The discovery of the CMB in 1965 by Arno Penzias and Robert Wilson became a turning point in the twentieth-century cosmology after it was interpreted by Robert H. Dicke and his coauthors. Penzias and Wilson were not looking for the CMB, which was, in fact, first predicted by George Gamow (Reference Gamow1946) and his collaborators (Alpher, Bethe, & Gamow, Reference Alpher, Bethe and Gamow1948; Alpher & Herman, Reference Alpher and Herman1948a, Reference Alpher and Herman1948b, Reference Alpher and Herman1949) almost two decades earlier. Penzias and Wilson initially investigated the signal as noise, a possible artifact of their antenna or surroundings, and were completely unaware of its relevance to cosmology. The signal turned out to be a milestone in the development of physical cosmology. It ended the period of heated cosmological controversy (Kragh, Reference Kragh1996) and started a period of increasing convergence on the Hot Big Bang cosmology (Peebles, Page, & Partridge, Reference Peebles, Page and Partridge2009). It quickly became perceived as a standard but also initiated an interpretational race.

Gamow and his students (collaborators Robert Herman and Ralph Alpher) made an early prediction of the CMB temperature. Although their prediction was substantially off in terms of the value of its temperature, and turned out to be a calculating conundrum (Peebles, Reference Peebles2014, 210–211), it finally resulted in a value of $5\text{ K}$ (Alpher, Bethe, & Gamow, Reference Alpher, Bethe and Gamow1948). One the one hand, these papers pioneered the use of computing numerical methods in cosmology. On the other hand, it appears Gamow himself believed the radiation would be undetectable (Sharov & Novikov, Reference Sharov and Novikov1993, 146).

It is easy to forget that these developments we see as a milestone today were the growing pains of cosmology, not a sudden revelation obvious to the entire physics community. Moreover, as Peebles noted:

The attempt was, to a great extent, a leap of faith.

Independent predictions by Andrei Doroshkevich and Igor D. Novikov (Reference Doroshkevich and Novikov1964) and Dicke et al. (Reference Dicke, Peebles, Roll and Wilkinson1965) were the result of much more thoroughly worked out models. In their 1965 seminal paper, Dicke and his collaborators suggested an interpretation of the CMB as a remnant of the primordial fireball, and this became the standard after Penzias and Wilson’s discovery (Dicke et al., Reference Dicke, Peebles, Roll and Wilkinson1965). This interpretation postulated that in the early universe, baryonic matter and radiation were in balance. Baryonic matter is composed of protons and neutrons and also includes atomic nuclei; but in the very early stage (the first $200\text{ s}$), it was only composed of free protons and neutrons. Thus, radiation consisted of continuously moving and scattering photons. This sort of balance, or equilibrium, of matter and radiation is similar to the one in human-made gas-discharge lamps; photons continuously “run around” and scatter off the matter. After the first light nuclei were formed about $200\text{ s}$ after the Big Bang, the universe was a hot, dense plasma composed of photons, electrons, and light ions, mainly protons and 4He nuclei. Now, the plasma was opaque to electromagnetic radiation – quite similar to fog – because of the strong Thomson scattering of light-photons by free electrons. Electrons were not yet tied to atomic structures but “floated” freely around. The mean free path each photon could travel before encountering an electron and scattering off it was very short, hence the fog-like opaqueness of the universe. Finally, a high ratio of photons compared to baryons ($~109$ times more of the former), starting with the early universe and extending to today, makes the Big Bang hot, that is, overwhelmingly heat- rather than matter-dominated (Figure 5.3).

As the universe was expanding and thus cooling, the formation of neutral hydrogen quickly became energetically favored. The fraction of free electrons and protons compared to neutral hydrogen decreased fairly rapidly, and the process was over about 400,000 years after the Big Bang. This process is labeled recombination, although electrons and ions met in the cosmological context for the very first time. As in other scientific fields, many key labels in astronomy are misnomers; the recombination was really the initial combination. In a way, such labels are devised to capture a particular view of a state of affairs. The term remains even though eventually the view changes. A more famous case of this sort that froze the key historical turning point in modern cosmology was the use of the label Big Bang. It was uttered as a joke by Fred Hoyle on a BBC show. As the key proponent of the steady-state universe, Hoyle devised a derogatory term that was meant to point out what he thought was a crucial weakness, even absurdity, of Georges Lamaître’s expanding universe model that had a singular beginning in the form of sudden expansion. Yet over time, as expanding universe models became much more favored than static universe models, the term acquired an appeal because of the plastic description of the key point of the models, that is, the expansion from a singular moment in space and time.

