Nassim Taleb’s False Dichotomy of Top-Down and Bottom-Up Approaches to Farming With Respect to GMOs

Nassim Taleb’s False Dichotomy of Top-Down and Bottom-Up Approaches to Farming With Respect to GMOs

Stuart Hayashi has returned to the blog with a treatise on Nassim Taleb’s precautionary principle working paper.

 Stuart is a freelance writer based in Hawaii. He is the author of The Freedom of Peaceful Action: On the Origin of Individual Rights and Life in the Market Ecosystem

Are you interested in writing about a science related topic? Then check out our information on becoming a guest writer.


A working paper coauthored by financier Nassim Taleb and several other scholars has gained influence.  It argues that the creation of GMOs through DNA-splicing is substantially more dangerous than producing new plant varieties through selective breeding. The paper issues these three main arguments for that proposition.

  1. The Top-Down Argument: DNA-splicing by multinational corporations — particularly using monocultures — is a top-down process, in contrast to the bottom-up processes of natural selection and selective breeding.
  2. The Fast Evolution Argument: DNA-splicing causes organisms to evolve at a dangerously rapid rate, much more rapid than what happens in either natural selection or selective breeding.
  3. The Extreme Change Argument: DNA-splicing is more radical than any mutation that can occur in any other breeding process. Genetic engineers splice a fish’s DNA into a tomato, and no other breeding technique alters DNA in any manner as extreme.

In the essay you presently read, I rebut each of the three main arguments of Taleb and company, devoting the most space to the first argument.  Addressing each of the three main arguments respectively, I put forth:

  1. The Top-Down Rebuttal: Taleb’s distinction between top-down DNA-splicing versus bottom-up natural selection and selective breeding is misleading. The same top-down and bottom-up processes occur with natural selection, selective breeding, and DNA-splicing alike. In all cases, there is a top-down determinant: the environment itself “selects” which organisms are fit to survive in it and which are not. Conversely, in all three processes, the organism must “earn” its survival through adapting itself bottom-up to the environment.
  2. The Fast Evolution Rebuttal: Contrary to Taleb’s assumptions, there are case studies of evolution happening quickly under selective breeding, thus undermining Taleb’s claim that the rapidity of DNA-splicing renders it incomparable to selective breeding.
  3. The Extreme Change Rebuttal: Taking DNA from one species (a fish) and splicing it into another (a tomato) is shown not to be very extreme when one considers the nature of DNA. DNA is a sequence of the same four molecules — guanine, adenine, thymine, and cytosine.  Every type of genetic engineering rearranges this sequence.  If putting a fish DNA sequence into a tomato’s DNA sequence amounts to a relatively small rearrangement of the molecules, it might not be as radical a rearrangement of DNA as what takes place when selective breeding randomly sorts thousands of genes.

For these reasons, Taleb and company fail to demonstrate that DNA-splicing techniques pose a danger to human beings or the ecosystem beyond anything that commences with selective breeding.


 Since 2014, the famed financial writer and New York University professor Nassim Nicholas Taleb and a number of his colleagues have raised alarm over genetically modified organisms (GMOs) in their working paper “The Precautionary Principle (With Application to the Genetic Modification of Organisms).” The paper invokes the idea of the Precautionary Principle to assert that in the present geopolitical-economic circumstances of modern farming, the introduction of GMOs into agriculture imposes a risk no less than unacceptable, and consequently implores public policymakers to act according to this conclusion (Taleb et al. 2014a, 9, 14). Scientists have offered numerous thoughtful critiques of both this paper and Taleb’s quite emotionally charged public comments on the matter. Still, another aspect of the paper’s argumentation merits addressing — the misleading manner wherein Taleb and company classify either a natural phenomenon or farming method as “bottom up” or “top down.” The distinction between “top down” and “bottom up” is crucial to Taleb’s argument, and yet, in consideration of the wider context, this distinction is one that proves nebulous at best.

