Woodpeckers and Diamonds: Some Aspects of Evolutionary Convergence in Astrobiology

1. Introduction

It has become “common knowledge” that evolutionary biologists have been among the fiercest opponents of the search for extraterrestrial intelligence (SETI). As with many other cases of such putative knowledge, the claim contains a kernel of truth, but is in general misleading. It is undoubtedly true that some distinguished evolutionary biologists have been among the most vocal critics and opponents of the SETI enterprise (and the wider search for “life elsewhere”) ever since its inception in late 1950s: most notably, perhaps George Gaylord Simpson, Ernst Mayr, François Jacob, and Peter Ward (Simpson, 1964; Jacob, 1977; Mayr, 1993; Ward and Brownlee, 2000; Ward, 2005).

The view of evolutionary biology as hostile toward SETI—sometimes implacably so—has not only been repeated and promoted by avowed opponents (Barrow and Tipler, 1986; Tipler, 2003) but also surfaced and resurfaced in various other contexts and fields: in philosophy (McMullin, 1989; Kukla, 2001), in biosemiotics (cross-disciplinary field dealing with the communication and sign interpretation in the biological realm; Ulvestad, 2002), and even astrobiology itself (Lineweaver, 2008). On the contrary, the fact that other distinguished evolutionists have been cautiously supportive of SETI, and of what now comes under the heading of “astrobiology,” has all too often been downplayed or outright ignored. Among these we find, for example, Stephen Jay Gould, David M. Raup, and Dale A. Russell (Russell, 1983, 1995; Gould, 1987; Raup, 1992). There are also those who might be said to have one foot in both camps; besides Gould (whose position on contingency of evolution implies a sort of SETI skepticism), we can mention his old opponent Simon Conway Morris (e.g., 2003a, 2003b, 2011) who shows a long-standing interest in astrobiology- and SETI-related issues. The situation is therefore far from simple or black-and-white, and only careful analysis of the merits and demerits of individual evolutionary arguments can bring us closer to a true assessment of the lessons of current evolutionary biology for the search for life and intelligence elsewhere. I have discussed several versions of Simpson's argument from evolutionary contingency in a previous study (Ćirković 2014, henceforth Paper I). In this study, a further step is taken toward a comprehensive assessment of the role of evolutionary criticism for astrobiological and SETI studies.

Another highly prominent critic of SETI has been the famous geographer, anthropologist, evolutionist, and synthetic historian Jared Diamond, whose provocative 1990 article “Alone in a crowded universe,” originally published in Natural History (Diamond, 1990), has been reprinted in the influential anthology of Ben Zuckerman and Michael Hart (Zuckerman and Hart, 1995). It also appears as a chapter in Diamond's well-known book The Third Chimpanzee, a best-seller still found in most good bookshops (Diamond, 1992, pp 205–215). Diamond argues that relying on evolutionary convergence is misleading and that contingency, opportunism, ecological factors, and lack of synchronization strongly suggest that SETI is unfeasible. In conjunction with some other pieces in the same anthology, it has helped to solidify the widely held view that evolutionary biology in toto is antithetical to SETI studies. Diamond's arguments to this effect are often cited in the literature in both evolutionary biology and astrobiology (e.g., Conway Morris, 2003a, 2015; Lineweaver, 2005; Primack and Abrams, 2007; Vakoch, 2011; Conway Morris, 2017).

I have chosen to analyze Diamond's narrative here since it is remarkably well written, unambiguous, clear, and—of course—well-informed. It is arguably the best that skeptical evolutionary biology has offered in the half-century since Simpson's study. In contrast to Simpson, Diamond takes into account some of the advances made since the heyday of the Modern Synthesis—notably the increased importance of the eco-evo-devo context in which evolutionary explanations are now sought. In addition, Diamond's enormous intellectual stature and the prestige of his name as a distinguished public intellectual and author of such seminal works as Guns, Germs, and Steel (Diamond, 1997) and Collapse (Diamond, 2005) ¹ exceed even Simpson and Mayr in building public perception that evolutionary biology—as well as sociobiology, anthropology, and even social science and history—has “debunked” SETI research. Yet, as I will argue, such a perception—at least to the extent that it depends on arguments such as Diamond's—is unfounded. Diamond did succeed in pointing to some extremely interesting issues for the philosophy of astrobiology. His objections are serious and should be vigorously explored and engaged with by SETI researchers in a critical debate—which, sadly, has not taken place so far. This, however, is a far cry from any perceived “debunking.”

Simpson based his anti-SETI argument on the contingency and opportunism of evolution in each particular lineage, coupled with some proto-rare-Earth thinking. While he has expressed some reservations about the alleged extremely high adaptive value of intelligence, ² the question of adaptation actually plays only a subordinate role in his argument. Even if the adaptive peak in the fitness landscape corresponding to Simpson's “humanoids” is high, the corresponding part of the morphological space is so minuscule that it is unrealistic to expect that the evolution anywhere climbed this Mt. Super-Duper-Improbable (to use a Dawkinsian metaphor) more than once. On that basis, Simpson concluded that we are indeed alone and that, consequently, SETI searches are a waste of time and resources. As I have argued in Paper I, among the many problems with Simpson's argument visible now—although not necessarily visible in the 1960s—are the static, atemporal nature of his space of all possible biological forms (morphospace), the neglect of convergence as well as of ecological and observation selection effects, and ambiguity in respect to the biosemiotic status of possible SETI targets.

