Sunday, April 18, 2010

"The Myth of Green Beef": Pseudo-logic in action

Check out this article: "The Myth of Green Beef".

The author, Helene York, provides a wonderfully clear example of "pseudo-logic", reasoning that is technically correct, but based on flawed or incomplete assumptions. Check out this quote:

Linked to cardiovascular disease and maligned for its industry's dependence on federal corn subsidies, it now has a reputation as the Hummer of foods—an excessive contributor to environmental ills including climate change, nitrogen blooms, pollution, and depletion of Midwestern aquifers—not to mention E. coli contamination that has sickened and scared thousands.

Hmmm, sounds like the root of the problem here is the federal corn subsidies. Bon-Appetit Management, where Ms. York is the director for strategic initiatives, ran the cafe at my former company. I know from personal experience that a good chunk of the food provided by Bon-Appetit is made possible by federal subsidies corn, wheat, and soy. And of course there is a tidal wave of scientific evidence emerging that said foods are more likely the culprit cardiovascular disease, via the metabolic disturbances they create. The evidence that red meat per se causes any disease has, to my knowledge, never risen above association (the E. coli issue is problem with factory farmed animals, and only then for people whose health is otherwise compromised, maybe from eating "healthy" soy goo and avoiding the sun).

Here's another classic:

Voluntary rancher fees from an industry association's advocacy program have underwritten pro-meat marketing campaigns, stipends for researchers to raise doubts (but not conclusive evidence) about scientific studies, and dissemination of talking points that are misleading at best. "Reducing intakes of meat and dairy would only lead to hunger," I read recently, and the headline of an industry newsletter stated, "Meat and dairy intakes not linked to climate change." These news items represent a disturbing trend: raise doubts, obfuscate the facts, and misinform.

Isn't that EXACTLY what Ms. York is doing here? What makes her "facts" better than those she criticizes? Why are her studies more "scientific" than those that contradict her "conclusive" evidenct? Talk about confirmation bias. This is the fundamental problem we face when turning scientific information (or more precisely, the lack thereof) into decisions. Humans seem to have a psychological propensity to gravitate toward "absolute truths", and their absolute belief in those truths are motivated more by social and emotional factors than any sort of actual accounting of the evidence. Indeed, people like Ms. York seem to get wound around some sort of moral axle that drives their reasoning process. Beef is "bad" in her world. That's a "fact". Thus beef must be bad for your health, the environment, at the root of the global economic meltdown, bad hair days, etc. And maybe I'm pessimistic, but I have a feeling that, more than anything, serving beef might be "bad" for Bon-Appetit's bottom line. I would guess it is cheaper to sling soy/corn/wheat processed food (where you can reap the benefit of less prep and less annoying middle men sucking off the teat of government subsidies).

But let's be optimistic, and presume Ms. York's motives are altruistic, that she really wants to save our hearts and our planet from the evils of a nice juicy steak. Does her reasoning hold water? I believe you would need to take the following assumptions as "facts" to support her conclusions:
  1. Human activity causes global warming.
  2. This warming trend will continue.
  3. Changes in human activity can reverse the trend.

This is where we run into trouble. The implication is that we have both a great enough understanding of global climate to make reliable predictions, and further that even if we had such detailed understanding, that behavior could be reliably extrapolated decades into the future. I'm no expert in global climate, but I know a thing or two about modeling complex systems, particularly in the face of uncertainty about the details. I seriously doubt that global climate models even begin to approach anything beyond a coarse representation of reality. There are plenty of aspects to the problem that we know we don't know, like the response of aquatic life to increased CO2 concentrations. There are significant uncertainties as well, e.g. solar and volcanic activity. And no doubt there's plenty of stuff we don't even know about, the "don't know what you don't know" category.

And it gets worse. Climate is basically just another word for weather. I don't know if you've noticed, but it's pretty hard to predict the weather even a week into the future, much less 50 years. And short-term weather modeling is much better understood for the simple reason that when examining a shorter time period, less variables are likely to have a large effect (e.g. large glaciers don't change enough in a week to affect your forecast significantly). Even so, the weather remains unpredictable, and this unpredictability is intrinsic. Weather is an example of a non-linear system, one which exhibits a phenomenon called deterministic chaos. A brief digression might be in order.