Here I shall call the technology that produces GMOs by the name of “DNA-splicing.”  DNA-splicing specifically refers to what genetic engineers do in a laboratory setting.  Namely, they splice a specific new sequence of DNA — sometimes from a different species altogether — into the genome of an organism. A new breed or variety of plant used in a domestic setting — whether it was developed through selective breeding or DNA-splicing — is known as a cultivar.

In response to the public raising alarm over biotech companies manipulating the genetic sequences of organisms, scientists commonly reply that humanity has already manipulated the genomes of other organisms since ancient times, through selective breeding. According to this understanding, the act of genetic engineers splicing a specific DNA sequence into a cultivar is simply an extension of this long tradition of genetic manipulation, except that selective breeding produces thousands of randomly sorted combinations of DNA sequences, whereas the results of DNA-splicing is more standardized and predictable. Insofar as one might fear that a new organism with a new genetic sequence may wreak havoc, scientists commonly assure those who hold such fear that DNA-splicing is more controlled and predictable than selective breeding, and therefore DNA-splicing is probably safer. In response, Taleb proclaims that the opposite is true: that selective breeding being more randomized and chaotic, which he calls “bottom up,” is what mitigates risk, whereas the more controlled aspects of DNA-splicing, which he dubs “top down,” nullify the natural forces that mitigate risk and thereby produce an unacceptable risk themselves. As I refer to the paper of Taleb and his colleagues, I use the page numbering provided in the PDF document of the paper to which Taleb’s Fooled By Randomness website links.


 The Models of Bottom-Up Versus Top-Down

According to the image projected by this paper and Taleb’s other online rhetoric, a monopolistic or oligopolistic transnational corporate agribusiness cartel or cabal — headed  by Monsanto — DNA-splices together a single model of crop, such as Bt corn or Roundup Ready soybeans, and then farmers on every inhabited continent are pressured to plant this very same standardized GMO cultivar. This business model exacerbates monoculture, which the paper defines as “the use of single crops over large areas” (Taleb et al. 2014a, 9).  Taleb refers to the usage of a single GMO cultivar throughout this “monoculture” as a TOP-DOWN process, as the corporate cartel supposedly imposes this one breed on farmers everywhere. As evidence of the ubiquity of these standardized GMO models: “In 2014 in the US almost 90% of corn and 94% of soybeans are GMO” (Taleb et al. 2014a, 9). As members of the same species within a population of crops are probably clones of one another, they lack genetic diversity, rendering an entire field of clone crops vulnerable to disease; if one stalk of GMO corn is unable to fend off a pathogen, the pathogen will probably infect and kill the other stalks, too (Taleb et al. 2014a, 9).
In contrast to the top-down nature of DNA-splicing, the paper classifies natural selection as a BOTTOM-UP process, and then it says natural selection still applies to selective breeding. Ergo, deduces the paper, selective breeding is also bottom-up (Taleb et al. 2014a, 9). The natural environment — this case, the field in which farmers plant their seeds — “selects” which cultivars are best-adapted to this very same environment.  Furthermore, Taleb’s paper goes on, the cultivar that emerges in a particular environment will be adapted only to the locality in which it was developed. It appears that under Taleb’s ideal agricultural system, there will be multiple types of crops grown near one another at the same time, and even among members of the same species of crop, there will be much genetic diversity, as they will not be clones of one another. Should one stalk of a crop be killed by a disease, other stalks nearby could still be immune against this pathogen.  In all, the main point is that if a cultivar bred through traditional selective breeding does any harm to the environment, the harm will be “localized” (Taleb et al., 2014, 1, 3-4, 9, 11), meaning the harm is confined to a limited geographic region. As the paper phrases it, selective breeding “still happens in a bottom-up way” comparable to natural selection (Taleb et al. 2014a, 9) and such “bottom-up modifications do not remove the crops from their long term evolutionary context…” (Taleb et al. 2014a, 10). Thus, if a “harmful variation” emerges through either selective breeding or natural selection, that harmful variation “will not spread” throughout the entire global ecosystem “but end up dying out through local experience over time” (Taleb et al. 2014a, 9).