In contrast to Simpson, the arguments invoked by Diamond make important use of the fitness landscape and ecological arguments, and he is at least aware of the importance of observation selection effects (although he invokes them in a curiously one-sided way, as we shall see below). More importantly still, Diamond acknowledges the role of convergence in evolution, and how convergence prima facie offers support to the underlying philosophy of SETI projects. The thrust of his arguments is directed not against convergence per se, but whether it is frequent enough to offset the contingency and opportunism of evolution. He follows Simpson in elaborating the ambiguous adaptive status of human intelligence, but in general his arguments are updated and more general than Simpson's, thus meriting a separate, detailed analysis. Another important reason to consider Diamond's skeptical arguments is that they clearly demonstrate the dangerous pitfalls of treating cultural evolution as a straightforward extension of the biological—an issue with obvious relevance to the behavioral and social sciences, as well as for future SETI research.

After briefly recapping some of the orthodox positions and reasons for believing that Earth is not unique, some fun with the alleged “flying saucers,” and some flirting with Fermi's paradox (“the silence is deafening”), Diamond turns to his central task—the criticism of convergence-based arguments in favor of SETI (p 158 ³ ):

The core of Diamond's argument rests on the example of woodpecking. He argues that woodpeckers are an extremely well-adapted species, whose complex adaptations enable them to fill a stable ecological niche with significant success (p 159):

Although there are many individual subtraits intrinsic to woodpecking, Diamond argues that the adaptive value of the whole combination is so high that one would expect woodpecking to be extremely frequent, both in terms of the number of independent lineages and geographical coverage. Thus, if we expect universally adaptive traits to be convergent, we have reasons to expect woodpecking to be convergent. And if we rely on the convergent evolution of radio communication in planning and executing our searches, as the argument of the pro-SETI crowd goes, we should also expect to see a convergent evolution of woodpecking. And yet, Diamond points out, the opportunism of evolution caused woodpeckers to evolve—to the best of our knowledge—once only, and in a limited geographic area. Woodpecking is sufficiently complex that we can discern different types of adaptations: for drilling in live wood, for perching on bark, extensible and insect-catching tongue, tough skins, and so on. We observe only individual components of complex traits in unrelated clades: the strong bills of honeyeaters, ripping of dead wood to find insects in striped possums, and so on. The full package evolved, apparently, only once in the history of life, even though the conditions in which the full package is adaptive persisted for a long time.

Even more pertinent is the fact that the geographical coverage of woodpecking seems opportunistic and contingent: woodpeckers radiated from the first population over land and have never colonized distant landmasses such as Australia, Madagascar, or New Zealand. Despite the fact that some members of the woodpecking trait set apparently evolved on these landmasses independently, no real analogues to woodpeckers appeared there. The conclusion that high adaptive value does not in itself lead to convergent evolution of a complex trait therefore seems inescapable—and our SETI optimism about still more complex traits such as the capability for radio communication or interstellar travel seems even less well founded. ⁴

Diamond also points out that the synchronization of transmitters and ourselves is quite a tricky business in contemporary SETI practice. To expect that our radio listening will be successful means that alien transmitters are present for a long time, comparable to evolutionary timescales, thus constraining not only the lifetime of an intelligent species but its cultural evolution as well. Diamond further notes that, sadly enough, the current state of human species does not offer too many reasons for optimism regarding the long-term stability of our transmitters and receivers. On the contrary, following Simpson (1964), Diamond questions our chances for long-term survival, adding to the worry about global nuclear war—predominant in Simpson's time—a new and timely concern about anthropogenic climate change. He thus concludes that, since constraints upon human-type intelligence are such that required synchronization is improbable, our practical SETI efforts are in vain. This is an auxiliary argument, and is subordinate to the main one dealing with the (im)probability of convergent biological and cultural evolution. Importantly, this auxiliary argument points out an inconsistency between our intuitive take on the relevant timescales for SETI on the one hand and cultural evolution on the other. ⁵

Overall, there are three basic problems with Diamond's narrative: (1) its neglect of the “other side” of observation selection effects in evolution; (2) its adherence to strict gradualism; and (3) its atomized view of traits, which is utterly inapplicable to products of cultural/technological evolution. Taken individually, each of these problems could be addressed, and Diamond's view could perhaps be patched up accordingly. Taken together, however, they severely undermine the credibility of his conclusions, especially the celebratory coda to the article (“What woodpeckers teach us about flying saucers is that we're unlikely ever to see one. For practical purposes, we're unique and alone in a crowded universe. Thank God!”—p 164). Intertwined with these themes is his failure to take into account niche construction and specialization as sufficiently important factors in the overall context of biological and postbiological evolution. In what follows, I will show how analysis of these themes undercuts Diamond's skeptical conclusions and leaves the door for unbiased, innovative SETI projects wide open. Again, it needs to be emphasized how such undermining of Diamond's narrative offers a more general opportunity for discussion of evolutionary contingency and convergence in specific context of astrobiology and SETI studies, which has not been considered much until recently.

There is another motivation for undertaking the present study. In recent years, there has been a surge of studies of evolutionary convergence and parallelism, motivated by quite heterogeneous concerns, from the structure of the protein space to experimental evolution to evolutionary genetics to ergodicity in biophysics to the attempted “neo-Gouldian” developmental account of homology versus homoplasy (Beatty, 2006; Dryden et al., 2008; Turner, 2011; Lobkovsky and Koonin, 2012; Pearce, 2012; Powell, 2012; McLeish, 2015; Orgogozo, 2015; Powell and Mariscal, 2015; Louis, 2016; O'Malley and Powell, 2016). Roughly speaking, most of these new results are strongly supportive of convergence—in more or less radical form—as the key feature of macroevolution.