Somewhat earlier, a process similar to recombination involving helium and an intermediate state of singly ionized helium $\text{He}þ$ started; in this process, interactions of photons with neutral atoms replaced the Thomson scattering of photons by free electrons. The substitution is equivalent to switching the gas-discharge lamp off. The former interactions are several orders of magnitude less energetic than the latter, so interactions quickly subsided; photons effectively decoupled from (baryonic) matter and freely “flew” away. Their mean free path became comparable to Hubble length, or roughly the size of our cosmological domain, and the universe became transparent to light. This occurred about 400,000 years after the Big Bang. Photons from this epoch, or those incoming from the surface of the last scattering, traveled and cooled down freely, without interacting with matter, until some were stopped by a horn-shaped antenna in Holmdel, New Jersey, in 1964.

Given that these cosmological photons predate photons from almost all other sources in the universe, they show up as a background radiation “hum” – they ought to be detected as incoming from all directions in the sky. As the matter in the universe is isotropically distributed on large scales (i.e., equally throughout the cosmic space), many expected the primordial radiation must be distributed the same way. As we have seen, according to the Hot Big Bang model, the background radiation originated in the opaque state of plasma, so it should not be modulated, and it should thus behave as a perfect blackbody radiation (see Figure 5.1). This has been valid up to the order of minuscule anisotropies reflecting the beginnings of the physical structure formation. While such small-scale anisotropies are relevant in current cosmological debates, they were unobservable between the 1965 discovery and the advent of the COBE (Cosmic Background Explorer) satellite measurements in 1990s and, consequently, played no role in the overarching interpretation. As we will see, even the alternative interpretations focusing on the apparent isotropy of the CMB did not address them; they either bet on deviation from the blackbody radiation shape or modified the hot beginning of the Big Bang by devising plausible ways of getting current isotropy from early fluctuations.

It was also immediately clear within this framework that the original temperature of the blackbody radiation during recombination was in thousands of kelvin and should have steadily decreased due to the expansion of the universe to a few, or perhaps a few dozen, kelvin. This parameter was arguably crucial in testing the model, solidifying the interpretation of the background radiation within it, and eliminating the alternatives.

As we mentioned earlier, George Gamow and his students Ralph Alpher and Robert Herman developed a relativistic Hot Big Bang model that also predicted a primordial relic radiation some 20 years before the discovery by Penzias and Wilson. And in fact, right after the publication of Lemaître’s (Reference Lemaître1931) speculative column in Nature on the possibility of the creation of the universe from a single quantum, the physicists realized any such scenario would result in leftover radiation of some sort, and this was commonly understood as a good reason not to accept such models (Gregory, Reference Gregory2005, 38). The account was ignored, and with it, Gamow et al.’s understanding that a very hot initial state must have been opaque to light and that the subsequent recombination would cause photon decoupling and the emergence of a cosmological photon reservoir with adiabatically decreasing temperature. A paper arguing this point was published in 1949 (Gamow, Reference Gamow1949), but these insights dated back to 1946. And Alpher and Herman (Reference Alpher and Herman1948a, Reference Alpher and Herman1948b, Reference Alpher and Herman1949) predicted that the temperature of the background is about $5\text{ K}$. Later on, this prediction became known as the “temperature of the universe.”

It is not surprising that Gamow and Alpher and Herman kept revising the value, as their procedures for determining it were inherently problematic. The two students used the equilibrium Saha approximation, but it included a poorly estimated range, between $5\text{ K}$ and $50\text{ K}$. Gamow used the Jeans stability criterion, which was unfortunately inapplicable in the simple form he used due to the presence of dark matter he did not know about. Generally speaking, all three pictured the transition from the radiation-dominated to the matter-dominated universe as occurring far too late and at far too low an equilibrium temperature, at least based on current insights, due to their assumption of the baryon density of the universe that was too high – it turned out much later that those baryons comprise only about 5% of the total mass-energy budget.

Their model and their prediction were well worked out given the available knowledge at the time, but they were not taken seriously until the emergence of the great controversy forced astronomers to search for observational tests of the models. One key reason why their model did not take off earlier within the community was that the model’s non-equilibrium calculations required a numerical solution of differential equations that could be realistically done only on a computer. Computer calculations were still uncommon in the 1940s and 1950s when they were developing their model. The required computer calculations were performed independently by Dicke and Jim Peebles at Princeton and Doroshkevich and Novikov in Moscow between 1960 and 1964 (Doroshkevich & Novikov, Reference Doroshkevich and Novikov1964; Dicke et al., Reference Dicke, Peebles, Roll and Wilkinson1965). The initially obtained values were still too high, at about $40\text{ K}$, and were reduced ten-fold with Penzias’ and Wilson’s discovery.

