On a Scientific Approach for Deep Time Investigations

I came to geology through a metaphoric back door and was gifted a series of visionary mentors who imbued me with an outside‐the‐box view of Earth history. Coupled with remarkable advances in microscale analytical capabilities that arose just as I was graduating into the field, that unorthodox view permitted my research group to explore an alternate conception of the longstanding paradigm of earliest Earth history. As I look back on how the field has evolved in my 40 year career, I see a mixed record of significant advances in our understanding alongside a persistent groupthink that discourages creativity and individual responsibility. I've come to understand this as reflecting different intellectual requirements of deep time geology relative to the study of modern processes. This difference between the general understanding of the scientific method and our needs has given cover to an unconstructive habit. My field is ripe for a fundamental change in the way we address the meager and almost certainly biased geologic record. This disruption could lead younger scientists and future generations away from poorly justified consensus models toward a more complete understanding of the darkest age of our planet's history.


Perspectives of Earth and Space Scientists
HARRISON 10.1029/2022CN000193 2 of 10 But ask five geologists the same question and you're likely to hear four or five different answers.My own path to geology is likely among the more tortuous.
Growing up in Canada, I was a lazy and indifferent student in elementary (e.g., a failing grade in Grade 3 science) and high school (where I chose the easier "humanities" track over "science").I slouched into the local community college which soon introduced me to the concept of "academic probation."At about that time, a pal planning to move to Australia suggested I join him.Figuring I didn't have anything better to do we headed off to a nation that, unbeknownst to us, was entering a protracted period of stagflation and high unemployment.After a stint picking pears in central Victoria, I answered an ad in the 9 February 1972 Melbourne Age for a "junior technical assistant" (i.e., a teenager) to prepare rock thin sections in the Melbourne University geology department.There I mingled with the grad students, including Andy Gleadow who would go on to become the doyen of fission track geochronology.I became enchanted with geology.I continued my travels the following year, ending up working as a technician in the Anglo-American/De Beers research lab in South Africa (Figure 1).This experience cemented my determination to study geology and I returned to Canada and community college to make up my deficiencies in the math and science courses I'd shirked in high school.Now highly motivated, I was admitted to the University of British Columbia graduating with a B.Sc. in geological sciences in 1977.My honors thesis, supervised by Dick Armstrong, marked the first time "cooling ages" were used to reconstruct the thermal history of a pluton-the realization of what had been a fanciful vision when Andy Gleadow laid it out to me in a Lygon Street pizzeria only 5 years earlier.Recognizing my interest in research but continued disinterest in formal learning environments, Dick recommended me to Ian McDougall for a Ph.D. at the Australian National University (ANU), which had no graduate coursework requirements, where I could continue my thermal history studies using the 40 Ar/ 39 Ar dating method (a destination likely facilitated by my Aussie wife Susan whom I'd met in Africa).As well as being a pioneering geochronologist, Dick was known for holding original views of earliest Earth evolution.I admit to not understanding much of what he was talking about back then but was made aware at ANU that his assumption that Earth had differentiated a core, mantle, and felsic crust immediately upon formation was regarded by most geochemists as verging on lunatic fringe.The paradigm of the day was clear: comparative planetology taught us that the first billion years of Earth history was a continent-, life-, and ocean-free hellscape (e.g., Cloud, 1972;Wetherill, 1972).