Conversely, they undermine various strong forms of evolutionary contingency upon which evolutionists' skeptical arguments regarding extraterrestrial life and intelligence have historically been based. For instance, Dryden et al. (2008) and McLeish (2015) argue that the accessible part of the genomic space is much smaller than conventional combinatorial wisdom suggests—and that evolution could have actually explored most of it so far! Together with Vermeij's (2006) study of observation selection effects in the fossil record of evolutionary innovations, which is discussed in the next section, these recent developments are very good news for SETI studies. For many reasons, among which conservatism and disciplinary segregation are certainly not the least, this good news has hitherto not been recognized as such. There are important lessons for the astrobiological and SETI community here as well, and Diamond's critique is perhaps the clearest battlefield on which to draw out these lessons.

2. Observation Selection Effects as an Explanation for “Uniqueness”

Reasoning about traits or properties that constitute observers, or that are in any way necessary for the existence of observers, is subject to observation selection effects and biases (for a comprehensive review, see Bostrom, 2002). For example, the number of asteroid and comet impacts in Earth's history is constrained by our existence at the present time (Ćirković et al., 2010). The importance of these anthropic biases has only recently begun to be fully appreciated.

Observation selection effects appear in judging incomplete samples in palaeontology—and hence in the history of life—as well. In particular, episodes of mass extinctions in Earth's history are strongly deformed in the fossil record by observation selection effects, as demonstrated in a famous study of Signor and Lipps (1982): what actually happened quickly and dramatically looks protracted and gradual in the record due to selective sampling. This has important adverse consequences for gradualism (see section 3 below), but also weakens the case against evolutionary convergence: if extreme ecological perturbations corresponding to mass extinction events have overturned entire faunas, switched macroevolutionary regimes (Jablonski, 1986), and introduced large amounts of contingency in evolution, then the remaining amount of contingency due to “normal” evolution within a single macroevolutionary regime has to be reduced in our explanatory hypotheses. In other words, if, for example, the asteroid impact at the end of the Cretaceous was an epistemically random event, hence effectively impossible to predict or retrodict, then subsequent Palaeogene evolution would have been less unpredictable if the impact had never occurred. Therefore, it would have made more sense to produce explanatory hypotheses based on extrapolation of the known evolutionary trends and processes. (This applies to our hypotheses about the evolutionary processes, not to the processes themselves, which are, of course, completely independent of our predictive/retrodictive powers and capacities.)

For the purposes of SETI-related discussion, lacking precise knowledge about correlations between properties of individual macroevolutionary regimes and the prerequisites for the evolution of cognitive capacities and technology, we should find that the latter is less contingent and hence likelier to be convergent and predictable barring a sudden switch of macroevolutionary regimes (cf. Ćirković, 2012, esp. Chapters 6–7).

Diamond argues that woodpecking is an example of extremely useful adaptation originating only once in the entire history of life. While ornithologists concentrating on woodpeckers may use this as a day-in, day-out working principle, it is entirely natural to ask: how confident could we be in such a conclusion? Obviously, we have not yet found fossil traces of independently evolving woodpecking, but this is a far cry from claiming that it never happened, or that we understand the reasons why it did not happen (or even why it could have never happened). This is a reasonably skeptical general stance, but when we also take into account observation selection affecting the fossil record, we can go some steps further. ⁶

In Paper I, it was argued that the correlation studies reported by Vermeij (2006) undermine the Simpsonian case for evolutionary contingency and, conversely, strengthen the case for convergence; I shall extend the same criticism here to Diamond's argument. Insofar as key evolutionary innovations are largely determined by universal principles of physics and economy, as Vermeij has persuasively argued, they lead to widely-to-universally useful designs. Their repeated appearances in the history of life tend to be hidden by the incomplete and sparse nature of the fossil record; allegedly unique innovations tend to be either very old or appear in large clades (so that independent reinventions in smaller clades vanished due to incompleteness). Few innovations are ever truly unique (Vermeij, 2006, p 1804):

Much here is applicable to the case of woodpeckers. First of all, we cannot be certain that woodpecking truly evolved only once—it might have evolved several times in very small clades of birds, all of which went extinct without leaving a trace in the fossil record. In that case, Diamond's insistence that a “wonderful niche” has been occupied by a unique, opportunistic accident, would be undermined. While this might sound as a weak argument, it is similar to the other arguments from selection effects (starting with Bacon's argument of 1620 against using shipwreck survivors to argue for the causal efficacy of prayer) in that it cuts overconfidence of the opposing side to measure. Until we establish with any measure of certainty that woodpecking did never evolve independently in a small lineage, we should refrain from making sweeping statements on the basis of perceived uniqueness.

Even more important for the present discussion is Vermeij's lesson regarding the “union and integration of previously independent components.” As Diamond himself emphasizes, the tool-making capacity essential for achieving radio communication depends on two key traits: intelligence (or whatever other measure of advanced cognition) and mechanical dexterity. ⁷ Isn't that a paradigmatic example of the “integration of previously independent components” Vermeij talks about? If so, could not both human technology and woodpecking be exceptions that confirm the rule, namely, that large-scale, complex integrations of many component traits are the source of apparent uniqueness in an otherwise convergent world?