On the other hand, Taleb adds, if the corporate agribusiness cartel’s standardized GMO model is adopted in multiple regions and there proves to be any weakness in the business model, the harm it causes will be globalized, meaning the harm will persist everywhere. The consequent hazards shall “spread uncontrollably,” and “their risks cannot be localized,…leading to irreversible system-wide effects with unknown downsides” (Taleb et al. 2014a, 9).

The paper adds that there are other factors compounding the hazards of DNA-splicing.  One is that DNA-splicing is a much more rapid form of evolution than what happens under selective breeding. This supposedly gives humans and the environment much less time to counteract any harmful consequences wrought by the GMOs (Argument 2: The Fast Evolution Argument).  Moreover, the Taleb paper goes on, DNA-splicing produces alterations in the gene pool that are more radical than any mutation transpiring under selective breeding (Argument 3: The Extreme Change Argument). These two arguments shall be addressed at the end of my essay.

Despite its abundance of statistical and other dry technical jargon, Taleb’s paper employs jarringly melodramatic verbiage to describe the hazards it pronounces inextricably attached to DNA-splicing. It insists that the introduction of DNA-spliced organisms into the ecosystem shall inevitably result in “total ruin” (Taleb et al. 2014a, 2) on a “global scale” (Taleb et al. 2014a, 5) — ruination that is “forever” and “infinite” (Taleb et al. 2014a, 3) and which amounts to “ecocide: an irreversible termination of life…” (Taleb et al. 2014a, 2). Taleb and company therefore conclude that any unintended adverse repercussions of the introduction of DNA-spliced organisms into the biosphere will be more disastrous than the meltdown of a nuclear power plant (Taleb et al. 2014a, 8-9, 14).

Of special note is that the paper does not fault genetic engineering, in isolation, as the great threat. Rather, it is “[m]onoculture in combination with genetic engineering…” (Taleb et al. 2014a, 9). In fact, Taleb’s paper characterizes “monoculture” per se as a house of cards waiting to collapse (Taleb et al. 2014a, 4), announcing, “Global monoculture itself is of concern for potential global harm…” (Taleb et al. 2014a, 9). Yet the paper does not focus on the ostensive weaknesses of monoculture alone. From this circumstance, one can infer that Taleb’s argument is that though monoculture is horrid for producing an inherently “fragile” system, DNA-splicing technology will provide the final shock that will devastate everything. Hence, Taleb tweets, the creation of genetically modified organisms should be restricted to applications other than agriculture and, even then, such use must be “limited.”

It is conspicuous that Taleb and his colleagues have not commented in this paper on if they think DNA-splicing technology would still be too risky if monoculture were not as prevalent in the global food system — if DNA-splicing technology were applied to farming even as farmers grew a much wider variety of cultivars and if, within a population of plants grown, there was more genetic diversity between members of the same species. Whatever the reason for this omission, it might be helpful to disabuse commentators of the assumption that the concepts of “DNA-splicing” and “monoculture” cannot properly be separated.  Each can exist without the other, and throughout much of the twentieth century since the New Deal era, one has been practiced while the other was not. Moreover, Taleb and other contemporary critics of modern agribusiness have still failed to demonstrate — despite their presumption to have succeeded at this long ago —that a continued combination of DNA-splicing, monoculture, and standardization will necessarily spell the doom they predict in the decades ahead.

I should mention that many critics of monoculture equivocate on the matter of crop rotation.  There is often the assumption that when farmers practice monoculture, that precludes them from engaging in crop rotation, which means planting a new type of crop for each planting season.  Scientist Steven Savage points out that it is common for farmers to practice both monoculture and crop rotation.  What this means is that for one planting season, the farmer plants multiple units of one type of crop and then, for the next planting season, plants multiple units of a different crop.  If a farmer planted only one type of crop during the season, and used the same type of crop for every season, that would be a “non-diverse rotation,” and yet monoculture is often critiqued as if it is synonymous with non-diverse rotation.

