I always enjoy when someone whom I respect and who cares about using careful and valid arguments disagrees with me. It is an opportunity for me to correct any mistakes I have made, to deepen my understanding of the topic, or at least tighten up my arguments.

Last week I wrote an article responding to a recent New York Times piece on GMOs (genetically modified organisms). Massimo Pigliucci, who is a friend and skeptical colleague, disagrees with my analysis. Massimo thinks that knee-jerk defense of GMOs is a problem generally in the skeptical movement, and uses me as an example. I disagree with him, but will discuss that toward the end.

I want to take the points that I make in my previous post one by one and see how they hold up to Massimo’s criticism, and may expand upon them and include other comments as well.

GMOs should not be considered as one thing.

I wrote in my previous article:

“Any meaningful analysis of GM technology has to consider each application unto itself. Further, the GM trait is only part of the picture – you also have to consider how it is being applied.”

I have consistently taken this position in my writings, and this is also the most common position I encounter when reading other skeptics writing about GMOs. It is not really meaningful to consider GMOs as if they are one thing, and this is a mistake that Hakim makes in the original NYT article.

Hakim talks about “genetic modification” as if it were one thing. Of course he later talks about glyphosate resistant crops and pesticide producing crops, but only after he has set up the premise that these two specific applications are essentially equal to “genetic modification.”

I maintain that this is meaningless. This is similar to talking about “drugs” as if they were a thing, rather than looking at the risks and benefits of each drug in the context of how those drugs are used.

Even worse, Hakim is attempting to answer a meaningless question – what is the net effect of “genetic modification.” Not only is this an incredibly difficult question to answer, it is inherently deceptive to even approach the question in this way.

The problems with this framing have to be understood in the context of the ongoing GMO debate. The anti-GMO crowd wants to equate genetic modification with increased chemicals in the mind of the public. That strategy has largely worked.

There are two big problems with this. The first is that “genetic modification” is not even one technology, but a suite of technologies. Deciding where to draw the line on what constitutes a GMO is somewhat arbitrary. Does this only refer to transgenic crops? What about cisgenic crops, gene silencing, directed breeding, mutation farming, or forced hybridization?

Further, this approach confuses a technology with specific applications of that technology (more on that below).

Massimo does not even address this point, which I think is a major omission in his commentary. This is a problem because he tacitly endorses the approach by accepting Hakim’s framing and not addressing this issue, either to agree or disagree with my position. Hakim, in turn, is tacitly accepting this anti-GMO framing. He gives no indication that he recognizes this, so I have no way of knowing if it is deliberate or naive.

Evaluating a technology by its specific applications.

I maintain that when evaluating a technology, which is explicitly what Hakim’s piece is doing, it is very important to separate out the technology itself from specific manifestations of the technology and how that technology is used.

About this Massimo writes:

Steve continues: “The promise of genetic modification, rather, is that it provides a tool for agricultural scientists to make more rapid and more specific changes to crop cultivars. The technology has completely fulfilled that promise. The technology works, it is safe with no demonstrable inherent risks.” Again, this is a rather strange way to put it. No, the promise of GMOs is not just to provide agricultural scientists with faster ways to change crops, it is to make life better for the farmers and the public at large. And yes, the technology works when “works” is narrowly defined, but whether it has delivered on its broader, and most important, promises, is very much at issue.

I have to fervently disagree. Hakim’s article is not really about early hype or marketing from the biotech industry, it is about genetically modified organisms. He begins his piece:

The controversy over genetically modified crops has long focused on largely unsubstantiated fears that they are unsafe to eat.

But an extensive examination by The New York Times indicates that the debate has missed a more basic problem — genetic modification in the United States and Canada has not accelerated increases in crop yields or led to an overall reduction in the use of chemical pesticides.

He characterizes these issues as a basic problem with genetically modified crops. Again, this is like asking, does surgery work?

If the piece were, rather, just criticizing the way in which the industry marketed their GMOs, and clearly distinguished this from GMOs themselves, or the technology, Massimo might have a point.

I deliberately defined the purpose of GMOs narrowly to make a very clear point – GMOs are a technology, they are not equal to their application. Hakim is making a category mistake.

