* This is the first installment in a new series of short essays called “Problematica.” It is written by Max Dresow…
America’s great confessional poet, Anne Sexton, once wrote that “Nature is full of teeth | that come in one by one, then | decay, fall out.” And thank goodness, too, for without this steady supply of teeth our knowledge of the history of life would be greatly impoverished. The fossil record is mostly a record of hard parts, of which the teeth of gnathostomes, or jawed vertebrates, are exemplary. Variously long and dagger-like, or flat, or conical, teeth are the most common element in the vertebrate fossil record. Yet they are also among the most informative. Many extinct creatures are known only from their teeth, as if commanded to leave behind a single part, they chose the one that would reveal the most about their life and times. Several months ago, the earliest gnathostome teeth ever described were announced in Nature. These teeth, or more properly, tooth whorls, look a bit like Sonic the Hedgehog entering a barrel roll. But their significance lies in their age: about 439 million years, or 14 million years older than any previous evidence of vertebrate jaws. This extends the reign of teeth nearly all the way to the evolutionary riot known as the Great Ordovician Biodiversification Event. “In nature nothing is stable,” Sexton writes, but a tooth whorl could give the fleeting world a run for its money.
Volumetric reconstruction of a tooth whorl from the Silurian gnathostome Qianodus duplicis. The entire whorl is only about 2.5 mm long. It is the only part of the animal that is known. Credit: Zhu et al.
Teeth are hugely interesting, but they are not my main concern here. Instead, I am concerned with two sets of inferences involving teeth, which together illustrate the power and pitfalls of the actualistic method. This is the method of using the present as “the key” to the past: of using observations of living animals to interpret fossils, say. It has long seemed the best, and perhaps the only, way of reconstructing the denizens of deep time. But for all its incorrigibility, it has proved a ticklish thing to use. Much hinges on finding the right comparisons, but comparisons between the present and the past are always approximate and are never foolproof. Sometimes they require supplementing. Other times they require correction. Always they involve an element of characterization, which can be controversial. You see the issue. With so many moving parts, the actualistic method needs to be applied carefully if it is not to lead up blind alleys.
The two examples I will discuss are separated by nearly two hundred years. The first is Gideon Mantell’s description of a large reptile from skeletal remains (including a puzzling set of teeth) in southern England. The second is a recent study of Otodus megalodon: an animal known mostly from its iconic chompers. Unsurprisingly, the more recent study reaches the more credible conclusions, but the purpose of this exercise is not to moralize about history. It is rather to demonstrate the persistence of a mode of reasoning and to explore how paleontologists navigate the risks associated with actualistic comparison. It is also to say, in simple words, what actualism really amounts to: a subject that has recently reared its scholastic head in the philosophical literature.
Example 1: The good doctor describes a monster
Gideon Mantell was a provincial doctor in Sussex, keen on establishing his credentials as a geological authority. So when he discovered teeth “of a very singular character” deep in the Secondary formations (today’s Mesozoic strata), he moved to capitalize on them (Mantell 1822, 54). Mantell suspected that the teeth belonged to an herbivore; and when Charles Lyell presented one to the master naturalist Georges Cuvier, the latter suggested that they may have belonged to a rhinoceros. But this was a remarkable suggestion. No large mammals were known from the Secondary period, which was then coming to be regarded as an “age of reptiles” preceding the “age of mammals.” Finding a large mammal in the Secondary formations would thus have been an anomaly, and one with serious implications for the practice of stratigraphy and the developing understanding of life’s history. But there was a second interpretation. Perhaps the teeth did not belong to a mammal at all but to a reptile. This was Mantell’s hunch, and it was hardly less remarkable, since herbivorous reptiles are rare, and anyway, all the reptiles that had been recovered from the Secondary formations were carnivores.
An Iguanodon tooth, which, according to most accounts, Mary Mantell collected in 1822
To show that the teeth belonged to an herbivorous reptile, two things needed to be done. The first was to resolve the stratigraphy of the so-called “Tilgate beds,” and as it happened this was rapidly accomplished, in part through Mantell’s joint fieldwork with Lyell. The second was to recover more fossil material, preferably enclosed in hard rock (as opposed to eroded out and thus potentially sourced from younger formations). This too was quickly done, by the tried-and-true method of bribing quarrymen. Among the fruits of this bribery were enough teeth to arrange in a growth series, which suggested that the flat crowns of the larger teeth were acquired by wear and tear in the manner of a plant-grinding herbivore. No less important, Mantell found evidence that the animal replaced its teeth during its lifetime. This suggested that it belonged to the “tribe of Lizards,” since living reptiles, but not mammals, replace their teeth as adults. The case for a giant herbivorous reptile in the Secondary formations was firming up. Even the great Cuvier was made a convert.
