Attack of the Clones


Woolly Mammoths, Charles R. Knight, 1915

by Leonard Finkelman

I need two things to start my average weekday. One of them is coffee. The coffee, of course, goes into a mug [1]. Mugs reflect our deepest-held values, proudly displaying the logo of a faceless corporate monolith or the title of that conference that you kind of remember attending two jobs ago. My mug features a mural of endangered animal species overlaid with some text. The text reads: EXTINCTION IS FOREVER. The second thing I need to start my day, you see, is a bit of light philosophy.

After I had finished my coffee this morning, a friend of mine shared a link to an article about the recent discovery of woolly mammoth blood. “We’ll probably have new woolly mammoths soon,” he noted optimistically. I glanced over at my mug and sighed. “Time to go to work, old friend,” I said, being the sort of person who talks to his mug when the mug is the only thing that truly understands him.

The article that my friend shared is the latest fuel to fire the current fad that’s sweeping the scientific community: “de-extinction.” This is actually the fad’s second wave: de-extinction was all the rage twenty years ago, too, even if it wasn’t yet being called by that name. (If only we could find some cause that might explain this behavior.) The first wave did not yield many tangible results; at least, I’m personally unaware of any zoos proudly displaying their mammoths. You might think that this second wave stands a better chance of success because of advances in theory and technology in the past decade. There’s a sense in which you might be right, which is why this wave of de-extinction seems to be attracting so much more attention than the last.

Unfortunately, there’s another — better — sense in which you’d be wrong to be more optimistic for de-extinction’s current prospects. It doesn’t work, it can’t work, and it shouldn’t be made to work. I do not choose my mugs lightly.

Those are admittedly some bold claims I’ve just made. I’m a philosopher, that’s part of my job. The other part is justifying the bold claims. This, then, is my plan: I’m going to explain what de-extinction is and then I’m going to explain why it’s problematic — on a practical level, on a theoretical level, and on a normative level. Then I’m going to have another cup of coffee.

Paleobiologist David Raup once estimated that more than 99% of the biological species that have ever lived are now extinct. To get some of those species back would be a boon to science — either by providing valuable new information [2] or by assuaging our collective guilt for those times we’ve been responsible for their extinction. Given advances in our understanding of genetics and development, along with attendant improvement of technology, this “de-extinction” of extinct species is (supposedly) possible [3].

The technology spurring de-extinction’s second wave of interest is a sort of reverse engineering. This method depends on one of the insights of evolutionary developmental biology: that the evolution of biological form can sometimes be derived from changes in the timing or composition of genetic switching mechanisms. Consider the development of bird wings (quick and dirty version™). Normally, the cells in a developing tetrapod forelimb have genes that direct the development of digits in five places; birds have those genes switched off in the places that would normally correspond to the ring and pinky fingers [4]. After the three remaining digits develop, another set of genes directs those digits to fuse together, forming the familiar melty-looking bird hand. The fusion switch is normally turned on during embryonic development, but is delayed past hatching in the species Opisthocomus hoazin, resulting in chicks with distinct fingers and claws. What this means is that we can reverse engineer dinosaur claws from, say, chicken wings: just switch on or off the relevant genes directing digit development and suppress that fusion switch indefinitely. Similarly, birds have structural genes that would code for long tails, but these genes are switched off early in development resulting in the stubby pygostyle; you can draw your own conclusions about what would happen if we switched those genes back on. Paleontologist Jack Horner drew those same conclusions and now proposes that the key to creating new non-avian dinosaurs is to engineer what he calls the “chickenosaurus.” Similarly, if an extinct species differs from an extant sister species as a result of regular developmental differences, or at a few recognizable genetic loci, then scientists could recreate the extinct species’ genome by the same method [5].

More “traditionally,” the primary method of de-extinction would be somatic-cell nuclear transfer (SCNT). SCNT is the way an insecure academic says “cloning” when he wants to sound impressive at a party; hence this essay’s title [6]. Most famously popularized by Mr. DNA, SCNT is a process that many people understand to have three steps: first, assemble DNA; second, Science!; third, collect goats to feed your newly-grown T. rex [7]. Unsurprisingly, the actual process is somewhat more elaborate: there’s quite a bit that gets packed into the vague and ambiguous second step. The full genetic sequence isn’t simply a book of instructions; the instructions are broken into chapters, i.e., chromosomes, and the division of those chapters can make an important difference to the effect they have. Assuming that the geneticist can correctly determine how to divide the sequence into chromosomes, those chromosomes must then be placed into a cellular nucleus along with the various enzymes and organelles that power DNA transcription. The nucleus is also insufficient: DNA contains instructions for protein synthesis, but those proteins get synthesized by cellular components that lie outside the nucleus. To get development going, the geneticist must borrow a donor egg from a member of the same species or a closely-related sister species, remove that egg’s nucleus, and replace it with the cloned nucleus (hence the “nuclear transfer” part of SCNT). The egg then gets zapped with a bit of electricity to trigger the start of cellular processes, after which point it can (in principle) be considered a viable embryo. Embryos, of course, don’t grow in test tubes; they have to be placed in the proper developmental environment, i.e., a mother. If an appropriate surrogate can be found (again, from the same species or a closely-related sister species), then the embryo is implanted in that surrogate and embryonic development can (hopefully) proceed apace. Only then, after all the implanting and hoping and spending of grant money, can anyone proceed to Mr. DNA’s third step.

