1. Some Pre-Carter Anthropic Ideas

The general meaning of the anthropic principle is that what we observe must be compatible with our existence or, more generally, with the existence of advanced life. Humans can occupy only a universe like ours, and this explains in a sense why the universe is as it is. Today the anthropic principle is often seen as a selection principle operating in the context of the multiverse, a view which goes back to Brandon Carter's original formulation nearly 40 years ago. The principle has a rich prehistory and anticipations of it can, if only with a considerable amount of hindsight, be found as far back in time as in ancient Greece (Ćirković, 2003; Barrow & Tipler, 1986). However, I shall limit my remarks to a few possible and not well known precursors of the twentieth century, including James Jeans and Arthur Eddington from the period between the two world wars.

Speculations about the role of life in the universe were a common theme in Jeans's many popular addresses. In a lecture of 1926 he argued that the liquid state, and hence the existence of ordinary water, required very special conditions. More generally he emphasized that "the physical conditions under which life is possible form only a tiny fraction of the range of physical conditions which prevail in the universe as a whole." Moreover, "In every respect – space, time, physical conditions – life is limited to an almost inconceivably small corner of the universe" (Jeans, 1926, p. 40). The fine-tuning necessary for life formed an important part of the later anthropic principle, but Jeans did not explicitly associate it with either the constants of nature or cosmic evolution. Indeed, by 1926 the expanding universe was still in the future.

Eddington was convinced that the laws of nature are indirectly imposed by the human mind which largely determines the nature and extent of what we know about the universe. In his method known as selective subjectivism he appealed to selection arguments somewhat similar to those later associated with the anthropic principle. He argued that the number of particles in the universe (N ≅ 10⁷⁹) and most other constants were determined by mental and therefore human factors, namely "the influence of the sensory equipment with which we formulate the results of observation of knowledge." This influence, he said, "is so far-reaching that by itself it decides the number of particles into which the matter of the universe appears to be divided" (Eddington 1939, p. 60). Eddington's anthropic-like reasoning related specifically to the human mind, not to advanced life in general.

It is often stated that Fred Hoyle applied anthropic arguments as early as 1953, namely in his famous prediction of a 7.7 MeV resonance level in carbon-12. In 1952 Edwin Salpeter had suggested a "triple alpha" process according to which three alpha particles would fuse into a carbon-12 nucleus under the physical conditions governing the interior of some stars. Hoyle recognized that Salpeter's process would only work if there existed a resonance level of about 7.7 MeV, which was subsequently found experimentally (Salpeter, 1955; Kragh, 2010). Since carbon is a prerequisite for life as we know it, we would not be here had it not been because of this particular energy level. Hoyle's successful prediction has often been quoted as "the only genuine anthropic prediction…" and "evidence to support the argument that the Universe has been designed for our benefit – tailor-made for man" (Gribbin & Rees, 1989, p. 247). However, Hoyle does not qualify as the originator of the anthropic principle. Although he did indeed predict the resonance level necessary for the production of carbon, and then life, his prediction owed little to anthropic reasoning (for details and references, see Kragh, 2010).

A few years after Hoyle's non-anthropic prediction, in 1958, the Russian astronomer Grigory Moiseevich Idlis published a paper in the Proceedings of the Astrophysical Institute of the Kazach Academy of Sciences in which he vaguely anticipated the anthropic principle. It is however unclear to what extent, since the obscure paper only exists in Russian language. In a later review Idlis (2001) claimed to be the true discoverer of the anthropic principle, a claim which was supported by his famous compatriot, the pioneer cosmologist Yakov Zel'dovich (1981). According to Idlis (1982, p. 357), in 1958 he argued that "properties of the Metagalaxy..." universe "… are, generally speaking, just the necessary and sufficient conditions for the natural origination and evolution of life to higher intelligence forms of matter, similar to man, finally aware of itself."