But what did this animal look like? How did it move? And, not the least important thing in an age enamored of spectacle, how big was it? To get a handle on these things, Mantell needed a standard of comparison: something that could ground the relevant inferences. Basically, he needed an analog for his animal in the contemporary fauna. And he found it—or rather, it was suggested to him by an assistant at the Royal College of Surgeons—in the form of an iguana. Iguanas are herbivorous reptiles, and crucially for Mantell’s purposes, they have “teeth possessing the form and structure of the fossil specimens” (Mantell 1824, 182). But iguanas are no one’s idea of behemoths. Going just by the teeth, Mantell’s animal was about twenty times larger than the iguana skeleton in the collection of the Royal College of Surgeons. So, taking this at face value, he inferred that the teeth “must have belonged [to] an individual upwards of sixty feet long,” a conclusion, he observes, “in perfect accordance with that deduced by Professor Buckland for the femur” (185). Later he would modify this estimate to 100 feet on the strength of a newly discovered thigh bone “no less than 23 inches in circumference!” (Mantell 1827, 77). Thus was born a creature “three or four times as large as the largest crocodile; having jaws, with teeth equal in size to the incisors of the rhinoceros; and crested with horns; such a creature must have been the Iguanodon!” (Mantell 1827, 83).
An iguana juxtaposed with its namesake, the Iguanodon. The image on the bottom comes from a famous watercolor, “The Country of the Iguanodon,” painted by the popular English artist John Martin (1837). Martin’s rendering of the Iguanodon would go exert a large influence on early representations of prehistoric life
Example 2: Jaws!
Sadly, the iguanodon was not a hundred foot-long behemoth. Current knowledge suggests that individuals are unlikely to have surpassed 40 feet in length: a perfectly respectable figure, but hardly the “extremity of a lizard” its discoverer had imagined. New fossils played a key role in this revision, in particular numerous complete and articulated skeletons unearthed at Bernissart, in Belgium. However, investigators are not always so fortunate in the provision of skeletal materials.
Take the megalodon. Like most sharks, this aquatic extremity is known primarily from its teeth. Tall and serrated, these nightmarish objects are staples of the collector’s market. A small one will set you back only about 60 American dollars, so you will have guessed that they are not exactly rare. Yet what is exceedingly rare is any part of the animal that is not a tooth. Even its iconic jaws, which frame a small huddle of people in numerous photographs, are only best guesses. The most famous entry in the genre has scientists from the American Museum of Natural History posed uncomfortably within the maw. But look closely, and you will see that the jaws are ringed by teeth of a uniformly large size. This means that the reconstructed gape is almost certainly too big. Likewise the estimate of the animal’s length: a staggering 98 feet, arrived at by scaling up the proportion of a living white shark to match the reconstructed gape. (For comparison, the shark in Jaws is about 25 feet long.)
Subsequent decades have seen many more attempts to work out the megalodon’s length on the basis of its teeth. And like the first, nearly all have been based on data from white sharks. For example, in 1973, John Randall used measures of “enamel length” (the distance from the base of the enamel to the tip of the tooth) to arrive at an estimate of 43 feet based on a comparison with great whites. Two decades later, Michael Gottfried and colleagues gave an estimate of 50 feet based on a formula relating a white shark’s length to the length of its longest tooth. (The longest megalodon tooth then known was 6.6 inches. We now possess teeth a full inch longer, giving a length estimate of 66 feet.) More recently, Kenshu Shimada has cast doubt on whether any of these approaches yield reliable estimates. Working from painstaking studies of white shark dentition, Shimada claims that white shark teeth are generally too variable to permit a linear relationship between a proxy variable and body length to be written down. The only exception, it seems, are measures of enamel length taken on the first two anterior teeth of the upper jaw. So, after analyzing data from these tooth positions and performing some statistical corrections, he estimated that O. megalodon grew to a length of between 46 and 50 feet. Poor megalodon seems fated to be pumped up and down like an air mattress, but perhaps Shimada’s work marks the onset of stabilization.