During my Ph.D. studies at ANU, as I watched Bill Compston and colleagues build the first Sensitive, High Resolution Ion MicroProbe (SHRIMP), I became convinced that I was seeing the future of geochronology.While an assistant professor at SUNY Albany working primarily on 40 Ar/ 39 Ar thermochronology in Tibet, I was stunned by the publication of Froude et al. (1983) who reported detrital zircons from Jack Hills, Western Australia, as old as 4.2 Ga.The idea that any rock or mineral older than ca.3.9 billion years (Ga) would ever be found on Earth was then close to geological heresy.Thus the SHRIMP group's report that they had documented Hadean zircons was revolutionary!How could zircon, a mineral characteristic of the continental crust, have formed and survived when Earth's basaltic surface was surely being pummeled from above and jostled from below?I was recruited to UCLA in 1989 with a mandate to bring a large radius ion microprobe to the northern hemisphere.Kevin McKeegan and I collaborated with CAMECA, then an independent French firm, to commission the first of what would be their ims1200-series instruments which have similar U-Pb geochronologic capabilities to SHRIMP but are also capable of analyzing negative secondary ions from electrically insulating samples (de Chambost, 2011).This latter capability reawakened the interest in early Earth that Dick Armstrong had kindled and we used it to investigate in situ, microscale carbon isotopic variations preserved in graphite inclusions in ancient apatite and zircon crystals.The results surprisingly suggested that life may have emerged much earlier (at ≥3.8 Ga; Mojzsis et al., 1996;at ≥4.1 Ga;Bell et al., 2015) than previously thought (i.e., ca.3.5 Ga).We then turned our attention to other information that could be teased out of the >4 Ga Western Australian detrital zircons, co-discovering evidence in them of liquid water at or near Earth's surface by ca.4.3 Ga (Mojzsis et al., 2001;Wilde et al., 2001).The availability of three large ion microprobes there, coupled with my considerable administrative duties, led to a strategic decision to build an unprecedented archive of Hadean zircons by automating the ion microprobe sample stages.The five year U-Pb dating analysis campaign of over 100,000 detrital zircons yielded ca.3,000 > 4 Ga grains from that Western Australian locality (Holden et al., 2009).
Subsequent geochemical analysis of those grains and their inclusions at ANU and UCLA revealed even clearer evidence of near-surface water (e.g., Watson & Harrison, 2005) and the suggestion of >4 Ga plate boundary interactions (e.g., Hopkins et al., 2008) during the portion of the Hadean eon informed by the detrital zircon record (i.e., 4.38-4.0Ga).Note that this interval excludes the period of lunar formation and magma ocean(s).While impacts were surely larger and more frequent through this period, they were still very rare (e.g., no significant impacts were occurring during 99%+ of that eon; Abramov et al., 2013).While the plate boundary interpretation remains controversial, the former-clearly antithetic to the old paradigm (e.g., Cloud, 1972)-is now widely accepted (e.g., Catling & Zahnle, 2020;Javaux, 2019;Rollinson, 2008;Shirey et al., 2008).This sudden reversal in view from a sterile, desiccated planetary surface to one supporting liquid water and perhaps life and modern-style tectonics led me to ask what had drawn our community to the hellish early Earth scenario in the first place, particularly given the lack of observational evidence supporting that view, and, as a consequence, why Dick Armstrong's ideas had been relegated to the intellectual wilderness (see Armstrong, 1991, for his death's door appeal for an evenhanded treatment of his views).As I looked back across the plate tectonic era (1965-present) at publications purporting to portray Earth's early history (Abe, 1993;Birch, 1965;Cloud, 1972;Ernst, 1983;Fyfe, 1978;Hamilton, 2007;Hargraves, 1976;Lunine, 1999;Moorbath, 2005;Smith, 1981;Solomon, 1980;Ward & Brownlee, 2000;Wetherill, 1972), three features stood out.The first is that none of these imposing authors (e.g., the vast majority elected to their respective national science academies) imagined a role for mobile lid (e.g., plate tectonics) behavior on early Earth, the second is the seeming confidence with which many of the more sensational opinions (e.g., no stable crust over the first billion years of Earth history; Wetherill, 1972) were pronounced, and the third is how little evidence to support those views then existed or has since arisen.What is it about our field that permits essentially unconstrained speculations to become cemented into a community convention?In what way do these papers represent the scientific method at work?Indeed, are they scientific?I return to these questions later.