Standardization and Monoculture Before DNA-Splicing

On the matter of monoculture existing independently of DNA-splicing, there is already some “factory-modeled” standardization in cultivars that were developed through selective breeding techniques — the very same selective breeding techniques the Taleb paper approves as acceptable even as it decries monoculture and standardization. Most apples on the market at the time of this writing — and these are not DNA-spliced Arctic Apples — are already “standardized” insofar as they are clones of one another. Should anyone grow an apple tree, the fruits sprouting from its branches will each vary in quality. Apple growers generally prefer more uniformity — consistency in quality — in the apples they produce. Hence, they want each unit to be as much like the others as possible. One method of getting closer to this goal is ensuring that each apple is genetically identical to the others. They therefore took the apple sport they found most suitable for their purposes — everything growing from this branch containing the exact same genomic sequence — and grafted that sport onto the trunk and roots of a different tree. This has become standard practice (Fedoroff and Brown 2004, 52-53). No matter how different the trees are from one another, the apples growing on them are on branches that were cut from that original favored branch. Two apples on the market can be genetically identical despite being grown on entirely separate continents in geographically dissimilar regions. The monocultures and standardization have not instigated the long-heralded catastrophe. Still, Taleb and other critics of modern agribusiness might reply that this is but one example of global agribusiness already going too far in exercising monoculture, and that the introduction of DNA-spliced organisms into this system will merely contribute to the already horrendous fragility of the global agricultural system.

Do Top-Down Governmental Impositions Alleviate or Contribute to Problems With Monoculture and Non-Diverse Rotations?

Should any person worry that too much top-down imposition of “monocultures” in the agricultural sector might soon imperil the ability to grow enough nutritious food to sustain the population, that individual might look to the most obvious sources of top-down mandates. Insofar as there may be a problem with monocultures or standardization or non-diverse rotations, one can study the degree to which regulations and taxpayer subsidies, mandated top-down from governments and often modeled after the U.S. system, give preference to some cultivars and farming methods at the expense of others, and thereby incentivize farmers to favor monocultures and non-diverse rotations more than they otherwise might. Yet Taleb’s paper is mysteriously silent on that issue. That is interesting, as Taleb has gained an astonishing following among free-market enthusiasts, on account of his praise for Nobel laureate Friedrich August von Hayek (Taleb 2010 2d ed., 134-35, 176-182) and for former presidential candidate Ron Paul.  It might seem that Taleb agrees with free-marketers that procedural norms and best practices emerge from the bottom up in markets, concordant with the free-marketers’ studies of self-governing market ecosystems and their critiques of attempts by various governments to manage entire economies from the top down. This is ironic, as, in lieu of asking readers to reconsider governmental mandates that might incentivize monocultures or non-diverse rotations, Taleb’s paper argues that there is a strong necessity for policymakers to impose their own edicts upon the biotechnology sector from the top down.

Given that Taleb’s paper repeatedly castigates the Food & Drug Administration as too lax in curbing the alleged risks of GMOs (Taleb et al. 2014a, 9-10, 14), faulting
“limited oversight” (Taleb et al. 2014a, 11) and foolishly acting on “limited existing knowledge” (Taleb et al. 2014a, 10), and given that Taleb’s paper proclaims itself to be aimed at vaguely defined “policy makers” (Taleb et al. 2014a, 1, 9) and “decision makers” (Taleb et al. 2014a, 1, 14), one can ascertain that the paper implies — though does not articulate outright — that Taleb intends for the paper to influence governmental regulatory bodies to impose far more stringent impediments to the development and growing of DNA-spliced crops than presently exist. Indeed, the paper judges that on account of the presumably apocalyptic hazards associated with DNA-splicing, it must urge that these policymakers be “extreme” in foisting such impediments (Taleb et al. 2014a, 9) — the explicit goal is that the Precautionary Principle “be used to prescribe severe limits on GMOs” (Taleb et al. 2014a, 1).  Hence, one bears witness to the irony of so many laissez-faire partisans hailing a man who notoriously recommends coercive measures that, if implemented, will result in the hobbling of an entire peaceful industry.
The very topic of top-down impositions reveals the fragility of Taleb’s own argument. The top-down-GMO-versus-bottom-up-selective-breeding dichotomy that Taleb posits is grossly overstated. This anti-GMO argument overlooks the manner in which the same top-down and bottom-up aspects remain applicable in natural selection, selective breeding, and DNA-splicing alike.