In an ironic twist of events, the possibility that a relevant signal of the background radiation was detected much earlier was raised by none other than Sir Fred Hoyle, the key proponent of steady-state cosmology. He stated the CMB had indirectly been observed as far back as 1941 by Andrew McKellar, a Canadian astronomer working at the Dominion Astrophysical Observatory in British Columbia, although he took it as a refutation of Gamow et al.’s proposal, taking into account the $50\text{ K}$ prediction in Gamow’s book, instead of the $5\text{ K}$ prediction of Alpher and Robert Herman (Peebles, Reference Peebles2014, 220). McKellar (Reference McKellar1941) observed the rotational excitation of the cyanogen (CN) molecules toward the star z Ophiuchus. The data clearly indicated that some external factor was exciting those rotational transitions. The subtraction of energy calculated from local excitation indicated excess energy and the corresponding temperature of the excess source. McKellar assumed it was a contact with the thermal bath and, based on this, estimated the “temperature of deep space” to be about $2.3\text{ K}$. Contemporary studies of the same sort have confirmed the uncanny coincidence of this estimate with the modern value for the CMB temperature (e.g., Roth, Meyer, & Hawkins, Reference Roth, Meyer and Hawkins1993). These studies have also found that the CMB indeed excites the lowest rotational levels of interstellar cyanogen across interstellar clouds. In fact, these sorts of measurements are another important line of evidence for the origin of the CMB and the prediction of the CMB temperature at earlier times. And as we will see, they played a role in defending both the orthodoxy and the alternatives. The temperature of the CMB should increase in a linear fashion as we observe more into the past, and the temperature of distant gas clouds should be detected in accord with linear increase, once the local kinetic energy is known and subtracted. In 2000, the spectrograph mounted at the European Southern Observatory (ESO) measured the temperature of such a cloud at $z=2.3371$ (Srianand, Petitjean, & Ledoux, Reference Srianand, Petitjean and Ledoux2000). The result was in excellent agreement with the COBE results and served as an independent source of evidence for the CMB’s temperature.

The first deliberate testing related to the model was performed in Florida in 1946 (Dicke et al., Reference Dicke, Beringer, Kyhl and Vane1946). It tested the prediction of Dicke and collaborators that at a temperature less than $20\text{ K}$, the CMB would not show much leftover radiation at radiometer wavelengths.

All these tests and discoveries were merely the first segment in the long road to acceptance of the Hot Big Bang model. It was opportune that a sufficiently sophisticated theoretical framework already existed when the background signal was undeniably discovered. This contributed to a fairly quick and wide perception of the model as an orthodox interpretation, making it a target of alternative accounts. Over the following several decades, this interpretation gradually became an integral part of the standard cosmological paradigm, to such an extent that abandoning it now would amount to a wholesale rejection of the entire edifice of physical cosmology.

Before we dive into the rich and diverse world of explanations of the CMB following Penzias and Wilson’s discovery, we will present the CMB and its properties from the standpoint of the now overwhelmingly accepted Hot Big Bang theory. The extraordinarily successful COBE mission provided a plethora of relevant data that resulted in decades of numerous observational and theoretical studies of the CMB. The satellite realized three objectives: detailed testing of the shape of the blackbody spectrum, mapping the CMB across the sky, and detecting diffuse background in infrared and millimeter range (possibly resolving early cosmic objects, stars, and galaxies). The goals of the mission were, in fact, a result of decades-long debates on the nature of the CMB and its basic properties (see Figure 5.3).

Following the observational studies of the acquired data, the first result concerning the first key property of the CMB essential to any attempt to explain its nature was that the background radiation is a blackbody to very high precision levels (Mather et al., Reference Mather1994; see Figure 5.1). Moreover, its temperature, the second key property, has been measured with unprecedented precision. According to the COBE dataset, it is $2.728\pm 0.004\text{ K}$ with a 95% confidence level (Fixsen et al., Reference Fixsen, Cheng, Gales, Mather, Shafer and Wright1996). A more recent COBE dataset with WMAP (Wilkinson Microwave Anisotropy Probe) recalibration suggests the value of $2.72548\pm 0.00057\text{ K}$ K (Fixsen, Reference Fixsen2009), while an even more recent WMAP and Planck dataset (Hinshaw et al., Reference Hinshaw2013), which is still being improved, suggests $2.7260\pm 0.0013$. All these results are quoted in the standard $T\pm \Delta T$ format, from which the relative amplitude of anisotropies $\Delta T/T$ is readily obtained.¹⁵

We should note that one of the most impressive recent results is the measurement of the CMB temperature by the European Southern Observatory (ESO) at the epoch corresponding to z of approximately 2.34, as mentioned earlier, using properties of the molecular hydrogen in a damped Ly-alpha absorption system in the spectrum of background QSO (Srianand, Petitjean, & Ledoux, Reference Srianand, Petitjean and Ledoux2000). The obtained result, although characterized by a large error margin, corroborates the predictions of the standard CMB interpretation.