Because of the significant role that physics and chemistry played in the enlightenment, the philosophy of science literature is largely a retrospective linearization of their histories (e.g., Gillispie, 1960) and thus has contributed relatively little to understanding the intellectual foundations of derivative fields such as historical geology (Gillispie, 1951).Although this neglect led to complaints from geologists, the implicit lack of respect is in some ways understandable.For example, at about the time Albritton (1961) and Macklin (1963) were criticizing the history of science discipline for ignoring geology, distinguished representatives of our field (e.g., all members/ fellows of either the U.S. National Academy of Science or Royal Society of London; Urey a Nobel-laureate) were advocating views that are close to magical thinking (e.g., geosyncline theory, Knopf, 1960;granitization, Read, 1957; the cold origin of the Moon, Urey, 1955; the permanence of the seafloor, Ewing, 1960).These examples are hardly anomalous; in the subsequent 60 years our field has a clear track record of coalescing around explanations of the past that are weakly or unconstrained by empirical evidence (e.g., a terrestrial hellscape until ∼3.5 Ga, Wetherill, 1972, Cloud, 1972, cf. Harrison, 2020; a Late Heavy Bombardment; Tera et al., 1974;cf.10.1029/2022CN000193 4 of 10 Hartmann, 2019; Mars-sized-impactor origin of Moon, Canup, 2004, cf., Canup et al., 2021; initiation of plate tectonics at 2.5 Ga, Condie, 2015, Dhuime et al., 2015, cf., Keller & Harrison, 2020) but yet widely accepted in the community (as evidenced by introductory textbooks) as geophysical facts.
So why are geologists attracted to consensus models of past Earth behavior that are in most cases innocent of observational constraint, only to be later embarrassed by their failure to hew to the "scientific method"?The answer, I think, is twofold.
The first lies in a flaw of human nature.As long as we've been thinking about the scientific pillars of our field, we've been repeatedly warned about one particular misbehavior: promulgation of a "ruling theory" (e.g., Bowen, 1948;Chamberlin, 1890;Gilbert, 1886;Johnson, 1933;Kitts, 1974;cf. Popper, 1968).Chamberlin (1890) described the descent of an impartial observer into an ardent defender of a singular viewpoint with such flourish that extensive quotations of his paper are routinely reprinted in discussions of geology's intellectual fundament.I follow that tradition here: "Love was long since represented as blind, and what is true in the personal realm is measurably true in the intellectual realm.Important as the intellectual affections are as stimuli and as rewards, they are nevertheless dangerous factors, which menace the integrity of the intellectual processes.The moment one has offered an original explanation for a phenomenon which seems satisfactory, that moment affection for his intellectual child springs into existence; and as the explanation grows into a definite theory, his parental affections cluster about his intellectual offspring, and it grows more and more dear to him…So soon as this parental affection takes possession of the mind, there is a rapid passage to the adoption of the theory.There is an unconscious selection and magnifying of phenomena that fall into harmony with the theory and support it, and an unconscious neglect of those that fail of coincidence…The search for facts, the observation of phenomena and their interpretation, are all dominated by affection for the favored theory until it appears to its author or its advocate to have been overwhelmingly established.The theory then rapidly rises to the ruling position, and investigation, observation, and interpretation are controlled and directed by it.From an unduly favored child, it readily becomes master, and leads its author whithersoever it will.The subsequent history of that mind in respect to that theme is but the progressive dominance of a ruling idea.Briefly summed up, the evolution is this: a premature explanation passes into a tentative theory, then into an adopted theory, and then into a ruling theory" (Chamberlin, 1890).Chamberlin's (1890) remedy for this disease is the use of multiple working hypotheses in which a detached investigator explores "every rational explanation of a phenomenon" and then "develops every tenable hypothesis respecting its cause and history."The investigator thus becomes "the parent of a family of hypotheses; and, by his parental relation to all, he is forbidden to fasten his affections unduly upon any one" (Chamberlin, 1890).While the method of multiple working hypotheses has been generally well received by the philosophy of science community (e.g., Platt, 1964), not everyone believes it is a complete description.In a paper entitled The granite problem and the method of multiple working prejudices, Norman Bowen (1948) famously noted that "when one views the present state of petrology, he cannot fail to wonder whether any individual is capable of such detachment."That all eight of the generalized pronouns in the quotations in this and the previous paragraph are he/him/ his suggests that about half of humanity senses that something else is missing.
As a side note, it is curious that the best known and most compelling arguments for the use of multiple working hypotheses in historical geology come not from those studying deepest time, where observational evidence is scarcest, but instead from Quaternary geomorphologists (Baker, 1999;Chamberlin, 1890;Gilbert, 1886;Johnson, 1933;Schumm, 1991).One might think that those studying the Precambrian would have the greatest incentive to keep all viable hypotheses in play, particularly given the diminished value of uniformitarianism as a guiding interpretive principle (e.g., Engels, 1939;Keller et al., 2017), but our history suggests otherwise.One plausible explanation is that the relative abundance of preserved Quaternary evidence trains investigators of such youthful features to expect a unique solution to a geologic problem.In such a mindset, alternate ideas may appear less threatening to achieving the satisfaction of a solved problem than when evidence is spotty…or non-existent.