If Idlis's paternity to the anthropic principle is questionable, so is the claim that a version of the principle was formulated even earlier by Abraham Leonidovich Zelmanov, a Russian physicist and cosmologist who in the 1940s did important work on inhomogeneous cosmological models (Zelmanov, 2006). Zelmanov argued for an intimate interdependence between observers and the observed universe, including the idealist claim that "If no observers exist then the observable universe as well does not exist." He also stressed that "humanity can exist only with the specific scale of the numerical values of the cosmological constants [and] is only an episode in the life of the universe" (Rabounski, 2006, p. 35). Zelmanov seems never to have published his anthropic considerations. Moreover, what we know of them lacks proper documentation. I believe the priority claims of Idlis and Zelmanov are unconvincing.

2. Large Cosmic Numbers: Dirac and Dicke

While Carter was unaware of Idlis and Zelmanov, he was to some extent influenced by the views of Paul Dirac, Robert Dicke and John Wheeler. The cosmological theory that Dirac proposed in 1937 was based on the relationship between two large dimensionless numbers:

_T0 denotes the Hubble time, M the mass of the proton, and Δt = e²/mc³ ≅ 10^-24 s is an atomic time unit; the other symbols have their usual meanings. According to Dirac (1937), the numerical relationship was a permanent feature, valid for any value of the Hubble time and not only for the present one. From this he inferred that the gravitational constant slowly decreased in time according to G ~ 1/t. He did not refer to life in his paper of 1937. More than twenty years later Dicke reconsidered the significance of Dirac's large numbers which he expressed in a somewhat different way:

On this [Dicke's], assumption habitable planets could exist only for a limited period of time. With my assumption they could exist indefinitely in the future and life need never end. There is no decisive argument for deciding between these assumptions. I prefer the one that allows the possibility of endless life.

Dirac had for a long time been devoted to the doctrine of eternal intelligent life in the universe, a doctrine that would later form the basis of the final anthropic principle (FAP). In private notes of 1933 he stated as his credo that "the human race will continue to live for ever and will develop and progress without limit." What he characterized as an "article of faith" was "an assumption that I must make for my peace of mind. Living is worthwhile only if one can contribute in some small way to this endless chain of progress" (Farmelo, 2009, p. 221).

3. Towards the Anthropic Principle

Australian-born Brandon Carter wanted to understand the role of microphysical parameters in cosmology. In 1967, while a PhD student at Cambridge University, he wrote a manuscript on the subject which circulated among a small number of physicists. Carter's purpose was "to clarify the significance of the famous coincidence between the Hubble age of the universe and a certain combination of microphysical parameters," that is, the same relationship N₁ ∼ N₂ which Dicke had considered in 1961. This coincidence, he said, "can be fully explained in principle … in terms of conventional physics and cosmology, so that revolutionary departures such as Dirac's hypothesis of varying gravitational constant, or Eddington's Fundamental Theory are not justified" (Carter, 2007, p. 1). Surprisingly, he was at that time unacquainted with the works of Dicke, which he only came to know about the following year (E-mail from Carter to the author, 18 February 2010). Although he was aware of Dirac's Large Number Hypothesis, which he knew from Hermann Bondi's textbook Cosmology, he did not mention it in his manuscript of 1967. It is also worth noticing that he was not yet acquainted with the works of Salpeter and Hoyle on the triple-alpha process generating carbon-12 in the stars.

The first time Carter had an opportunity to present his ideas of what would become the anthropic principle was at a meeting in February 1970 in Princeton organized by Wheeler. At this meeting, a commemoration of the works of the English nineteenth-century mathematician William Kingdon Clifford, he gave a talk on "Large Numbers in Astrophysics and Cosmology" in which he argued that the universe could only be understood if life and observers were taken into account. He did not, at that time, speak of an anthropic principle. Carter's ideas were known to a small group of physicists and cosmologists before they finally appeared in print in 1974. Wheeler, Stephen Hawking and Freeman Dyson had participated in the Clifford memorial meeting in Princeton and discussed Carter's presentation. They were receptive to his ideas, which they found to be fascinating, such as shown by their published responses (Rees et al. 1974, p. 307; Hawking, 1974; Dyson, 1972). As early as 1972, in a review of the possible time variation of the constants of nature, Dyson referred to Carter's "principle of cognizability." Dyson considered Dicke's numbers N₁ and N₂, which he rewrote as

