In 1987, Xianguang Hou, then one of the first palaeontologists to work on the Chengjiang biota—the older Chinese equivalent of the Cambrian Burgess Shale—published a benchmark discovery: Fuxianhuia protensa, an apparently unclassifiable new arthropod that seemed to have a very short and simple head. This idea was developed over the years, and Fuxianhuia became both a literal mascot of the Chengjiang fossil site and a candidate as the earliest “true” arthropod (euarthropod), with articulated legs and segments but with a very reduced head and little-differentiated body. Worth noting, however, is that for a long time this view wasn’t backed-up by any phylogenetic analysis. One of the earliest cladistic investigations specifically trying to place Fuxianhuia suggested the taxon could not unambiguously resolve as the most basal euarthropod, because there were no characters in the sister groups that would have made Fuxianhuia more ancestral than other of the ‘stem’ forms of the Cambrian, as they were known at the time. As a matter of fact, there have always been clues that Fuxianhuia, and its relatives, described some time later, were more derived and closer to modern clades. 

Nevertheless, early criticism was quasi inexistent and did not take hold, while the scenario of a Fuxianhuia emblematic of the earliest euarthropods sank in. Foundations for this scenario were mostly implicit, but slipped through as evident. The idea also became convenient, so much that it served as the basis for an emerging paradigm of early arthropod evolution, as illustrated by the fact that entire syntheses largely based on this perspective were written. Acceptation of basal fuxianhuiids as a fact was granted, and, importantly, defended, to the point that contradicting evidence was either overlooked, horseshoed ad hoc inside the dominant theory, or outright kept from being published through more or less consciously-biased peer-review. 

The latter was the case of a piece which describes fuxianhuiids as mandibulate arthropods, and now published in the Journal of the Geological Society by my co-authors from the Nanjing Institute of Geology and Palaeontology and myself. First submitted in 2017, and presented at the International Palaeontological Congress in Paris in 2018, the manuscript has been repeatedly sent for review but relentlessly rejected by our peers. Symptomatically, however, the content was seldom targeted and has itself changed little over these three years. The story of this paper as a whole represents an archetypic example of the importance of the human factor in understanding the evolution of scientific paradigms today. Among other things, it should serve as a reminder that scientific progress does not work as ideally as it is often depicted. 

The scientific paradigm: Between theory and reality 

Thomas Kuhn famously wrote The Structure of Scientific Revolution from his office in Harvard, and published it in 1962. From a general perspective, the theory is simple. Corroborated hypotheses coagulate and form ensembles of ideas that constitute the dominant theory at a given time. Once in a while, “anomalies” emerge, challenging the dominant theory. Most of the time, these anomalous reports are isolated and discarded, but occasionally they generate what Kuhn coined “paradigmatic crises”, events characterized by the accumulation of numerous- and consistent-enough anomalies to convincingly threaten the "normal science," as Kuhn put it. Challenging theories can replace old ones, but can also give rise to entirely new syntheses. Kuhn drew a great deal of his inspiration from the rise of relativity amidst conventional Newtonian physics, which resulted, not in the replacement of one paradigm by another, but by the coexistence of both in different subsets of reality. 

The Central Dogma of Francis Crick is often taken as an example of scientific paradigm. In this case, the choice of word is interesting in its confusion, because, conceptually, a paradigm is defined precisely by not being a dogma, meaning that it always remains refutable, as the core of its scientificity, irrespective of the wealth of inductive evidence supporting it. Nonetheless, veracity (“truth”) is always a variable of language, and therefore certain tenets—say, heliocentrism—are “emendable” rather than refutable per se. In that, a paradigm can resemble a dogma. But it is fundamental that criticism to any paradigm be allowed and evaluated objectively, as long as it is based on evidence and a hypothetico-deductive approach. 

This is one of the principal assumptions behind Kuhn’s system of scientific revolution, but, unfortunately, like all other human factors, it remains implicit throughout the The Structure. Dismissal of the human context is arguably the major shortcoming of Kuhn’s book, and what makes it an idealistic representation, because, in reality, this human context is a true driver of scientific progress. Economical, historical and political factors are behemoths behind scientific change, but they are external to a paradigm. How, when and by whom science is diffused are attributes more largely dependent on scientists themselves and publishing houses—which should be scrutinized separately. 