A faction of the philosophy of science discipline has long argued that the conduct of science is little different from other social activities (e.g., Feyerabend, 1975;Kuhn, 1962).If true, then a second possible explanation for deep time geologists' continuing failure to prosecute a "scientific method" is that there isn't such a one-size-fits-all practice (see the quotations with which this paper began).Schumm (1991) argued that the diversity of challenges faced by Earth scientists precludes a single scientific method.Instead, our efforts can be understood in context of a "scientific approach."The features of this approach are the absence of bias, an intellectual honesty, and an avoidance of metaphysics within a systematic investigative framework.You may have, as I did when I first read his proposal, thought that Schumm's (1991) approach isn't just limited to scientific matters but also describes how you'd want your physician or auto mechanic to conduct their professional life.Chamberlin (1890) anticipated this conclusion-and our currently fraught socio-political environment-in stating that application of multiple working hypotheses "to the affairs of social and civic life would go far to remove those misunderstandings, misjudgments, and misrepresentations which constitute so pervasive an evil in our social and our political atmospheres."The notion of a one-size-fits-all "scientific method" that separates us from other rational avenues of human endeavor is an illusion that prevents those within and outside the realm of science from recognizing the unity of all diagnostic investigations.While we scientists have a shot at actually extending human knowledge, the vast majority of our time spent in "puzzle-solving" (Kuhn, 1962) mode is effectively indistinguishable from, say, judicial, mechanical, or medical analysis.What deep time geology needs isn't more and deeper philosophic ruminations about what makes us special; it needs rational, knowledgeable, fair-minded, and accurate observers and interpreters of the meager and almost certainly biased geologic record.

Early Earth: When Ruling Theories Ruled
I earlier laid out how speculations of early Earth history remained largely static for more than a half century, and now focus on the 20 years that postdate the introduction of the Hadean "waterworld" hypothesis and that axiomatically is the generation in which we've enjoyed the most sophisticated observational capabilities.I emphasize first that our knowledge of the growth history of continental crust is remarkably sparse, largely reflecting the lack of a pre-4 Ga macroscopic rock record and the unknown amount of recycling of that crust back into the mantle over Earth history (see review in Harrison ( 2020)).That is, our current geochemical knowledge is equally consistent with between ∼1 and 100% of the mass of continental crust today being present at 4 Ga (see review in Korenaga (2021)).Two variants of the "ruling hypothesis" (i.e., of an early and sustained mafic crust) have flourished in this ambiguous environment over the past generation, each reflecting the nature of the available analytical tools.The first emphasizes the lack of a mechanism to make early felsic crust and the second the lack of it.
The first is the domain of isotope geochemists studying short-lived (now extinct) radioactivities in mantle samples.Take, for example, 142 Nd which forms from the decay of 146 Sm (t ½ = 103 Ma).Whether the Hadean eon saw a form of mobile (i.e., Earth today) or stagnant lid (i.e., Venus today) tectonics has first order implications for Earth's thermal evolution (e.g., Korenaga, 2021), the growth of continents (e.g., see review in Harrison, 2020) and climate evolution (e.g., Foley, 2019).Caro et al. (2017) interpreted anomalous 142 Nd in Archean rocks from northern Quebec as evidence of the stabilization of Hadean plates whose long-term immobility implies inefficient lithospheric recycling and thus stagnant lid tectonics throughout that eon.Others have made similar arguments on the basis of 142 Nd signals preserved in ca.2.7 Ga (Debaille et al., 2013), 3.4 Ga (Roth et al., 2014), and ca.3.7 Ga rocks (Saji et al., 2018).Such papers are co-dependently fed by mantle convection modeling that claim to predict a transition from stagnant to mobile lid tectonics from first principles (e.g., O'Neill & Debaille, 2014;cf. Korenaga, 2021).Oddly, none of these publications offer an explanation for how Hadean zircons could have acquired the full range of their geochemical characteristics in such an environment.So, does mantle isotopic inhomogeneity uniquely require long-term sluggish mantle circulation and thus the lack of mobile lid tectonics?In fact, plate tectonics doesn't imply a singular degree of mantle convective vigor (Rosas & Korenaga, 2018) and the billion-year-scale for the apparent disappearance of 142 Nd signals in mantle rocks could possibly reflect crustal recycling (e.g., Armstrong, 1991;Korenaga, 2021).Furthermore, significant isotopic heterogeneity observed in the modern mantle (Zindler & Hart, 1986)-developed in the presence of plate tectonics over at least the last billion years-belies the assertion that long term preservation of mantle isotopic domains is exclusive of mobile lid behavior.Indeed, more recently, Hyung and Jacobsen (2020) found that 142 Nd data from ancient rocks are consistent with a mantle stirring time of about 400 Ma since the early Hadean, suggesting that Earth's thermal and chemical evolution may have been regulated by plate tectonics for most of its history.