This is a critical distinction because, getting back to the context of the public conversation about GMOs, the question is whether or not we as a society should continue to use various tools of genetic modification or should we abandon this technology? There are those who want to ban the technology completely.

One could even reasonably argue that even if there weren’t a single application of genetic modification that has had a positive result, we should not necessarily blame the technology or abandon it, just seriously reconsider how that technology is being used.

Another way to look at this is – are the alleged failures of the application of a technology inherent to the technology itself, or are they solely a manifestation of how the technology is currently being used? Hakim does not even ask this question, he just equates the technology with some of its uses.

Is yield the best way to evaluate GMOs?

I argued that it was unfair to use yield as a primary criterion in assessing the impact of GMOs because there aren’t even any GM traits on the market that are specifically designed to increase yield. Why pick yield, when we could also look at predictability of crops, loss reduction, reducing labor, improving profits for farmers, reducing tilling of soil, environmental outcomes, and improved nutritional and other qualities in the food themselves?

Hakim justifies this choice because he argues that this was the way in which GMOs were originally sold by the biotech industry. Massimo agrees, writing about my position:

This makes little sense in light of the fact — as clearly substantiated in the NYT article — that GMO companies made precisely those promises, very clearly, and very early on, presumably in order to gain both public support and regulatory approval.

Actually, although I did not address this in my previous article, the NYT article does not clearly establish this premise. Hakim, for example, gives this quote:

“It’s absolutely key that we keep innovating,” said Kurt Boudonck, who manages Bayer’s sprawling North Carolina greenhouses. “With the current production practices, we are not going to be able to feed that amount of people.”

That certainly implies increased yield, but it is a much more subtle statement. The word “yield” does not occur. They are arguing generally for biotech to improve food production (which has a lot of components, not just yield) and that is still a reasonable position, whether or not current GM traits have specifically increased yield.

But that is a minor point – even if we accept Hakim’s premise that the biotech industry made promises about yield they did not keep – so what? Again, that is a legitimate point if you are writing an article about how industries use hype and empty promises, but this article is ostensibly about GM technology.

Judging technologies in general.

Leading from this last point, is the broader point I made that it is misleading to use early expectations, even specific industry claims, to judge a technology itself. To be clear, I frequently criticize industry hype, over hyping a new technology, making unwarranted predictions about technology,  and making premature claims for technology. Those are a staple of this blog.

But a technology should be judged based upon its inherent vices and virtues, not how it was hyped, and not necessarily how it is currently used. These are distinct things that can and should be teased apart in any meaningful analysis.

I gave a couple of example that Massimo did not accept. He writes:

The first one concerns the human genome project, which he points out hasn’t cured any major human disease, contra to the hype that helped promote the original effort. Right, but the human genome project was largely an exercise in basic science, and it was funded as such, not because it promised to cure cancer (the National Science Foundation, for instance, clearly separates the actual immediate scientific merits that justify funding a given project from the possible “broader impact” of the same project).

I don’t follow Massimo’s logic here. It seems irrelevant to me that the human genome project was a basic science project. I know that, that is actually a key part of my point. But Massimo is glossing over the key part of the analogy – the project was absolutely sold to the public for its potential to lead to cures to serious diseases. Sure, the NSF knows that funding basic science is not always about later applications, but that was not the public hype.

My point was clear – early hype is not a valid criterion for judging a technology. The human genome project was a basic science success, and still has the potential (20 years later) to lead to specific disease applications, but it hasn’t yet.

Massimo also rejects my computer analogy:

The second example brought up by Steve is that of personal computers: “One might have argued (and many people did) even into the 1990s that computers did not improve productivity. They did not reduce use of paper (remember that one – the paperless office).” No, they didn’t, but I adopted personal computers at around the time Steve did, and they delivered immediately exactly what they promised: they helped me run some nice statistical analyses for my thesis, to produce graphics to accompany the same thesis, and to write the thesis to begin with.

Again, Massimo misses the point. He actually strengthens my analogy. The computer industry hyped their products to companies with the explicit promise that they would improve productivity and reduce paper. It took a long time to deliver on the productivity end, arguably two decades, and even still we have to take this case-by-case. They never delivered on the reduced paper claim, and in fact we use more paper than ever.