The Same Bottom-Up and Top-Down Factors Exist With Selective Breeding and DNA-Splicing

First to be addressed is the top-down pressure that commences in each of these phenomena. When any environment houses an entire population of organisms of the same species, there will be variations in their individual genetic make-ups. The genes in each selectively bred individual are randomly assigned. That configuration arrives through the sorting of DNA sequences in sexual reproduction, and mutations provide new arrangements of the nucleotides comprising these DNA sequences. That is the bottom-up aspect. Whichever genetically induced physical traits best help the organism survive and procreate in that environment will be the physical traits that come to be most widespread in that population in that environment.  This then raises the question, “What determines which such traits will proliferate and which will be discontinued?” The environmental conditions themselves determine as much, and the environmental conditions themselves are imposed top-down by the natural terrain itself. This leads to the reasons why DNA-splicing is not as dissimilar from selective breeding as the Taleb paper insinuates.

Imagine I am a farmer trying to grow apples in one orchard, Environment 1. I prevail at this. When I then attempt to apply the same growing techniques in another orchard, Environment 2, I face unanticipated complications, as there are crucial differences between Environments 1 and 2 that I had not considered upon entering Environment 2. Upon some reflection, I undertake adjustments to my cultivar so that I can produce apples that do grow successfully in Environment 2. Such adjustments might result in the apples I grow in Environment 2 not being exactly the same, physiologically, as those I continue to raise in Environment 1.

That is what the Taleb paper and a New York Times op-ed coauthored by Taleb both overlook. A setting in which crops grow is a setting that applies naturally selective pressures on those crops from the top down irrespective of whether those crops originated through natural selection, selective breeding, or DNA-splicing.

It might be the case that my selectively bred apples prosper in Environment 1 but that I must originate a new version — also through selective breeding — to have other apples thrive in Environment 2. Likewise, it may be the case that my GMO apple prospers in Environment 1 but that I must employ even further DNA-splicing to produce a new version of apple that thrives in Environment 2. The cultivar’s ability to thrive in an environment is always at the mercy of the environment, and that principle applies to cultivars regardless of whether they came about through selective breeding or DNA-splicing. In that respect, a farmer’s attempts to produce a DNA-spliced cultivar that produces high yields in its environment remains a bottom-up trial-and-error process. Here, some DNA-spliced cultivars are “selected” by the environment for survival as other DNA-spliced cultivars remain unable to adapt and are thereby phased out. A cultivar’s successful adaptation to its environment — the farmer’s field or grower’s orchard —remains a bottom-up endeavor in both cases.

That is, in both cases of selective breeding and DNA-splicing, the same respective bottom-up and top-down pressures remain present:

  1. The bottom-up experimentation of adding or removing specific physiological traits to cultivars to ascertain which traits most suitably adapt the cultivars to the environment.
  2. The top-down “selection” by the environment of which traits cause the cultivar to flourish and which traits are to be discontinued.

What If Monsanto Alters the Environment From the Top Down?