Finally, if woodpeckers were chosen as a generic example, this may be regarded as slightly unfair, because they occupy a niche that is highly evolved in itself, and must be analyzed through a complex eco-evo model that includes its coevolution with trees, insects, and so on, with myriad short-, medium-, and long-term feedback loops. This conclusion is, indeed, independent of our choice for the predominant evolutionary mechanism and is fully compatible to assigning much bigger importance to developmental constraints than the run-of-the-mill adaptationists usually do. It would be much fairer (and to the point) to discuss more general traits instead, such as sight or locomotion (e.g., the capability of flight). And these did come up over and over again, in widely divergent lineages.

The criteria developed by Powell and Mariscal (2015) for sorting different instances of convergence (specificity, scope, evolutionary significance, etc.) could be immensely useful for clearing the confusion here. Traits such as sight or locomotion generally score very high on this checklist due to great scope, high evolutionary significance, and so on. This seems to apply to intelligence as well, in a more relaxed and less anthropocentric sense appropriate for zoopsychology; conspicuous levels of intelligence likewise have evolved independently in various lineages—in cephalopods, corvids, psittacines, marine mammals, elephants, as well as hominids. Woodpecking, on the contrary, would have been rated more modestly on the Powell and Mariscal criteria: highly specific, but of obviously very limited scope and at best modest evolutionary significance.

3. Reinventing Gradualism?

Traditional gradualism of 19th century vintage has been summarized by the motto: The present is the key to the past. As such, it resists any sudden—or, on geological, evolutionary, or astrophysical timescales, relatively quick—changes in the size of the morphological/design space. ⁸ Today we understand very well how brief and violent episodes of mass extinction represent effective shrinkages of the morphospace, and in some cases correspond to a change of the macroevolutionary regime and a major reorientation of the “vector of progress” (e.g., Gould, 1985; Jablonski, 1986). Since the 1980s, we have witnessed the resurgence of (neo)catastrophism as the most successful theoretical framework to account for large-scale features of both the Earth system and its history, and of other well-studied systems such as those associated with the Solar System planets. (Neo)catastrophism, as opposed to strict gradualism, need not restrict itself to great cataclysms or extinctions; it might just as well encompass Tolkienesque eucatastrophes (“good catastrophes”—Tolkien, 1983). This would correspond to a brief, dramatic expansion of the morphospace/design space available to evolution.

This is not a particularly new insight. Each of the “major transitions” (Maynard Smith and Szathmary, 1997) or “critical steps” (Carter, 1983) in evolution opened up huge new parts of the total morphospace—parts that without the transition in question would have been not only unreachable but perhaps inconceivable as well. They do not only represent branching in the pre-existing morphospace, rather, they also create new morphospaces—or, equivalently, they create accessible subspaces of some abstract total morphospace. In this sense, they represent eucatastrophes—and they violate the kind of ultragradualism implied by Diamond's account, which compares the adaptive value of woodpecking with that of radio technology. The two are simply incommensurable—they do not refer to the same morphological/design space. The “evolution” of radio, as well as any other aspect of the cultural evolution of homo sapiens, was made possible by the eucatastrophe of the emergence of advanced cognition/language/intentional tool-making/culture—whatever label we choose for the most recent major evolutionary transition.

Opening up this huge design space of cultural evolution immediately changed the properties of all other parts—even very ancient/ancestral ones—in the total accessible morphological/design space. The fitness of other life-forms is, obviously, changed by the emergence of human civilization on Earth, for better or for worse. Some species have found new and particularly valuable niches directly related to human cultural activities, such as rats or pigeons; most have felt the adverse impact of human meddling such as deforestation and other forms of habitat destruction. To ignore this is to commit an error of oversimplification.

In short, Diamond's comparison between woodpecking and radios is based on the erroneous assumption that their fitness is measured within the same morphological/design space. If we accept that dramatic changes of the accessible morphological/design space are not only possible, but true historical features of evolution and are likely to occur on any inhabited world as well, the case against convergence as a motivating rationale for SETI optimism is undermined. Of course, an alternative available to Diamond and his supporters is to reject the reality of major transitions and in particular those pertaining to the human society/language/technology (e.g., McShea and Simpson, 2011). The cost looks steep, though; there is a huge amount of evidence for sudden changes, faunal overturns, and other phase-transition-like phenomena in the fossil record, and the dramatic change brought about by the emergence of human cognition is perhaps best ratified by the introduction of the term “Anthropocene” for the current geological unit of time (Crutzen and Stoermer, 2000; Frank and Sullivan, 2014; Zalasiewicz et al., 2014).

4. Trait Atomization and Astroengineering

Leaving aside for a moment the very pertinent issue of the extent to which SETI is in general terms dependent on radio technology, let us consider the following aspect of Diamond's argument (p 162):

If Diamond wanted to shoot down the straw-man notion that radio, like woodpecking, is indeed increasing the reproductive success of a species, he undoubtedly succeeded. ⁹ No SETI scientist has ever claimed such a thing. At most, the vague claim that radio communication, given enough time, is a necessary consequence of highly adaptive traits such as intelligence and tool making was suggested by some of the early SETI pioneers (e.g., Cocconi and Morrison, 1959; Shklovskii and Sagan, 1966). Such a claim is problematic, but not for the reasons Diamond explicates. Most importantly, as Gould (1987) among other critics points out, the claim is unabashedly anthropocentric. That radios are a quirk of particular human history, and that we can easily imagine an advanced technological civilization that would minimize the use of radio and prefer (say) optical lasers or gamma rays or neutrinos or gravitational waves or whatever other medium of communication, is obvious. On the contrary, Diamond seems to be indirectly defending the claim from the charge of anthropocentrism, since he implies that there is an important universal adaptive value to be found in radio communication. While radio has some physical advantages in locations similar to ours, it is not at all obvious that such advantages persist in an even slightly different environment (e.g., a habitable planet around an M-dwarf star or any other place with higher levels of natural radio noise). Radio communication is no more inherently connected with any particular adaptive peak than is communication via gravitational waves (or would be, could we achieve it).