Figure 5.1 Blackbody spectrum of CMB as established by the FIRAS experiment on board WMAP.

From Fixsen (Reference Fixsen2009). (The copyright obtained from the author and the publisher, ACS.)

Finally, precision measurements of isotropy, the third key property of the CMB, have been obtained. Thus, the CMB is fairly uniform over the sky. Except for the already detected dipole anisotropy due to the motion of the observer together with the local galactic group (see Appendix B), the abovementioned very small anisotropies have been detected only recently at large angular scales ($7°$ and larger) and at the extremely faint level of $\Delta T/T\sim {10}^{-5}$ (Smoot et al., Reference Smoot1992; Hinshaw et al., Reference Hinshaw2003, Reference Hinshaw2013; see Figures 5.2 and 5.3).

Figure 5.2 The advancement of space-based CMB observatories: While COBE discovered intrinsic anisotropies in the CMB (those which are not consequences of our motion), the WMAP and Planck missions obtained insights into the map of the CMB.

Credit: By NASA/JPL-Caltech/ESA – http://photojournal.jpl.nasa.gov/catalog/PIA16874 (direct link), Public Domain.

Figure 5.3 WMAP all sky survey of CMB anisotropies. The Internal Linear Combination Map minimizes the Galactic foreground contribution to the sky signal. It provides a low contamination image of the CMB anisotropy, which translates into the angular-scale power spectrum of primordial inhomogeneities. It is, arguably, the major tool of contemporary cosmologists.

Credit: NASA/WMAP Science Team.

These three crucial properties – the shape of the spectrum, temperature, and isotropy – are not completely independent. We could not speak of the global CMB temperature were it not for its blackbody spectral shape and unusual isotropy. In fact, if we follow the standard interpretation, the spectral shape and isotropy are the primary properties of the CMB, while the temperature is an epiphenomenal property that, theoretically speaking, is adjustable, contingent on other nonessential factors. In one of the particularly serendipitous moments of modern history of cosmology, three decades before the discovery of the CMB, Richard Tolman (Reference Tolman1934), in his influential book, proved that the Hubble expansion of the universe would preserve the blackbody shape of any initially present blackbody radiation, with only the temperature decreasing linearly with the scaling factor. As we will see in due course, this physical fact makes the standard CMB interpretation seem “natural” but interferes with some of the attempts to interpret the radiation as a patchwork of sources thermalized at different epochs.

A significant and frequently cited consequence of the orthodox interpretation of the CMB that figured prominently as a target of the alternatives is the limit the temperature of the background radiation sets on the fraction of the universal density in the form of baryonic matter. The simplicity of the physical view that results in a predicted limit is part of the appeal of the standard model. Thus, if we leave aside negligible, very slow, physical processes affecting particles, like potential proton decay, and approximate at timescales comparable to the Hubble time, the baryonic number turns out to be a conserved quantity, while the vast majority of photons currently existing in the universe are the photon relics detected as the CMB. Therefore, the photon-to-baryon ratio today is essentially the same as it was at the time of decoupling – a remarkably simple trait of the universe. Finally, if the limitations on the baryon-to-photon ratio in the early universe, which are based on the theory of primordial nucleosynthesis (Copi, Schramm, & Turner, Reference Copi, Schramm and Turner1995; Schramm & Turner, Reference Schramm, Michael and Turner1998), are coupled with the fixing of the photon density per co-moving volume, we obtain a unique handle on the total cosmological baryon density Ω_b.

As we have pointed out, this value played a major role in the formulation and defense of the alternatives. We will look at the details of these arguments, but we should note that it is obvious that this value could have taken cosmologists in various directions. And already in 1968, John R. Shakeshaft and Webster (Reference Shakeshaft and Webster1968) demonstrated that the energy density ratio of primordial to non-primordial radiation is about 400 to 1. They drew this conclusion independently of the interpretation of the CMB. Moreover, even in the steady-state universe, the total number of photons emitted by conventional sources, such as stars within a sufficiently large co-moving volume, diverges. So does the number of thermalized photons originating with a hypothetical early stellar population (usually called Population III, although the term is occasionally used to denote any object with primordial chemical composition or zero metallicity), which provided both the first metals and the energy of the CMB. It is not surprising, then, that Sir Fred Hoyle repeatedly used this “coincidence” to argue for the Population III origin of the CMB (e.g., Hoyle, Reference Hoyle1994).