A second form of ruling hypothesis is practiced by trace element geochemists and petrologists interested in the evolution of continental crust-or at least in demonstrating its late emergence.In this interpretive mode, the absence of preserved evidence of a process in the rock record is seen as evidence of absence (e.g., Brown, 2006;Taylor & McLennan, 2009) rather than due to preservation biases (e.g., Morgan, 1985) or secular changes (e.g., Palin & White, 2016).A variant of this attitude, informed in part by the problematic interpretations of 142 Nd data noted above, includes asserting certain knowledge of the timing of tectonic transitions in the absence of uniquely interpretable evidence (e.g., "No one method should be used to identify ancient tectonic settings, and this is especially true for rocks older than 2.5 Ga, when Earth was changing from stagnant lid to plate tectonics"; Condie, 2015).This appears to go a step beyond a myopic focus on the ruling hypothesis to something approaching revelation.

Culpability Assessment
So, when did we get into this mess?T.C. Chamberlin (1890) was clearly exasperated by his colleagues' narrowmindedness over 130 years ago.His lack of specific examples, however, leaves us to only speculate what, if any, behavioral similarities there were to the concerns I've documented here.I think there is a case to be made that in the "modern" era, Harold Urey's series of publications in the 1950s (Urey, 1951(Urey, , 1952a(Urey, , 1952b(Urey, , 1955) ) arguing that Moon is an undifferentiated carbonaceous chondrite provided a template for applying sheer speculation based on physical principles to deep time processes.His "ruling hypothesis" did not long survive the return of lunar highland samples, which provided evidence of an early magma ocean (see Mitroff, 1974 for a sense of how high feelings were then running against an unnamed senior scientist that was almost certainly Urey).But ironically, given Urey's influence in persuading post-Sputnik American governments that a sample return mission to Moon held the key to the origin of the Solar System and perhaps life (Brush, 1982), it's arguable that his poorly constrained conjecture ultimately fueled our current understanding of lunar origin.I asked earlier if those publications spanning the past 60 years that I singled out for their premature support of the protracted terrestrial inferno scenario were scientific.Taken individually, speculating about the past from physical principles alone (such as Urey did) seems fair game provided their underlying assumptions and all contrary evidence are clearly identified.Indeed, inductive reasoning is part of the classic prescription for generating new hypotheses.But viewed together, these authors weren't so much generating new ideas as simply proposing variants of a single, sustained, ruling hypothesis.