Consider a simple experiment, say measuring the time it takes a marble to fall from a height of one meter. We call such a system "deterministic" because the equations used to model it have no uncertainty. Given a particularly position and velocity for the marble, we can calculate the precise position and velocity an instant later. And this is a good approximation in our experiment. We might induce a little uncertainty in how our hand releases the marble, some from the measurement of the height, maybe some from air currents, etc. But we can repeat this experiment and get the pretty much the same results every time. In other words, small errors in our information about the marble's state translate into small errors in our predictions. The more accurate our information about the marble, the more accurate our prediction of the time to fall one meter.

A system exhibiting deterministic chaos is deterministic in the strict sense of the term: given precise knowledge of it's state, we can predict exactly what will happen next. But unlike our marble experiment, chaotic systems amplify uncertainty. In other words, even small inaccuracies in your information about the system quickly become large. Worse yet, this amplification is exponential in time, so getting more accurate information might make them predictable for a slightly longer period, but it's still going to fall apart on you pretty quickly. Chaotic systems are predictable only in principle, but in practice your information is never perfect, and predictability drops exponentially with time. Deterministic chaos as we now think of it was "discovered" by Edward Lorenz, who was modeling (you guessed it) global weather.

So, even assuming that the East Anglia boobs, with their lost data and bogus statistical analyses, were "right" about there being a significant increase in mean global temperatures, how does that help us predict the future behavior of a complex chaotic system where are models are incomplete and full of uncertainties?

Now when I drop this line of argument during discussions of global warming, the AGW crowd (after a bit of cognitive dissonance induced brain paralysis) come up with something like the following argument: human activity MIGHT be causing global warming, and since the downside has a value which is essentially negative infinity (extinction of the human race), we have to do everything possible to avoid it. Such an argument is more pseudo-logic, in this case by excluding the most likely scenario. AGW arguments center around whether or not the (supposedly) observed warming trend is caused by humans, and extrapolate that to conclude that humans might be able to reverse said trend. But this ignores the most likely scenario, which is that the climate will undergo a significant shift regardless of anything humans have done or will do. Why do I say this is the most likely scenario? Because it has happened many times in the past, and given the chaotic nature of climate, it is unlikely to stay in the current meta-stable state for long (many argue that the rise of civilization was made possible by unusual relative stability of climate). Arguments such as those put forth by Ms. York completely miss the point. We don't need to be worried about whether eating less hamburgers can affect the climate, we need to start hedging our risk that the climate will change regardless of what we do. It's the short-sighted thinking and associated bad decision-making of individuals like Ms. York that will doom us, missing that the forest around them is burning down while hugging the tree right in front of their face.

But back to the main thread. So the whole global warming argument is bogus. That's about as close to a fact as you're going to get, since it's really just mathematics (climate is chaotic, our knowledge of it is uncertain). Let's wander from math to the realm of science, where we consider evidence. The more detailed assumption underlying Ms. York's proposal is that cattle farming is particularly bad for the environment. She's basically equating cows with global destruction. This begs the question of how the Earth managed to survive millions of years of grazing animals, all of whom presumably had the same basic digestive strategy of modern plant eaters.
  1. Possess large gut full of bacteria which can break down cellulose.
  2. Eat plants, and lots of 'em.
  3. Bacteria eat the cellulose, make CO2/methane/etc. as by-products.
  4. Fart voluminously to avoid exploding.
I'll close by noting that there's nothing particularly "green" about any large-scale agriculture. Grain-fed cattle are no doubt the caboose on the train to ecological destruction, as cattle inefficiently convert grain into food (compared, say, to a chicken). But that whole process, like most modern agriculture, represents a massive perversion of a natural process, requiring considerable human intervention. And that perversion goes right back to the fields of corn/wheat/soy grown to feed both cows and people. The ecological issues have been described elsewhere (my favorite discussion is in Lierre Keith's fantastic book The Vegetarian Myth), but the bottom line is that the sort of large-scale monocultures we see today are an ecological dead end. Sooner or later you deplete the resources required to grow specific foods in high densities: soil, water, and petroleum for chemical fertilizers. Feeding people directly with that food instead of "wasting" on cattle might delay the end a bit, but you still get there, and if history shows us anything it's that making more food just winds up making more people who will wind up starving when the train crashes Thanks a billion, Norm Borlaug, for "saving a billion people from starvation" by setting the stage for multiple billions to starve. This is the poster child for why the sort of short-term thinking displayed by Ms. York is to be avoided.