As we write these words, although some doubts are voiced from time to time (e.g., Baryshev, Raikov, & Tron, Reference Baryshev, Raikov and Tron1996), the astrophysics community has pretty much converged on the orthodox interpretation of the CMB. Moreover, the standard cosmological model is firmly founded on both available evidence and theoretical studies. And the 1965 discovery was clearly a watershed moment for both of these developments. Its magnitude makes it really difficult to resist the view that the now-standard interpretation of the discovery of the CMB, as a discovery of a remnant of primordial fireball, was an inevitable, fully transparent, and even predictable moment. Without dwelling on the details of forgotten history, it is very hard to fathom that any alternative interpretations would have been offered, seriously or even semi-seriously, by any distinguished cosmologist at the time. As we generally prefer to focus on and remember a triumphant scientific moment rather than ponder the winding and rocky road that led to it, with all its tributaries and blind alleys, the impressions of the inevitability of the current view and the shiny moment of 1965 are widely shared by astronomers and laypersons alike.

In fact, the textbooks reinforce such a view. Two of the best cosmology textbooks available, by Peter Coles and Francesco Lucchin (Reference Coles and Lucchin1995) and John Peacock (Reference Peacock1999), are great examples of such reinforcement. For example, Peacock formulated a mind-bogglingly simplistic, if poetic, statement about the CMB: “The fact that the properties of the last-scattering surface are almost independent of all the unknowns in cosmology is immensely satisfying and gives us at least one relatively solid piece of ground to act as a base in exploring the trackless swamp of cosmology” (Peacock, Reference Peacock1999, 290). Using a less flourishing style and adding a dose of reflection on the history of the field, in his commentary on the re-edition of Penzias and Wilson’s (Reference Penzias and Wilson1965) paper, Peebles wrote:

Albeit unwittingly, the passage sets the stage for a deeper philosophical-historical analysis by pointing out two key factors. First, the “willingness to believe” the standard model was certainly a major motivator of the push to develop the orthodoxy and was also the reason for a subsequent lack of reflection on the roads that led to it. Second, the discovery of the CMB led to a subsequent lack of confidence in the pre-1965 cosmological research, and this significantly contributed to a streamlined view of the history of physical cosmology (Peebles, Reference Peebles2014). These two factors merged into a widespread impression that the microwave noise detected serendipitously by Penzias and Wilson suddenly threw us into an epoch of serious, quantitative cosmology, and the essential validity of the Hot Big Bang paradigm occurred pretty much instantly and has remained unchallenged ever since. The following sentence by Coles and Lucchin summarizes this attitude: “it is reasonable to regard this discovery as marking the beginning of ‘Physical Cosmology’” (Coles & Lucchin, Reference Coles and Lucchin1995, xiii).

There is a dearth of good historical studies on the recent stages of our cosmological adventure, and there are multiple reasons for this. One reason touches on a long-standing controversy about what has been called “the Whig interpretation of history”: the tendency to portray the past as the inexorable march of progress toward the present enlightened and desirable state. The crux of the matter is Herbert Butterfield’s injunction:

“Making the past our present” has become a slogan of anti-Whig mainstream historians. A useful analogy would be travel in space where those vistas and customs most distant from our own are the key object of interest for a traveler; it would be a poor traveler who would denounce the denizens of distant regions for the dissimilarity of their customs. Similarly, Butterfield invited us to regard historical research as something similar to traveling to another time with the help of a Wellsian time machine, without prejudicing or denouncing the denizens of those times. This position is uncontested in political or economic history, as well as history of art; perhaps the only field where it is seriously doubted or even rejected is history of science, since there is no reasonable construal on which science has not made objective and often measurable progress (e.g., Hall, Reference Hall1983; Harrison, Reference Harrison1987; Mayr, Reference Mayr1990; Jardine, Reference Jardine2003; Alvargonzález, Reference Alvargonzález2013; Oreskes, Reference Oreskes2013; Giunta, Reference Giunta2022).