To be sure, each form of de-extinction faces practical difficulties. Reverse engineering can only work on relatively young species: the extinct species and the extant sister species (from which the extinct species is re-engineered) must have genomes that are largely identical, meaning that the taxa cannot have diverged very long ago. SCNT requires not only a complete genome, but also information about chromosomal divisions (which may not be easily inferred from the genome itself) and extrinsic developmental inputs; even given all that, the extinct species must have extant relatives that are similar enough to bear one of the extinct species’ embryos successfully. Recent attempts to clone the extinct species Thylacinus cynocephalus have failed for these reasons. And even if attempts in the near future prove more successful, most clones could be nothing more than lab animals or zoo curiosities since the cost of breeding a sufficient number of organisms to keep the de-extinct species above a minimum viable population size are prohibitively expensive.

The march of scientific progress will eventually trample these practical obstacles. Some — perhaps many — organisms will remain beyond our capabilities to recreate, but I’m optimistic that we’ll someday see living birds that look for all the world like dodos or shaggy (woolly, if you will) elephants. Even so, I don’t think humans will ever again see dodos or woolly mammoths, because the problem with de-extinction is not a practical one. The problem lies on my mug.

Extinction is forever. Even the lightest philosophy can be a burdensome load.

Here’s an embarrassing secret that few biologists admit: for all our worry over it, “extinction” is a very poorly-defined term. On the one hand, extinction intuitively accompanies the death of the final member of the species, or endling. On the other hand, extinction is the species’ analogue of an organism’s death. But these two definitions are inconsistent. The death of an organism follows from breakdown of that organism’s functional integration; parts of the organism, such as individual cells, may survive beyond the clinical time of death. By this standard, endlings (or at least endlings in sexually-reproducing species) would be surviving members of already-extinct species. So if extinction is analogous to death, then our intuitions about extinction are wrong; if our intuitions about extinction are right, then extinction is in a sense worse than mere death for the species.

Extinction is so poorly defined because it is a property of species, and neither biologists nor philosophers can agree on a single species concept. In particular, the permanence of extinction depends on whether or not species are natural kinds.

Because natural kinds are defined by essential properties, a natural kind may reappear after its last representative perishes. Gold is a natural kind, defined by the atomic number 79; we could destroy every gold atom currently in existence, but the kind would reappear as soon as fusion processes generated another atom with 79 protons in it. If species are natural kinds, then they could reappear following extinction.

The conceptual problem with de-extinction is that there are very good reasons to deny that species are natural kinds. Members of a species are fundamentally variable; variation is the fuel necessary to power evolution by natural selection. With few exceptions, biologists and philosophers therefore deny that species have essences, preferring instead to conceive species as nominal groups bound together by ancestry. It is for this reason that Darwin wrote in the Origin: “When a group has once wholly disappeared, it does not reappear; for the link of generation has been broken.”

Paleontologist Louis Dollo proposed that the irreversibility of extinction (along with the irreversibility of single-trait evolution) ought to be considered a law of evolution. Dollo’s “law of irreversibility” has a number of different interpretations, all of which have been fiercely debated. But I think we all ought to agree on at least one point: de-extinction does not turn back the evolutionary clock. Both reverse engineering and SCNT rely on the modification of extant genomes and developmental environments to create organisms that bear a very strong — perhaps even exact — resemblance to organisms from an extinct species. But to qualify this as the resurrection of an extinct species is as silly as claiming that my (alas, hypothetical) clone and I are numerically identical [8].

It seems, then, that the de-extinction of species that evolved by natural selection is a logical impossibility [9]. In my capacity as a philosopher I therefore consider the case closed. But most people aren’t philosophers, and refutations from logical impossibility rarely carry the weight that I would hope. In presenting the above argument to friends who work in the life sciences, I’m inevitably confronted with either of two responses. Some say, “you’re thinking too much!” Some say, “this is just semantics!” So I have to admit that something beyond conceptual analysis is probably necessary to get the (relative) masses off their de-extinction high.