But isn't this what Diamond claims all along? No. Rather, he uses radio communication as a placeholder for the totality of our SETI efforts. After all, his criticism claims to be much stronger and to conclude that we are alone (as per his title, among other things) and not just that our radio searches will fail for technical reasons or lack of generality. The structure and justification of “we are alone!” claim differ drastically from the structure and justification of the claim (say) “a bunch of people doing SETI in California got it wrong.” Indeed, here we are approaching the crux of the problem with almost all sweeping criticisms of SETI: critics persistently fixate on a weak point and use it to argue against SETI as a whole. That strategy would make sense if the ideas and concepts of SETI had the topology of a chain; but if, for instance, it looks more like a bush, the strategy is just rhetorical and fails to sustain the critics' platform. There is no “weakest link,” the breaking of which could obviate the entire SETI research program. So even if Diamond is correct, in the specific case of radio, to conclude that it is contingent, improbable, and unsynchronized, it neither follows (1) that there are no meaningful SETI targets nor (2) that a different, nonradio approach to practical searches (e.g., optical, infrared, gamma-ray, neutrino) is unlikely to yield results.

Part of the problem here is an overly cavalier approach to the component subtraits. Although Diamond does identify a number of subtraits necessary for success in woodpecking (stiff tails, brain-protecting helmet, etc.), as well as radio communication (having transmitters, receivers, etc.), most of the time he considers them as isolable, atomistic traits. If all subtraits are necessary for woodpecking, one would naively expect the probability of woodpecker evolution to be equal to the product of the probabilities of all individual subtraits: the probability of evolving stiff tails times the probability of evolving brain-protecting helmets times, and so forth. Of course, given enough subtraits such as these, the probability of the desired outcome might be very small indeed. This seems to be in conflict with Diamond's assertion that since woodpecking is such a good adaptation, the probability of its evolving—and repeatedly evolving—should be high. (In fact, the antithesis goes further, since Diamond notices that woodpeckers' nests are so good that other birds often usurp them, thus clearly limiting the adaptive value of the corresponding traits in the realistic ecological context.) These subtraits are not, however, independent. Each has its adaptive value only in clear, well-defined ecological conditions; that we might not be entirely aware of them does not make them any less well-defined. Once a species or a group is able to access the niche—in this case, to eat grubs under the bark of trees—the remaining adaptations simply do not matter, since they are adaptations elsewhere; for all adaptation is exclusively defined at the time of response to selection pressure. Diamond thus conflates adaptation with adaptive pattern; it is the latter that forms the basis of convergence. There are correlations among various utilities of a subtrait, so the probability of their conjunction is not simply the product of the individual probabilities; in fact, it is bound to be larger. How much larger one cannot say without careful examination on a case-by-case basis and, lacking such an examination, we should not claim that their integration is convergent. For our purposes, in this study, it is enough to show that Diamond's claim that such integration cannot be convergent is spurious.

This point is applicable with equal force against Diamond's biogeographical argument. The fact that we are not aware of any independent evolution of woodpecking in Australia or Madagascar could simply indicate a lack of ecological prerequisites, not that convergence is inherently rare. Imagine, counterfactually, that there were no trees on Madagascar; would the absence of woodpeckers have surprised anybody? Evolution works with what is available here and there in both abiotic and biotic terms. As Darwin intimated in the famous closing words of On the Origin of Species, it is not a process like the law of gravity nor is it like the expansion of the universe, or even the diffusion of gases. At best it could be compared to phase transitions which, although broadly predictable (“water will eventually freeze at particular temperature and pressure”), are strongly dependent on the available external conditions, and thus intractable to follow in detail. ¹⁰

It is highly ironic that Diamond—who, like Simpson and Mayr, argued that the naive optimist view of SETI was created by physical scientists owing to selfish reasons and educational deficiencies, which biologists should correct—adopts what essentially is an extreme physicalist view of biological characters. Putting woodpecking and radio communication on the same footing looks surprisingly similar to those old jokes in which physicists say things like: “Consider a spherically-symmetric cow.…” The items are on entirely different tiers of complexity, so even if the parallel involving woodpeckers' adaptation were free from selection effects and took into account eucatastrophic expansion of the morphological/design space, it would still fail to capture the necessary complexity of cultural evolution and its artifacts such as radios. As Stephen Jay Gould wrote in one of his most famous essays on natural history (Gould, 1987, pp 34–35):

While a radio-communicating, interstellar-traveling civilization (the “flying saucers” mentioned by Diamond, or high Kardashev's types) might indeed be a Gouldian “theoretical optimum” for evolved intelligent species, we should not expect it to evolve spontaneously everywhere. There simply isn't enough time or sufficient “material available.” It is not the classic question of Allmacht versus Ohnmacht of natural selection (plus other evolutionary mechanisms); it is rather a question of external constraints, tracking back to the cosmological inventory of the world, and to the laws and regularities of physics and economy.