A couple of pages later, Butterfield stated another point very salient for our present purposes: “The things which are most alien to ourselves are the very object of his [historian’s] exposition” (Reference ButterfieldButterfield, [1931] 1959, 18). Unexplored – or only weakly explored – alternative explanatory hypotheses agree with this goal. So in a sense, by doing history of recent science, we strike a narrow compromise, something mentioned by the distinguished astronomer Edward Harrison in his well-known critique of Butterfield’s approach:

This is indeed a serious problem, not significantly discussed since the publication of Harrison’s paper. Our decision in the present book is to entirely ignore “the opprobrium associated with the Whig epithet” and to employ various methods of historical and philosophical research, disregarding more or less fashionable labels and derisions. In studying a relatively novel phenomenon like the CMB (from the point of view of human science, it is not intrinsically novel), there is simply no escape from keeping an eye on the standard interpretation and powerful theoretical framework of modern physical cosmology. To do otherwise would be a dereliction of duty to the prospective audience that, similar to soccer audiences, is fully entitled to know the score of the match. Since everyone is likely to concur that history of science is not written just for historians of science – especially not just those historians of science steeped in the non-Whiggish gospel – this duty should be taken seriously.

We could go a bit farther and suggest the very definition of the topic is impossible under the strict construal of non-Whiggish history of science. Harrison argued from a cost-benefit analytic viewpoint (“better to have eye-witness accounts, etc.”), but we could adopt a stricter stance and ask, “Eye-witness accounts of what exactly?” The boundary of a topic or a subfield cannot be properly set following the strictures of non-Whiggism, not only since the particulars of our scientific knowledge, or even just scientific speculation about a particular phenomenon are biased by our present-day dominant scientific paradigms, but also because the very existence of the phenomena under scrutiny is often part and parcel of those dominant paradigms. Since history of science, as any historical discipline, depends on the process of selecting available evidence on the basis of relevance (as well as some further criteria), this process cannot even begin without present-day scientific knowledge, alleged inadmissible by the non-Whiggish fundamentalists.

Nowhere is this seen better than in the case of the CMB origin and properties. Sir Fred Hoyle brought the attention of astronomical audiences to Andrew McKellar’s Reference McKellar1941 rather obscure publication on the “rotational temperature” of interstellar space as a precursor to the discovery of the CMB, but his act could only be understood when the importance of the CMB photons was realized, and that came later. There is no way to properly assess the relevance of Hoyle’s act – notwithstanding his motivations – without both the factual input about the “official” discovery of the CMB in 1965 and the background understanding of the crucial importance of this discovery in the context of cosmological controversies.

We give many such examples throughout this book. For that reason, we regard the book as a contribution to the perennial debate on the nature of historical knowledge and historical analysis, especially Whiggish versus non-Whiggish histories.

Our central goal, then, is to understand the emergence of the consensus, the exact role alternatives and their failures played in its formation, and the epistemological attitudes that drove all this. We develop our historical account of the discovery of the CMB and the great controversy that preceded it with that central aim in mind. In effect, we provide the look probably at the first prolonged debate, a fruitful scientific controversy, so to speak, documented according to the current standards of currently desired historical detail. The exhaustive publication practices of today had just caught on around the time of the controversy’s emergence, so every theoretical and empirical detail and idea is well documented and can be put into a larger historical context.

Yet we are not offering a comprehensive history of the discovery of the CMB. The historical accounts the reader may want to consult are those by John North (Reference North1994) and Bruce Partridge (Reference Partridge1995). A history of measurements of the CMB temperature up to the early 1980s appears in Chapter 12 of Jayant V. Narlikar’s Reference Narlikar1983 work. General accounts of modern cosmological paradigm can be found in many advanced textbooks, but we most frequently consulted Peebles (Reference Peebles1993).¹⁶

Having stated our caveats, we can now summarize our account of the gradual emergence of the orthodoxy. There were four distinct stages on the road to the full-scale establishment of the orthodox interpretation of the CMB. First, the theoretical model of the Hot Big Bang was developed, starting in the 1940s. Its various implications and parameters were, at least in principle, observationally testable, although it took decades before astronomers were convinced of realistic chances of such tests. Second, in the 1960s, independent groups of astronomers and physicists performed crucial calculations very precisely with the use of computers of the model’s empirical implications, including the properties of the background relic radiation. Third, the discovery by Penzias and Wilson in Reference Penzias and Wilson1965 was quickly and increasingly seen as a successful test of the Hot Big Bang model. At this stage, a moderate convergence of agreement on the model and the interpretation of the discovery within its framework emerged quickly yet left a sizeable domain for the development of various alternative explanations over the following two and a half decades. The dynamics of the development of the alternatives was sensitive to the surprisingly varying results, first on the shape of the CMB spectrum (especially in the late 1970s and early 1980s) and its apparent deviations from the shape of the blackbody spectrum, and second on the measurement of its isotropy. The number and the variety of the results during that period were such that in a 2009 publication of the WMAP milestone results, Dale J. Fixsen noted, with a tinge of after-the-fact superiority, “[t]here were many publications of measurements of the CMB temperature from the late ’60s and ’70s, but the uncertainties are large and the systemics were not well understood.” Thus, they could be excluded from the combined values of more recent and more precise measurements he then presented. The fourth stage, in the early 1990s, was marked by the COBE satellite observations and their unprecedented precision. The results were quickly and overwhelmingly interpreted as truly identifying the origin of the CMB, thereby cementing a very wide convergence of agreement on the Hot Big Bang model and the interpretation of the CMB within it. Further “alternative research” has quickly become difficult, if not impossible in practical terms. Certainly no institutional support of appreciable direct funding has been forthcoming.¹⁷