Taleb will likely reject the proposition that a DNA-spliced organism is just as subject to the top-down pressures of its environment as are selectively bred organisms, because the environment to which the DNA-spliced organism is introduced is itself altered by the DNA-spliced organism. As Taleb’s paper asserts, the addition of every new DNA-spliced organism into the environment changes that environment. Worse, any “cross-breeding of wild-type plants with genetically modified ones prevents their disentangling…” (Taleb et al. 2014a, 9)  That is, every new DNA-spliced organism into the ecosystem compounds the dilemma.

Supposedly, the DNA-spliced organism no longer needs to accommodate itself to the environment; the DNA-spliced organisms will force the environment to accommodate them. Likewise, Monsanto might reshape the environment to accommodate the DNA-spliced crops it has patented. The fallacy in such an assertion, though, is that it overlooks that selectively bred organisms, too, already alter the environment in their own efforts to survive. Organisms that arose purely through natural selection have drastically restructured the environment to which they simultaneously adapted themselves. Phytoplankton affect a considerable proportion of the earth’s atmosphere. In that same spirit, human beings have altered the landscape for their own purposes even antecedent to the advent of agriculture. Some hunter-gatherers engaged in controlled burning of large sections of forest. Of note is that this paper’s authors seem to forget that a cultivar produced through selective breeding is likewise a contender for cross-breeding with wild plants.  Considering the massive extent to which human beings have changed the biosphere prior to any DNA-splicing, it is counterproductive to single out DNA-splicing as especially cataclysmic, somehow more dramatic in scope than any other technology. Ultimately, it is not correct to proclaim that Monsanto-patented GMOs artificially alter the crops’ environments from the top-down whereas selective breeders operate in environments not altered so substantially. Both DNA-spliced crops and selectively bred crops grow in environments that have been fundamentally altered by human beings from the top down. For this reason, Taleb’s top-down-versus-bottom-up dichotomy remains inadequate.

Even if the planting of DNA-spliced crops does significantly alter the landscape in the long run, it remains additionally misleading for Taleb’s paper to presume that a DNA-spliced organism might necessarily proliferate in almost any environment whereas a potentially harmful organism wrought through the bottom-up processes of natural selection or selective breeding “will not spread” across the globe but will “end up dying out due to local experience over time” (Taleb et al. 2014a, 9). To be sure, where I live —Hawaii — there are numerous native bird species that would probably die out if transported to any continent. However, before any DNA-splicing was done, international exploration, trade, and conquest during the Renaissance period were sufficient to transport other, more potentially troubling organisms from one locale to the other. Some of these species were intentionally transported across continents by traders. Other species stowed away on ships. In the latter event, rats — rats not originated through DNA-splicing — were able to adapt themselves to six continents, particularly in urban settings. Pigeons have adapted equally as well, as have cockroaches (Bettag prod. 1999; Jablonski 1991, 755; and Quammen 1998). Some descendants of domesticated cats and dogs, having lost any human master, become “re-wilded” and produce resilient feral offspring. More devastatingly, Europeans transported the smallpox virus to the Americas, and the Black Plague may have arrived in Europe through trade with the Mongol Empire (Weatherford 2005, 243-245). Marijuana is nicknamed “weed” because of how easy it is to grow throughout the world. Somewhat contradicting its other statement, the Taleb paper acknowledges the role “global transportation” played in the proliferation of “invasive species” (Taleb et al. 2014a, 6). Yet if such organisms that emerged through natural selection were able to adapt and vex other creatures, it is logically inconsistent to belabor the proposition that DNA-spliced organisms inexorably carry a special risk beyond anything that has occurred with natural selection or selective breeding.


Evolution:  Fast Versus Slow

To such an argument, Taleb and company might reply that their classification of selective breeding as safer than DNA-splicing holds up, due to an additional distinction. Taleb’s argument pronounces that natural selection and selective breeding alter the genomes of a population of organisms at a much slower pace than when it comes to DNA-splicing. Therefore, if a substantial menace arrives through the emergence of a new type of organism through natural selection or selective breeding, “policy makers” and other organisms have much more time — a grace period — to ameliorate the risks than they would if a fierce new DNA-spliced organism is introduced into the environment. When it comes to selective breeding, one finds “a limited rate at which variations can be introduced,” whereas farmers planting DNA-spliced seeds procured from Monsanto forces “rapid changes in organisms…inconsistent with this process” (Taleb et al. 2014a, 9).