A recent article by Wagner and Rosen (2014) considers points of contact between evolutionary and technological (i.e., cultural) innovation in a notably “Diamondian” manner. These authors list no less than nine distinct aspects of innovation, common to both biology and technology (such as trial-and-error, exaptation, and horizontal information transfer). These commonalities seem to justify a unified approach (p 8, italics in original):

At first glance it might seem that Wagner and Rosen legitimize Diamond's treatment of radios and woodpeckers. A closer look, however, reveals many difficulties. Multiple innovations in both biology and technology offer, for Wagner and Rosen, strong support to convergence; singular innovations are signs of “frozen accidents”—highly visible, but ultimately unimportant in the grand scheme of things. Furthermore, both living and technological systems dramatically transform their ecosystems—and those changes, be they adaptive or maladaptive, present significant opportunities for SETI activities.

More irony is contained in Diamond's argument—again following Simpson's lead—that intelligence might turn out to be maladaptive, in particular if we destroy ourselves through nuclear war or anthropogenic climate change. While all reasonable humans should take these existential threats very seriously, and arguably even regard their prevention as the supreme moral imperative (Bostrom, 2013), it is easy to miss the point that we are now in the Anthropocene, and that our influence on our environment may not only outlive us, it might actually increase in the course of our final act. In other words, if we destroy ourselves, we might make it even easier for alien astrobiologists and astroarcheologists to infer our existence. This is the conclusion of several recent studies that investigated the detectability of phenomena such as rampant industrial pollution, greenhouse gases, or atmospheric nuclear explosions on extrasolar planets (Lin et al., 2014; Stevens et al., 2016).

Apart from these extreme cases, there are many other activities and artifacts envisioned so far in the literature—constrained, of course, by the limitations of our still rudimentary knowledge and rude imagination (cf. Cohen and Stewart, 2002, 2009)—as the signposts of advanced technological societies. Since the epochal work of Dyson (1960), we have been aware that some of these are detectable over interstellar distances and present entirely legitimate targets of SETI searches. Apart from the eponymous Dyson spheres, one might search for transits of artificial objects across bright stars (Arnold, 2005; Wright et al., 2016), or nuclear waste disposal by advanced civilizations (Whitmire and Wright, 1980), or annihilation signatures of bulk antimatter burning (Harris, 1986, 2002), or artificial ministars (Ćirković, 2016). These are only a few items from a long list of possibly detectable traces and manifestations of extraterrestrial intelligent life, constituting at least as legitimate search targets (Bradbury et al., 2011) as the intentional radio emissions that Diamond uses as prototype of SETI activities. The emerging discipline of interstellar archeology (e.g., McGee, 2010; Carrigan, 2012; Davies, 2012; Voros, 2014) constitutes an attempt to reach the level of arguably more convergent set of features than the culturally ephemeral “radios and flying saucers.”

It is unlikely that Diamond has been entirely unaware of this issue, especially given his choice of epigraph for Collapse—“Ozymandias,” the famous 1817 poem of Percy Bysshe Shelley, which evokes the monumental ruins of a long-vanished civilization. What he seemingly misses, however—and this in common with many other outspoken critics of SETI—is that interstellar detection signatures may, in principle, be many orders of magnitude more durable than the archeological signatures of ancient civilizations on Earth. This enables the sampling of a much larger spatiotemporal volume and gives a priori better chances of success than conventional approaches that are limited to searching for intentionally emitted communication signatures.

5. Diamonds in the Sky: Consequences for Astrobiology and SETI

Consider a relatively simple physical system by any definition—a diamond crystal, say. Its cube-centered crystal lattice is one of the simplest such structures in the world of solid-state physics, with the representative unit in the famous form of a tetrahedron made of four carbon atoms. This structure is characterized by a crystallographic group of symmetries usually denoted by \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document}$$Fd { \bar 3} m$$ \end{document} , and the great triumph of solid-state physics is the fact that many particular macroscopic properties of diamond—what actually makes it diamond—can be derived from this geometric fact. (This triumph, in turn, has strongly boosted the use of geometric methods in other domains of physics.)

In contrast to the simplicity of diamond, other elemental crystals (identical atoms in all lattice vertices) achieve astonishing complexity. For instance, the high-pressure phase Ba-IVc of barium crystallizes as an extremely complex structure with no less than 768 atoms of barium in the representative unit (Loa et al., 2012). Of course, compound crystals with many kinds of atoms can be even more complex, while still obeying the transformation laws of one or another crystallographic group. All crystallographic groups (or space groups) are well-studied and understood—and the very fact that there is a finite number of them—230 in three spatial dimensions—imposes profound constraints on the properties of the physical world. The fact that the structure of Ba-IVc is extremely nonintuitive even for experienced solid-state physicists and crystallographers does not affect our capacity for explaining and predicting its properties in the least.

Suppose we do not know anything about any other crystals, crystallographic groups, symmetries, and so on, but are free to investigate diamonds with the best laboratory equipment available, for an arbitrary amount of time. Could we infer the existence of complex crystals such as Ba-IVc from knowledge gathered solely from diamond crystals, could we delineate the accessible space of their macroscopic properties without ever having in our hands the “real” samples of the stuff, and, most importantly, would we be justified in treating diamond as an entirely naturalistic, intelligible, and generic outcome of the same set of processes that give rise to ever more complex configurations of atoms?