As aspects of the CMB were discovered over the next several decades, the postulation of alternatives became increasingly constrained, especially after the COBE discoveries. Yet between the initial, moderate convergence and the later wide convergence some two and a half decades later, a number of alternatives were developed, defended, criticized, and repositioned. These developments were often reactions to the key content developed in the four crucial stages of the emerging orthodoxy or prompted by incoming data. Thus, to explicate and assess the alternatives, we will identify the crucial aspects of the orthodoxy as they relate to the understanding of the CMB in each instance.

Given the constant stream of observations relevant to cosmology provided by big science astrophysics, namely the Hubble Space Telescope and its recent successor the James Webb Space Telescope, we may perhaps overlook the fact that cosmology is the youngest of all physical sciences. Like other subdisciplines of physics, astronomy has experienced bouts of big science across centuries (e.g., Tycho Brahe’s Hven Island project, old Mongolian, Indian, and Mayan observatories, and possibly Stonehenge). Physical cosmology, however, entered the family of observational sciences in a substantial way in the 1960s with the discovery of quasars by Maarten Schmidt and with Penzias and Wilson; arguably, it fully entered the big science network only in the late 1980s and early 1990s. Until then, it was a field developed in small, often interdisciplinary circles composed of intellectually ambitious scientists ready to swim against the mainstream by applying the latest achievements of physics to daring cosmological questions.¹⁸

In their ground-breaking report, Penzias and Wilson provided the values of the key observational parameters and prior expectations of the values and carefully drew a potential theoretical-observational inference:

It is indicative of the things to come that Robert H. Dicke and coauthors provided the first discussion of alternative interpretations of their model. After an elaboration of the Hot Big Bang model emphasizing the presumed isotropy and uniformity of the universe, they explained why an oscillating universe alternative or open universe cannot be accepted, given the temperature of $3.5\text{K}$; they developed a fairly elaborate account of the possibilities of a closed universe and its problems, concluding that the required leptons-to-baryons ratio is too high (Dicke et al., Reference Dicke, Peebles, Roll and Wilkinson1965, 418–419).

If the discovery by Penzias and Wilson was a key step for cosmology as an observational science using regular astrophysical methods of observation, thus ceasing to rely exclusively on abstract mathematical or physical modeling, then its first step into the territory of big science was the experiment with the CMB performed in 1977 on a U-2 military aircraft by George Smoot, Marc Gorenstein, and Richard Muller (Reference Smoot, Gorenstein and Muller1977). The apparatus (Figure 7.1) was calibrated and collected data over eight flights of the aircraft. First, the atmospheric background and anisotropy due to the motion of the earth were eliminated by measurements with a twin antenna, while spurious anisotropies were eliminated by the aircraft changing flight direction every 20 minutes. The anisotropy due to the Doppler shift was expected and finally measured at $33\text{ GHz}$: “The cosine anisotropy is most readily interpreted as being due to the motion of the earth relative to the rest frame of the cosmic blackbody radiation” (Smoot, Gorenstein, & Muller, Reference Smoot, Gorenstein and Muller1977, 899). The researchers also detected a surprising velocity of the Milky Way and the local group of galaxies (Smoot, Gorenstein, & Muller, Reference Smoot, Gorenstein and Muller1977, 900).

Figure 7.1 Anisotropy detector mounted on a U-2 spy plane. Reprinted with permission from Smoot, G. F., Gorenstein, M. V., & Muller, R. A. (Reference Smoot, Gorenstein and Muller1977).

Copyright (1977) by the American Physical Society.

We will keep returning to these ground-breaking results in the study of the CMB. For now, we note how striking it is that the authors of the experimental report were open to the alternative interpretations of discovered cosine anisotropy as possibly inherent to the CMB, even in 1977. They wrote: “We cannot eliminate the possibility that some of the anisotropy is due to an intrinsic variation of the cosmic blackbody radiation itself” (Smoot, Gorenstein, & Muller, Reference Smoot, Gorenstein and Muller1977, 901). While they were probably aware of the seminal paper of Rainer Sachs and Arthur Wolfe (Reference Sachs and Wolfe1967) proposing the mechanism for generating small-scale CMB anisotropies from early overdensities – thus linking the temperature map with the process of structure formation and evolution – it is doubtful that they recognized its ramifications. Citation analysis shows Sachs and Wolfe garnered about dozen or so citations per year for the first 25 years (!) after publication, but this exploded by about tenfold in 1992, after COBE.