If evolution — the alteration of the genetic makeup of a population of organisms — never occurred any faster than millions, thousands, or hundreds of years, then Taleb’s might seem a commonsense point. Yet evolution and speciation within a population of organisms can actually occur at a much faster pace in the absence of any DNA-splicing. Despite the impression Taleb’s paper conveys — that, prior to DNA-splicing, it necessarily took multiple “generations” of farmers to alter one species (Taleb et al. 2014a, 9) — there are case studies of entire new breeds of organism being produced within a single person’s lifetime. Whether or not that happens is influenced by the environment and by other factors doing the “selecting.” In fewer than eleven years, Russian geneticist Dmitry Belyaev produced new breeds of domestic fox through selective breeding. Absent of any consciously intended efforts on the part of human beings, some new strains of bacteria can emerge in less than a decade. In the early 1900s, Hermann Muller achieved important discoveries about genetics by exposing large quantities of fruit flies to radiation, thereby mutating their DNA, and then breeding them selectively — all this within a single man’s lifetime. Although such instances of rapid evolution and speciation might pose problems for human beings, such precedents once again reveal the failure of critics to demonstrate any risks posed by DNA-spliced organisms that do not presently persist with those that have been selectively bred.


Taleb’s Fish-Tomato Gambit

In order to convey that modern DNA-splicing is more extremist than anything preceding it, though, the paper expresses a common equivocation: “There is no comparison between tinkering with the selective breeding of genetic components of organisms that have previously undergone extensive histories of selection and the top-down engineering of taking a gene from a fish and putting it into a tomato” (Taleb et al. 2014a, 9). The assumption here is that taking a DNA sequence from a salmon that allows it to survive in the cold, and then splicing that sequence into the genome of a tomato, is necessarily a more radical recombination of elements than anything that can happen in selective breeding. After all, a fish and a tomato have nothing in common.

The fallacy in this argument is the assumption that because fishes and tomatoes look and behave so differently, it follows that putting fish DNA in a tomato will produce a more radical change in the tomato population’s gene pool than would simply breeding one tomato plant with another tomato plant. While the “expressions” of genes – that is, what physiological attributes and behaviors organisms develop – often appear to be obviously varied to human observers, the chemical structure of one species (a fish) and another (a tomato) do not appear so obviously distinct, under the microscope, to the untrained eye (or even many well-trained scientific eyes). The genomic sequences of all organisms, from mushrooms to sea stars to Venus flytraps to beetles to lizards to blue whales to humans, consist of the exact same four molecules — the nucleotides named cytosine, adenine, thymine, and guanine, normally abbreviated, respectively, as C, A, T, and G.

Many of the sequences in which these four molecules are arranged are nearly identical between organisms that, on the outside, seem terribly different. More than a third of the human genome is the same — that is, there are sequences where those four nucleotides are sorted in the same order — as that of a banana plant. That more than a third of my genomic sequence is identical to a banana’s does not mean I am a monster that is part human and part banana. Likewise, scientists splicing a DNA sequence from a fish into a tomato did not produce a part-tomato, part-fish monster. In practice, it simply involves lab technicians employing chemical processes to rearrange the same four molecules to be in a sequence different from the one they started with. The tomato they ended up with still looked and grew like a regular tomato. DNA-splicing rearranges the same four molecules in a controlled and relatively predictable manner whereas selective breeding rearranges those same four molecules in a more randomized manner.  If putting a fish DNA sequence into a tomato’s DNA sequence amounts to a relatively small rearrangement of the molecules, it might not be as radical a rearrangement of DNA as what takes place when selective breeding randomly sorts thousands of genes.