Knowledge about symmetries of diamond crystal would, arguably, lead to understanding of the role of symmetry in emerging macroscopic processes. Furthermore, this would lead to revision of our prior very low (or zero) probabilities of very complex systems actually existing in the physical world. When we discover them, we do not argue about whether their formation is contingent or convergent (except, perhaps, in the very parochial and trivial sense of whether it is contingent or convergent that they formed at a particular place, say, in a specific sediment at a particular time). Instead, we shall ask questions about their structure and place in the overall crystallographic system. A prototype, such as diamond, gives us key information about the entire parameter space.

Why should it be so different with life-forms (Davies and Walker, 2016)? To immediately discard conclusions based on a sample of one (life on Earth) does seem premature, at least so long as we stick to methodological naturalism and materialism. This is not to argue, rather naively, as do some overenthusiastic SETI optimists (together with most of Hollywood), that alien life-forms will look like us and speak English. Rather it is to argue that, insofar as convergence is important within the framework of terrestrial evolution, it is at least reasonable to assume, as a working hypothesis, that it will be important within the framework of the Galactic Habitable Zone. Of course the products of such evolution will be different, in morphological terms, in different locales, but those differences will reflect more the differences in “boundary conditions” than in “dynamics” (to borrow a Newtonian idiom). Evolutionary “dynamics” is likely to be similar everywhere, just as crystallography is likely to be similar everywhere.

So while the relationship between biological and cultural evolution is obviously the key topic for SETI, Diamond's parallelism between complex biological and complex cultural traits is oversimplified, and downplays key differences, such as intentionality, long-term planning, and purposeful changes in the physical environment. On the contrary, his skepticism toward convergence is unwarranted. Indeed such skepticism is surprising coming from an author who famously argued—in his Pulitzer Prize-winning Guns, Germs, and Steel—that the differences between human societies on Earth have nothing to do with innate, biological differences between populations, originating rather in initial environmental conditions (Diamond, 1997). In other words, in his later book, Diamond seemingly implied convergence as far as the dynamics of cultural evolution is concerned.

Some years after the original publication of Diamond's article (but before its reprinting in Zuckerman and Hart, 1995 ¹¹ ) a short article by David Raup—a distinguished palaeontologist and evolutionist—was published in Acta Astronautica (Raup, 1992), throwing a completely different light on Diamond's central metaphor. Raup, one of the major contributors to our understanding of the great episodes of mass extinctions of species in Earth's past, speculates that animals on other planets may have evolved, by natural selection, the ability to communicate by radio waves. ¹² Our own SETI projects are still mostly radio-based. Radio communication in such nonintelligent organisms as proposed by Raup would persist much longer than radio communication developed by intelligent beings, which would be ephemeral due to cultural changes. Such adaptive radio sources can therefore be predominant in the total set of radio sources related to life. Strategies for SETI should take the possibility of such radio communication into account—the sensitivity of our detectors will immensely improve in the near future, but there is a real possibility, consistent with everything we have learnt about biological evolution on Earth, that we will most likely detect only the meaningless buzz of some noisy radio insectoids (say). The latter would be another source of confusion and noise, not the sublime messages of advanced minds.

Raup's warning seriously undermines Diamond's argument about stable and universal adaptive niches. This occurs on two levels. On the first level, Raup's hypothetical radio insectoids (or radio whatever) would be even more wonderful adaptations than woodpecking; ironically, the arguments of the proponents of orthodox SETI wisdom in favor of radio, if we take them seriously, help to increase the adaptive value of any organism that evolves to use it. Radio waves are fast, inexpensive, long-range, and so on—all that old-fashioned stuff from the dusty pages of SETI books and articles from the 1960s. Thus, Diamond's wisecrack about woodpecking “offering much more food than do… radios” (p 159), while true on Earth, becomes false in Raup's compelling thought experiment. On the second level, the very idea that one and the same outcome (radio communication) can be construed as the outcome of both biological and biological-and-cultural evolution suggests that we cannot use individual traits in conceiving and planning our SETI projects. Since we cannot rely on the successful outcome of the search for a single specialized trait (such as radio), we cannot predicate the wider search for cognition in the universe on the adaptive value—or the lack thereof—of a single trait or small number of traits.

Diamond's concern about the synchronization of radio communication is also addressed—and undermined—by Raup, who justifiably states (p 260):

Clearly, this is not what Diamond had in mind when comparing radio and woodpecking: while it is true that woodpecking has existed for at least 7 million years, and possibly much longer, his intention could not be to argue that SETI would be fine if only it switched from radio-detecting to woodpecker-detecting techniques and methods. Notice that Raup does not prejudge anything with regard to the convergence versus contingency debate here. For the argument to work, it is enough that physics (and the “laws of form,” to use an older idiom for structural and developmental constraints) allows for the evolution of radio-communicating animals.