Another two measurements turned out to be milestones. One was performed by David P. Woody and Paul L. Richards (Reference Woody and Richards1978, Reference Woody and Richards1979) from a hot air balloon and first published in 1978 as a preprint, and the other one by the COBE mission and published by Mather et al. in Reference Mather, Cheng, Eplee, Isaacman, Meyer and Shafer1990. The measurement by Woody and Richards was initially corroborated by Herbert P. Gush’s Reference Gush1981 measurement from a rocket (Gush, Reference Gush1981) and then by a collaborative measurement involving Richards from a rocket a few years before the COBE mission (Matsumoto et al., Reference Matsumoto, Hayakawa, Matsuo, Murakami, Sato, Lange and Richards1988). Mather et al. (Reference Mather, Cheng, Eplee, Isaacman, Meyer and Shafer1990) provided the first comprehensive COBE results, using essentially the same apparatus, but this time mounted on a satellite. This second string of measurements illustrates the progress in measurement techniques, the power of the results of measurements to sway theoretical attitudes, and the slow convergence process with all the twists and turns that were later forgotten, or at best diluted and distorted to create the appearance of smooth sailing.

Woody and Richards used a spectrophotometer mounted on a balloon and cooled it by liquid helium. Their results showed substantial deviation of the CMB spectrum from the Planck curve, from 10% to 20%, depending on the area of the spectrum. The measurements were performed at various zenith angles and floating pressures to avoid artifacts, and the authors estimated the confidence level at 85% (Woody & Richards, Reference Woody and Richards1978, 6; Figure 7.2). A number of the alternatives to the Hot Big Bang model were invigorated with these results, as they were the most precise at the time. We will encounter this observational milestone across different categories of the alternative interpretations.

Figure 7.2 The spectrum of the CMB obtained with a spectrograph mounted on a balloon in 1979, exhibiting deviations from the shape of the blackbody spectrum. Only the advent of the COBE mission unequivocally eliminated this deviation. Reprinted with permission from Woody, D. P., & Richards, P. L. (Reference Wright1979).

(Republished with the permission of one of the authors and the publisher.)

It should be noted that this was not the first of this sort of measurement. The ground measurements were exposed to thermal conditions of the lower atmosphere that were much warmer than the signal they were seeking. One way around this impediment was to repeatedly switch between pointing antennae at the cooled helium bath and at the sky. Yet the key problem was that the upper layers of atmosphere interfere precisely in the part of the spectrum (less than 1 cm in wavelength) that is crucial to measure to see whether the CMB fits the shape of the blackbody spectrum, or whether it complies only partially (following the Rayleigh–Jeans law capturing wavelengths up to the microwaves, but not Planck’s law). Thus, in the late 1960s and early 1970s, several ground, balloon, and rocket measurements were taken, but the results varied substantially. The proponents of the orthodoxy expected divergences from the blackbody spectrum shape to coincide with the “unexpected sources of radiation at high altitude” (Weinberg, Reference Weinberg1972, 515), while those keen on developing alternatives found this improbable. These uncertainties could be resolved only with infrared measurements from a satellite, outside the atmospheric and other thermal interferences, by cooling the equipment. In 1978, Woody and Richards introduced the cooling of the equipment, thus adding to the credibility of their results.

Fast forward 22 years: the same measurement technique, with essentially the same detector Woody and Richards used, but this time mounted on a satellite (COBE) devoid of atmospheric interferences, removed any reasonable doubt that the CMB spectrum was deviating from the blackbody curve by any value larger than 1%. The authors, in fact, cited the work of Woody and Richards when describing their own equipment, pointing out similarities, but also mentioning “several improvements” (Mather et al., Reference Mather, Cheng, Eplee, Isaacman, Meyer and Shafer1990, L37). Moreover, the Far InfraRed Absolute Spectrophotometer (FIRAS) mounted on COBE addressed the possibility of unresolved sources – the key hypothesis for a few alternative interpretations – and refuted it with unprecedented precision. We will be returning to this epoch-making measurement throughout the book as well. In between these two milestone results, George Smoot and colleagues’ (Smoot et al., Reference Smoot, Bensadouin, Bersanelli, De Amici, Kogut, Levin and Witebsky1987) measurements – part of a large international collaboration – reduced the divergence from the blackbody shape of the spectrum to less than 6%.