Fine. Let’s grant that at least one of the two de-extinction technologies can produce viable organisms in sufficient numbers to support a natural population, and that this population is in fact the same species as one that had previously been extinct. Let’s suppose that we will actually have new woolly mammoths tramping about the arctic tundra. Even then, the pursuit of de-extinction would still be prospecting for fool’s gold.

Conservationists are particularly worried about de-extinction. They should be, too. One of the primary justifications for the conservation of endangered species is that we can’t get back what we’ve lost. If de-extinction becomes commonplace enough, then there is less reason to preserve natural habitats, or to combat global climate change, or to alter ecologically harmful dietary habits. This is all true enough, I think. And derivative effects are potentially disastrous: even relatively small, temporary changes to an ecosystem can do irreparable harm, and the extinction of even a single species is potentially a horrendously big change.

These concerns are primarily political and could be resolved politically. Certainly, the cost of conservation is currently lower than the cost of de-extinction, the latter of which has been estimated in the high tens of millions of dollars for just a single viable organism. The average Joe may not worry about the extinction of species that can be resurrected, but one hopes that the Joes who actually wield power might recognize the problems just noted.

For my part, I think that the greatest problem with de-extinction is that it offers little actual utility to science. Consider the example of Dolly, the most famous of non-Star Wars clones. The cloning process that yielded a single viable sheep also produced hundreds of ultimately inviable embryos, and even the one viable organism that resulted — Dolly herself — suffered from a number of aberrant physical and behavioral problems. We could identify these problems because we have extant sheep against which we could compare Dolly. How could we ever know if a cloned mammoth behaves as extinct mammoths did, or if it is developing within the same parameters (growth rate, intellectual development, etc.) as previous mammoths? We cannot: we have no standard of measurement since there are no extant mammoths. De-extinct species have limited scientific utility simply because information gleaned from the de-extinct species cannot justifiably be extrapolated to the extinct species from whence it came.

Heck: de-extinction might actually harm science. As noted above, the cost of de-extinction is currently astronomical and would likely yield a single individual that can serve as little more than a scientific novelty. Funding in the sciences has become increasingly limited: the tens of millions of dollars spent on that work is tens of millions of dollars not being spent on research that might produce useful new technologies or important theoretical advancement. Given all of the problems noted above, to draw resources away from other scientific pursuits in favor of de-extinction only slows the progress that we might make otherwise [10].

I don’t know if I’ll ever get to see a live mammoth. The biophile in me hopes that I will. I have literally dreamed — repeatedly! — of seeing extinct animals in the flesh. But if I hope in one hand and hold my coffee mug in the other, the latter is the only one that will get me through any given morning.

Piece originally published at Rationally Speaking | Creative Commons License


[1] The mug is not the second thing because it is implied by the coffee. I have just enough sense not to stick my open mouth below a brewing spigot.

[2] Just how did dinosaurs, y’know, do it? Imagine the special issue of Cosmo we’re missing out on!

[3] Let’s get this out of the way: “de-extinction” is a ridiculous term, saddling an advanced theoretical concept with the sort of name a third grader would give to an ad hoc superpower used to avoid losing a game. My proposal is to go with “extantion.” The species that populate current ecosystems are referred to as extant species, but the process by which those species naturally become extant is speciation. “Extantion” — literally, “to make extant” — is just sitting there unused, waiting to serve as a linguistic counterpoint to “extinction.” It has the added virtue of sounding like “extension,” which is what we’d be doing for the lifespan of formerly-extinct species. Also, I came up with it.

[4] In the interest of full disclosure, it’s worth noting that many embryologists believe that the “reduced” digits in the bird hand are the first and last (normally labeled I and V), rather than the last two (IV and V). The debate over the exact numbering of bird digits is extensive, but largely beside the point here.

[5] To be clear: chickenosaurus would not be a de-extinct species. It would be an entirely new species, if one created by genetic engineering rather than by natural selection. De-extinct species resurrected by reverse engineering, however, are ostensibly a different story.

[6] It is certainly not meant to be associated with a different movie. How could it? Any such movie would be so bad that I could no longer acknowledge its existence without suffering residual mental anguish, obviously.

[7] A fourth step — run and scream — is optional.

[8] The analogy is made even stronger by the fact that many philosophers of biology consider species to be nominal individuals rather than nominal groups, in which case Mammuthus primigenius and a cloned woolly mammoth would be precisely like me and my hypothetical clone. I’m no fan of the individuality thesis myself, but it does drive the point home: resemblance does not make for identity, especially given separation in space and time.

[9] Which is the worst kind of impossibility, when you get right down to it.

[10] And this is to say nothing of what might happen if any of those de-extinct organisms ever learn how to open doors.