So if there's any "myth" to be dispelled here, it's that there's anything "green" about the food industry, which includes the erstwhile Ms. York and her employer, Bon-Appetit Management.

And winding back to the usual topic of this blog, we shouldn't forget the health consequences of a diet consisting mostly of processed soy/wheat/corn. There is plenty of evidence from all corners indicating that "diseases of civilization" arise from said foods. I'm still waiting for someone to detail the metabolic pathways by which eating a steak leads to diabetes, cancer, and heart disease. Any takers?


Anonymous said...

Nice to see you posting again! On a different topic from this one, have you ever looked at this study:

They found people who aren't insulin resistant lose more weight on a low fat diet than low carb. The data seems pretty solid, but I'm no expert.

LeonRover said...

The Spark of Reason, the Voice of Empiricism, thank you Dave Dixon for a nice post on chaos, modeling and the scientific method.

Via Physics Math and Statistics, I also ended up in world of business, but never left behind the idea that our conclusions be driven by facts, by data and by investigations and critiques which note where there is no predictability nor certainty.

I applaud your use of the scalpel of scepticism and the Razor of Occam.

Dave said...


Thank you.


Not an entirely surprising result. Obviously if you're not insulin resistant you tolerate carbs better (lower insulin response required to stabilize blood sugar), and we'd expect a low-fat/high-carb (percentage wise) calorie restricted diet to do better for those who are insulin sensitive vs. insulin resistant.

That said, I'm thinking that low-fat vs. low-carb misses the point, not the least because the experimental definitions of what "low-fat" and "low-carb" mean seem to be all over the place. Manipulating diet and watching the outcome is a "black box" approach to the problem. But we can "open the box", and measure things like insulin response.

I think it would be much more informative to do experiments that measured fasting and post-prandial insulin and glucose response as a function of diet composition, and then correlate THAT with weight loss. Since we know that insulin and glucose play a causal role in fat storage vs. release, we get more information. Such studies open the door for tailoring diet treatment for individuals, since I'm guessing that the lower insulin/glucose, the better the results. Specifics of insulin/glucose response are probably at least partially genetic.

Aaron said...

hmmm -- couldn't iron overload lead to diabetes, cancer, and heart disease.

however, it was just a though, I usually side with a paleo-type diet.

Dave said...


What's the mechanism by which iron might have these effects?

Tom Naughton said...

Excellent analysis. But of course, you know what's going to happen when you apply logic to an emotional topic like global warming ...

Dave said...

@Tom Naughton,

Anger, leading to increased respiration, greater output of hot CO2, and even more global warming?

Aaron said...

These are just some of the links of many that might show some relevence.

Albert Ciampa said...


Off topic but I'm not sure where else to post.

People with severe metabolic syndrome - lots of years, coupled with lots of bodyfat gain - do you think these types of people are forever off of any type of carbohydrate, save for the occasional green vegetable?

Also, what about adipose tissue - if I have someone who gained a lot of bodyfat, I am to assume that not only did adipose cells increase in size, but new cells formed. Given this (plus metabolic dysfunction), could that person's new homeostatic bodyfat level be higher even if they implement a proper low carb paleo plan - so losing only so much bodyfat under normal low-carb paleo conditions is within reason. In other words, this person will only get so lean without calorie restriction.

I've seen what I think fits this, many times, and am wondering exactly what is going on.

I'm asking because of your history and that some paleo guys are beginning to preach some form of grain or starch, and it seems that a "rebound" is starting to take effect.

Thanks in advance.


Dave said...

Hi Al. Excellent questions, many along the lines of topics I've been pondering lately. I'll see if I can squeeze off a blog in the next couple of days to discuss some of this.