This undermines another claim of Taleb and his coauthors. On the topic of rapid evolution, one should notice that throughout the twentieth century, breeders have deliberately exposed plants to relatively large concentrations of radiation to induce radical mutations. This process is known as mutagenesis. The Star Ruby grapefruit and Rio Ruby grapefruit resulted from that. When the scientist Trevor Charles inquired as to whether Taleb and his coauthors consider mutagenesis cultivars to be unacceptable GMOs (Charles 2014), Taleb and his coauthors replied with the arbitrary and vague assertion that DNA-splicing is somehow more extreme than mutagenesis. They said that “inserting functional genes from distantly related species quite generally has an immediate effect on gene expression products and regulatory networks, while point mutations (by mutagenesis and radiation) have an immediate genetic effect but not necessarily an immediate expression or regulatory effect” (Taleb et al. 2014b, 2). This pedantic and false distinction, Trevor Charles observes, “is nonsense.”

Therefore, glib rhetoric aside, there is no reason to presume that combining the DNA of two dissimilar species is a more radical alteration of DNA sequences than is selective breeding. Thus, Taleb’s claims that DNA-splicing technology (1) speeds up the evolutionary process too much and (2) alters the gene pool far too radically, fail to demonstrate that DNA-splicing is any riskier or more “top-down” of process than what occurs with natural selection and selective breeding.


It is legitimate for statisticians and actuaries to study and comment upon the degree to which they judge that new technologies, such as DNA-splicing, might introduce risks to human health or the ecosystem that otherwise might not be present. Thinkers from outside the disciplines of botany and genetics can indeed bring insights that entrenched members of the field had not yet countenanced. Unfortunately, neither the “Precautionary Principle” paper in its present form, nor Taleb’s outbursts countering scientific evaluations of it, have added to a sound understanding of the risks of genetic engineering and the context in which they must be weighed. The misleading rhetoric about an unbridgeable chasm between supposedly top-down versus bottom-up approaches to farming turns out to be emblematic of this. It is not by exaggerating DNA-splicing’s differences with older technologies, but by carefully examining where and how this breeding technique is similar to them, that we can realistically assess the new risks associated with such a technology and find a way forward.



  • Bettag, T., prod. 1999. “The Sixth Extinction?” Nightline in Primetime: Brave New World. August 26. Robert Krulwich, presenter. Television broadcast: ABC News.
  • Charles, T. 2014. “Review of Taleb et al., The Precautionary Principle (With Application to the Genetic Modification of Organisms.” PDF on the website of the New England Complex Systems Institute (NECSI). Cambridge, Massachusetts: New England Complex Systems Institute. November 27. <>.
  • Fedoroff, N., & Brown, N. M. 2004. Mendel in the Kitchen: A Scientist’s View of Genetically Modified Food. Washington, D.C.: Joseph Henry Press.
  • Jablonski, D. 1991. “Extinctions: A Paleontological Perspective.” Science vol. 253 (no. 5021, August 16): 754-57. Available as PDF at < 1991 Extinctions.pdf>.
  • Quammen, D. 1998. “Planet of Weeds.” Harper’s Magazine. October. Reproduced online at <>.
  • Taleb, N. N. 2010 2d ed. The Black Swan: The Impact of Highly Improbable Fragility. New York: Random House.
  • Taleb, N. N., et al. 2014a. “The Precautionary Principle (With Application to the Genetic Modification of Organisms).” NYU School of Engineering Working Paper. PDF on Nassim Nicholas Taleb’s website <>.
  • Taleb, N. N., et al. 2014b.  “Response to Review By Trevor Charles Re: Precautionary Principle.” PDF on the website of the New England Complex Systems Institute (NECSI). Cambridge, Massachusetts: New England Complex Systems Institute. November 28.  <>.
  • Weatherford, J. 2005. Genghis Khan and the Making of the Modern World. New York: Crown.
Image Credit: Richard Green