Finally, while Diamond's narrative presents a strong case against the orthodox SETI approach, based as it is on the search for intentionally emitted radio signals, to claim that his case holds against SETI in general is to commit a pars pro toto (i.e., the fallacy of mistaking a part for the whole). On the contrary, there are many strands of SETI thinking, which neither rely on radio communication nor accept the underlying orthodox methodology of searching for our cognitive peers (or moderately more advanced galactic civilizations). In particular, so-called Dysonian SETI puts strong emphasis on detecting traces and manifestations of hypothetical advanced civilizations in all parts of the electromagnetic spectrum, but mostly outside of the radio domain (Bradbury et al., 2011; Wright et al., 2014a, 2014b; Zackrisson et al., 2015). Instead of only expecting intentional signals at wavelengths reflecting our parochial biochemistry (e.g., the “water hole” searches being often associated, apart from the issue of noise, with the importance of H₂O for us), we should develop more flexible and universal strategies relying on laws of physics and rules of economics and technology. In particular, the Dysonian SETI calls for (1) the search for technological products, artifacts, and signatures of advanced technological civilizations; (2) the study of postbiological and artificially superintelligent evolutionary trajectories, as well as other relevant fields of future studies in conjunction with SETI studies; (3) the expansion of admissible SETI target spectrum; and (4) tighter interdisciplinary contact of SETI with related magisteria such as computer science and evolutionary biology (Bradbury et al., 2011).

Undermining Diamond's arguments should thus also impact and modify our view of SETI studies themselves. We should be more open to the existing SETI alternatives, which are more robust against old-fashioned criticisms, evolutionary or otherwise. If the case can be made that the philosophical assumptions and practices of orthodox SETI approaches are vulnerable to such criticism, those assumptions and practices should be either reformed or rejected.

6. Conclusions

As in the case of Simpson's argument—and even the patched Fiasco argument I consider in Paper I—what evolutionary theorists are trying to tell us is not that we ought to give up the search for extraterrestrial civilizations but rather not to put too much faith in communicating with them. Apart from emotional responses fueled by (bad) science fiction dramas, there is no reason why a SETI scientist could not entirely concur. As emphasized in Bradbury et al.'s (2011) discussion of Dysonian SETI, there is no a priori reason why the existence of extraterrestrial intelligence would imply communication in the usual, vernacular sense. If extraterrestrial intelligent species were all extinct—as are the historical human civilizations of ancient Egypt, Sumeria, and the Nazca valley—we would still be entirely justified in trying to find their traces and artifacts and investigate them. Neither communication nor synchronization between them and humanity—two points that Diamond insists are improbable—is necessary to the viability of the SETI enterprise. We need to clearly separate our dreams and desires about SETI from what it might be possible to achieve in reality. Heuristic tools such as the SETI parameter space (Barlow, 2013; Ashworth, 2014) or astrobiological landscape (Ćirković, 2012; Ćirković and Vukotić, 2013) can immensely help in this regard. In this spirit, Diamond's auxiliary argument about the lack of synchronization remains somewhat more valid, but cannot be considered decisive either. Namely, there are SETI programs that look for intermittent transmissions, and even if the lifetime of a transmitting society is relatively short, the generation rate of new such societies might be high enough to compensate for this. Clearly, a lot of work needs to be done in this area of SETI studies, but the prospects are not clear-cut gloomy as Diamond tends to present.

When we discard all these prejudices, the remaining evolutionary arguments of the SETI skeptic look entirely different—and rather unpersuasive. What we are left with is the rather bland and unprovable assertions that the part of the morphospace containing any kind of intelligence is vanishingly small, and the auxiliary assertion that it has to remain vanishingly small at all times, along with the assertion that even if intelligent beings are to be found elsewhere, they are bound to either self-destruct or else become locked in stasis, without evolving any radically novel characteristics or in any other way moving away from the ancestral form. ¹³ Irrespective of what we think about our own future and the likelihood that humanity will drive itself to extinction, to insist that these claims are truly universal properties of evolution is far-fetched at best. On the contrary, the tide has seemingly turned, and a host of new results tend to show that convergence on at least some universally valuable traits or components is predominant in the real world, and that the role of contingency is diminished. This is entirely in accordance with novel approaches to the fundamentals of biology such as the information-theoretic approach favored by Davies and Walker (2016), as well as Janković and Ćirković (2016). In conjunction with the confirmation of huge numbers of potential habitats (Petigura et al., 2013) offsetting at least, in part, the remaining improbability of convergence, this is truly a piece of good news for astrobiological and SETI research.

There is a potentially fruitful, wide-reaching research program for astrobiology and SETI studies hidden in the study of convergence, as hinted by Conway Morris (2003b, 2011). If convergence is indeed ubiquitous, Diamond's own woodpecking argument, as well as other pieces from the SETI skeptical literature (e.g., Simpson's and Ulvestad's linking of morphology to semiotics), would imply that the difference between outcomes ( = intelligent species, potential SETI targets) is mostly due to different initial habitat conditions. [“Initial” could be understood as “sufficiently early”; cf. Chopra and Lineweaver (2016).] Those conditions would fall within the domain of planetary science and would, in principle, be more amenable to remote sensing and analysis. Building models of the convergent “channels” of evolution for a particular planetary environment, among myriads of potentially habitable variations, is likely to be a highly challenging multi- and interdisciplinary endeavor, an effort ultimately rewarded by a cornucopia of deeper ecological and evolutionary insights.

All in all, while Diamond's ingenious anti-SETI arguments have contributed to the discussion of the relationship between evolutionary biology, cultural evolution, and SETI studies, they have not been—nor indeed by their intrinsic structure could they be—decisive in any way. There remains a vast uncharted space of possibilities for open-minded researchers in theoretical biology, SETI studies, the philosophy of (astro)biology, and other disciplines to build a new synthetic view of the place of life and mind in the universe. This is an opportunity that should not be squandered.