So far I have written as if megalodon is known entirely from its teeth, but this isn’t quite right. In addition, two vertebral columns have been found, the more complete of which hails from—where else?—Belgium. Consisting of 141 vertebrae, this specimen has recently been used to construct a three dimensional digital model of the animal. To do this, researchers first scanned individual vertebrae, which they subsequently arranged into a vertebral column based on described intercentrum distances in living sharks. Then they built a model of the jaws from tooth scan data, again using descriptions of living (white) sharks to infer intertooth distances. Finally, the team performed 3D scans of two living great whites “to digitally reconstruct the body outline of O. megalodon” (Cooper et al. 2022, 9). This is not because white sharks are believed to be the descendents or sister taxon of O. megalodon. (They were, but not anymore.) Instead, the stated rationale is a mixed-pragmatic one: the great white “is the largest, most well-studied species, and has the most similar dentition to O. megalodon.”
A representation of the modeling procedure used by Cooper et al. to construct their megalodon. The top row shows some vertebral centra (A) positioned above the reconstructed vertebral column (B). (C) shows a sample of teeth, which were scanned (D) along with the chondrocranium of a great white (E) to build a model of the megalodon’s head (F). Finally, a 3D scan of a great white’s body (G)–(K) was used to build the complete megalodon model (M)–(Q). Credit: Cooper et al.
Of these criteria, dental similarity is doubtless the most important. This is because it is dental morphology that establishes the plausibility of the white shark comparison, and with it a whole string of inferences that would not otherwise have been possible. With this comparison in tow, the model is able to make predictions on matters ranging from body mass (about 62,000 kilograms) to cruising speed (faster than any living shark) to feeding ecology (“transoceanic super-apex predator” status). All these depend critically on data from white sharks, and while the team was careful to employ “conservative estimates and cautious interpretations,” the model remains in thrall to its assumptions (Cooper et al. 2022, 8). “Garbage in, garbage out,” the old saying goes. Or to modify this to suit the present example, “White shark proportions and physiology in, giant white shark out.” This is the basic actualistic gambit, unchanged since the days of Mantell but applied with much greater sophistication than the good doctor could have imagined, still less mustered.
Actualism as piecemeal comparison
As I noted above, the motto of the actualistic method is “the present is the key to the past.” This has meant a variety of things to researchers since it was first uttered in 1905, but at its heart is the idea that comparisons with the present world provide a privileged way of interpreting the geological record. Only by assuming a certain constancy between the present and the past can we make inferences about history on the basis of surviving material evidence. Or, as the geologist-philosopher David Kitts puts it, “[in] terms of the way a geologist operates, there is no past until the assumption of uniformity [between the present and the past] has been made” (Kitts 1977, 63, emphasis added).
The trick has always been deciding exactly how the present resembles the past. After all, it does not resemble it absolutely. Material configurations have shifted, some causes have come into existence or ceased operating (consider human intentionality), and processes have changed in their intensity and complexion over time. To reason well, these changes need to be taken into account. But that is easier said than done. It may seem obvious to us that Mantell’s “iguana-in, iguana-out” logic is defective. Dinosaurs are not particularly closely related to iguanas, and reasoning strictly on the basis of tooth morphology ignores the possibility that rather different kinds of animals may have similar kinds of teeth. However, Mantell was writing before the taxonomic category “dinosaur” existed and when a belief in the laws of animal economy (these kinds of teeth go with these other features) had all the backing of Cuvier’s immense prestige. He could hardly have known that dinosaurs possess a number of characters not shared by living reptiles, in part because only one similar creature had been described, and this from fragmentary evidence. What this means is that the choice of a comparative model is often far from obvious. Present organisms resemble past ones only so well, and to evaluate the suitability of a model a great deal of knowledge is typically required, some of which may be difficult to gain (or in Mantell’s case, simply unavailable).
But even when a suitable comparison has been identified, one’s troubles are not over. White sharks have long been the comparison of choice for understanding all aspects of megalodon biology. However, this near-consensus on the comparative model has not made it easier to arrive at a stable estimate for overall size. Once again, auxiliary knowledge is key, like knowledge of what proxy variables are reliable, and beyond this, of what specific body parts the proxy variables should be measured on. But even when these matters have been settled, uncertainty remains; and this uncertainty is compounded whenever researchers move beyond relatively simple inferences to more complex ones. How much can we really know about the feeding ecology of O. megalodon on the basis of its teeth and some disarticulated vertebrae? Quite a bit, potentially, but these inferences are delicate and associated error bars are considerable.
In light of these difficulties, it may be worth asking whether there is a way of bypassing this complexity. Here is one possibility. At least since the 1965 edition of Arthur Holmes’s textbook, The Principles of Physical Geology, it has been a commonplace that “actualism” has to do with natural laws. Material configurations change, the thought goes, but natural laws never do. Thus, insofar as inferences about the past are based on unchanging laws they are secure. “Material evidence + natural law in, reliable reconstruction out.”