Looking at my own motivations, I was struck by Giere's (1988) observation that the "virtues of multiple working hypotheses tend to be advanced by scientists attempting to gain a hearing of a new view," suggesting that my endorsement of Chamberlin's (1890) method might be self-serving.In retrospect, my early advocacy of Hadean plate boundary processes (e.g., Harrison, 2009) does come across as narrowly championing an underdog model that the author felt needed the publication equivalent of a signal flare.Later (Harrison, 2020; Chapter 8) I systematically evaluated all published hypotheses that attempt to explain the geochemical features of Hadean zircons in terms of an origin-in Icelandic-type rhyolites, mafic igneous rocks, impact melts, sagduction, heat pipe tectonics, terrestrial KREEP, intermediate igneous rocks, and convergent plate boundaries.That effort was complicated by a tendency of some (see Harrison, 2020, p. 281 last paragraph) to ignore entire lines of evidence (a phenomenon interpreted in a generous fashion by Chamberlin, 1890).While I found only the latter two mechanisms to be consistent with the geochemical evidence, more recent work on Hadean zircons from southern Africa (Drabon et al., 2021) suggest their derivation under hydrous conditions more directly from the mantle.This interpretation of course doesn't impact the hypothesis that other Hadean zircons formed in arc-like settings (e.g., Carley et al., 2022) but rather suggests a planet more similar to the breadth of magmatic settings we see on modern Earth.More importantly, it potentially permits additional hypothesis generation and testing of their relative importance with regard to crust generation.To me, the most remarkable thing about attempts to explain Earth's formative era is not their accuracy; it's that it can be done at all given the lack of a macroscopic rock record representing its first 500 Ma.

Concluding Remarks
My journey to understand what attracted our community to the hellish early Earth scenario in the absence of any supporting observational evidence has led me to see science as a very human construct that is strongly influenced by perceptions of social order and subjective selection (Kuhn, 1962).How else to explain why less than half, and often much less than half, of all psychology, cancer, and pharmaceutical research studies can be replicated (e.g., Aarts et al., 2015;Begley & Ellis, 2012;Ioannidis, 2005;Prinz et al., 2011) with human fallibility often a significant contributing cause (Collins & Tabak, 2014;NASEM, 2019).My criticism here is not that the lack of a definable scientific method supporting geological research necessarily limits the value of its products, because we've come to an extraordinary understanding of Earth doing whatever it is we do.My concern is that our inability to signal the difference between a meaningful advance in our understanding of deep time and an overstated interpretation of an incremental addition to knowledge confuses the broader community and erodes efficiency (see Lenardic et al., 2022).How much more would we know about the world today if we had not uncritically adopted, for example, the protracted hellscape paradigm?And this criticism holds even if that scenario ultimately proves to be correct.How many young scientists would have been inspired to pursue alternate explanations for lunar formation, or the origin of continental crust, or the lunar bombardment history had we not told them that these were solved problems?How can we help ourselves out of this unvirtuous cycle?The limitations imposed on us by a profoundly incomplete rock record likely strewn with preservation biases may require more than multiple working hypotheses to overcome.Perhaps a system tailored to the paucity and type of evidence available is needed.In the absence of general agreement on what that should be, Harrison and Lenardic (2022) promulgated etiquette in which to frame speculations about earliest Earth history.They focused on the tendency of geologists to look past the possibility that preservation biases may have skewed the rock record to a limited number of refractory settings and their assumption that if a rock type has not yet been documented in ancient terranes, it and the processes that made them never existed (e.g., Brown, 2006;Condie, 2015;Taylor & McLennan, 1995, 2009).Their system emphasizes the importance of minimalism (i.e., giving weight to simpler models that are more easily refutable) and specifically "that the base level tectonic evolution model, the null to be disproved before more complex models are invoked, is continuous plate tectonics" since global silicate differentiation (i.e., "Burke's Law").Hopefully their call will inspire a community discussion that balances the sometimes rival competitive and cooperative aspects of science (Lenardic et al., 2022) and provide a uniform and more satisfactory style of communication to the general public than presently encouraged by weekly, international science publications.In an era that otherwise appears to value disrupters, it is surprising how little of that fashion has crept into the realm of science.But every discipline endures varying degrees of reactionary behavior and I should make it clear to the reader that my career didn't unduly suffer from my outsider status with respect to the once ruling hypothesis (although had I belonged to the previous generation I suspect I'd be as disaffected as I sense Dick Armstrong was at his premature passing).But this doesn't mean that others haven't been harmed or will be moving forward if poorly reasoned arguments about deep time continue.Perhaps the greatest value of this article is to inform young scientists that, despite 50+ years of serial, monotonous assertions about early Earth, we remain profoundly ignorant of its evolution and require their talents and open minds to truly understand the darkest age of our planet's history.

Figure 1 .
Figure 1.The author in 1973 shown with the JEOL JX-5A electron microprobe he was employed to operate by the Anglo-American/De Beers research lab located southwest of Johannesburg, South Africa.