curve: initially the number of discoveries rises steeply as the field arouses great interest and new ideas and instruments are developed. Ultimately, the number of phenomena that have not yet been discovered decreases so much that the frequency of detection decreases accordingly and general interest decreases. If such a course can be transferred to astronomy, according to Harwit, around 90 percent of all phenomena should have been found in the year 2200.

Figure: Will astronomers and astrophysicists one day have a similarly complete picture of the cosmos as geographers do today of our earth?

Harwit developed these considerations 30 years ago. If you hold back at first with fundamental criticism of your arguments, it is therefore natural to ask what has happened in astronomy since then. Have Harwits predictions been confirmed? In 2001, Virginia Trimble and Markus J. Aschwanden took stock and found that fewer new phenomena were discovered than Harwit had hoped, but without substantiating this observation in detail. In 2012, the problem was taken up again by astrophysicist Steven J. Dick, who today counts 82 classes of phenomena, which would confirm that fewer discoveries were made than predicted. The details of this count are expected to appear in Dick's book, Discovery and Classification in Astronomy, next year. Harwit himself recently confirmed on request that he still considers his estimate to be valid. In the meantime, not only have new phenomena been discovered, but old ones have also disappeared or merged.

The idea of ​​a finite space for possible observations appears elsewhere in astronomical research. The “Virtual observatory” as an online collection of astronomical data from various observatories has as one goal the systematic development of precisely that observable parameter space, in particular to enable the discovery of new astronomical objects and phenomena. The astrophysicist S. George Djorgowski and colleagues explicitly named Harwit's considerations as a historical reference in their 2001 description of the Virtual Observatory.

Figure: The Virtual Observatory - online collection of astronomical data. Screenshot https://www.virtualobservatory.org/whatis/

So it cannot be said that Harwit's provocative-sounding speculations about the future have been dismissed as crazy by the astronomical community. If you leave aside the perhaps somewhat shaky-looking statistics of the phenomenon estimation, at least the existence of a finite space of possible astronomical observations in the physical world view sounds convincing. This space would define the limits of our empirical astronomical knowledge. With this assumption, however, it is only a small step to the continuous “bucket theory of knowledge”, the idea that we simply know more and more over time. So are the philosophers wrong with their scientific progress skepticism in the case of astrophysics?

In fact, from a philosophical point of view, there are probably a few points that should be questioned more precisely in Harwit's argument. First of all, it is not all that clear that the difference between observations and experiments is actually that fundamental. After all, the idea that an observation is a more or less passive recording of a signal characterized by a small number of parameters, which then only needs to be interpreted, is a fallacy. Rather, the recorded raw data is only the first step in a long chain of data processing, data analysis and modeling processes. Even if all possible astronomical observations had been carried out, a new space of possibilities would open up as to what exactly to do with the data. At this point there is room for statistically difficult to take into account cultural-sociological, historically variable influences that make up what is called “current research practice” at a certain time.

Confusing data and phenomena is also dangerous, a distinction made in 1988 by the philosophers James Bogen and James Woodward. Science makes statements about certain phenomena, but we do not observe these phenomena directly; we first have to derive them from the available data. And that doesn't just work by dividing data packets into scale segments with a factor of 1000 and calling the result a phenomenon. If one takes seriously the distinction between data and phenomena, Harwit's counting statistic appears to be an oversimplification of a much more complex interrelationship between what we measure and what we learn from it about the universe.

Harwit's own interest has expanded in a similar direction in recent years, increasingly devoting himself to the cultural influences on science. In 2011, for example, following the philosopher Andrew Pickering, he asked: "Are physicists and astronomers just constructing appropriate representations that have little to do with the inherent structure of the universe in which we live?" Just as Pickering can see the Standard Model of particle physics as a mere preliminary theory that will probably need to be revised at some point, in the case of astrophysics, according to Harwit, one could speculate that at some point there will be a theory in which dark matter and dark energy turn out to be natural consequences result: "Such a description may then be just as insanely different compared to what we understand today, as Einstein's postulate was when he first announced that the speed of light always appears the same (...)." Against this background, it seems questionable whether, with such a description, when counting cosmic phenomena, we would arrive at the number 82 obtained by Dick.

Figure: The planned Laser Interferometer Space Antenna (LISA) for measuring gravitational waves - future-oriented technologies as a factor that determines progress, source: NASA

It is probably no coincidence that Harwit chooses an example from cosmology for his example of a foreseeable Kuhnian revolution, because the science of the universe as a whole resembles within the astrophysical disciplines what can be described as theory-driven science could. The idea of ​​abrupt breaks in scientific understanding for those branches of science whose progress is primarily connected with the development of new theories can be thought of as analogous to the historical examples that Kuhn and others have evaluated from an epistemological perspective. Harwit's 1981 consideration, on the other hand, was based on the primarily technology-driven areas of astrophysics. Such a more experimentally focused view of scientific research can be found in the more recent epistemological debate since the 1980s, associated with names such as Ian Hacking, Allan Franklin or Peter Galison. If one focuses on the technological development of scientific research, sudden discontinuous breaks seem far less natural. It seems difficult to imagine that 50 years from now we will be able to recognize that our radio telescopes today could have essentially provided incorrect information, even if our handling of the telescopes and the data generated may change significantly over time. In this perspective, the willingness to invest in future-oriented technologies appears to be the main factor that determines progress.However, it is important to see that both the theory and the technology side together make up what we understand by science today, and both sides strongly influence one another (see also Freeman J. Dyson in “Science”). Hence, both the assumption that science is predictable and the notion of the utter unpredictability of scientific research seem to be inappropriate extremes.

The bottom line is that looking into the future in detail is not as easy as one would sometimes wish. But predictions of the future, apart from their actual forecasting performance, have a very subjective additional benefit that is known from the annual horoscopes of the magazines: the pleasant shudder at the thought game that it could really come that way - only surpassed by the fun of arguing why of course it will turn out very differently.

Illustration: The view into the crystal ball "The Crystal Ball" 1902, John William Waterhouse



Dear readers, if you want to make us happy at the end of the year, then vote for Planckton as Science Blog of the Year 2012 at https://wp.me/p1XAlm-jB. We were nominated there as one of 20 blogs. Many Thanks! Sibylle Anderl



Keywords: astronomy, astrophysics, philosophy, physics, scientific theory
To press
Send post by email

Critique of Pure Physics (7): The Prophecy of Astronomy

From Sibylle Anderl

An error has occurred. Please check your entries.

Send post by email

Many Thanks
The post was sent successfully.