4. Carter's Address of 1973

On 10-12 September 1973 the International Astronomical Union held a meeting in Cracow, Poland, dedicated to the 500th anniversary of the birth of Copernicus. One of the sessions was chaired by Wheeler, on whose suggestion Carter gave a presentation of his anthropic considerations (Longair, 1974, p. 289). An extended version of the lecture appeared the following year in the proceedings of the conference edited by the Cambridge astrophysicist Malcolm Longair. The lecture and the corresponding article entitled "Large Number Coincidences and the Anthropic Principle in Cosmology" marks the true beginning of the anthropic principle. Hawking was also present at the Cracow meeting, where he gave a talk in which he suggested that the isotropy of the universe followed from anthropic arguments. He did not refer to the anthropic principle, but to the "Dicke-Carter idea" (Hawking, 1974).

Appropriately, at a meeting celebrating Copernicus, Carter introduced his topic with a reference to the so-called Copernican principle, the doctrine that we do not occupy a privileged position in the universe. But it was a critical reference: "Unfortunately there has been a strong (not always subconscious) tendency to extend this to a most questionable dogma to the effect that our situation cannot be privileged in any sense" (Carter, 1974, p. 291). Carter objected that there is indeed a sense in which our situation is privileged, namely a temporal sense. We obviously live in the epoch of life, which is far from random.

The large number coincidences considered by Eddington, Dirac and Dicke should be understood in terms of "what may be termed the anthropic principle to the effect that what we can expect to observe must be restricted by the conditions necessary for our presence as observers." It is necessary to take into account our special situation and properties when interpreting observational data, just as it is necessary to take into account the special properties of measuring instruments.

The name that Carter introduced in 1974 has proved highly successful, but it is generally recognized, and admitted by Carter himself, that it invites associations that are both unfortunate and unintended. At a conference in 1989 he proposed as a more informative and less ethnocentric name "observer self selection principle," thereby stressing that what is important to most applications of the anthropic principle is observers and not human beings (Carter, 1993, p. 38). Neither this name nor other alternative names, such as the "biophilic" principle, have caught on.

In addition to this "weak" anthropic principle (WAP), Carter introduced a "strong" version (SAP) according to which "the Universe (and hence the fundamental parameters on which it depends) must be such as to admit the creation of observers within it at some stage." Whether in one form or other he thought that the new antropic principle would only have explanatory power if associated with the idea of a world ensemble, the assumption that there exists numerous other universes. These other universes would be really existing and characterized by all possible combinations of initial conditions and fundamental constants. However, contrary to most later ideas of the multiverse, he only considered as real those universes which can accommodate observers of some kind. Carter felt that his suggestion of an anthropic multiverse (a word not yet coined) was justified by Hugh Everett's many-worlds interpretation of quantum mechanics to which "one is virtually forced by the internal logic of quantum theory" (Carter, 1974, p. 298). At the time this was a view shared only by a minority of physicists. According to Carter's argument our universe is special by providing conditions for intelligent life. Since no reasons can be given for the non-existence of numerous other universes without observers, he found it natural to assume that they do exist.

Carter was well aware that the new anthropic approach to cosmology was unorthodox and problematic from both a scientific and philosophical perspective. Realizing the speculative and non-testable nature of the strong principle he called it "rather more questionable" than the weak one. Did anthropic explanations, relying as they did on the assumption of other universes, qualify as bona fide explanations? "I would personally be happier," he admitted at the end of his paper, "with explanations of the values of the fundamental coupling constants etc. based on a deeper mathematical structure in which they would no longer be fundamental but would be derived." He reckoned a similar kind of problem in relation to anthropic predictions, which differed fundamentally from those ordinarily used in physics. Predictions based on the anthropic principle, he wrote, "will not be completely satisfying from a physicist's point of view since the possibility will remain of finding a deeper underlying theory explaining the relationships that have been predicted." Indeed, over the next decades questions of this kind would be hotly debated, both by cosmologists and philosophers. Currently they are mostly discussed within the framework of the string-based landscape multiverse (Carr, 2007), but they have their origin in Carter's address of 1973.