So, Hakim could write an article titled, “Doubts about the promised paperless office and increased productivity of computers.” Massimo points out that computers are great for statistical analysis, and added later for playing video games. Right, Massimo, that is exactly my point. Computers are useful for all sorts of things. (And GMOs saved the Hawaiian papaya industry and brought back the American chestnut.) Don’t judge them narrowly by two applications that did not meet early industry hype used to sell the public on the idea of personal computers.

I will add another good analogy – microwave ovens. They were initially sold with the claim that they would revolutionize cooking. The industry made microwave cookbooks, and sold them as cooking tools. It turns out that microwaves are completely useless as cooking tools. But they became indispensable heating tools.

If we are asking the question, was early microwave industry hype accurate, the answer is no. If we are asking the question, is it useful to have a microwave, the answer is yes.

Massimo reinforces his position when he writes:

The discussion concerns the use that the industry has made of that technology, therefore judgment very much needs to include a consideration of whether what the industry promised has been delivered or not.

Again, I have to completely disagree. Massimo is missing the real context here. The discussion is about whether or not GMOs should be labeled, demonized, banned (all GMOs) vs used as a useful tool among many in biotechnology.

The use to which the industry has put the technology is a valid question, but it is actually much more complex than saying it is all herbicide resistance and pesticides, and evaluating them is more than about yield. I maintain that industry hype (while it may be a legitimate point unto itself) is irrelevant to the current questions of GMOs.

Conflating herbicides and insecticides. 

Massimo writes:

Next, somewhat puzzlingly, Steve claims that “Hakim reproduces a common anti-GMO trope to combine Bt crops with herbicide resistant crops — two completely different applications.” But in fact Hakim does no such thing. The NYT article very clearly distinguishes the two, acknowledges that the use of insecticides has gone down in post-GMO USA, and then focuses on herbicides.

I, too, am puzzled. The opening two paragraphs of Hakim’s article refers to “pesticides” as one category. This is also a confusing term, and we have to keep reminding people that “pesticides” equals herbicides, insecticides, and fungicides. So Hakim conflates them in the opening of his article. He continues this strategy when he writes:

“The potential harm from pesticides, however, has drawn researchers’ attention. Pesticides are toxic by design — weaponized versions, like sarin, were developed in Nazi Germany — and have been linked to developmental delays and cancer.

and

The industry is winning on both ends — because the same companies make and sell both the genetically modified plants and the poisons.

My criticism remains valid. Hakim talks about pesticides as a category. Yes, deep in the article he does talk specifically about insecticides and herbicides, and the attached graphs split out the data as well, but the framing is clear and misleading. At the very least this is poor communication. The bottom line given in the first two paragraphs will have a much more dramatic effect on the average reader than complicated data given deep in the article.

To be fair I will give Hakim partial credit for this one since he does break out the data eventually, but the problem with his narrative is clear and remains a problem.

As I stated, insecticide use has decreased as a result of Bt crops. This is undeniable and a very positive thing for the environment. Herbicide use is a much more complicated story, as I have always stated. Much depends on how herbicide resistant crops are used, more so perhaps than the crops themselves.

When evaluating the net effect of GMOs, different GMO should not be combined, and herbicides and insecticides should never be combined. It is just misleading.

What are the net effects of GMO on yield and chemical use?

Here is where I think Massimo’s criticisms are most legitimate. To be fair, however, I never intended to do anything near a thorough treatment of this question. Reading back I don’t think I made my point clearly, but I was only trying to say that Hakim is cherry picking his data. I gave representative references showing that the full story is much more complex. These references are not definitive and have their own problems, but I only made the point that Hakim does not even mention such data.

I did point out that the NYT analysis of the UN data was not peer reviewed. This does not mean it is useless, and I never said so, but was making the more limited point that this is not sufficiently rigorous data and analysis to use as a primary basis for the major conclusion of a NYT article.

The primary focus of my first article was the narrative used by Hakim, not a review of the data on yields, herbicide use, and insecticide use. However, others have done that more detailed technical analysis. Nathaniel Johnson did a review of the article which focused on this question. He concluded:

Hakim cites the report where it supports his conclusions, but not in the places it contradicts them. He writes that the report found “‘there was little evidence’ that the introduction of genetically modified crops in the United States had led to yield gains beyond those seen in conventional crops.” But Hakim doesn’t mention that the report also noted that genetic engineering increased yields “where weed control is improved” and “when insect-pest pressure was high.” He doesn’t mention the report found that insect-resistant GMOs reduced insecticide use “in all cases examined.”