The problem is that natural laws are hard to come by, and anyway, paleontologists seem not to need them for most of their applications. What they need are local generalizations that hold for just those domains relevant to their interests. So, scientists interested in megalodons need a proxy for body length, measurable on a tooth, that holds for the entire clade comprising lamnid sharks and Otodontidae. This won’t be a law of nature unless we adopt a libertine attitude towards natural laws, but who cares? Gottlieb, Shimada and others are reasoning actualistically and by all appearances they are reasoning well. It is no argument against their practice that it fails to incorporate a law of nature.
What this means is that there is no way out of the tangle. Uncertainty and risk are baked into the actualistic method, which is a matter of muddling through as best you can with the information you can get your hands on. Lots of things can derail an actualistic comparison: not just “known-unknowns” (to borrow an expression from the former defense secretary) but also “unknown-unknowns.” The challenge for actualists is to account for these factors as best they can. Needless to say, this is hardly a trivial challenge, and so actualism must be a provisional and piecemeal affair. Yet to say that something is provisional is not to say it is unreliable. By repeatedly scrutinizing the basis of a comparison, researchers are able to give their inferences empirical teeth, rejecting faulty analogs and narrowing-in on relationships of relevance. Iterative application, then, is the key to actualistic reasoning, and in its characteristic dynamic of failure and adjustment lies much of the drama of the history of paleontology.
* * *
“As I write this sentence | about one hundred and four generations | since Christ, nothing has changed | except knowledge… ” Sexton’s poem is a somber meditation on death and how we should face it. But while we’re here, let’s take pleasure in the fact that in only a few generations, we’ve cut the mighty iguanodon and fearsome megalodon nearly in half. Knowledge indeed.
References
Andreev, P.S., Sansom, I.J, Qiang Li, et al. 2022. The earliest gnathostome teeth. Nature 609. https://doi.org/10.1038/s41586-022-05166-2.
Cooper, J.A., Hutchinson, J.R., Bernvi, D.C., et al. 2022. The extinct shark Otodus megalodon was a transoceanic super-predator: inferences from 3D modeling. Science Advances 8. https://doi.org/10.1126/sciadv.abm9424.
Gottfried, M.D., Compagno, L.J.V., and Bowman, S.C. 1996. Size and skeletal anatomy of the giant “megatooth” shark Carcharodon megalodon. In: A.P. Klimley & D.G. Ainley (Eds.), Great white sharks: the biology of Carcharodon carcharias, pp. 55–66. San Diego: Academic Press.
Kitts, D.B. 1977. The Structure of Geology. Dallas: SMU Press.
Mantell, G.A. 1822. The Fossils of the South Downs, or, Illustrations of the Geology of Sussex. London: Lupton Relfe.
Mantell, G.A. 1824. VIII. Notice on the Iguanodon, a newly discovered fossil reptile, from the sandstone of the Tilgate Forest, in Sussex. Philosophical Transactions of the Royal Society 155:179–186.
Mantell, G.A. 1827. Illustrations of the Geology of Sussex: Containing a General View of the Geological Relations of the South-Eastern Part of England. London: Lupton Relfe.
Randall, J.E. 1973. Size of the great white shark (Carcharodon). Science 181:169–170.
Shimada, K. 2019. The size of the megatooth shark, Otodus megalodon (Lamniformes: Otodontidae), revisited. Historical Biology 33. https://doi.org/10.1080/08912963.2019.1666840.
For more information about Mantell and iguanodon, see:
O’Connor, R. 2007. The Earth on Show: Fossils and the Poetics of Popular Science, 1802–1856. Chicago: University of Chicago Press.
Rudwick, M.J.S. 2008. Worlds Before Adam: The Reconstruction of Geohistory in the Age of Reform. Chicago: University of Chicago Press.
And this, from the Natural History Museum in London.
For more on the reconstruction of megalodon, see:
This nifty page, from the ReefQuest Centre for Shark Research.
And for a recent (and I hope constructive) exchange about the meaning of “actualism” in the historical sciences, see:
Dresow, M. Forthcoming. Actualism and uniformitarianism: from abstract commitments to forms of practice. Philosophy of Science. [This is a short response to Meghan Page’s paper, listed below. Here’s a link to a paywall protected version.]
Page, M.D. 2021. The role of historical science in methodological actualism. Philosophy of Science 88:461–482. (Here’s a link to a paywall protected version)