What do I mean, then, by “human context”? Affinities between scientists themselves, pressure to publish, perceived prestige of a journal, personal influence, financial considerations—everything orbiting science and which can influence its visible outcome by polarizing certain choices for individual benefits rather than for the betterment of science itself. In philosophy, the antagonism between “consequentialism,” which pertains to what is good (that is, has a favourable outcome), and “deontology,” which pertains to what is right, is appropriate to describe this pervasive ethical dilemma scientists are exposed to when presenting their work or commenting on the work of others. Of course, siding for or against a paradigm needs not necessarily come out of foul intentions, but simply out of pride—which may lead to actions that are just as reprehensible. 

The real challenge is not to identify these problems, because they are common knowledge among scientists. It is, first, to stop their growing prevalence and then to flush them out progressively from the scientific process. That is not to say that science should be reduced to a cold, pragmatic, purely mechanistic endeavour—no question would drive it and it would rarely advance in many disciplines (including palaeontology) if that were the case and imagination were not a factor—but, in fact, quite the opposite: every scientist should also use their ethical judgment and strive to have a deontology, which by definition has moral roots in the broader philosophical realm of science. 

Dogmatic fuxianhuiids 

In 2012, an atypical paper garnered attention from across the palaeontological community: it described remains of neural tissues in Fuxianhuia, a 518-million-year-old fossil, and concluded that the brain structure was unexpectedly advanced for a Cambrian arthropod—for Fuxianhuia in particular, supposedly very basal in the arthropod tree. Apart from the debate on tissue resilience that this study fueled, what is of special interest is the way the results were interpreted in the paradigmatic context. Instead of prompting a re-evaluation of fuxianhuiid affinities, and a novel phylogenetic investigation, the observation that Fuxianhuia possessed neural masses in its eyes akin to that of mandibulates led to say that some of the most basal arthropods had developed complex brains. As mentioned in the introduction, this basal position of fuxianhuiids had been by no means secured, and was largely based on early descriptions of a “short head” as well as more implicit views considering this animal as “ancestral,” an interpretation therefore not sitting behind thick walls. In that sense, one may say it was a “soft” paradigm, inflated somehow by the collective inertia of the scientific community, rather than a well-established phenomenon supported by a series of reports. 

The significance of fuxianhuiids was further put to the test five years later, somewhat indirectly, when a study challenged the conception that other Cambrian arthropods, hymenocarines, were also characterized by short heads with one or two pairs of appendages—another facet of the emerging paradigm fuxianhuiids were a part of (read the relevant article here). The central implication was that “short heads” in Cambrian arthropods were essentially an artefact, being either concealed by carapaces or indistinguishable amidst other superimposed remains. Because the view that fuxianhuiids were ancestral arthropods relied heavily on this “short head” concept, these new results should have pushed specialists in the field to cast doubt on the assumed place of these fossils in the evolutionary history of arthropods, especially after the brain paper of 2012. Instead, the heavily-anchored fuxianhuiid dogma led some authors to reject the evidence presented in the hymenocarine studies. Retrospectively, this may seem like backward reasoning and an excessively conservative approach to science, but, for whomever holds the paradigm as a pillar of their system of knowledge in a given domain, this reactionary behaviour is, paradoxically, only the result of the scientific process, insofar as new ideas must be kept in check. This kind of situation is as old as science itself, but, regardless of the amount of subjectivity or pathos associated with it, doesn’t constitute the core of the problem—simply the dawn of Kuhn’s state of “crisis,” or, at least, a sparkle of it. 

A few months later, a paper came out documenting what is arguably the most stunning fuxianhuiid material published so far. Preserved ventrally and in great detail, particularly appendages, these specimens offered the opportunity to clarify the morpho-anatomy of these arthropods and to move forward. In particular, these fossils showed thorough resemblances with the previously-redescribed hymenocarines, both in their head and limbs further down the body. But all this information stayed at the door of the only possible tenants: ancestral fuxianhuiids, emissaries of the earliest true arthropods. In part, evidence was interpreted in a way that satisfied the current emerging paradigm, in part some evidence one could consider conspicuous was overlooked. I believe this did not stem from any dishonesty on the part of the authors—in fact, this is a classic phenomenon, for scientific discoveries diverging from the dominant line of thoughts often point to pieces of evidence that were previously available but could not be deciphered. However, in this case, it is also quite certain that the eyes and minds of these researchers were conditioned by what they chose to believe and by the fact that they remained willingly opposed to putting their tenets into question. This is no less common, and, at healthy doses, is needed, because researchers giving up on their convictions too easily would make the scientific process all too chaotic. 