The main point is that if you take a zoomed-out view, and can choose which zoomed out view to pick, you can support whatever narrative you want. Again, this is like asking the useless question, does medicine work? A closer to home analogy would be comparing European medicine to American medicine and then making broad conclusions about medical systems, ignoring all the many variables that could influence such data. Johnson points out as others have as well that comparing North America to Europe is very complicated and does not lend itself to reliable conclusions.

Many critics also zinged Hakim for using France to compare insecticide use, when their prior use was unusually high and was just coming down to more average levels. This is therefore a highly misleading comparison.

Massimo rejects my references because they are not analogous to Hakim’s data, but that is irrelevant. My point is, there is nothing special about Hakim’s comparison, other comparisons may be as or more valid. Perhaps GMOs have more impact in undeveloped countries. Again, so what? That’s the problem when you frame your question about “GMOs.” Everything is on the table.

You need to zoom in and see how specific GMOs are being used in specific contexts. What specific disease populations are taking what specific drug at what dose?

Incidentally I can’t help but notice that Johnson independently came to the same computer analogy that I did, writing:

Back in the ’90s, businesses were buying computers like crazy, but overall productivity numbers “failed to suggest that anything unique was occurring in the workplace,” according to the St. Louis Fed. Computers didn’t bend the trend lines for decades.

Are skeptics biased?

Massimo concludes:

One thing, however, I am fairly confident of: the NYT article was not a “hack piece with a biased narrative.” If it needs to be criticized, by all means let’s do it, but we need to have good arguments and carefully laid out evidence, and Steve’s piece simply does not provide enough of either.

He also writes in the comments to the piece:

Yes, I do think that there is a bit of groupthink within the skeptic movement. Steve is one of the best, but we all have our blind spots (yes, yes, including yours truly). In this case, the attitude seems to be that any criticism of GMOs, no matter how qualified and circumscribed, is an “attack” that helps those irrational bozos who want to reject the technology.

I disagree on both counts. Hakim accepted a framing of the GMO issue that was crafted by the anti-GMO movement, namely treating GMOs as one thing. They are not one thing technologically, and the organisms themselves have to be evaluated individually. He then chose to evaluate GMOs based on two biased criteria, yield and overall pesticide use. He also chose to use cherry picked data for his analysis.

I agree in that I don’t really know if Hakim is biased or a hack, but he is some combination of the two in this article. My sense is that he is coming to this party late, is not up to speed, bought into narratives that are biased at the outset, and so his entire endeavor was doomed from the beginning.

I hold the NYT to a high standard, which I think is legitimate. At the end of the day his piece did not add anything useful to the narrative on GMOs. He made some rookie mistakes and the take home of his article was very misleading. It will be fodder for the anti-GMO movement, which does not make it wrong, just irresponsible (or biased).

On the latter point about skeptics and knee-jerk defense of GMOs, I will just say, having read pretty much everything out there on this issue, that is not what I have seen. Mostly I find well reasoned articles on GMOs that nicely deconstruct the propaganda that has dominated the conversation from the anti-GMO side for the last two decades.

As I have many times, as Kevin Folta does very carefully, and as other writers who typically defend GMO have done, we acknowledge that GMOs are not a panacea. We recognize that GMOs need to be regulated, that industry sponsored research may be biased, and that the data on the affects of specific GMOs are often imperfect and complicated. My own writings on this topic are here for anyone to see.

It could seem that we are defensive because there is so much anti-GMO misinformation out there. They have controlled the narrative, by blatant lying, and by many more subtle deceptions. Treating GMOs as if they were one thing is one biased anti-GMO framing. Treating GMOs as if they all have to do with chemical use is another. Saying that GMOs have not met their promise is yet another.

Hakim falls for all of them.

I honestly don’t know how much Massimo is aware of this context or how much he has engaged on the GMO issue before. While he makes some fair points on some of the details of my first article, I think he missed this overall context, and was simply wrong about my main points.