A deontological line is crossed whenever the scientific dialog is impeded, in any form. This is exactly what happened when my colleagues and I tried to provide our own evidence and discussion on the topic, and when a criticism of the above study was attempted. For three years, our manuscripts had been swallowed by reviews ranging from dismissive to aggressive, without ever demonstrating that the core of our study was flawed. Scientific peer-review is a pitfall of its own, and it is no coincidence if attempts at a more transparent peer-review process are piercing through. Open peer-review can help, but more importantly the duty falls on the editors to be critical about the reviews they receive: more than ever, their role of judge is crucial, and must be infallibly deontological. In any event, and whatever the reasons for providing such reviews may be—certainly strong egos and personal or institutional affinities are never excluded—they are the true police of a given paradigm. A critical paper that is never published never exists. The era of pre-prints opens other perspectives, but whether a pre-print is taken seriously or even read at all, in the context we face, will largely depend on non-scientific factors—reputation, for instance—for a large part of a given community. In addition, a pre-print is what it is, a non-peer-reviewed document, and, as such, arguably so, an informal scientific statement. But, if there can be “guard hounds” of the conventional publication system, isn’t there a real risk that scientific advances develop informally as never-published pre-prints? 

A Spectacle of Science? 

The question hence raised is: How deceivingly robust the resilience of a paradigm can become if the system is rigged in such a way that challenging ideas supported by tangible evidence can be kept out of the publication process by subversive means? The prospect is even more worrying considering that Science, as a system, has become much like any other market, where its content is seen more and more as a commodity, a means for profit; where scientific importance is, much like other aspects of society, increasingly dependent on appearances, on the purported prestige of journals and publishers, some of them putting a heavy strain on researchers by asking for colossal publication fees, and thereby dangerously correlating scientific significance with financial means. This is, in situationist terms, the recipe of Spectacle, shadows dancing on the wall of the cave. The strict opposite of the philosopher’s journey.   

Of course, the work that we published can itself be refuted, otherwise none of what was written here would make any sense. But whether the burgeoning fuxianhuiid paradigm was actually more accurate is of little importance here; what matters is that the scientific process embraces its deontological self, and, fundamentally, that it does so beyond appearances—for it is easy for leagued experts to pretend to deontology when in fact protecting their own interests. 

Yes, in the end, our article was published, and science prevailed, because an editor and some of our peers took a deontological approach, but also because of circumstances—this article was, notably, an invited contribution to a special issue of this journal, and, as such, the editor had fewer reasons to discard it because of its controversial status. The risk is high and real that science becomes its own “game of thrones” or, for the pun, “chairs.” To that, one might ask if the reality has ever been different, rather than something of an entelechy, thinking for instance about how Cuvier used to crush Lamarck under his weight down to the literal basement of the Museum National d’Histoire Naturelle in Paris. 

What, then, is to be concretely undertaken? As mentioned, the current system leans heavily on the shoulders of editors. It could be argued that, depending on the journal, editors might not have the expertise to sort the wheat from the chaff, and have to rely on the reviewers’ comments, even biased ones. However, in many cases I have experienced, unfair attempts at protecting the paradigm were so obvious either in their twisted argumentation, well away from the content of the study, or just in their destructive tone, that difficulty of assessment cannot be a generalized explanation. Other factors are involved, one of them being perhaps an unwillingness to spend time deciding internal quarrels, because it lengthens overall publication time and—as silly as this may sound—reflects poorly on a journal’s statistics. But it is precisely through these impartial decisions based on scientific robustness, and aside from perspectives and alliances proper to the field in question, that science advances. 

Regardless of what all these non-scientific factors influencing the rejection of papers can be, the solution to deontological peer-review problems should start with editors properly assessing the quality of reviews they receive, and not shying away from discarding reviewers who could not provide an objective, structured, and relevant assessment. Even during the same peer-review process, if needed. This can be helped in part by making reviews transparent, but reviews should be allowed to remain confidential nonetheless. Beyond that, a better system than the current appeal, which many journals do not take seriously, should be established. Again, certain journals have started adding statements to that effect, proposing to have reviewers reviewing reviews... in the spirit of the “peerage of science” system. Like the King of Plato’s Republic, the best decision maker is perhaps the one who has the least interest in the subject being debated, so editors outside of the strict field of consideration might be more apt to make impartial choices, but this, in turn, increases the chances that the editor might be influenced by arguments of authority more than by scientific ones. This is why an appeal should be transferred to a different editor rather than having its validity decided by the same editor. It is also why a progressist initiative like community-based open peer review services could unfortunately suffer from its localized concept, where people of the same field are more likely influenced by the same paradigms. Increasing the robustness of the evaluation of appeals might also increase cluttering of the editors’ work, and many may consider submitting to another journal a more efficient option (as is common practice nowadays), but it would be the price to pay to also ensure a fairer scientific output.

Last edited: 1st May 2021