A Chemical Hunger – Part VIII: Paradoxical Reactions


We come up with theories to try to make sense of the world around us, and we start by trying to come up with a theory that can explain as much of the available evidence as possible.

But one of the known problems with coming up with theories is that sometimes you are overenthusiastic, and connect together lots of things that aren’t actually related. It can be very tempting to cherry-pick evidence to support an idea, and leave out evidence that doesn’t fit the picture. It’s possible to make this mistake honestly — you get excited that things seem to fit together and don’t even notice all the evidence that is stacked against your theory.

But sometimes noticing that things seem to fit together is how an important insight comes to light. The theory of continental drift was invented when Alfred Wegener was looking through a friend’s new atlas and noticed that South America and Africa seemed to have matching coastlines, “like a couple spooning in bed”. He wasn’t even a geologist — at the time, he was an untenured lecturer in meteorology — but he thought that it was important, so he followed up on the idea. “Why should we hesitate to toss the old views overboard?” he said when his father-in-law suggested that he be cautious in his theorizing. He was criticized by geologists in Germany, Britain, and America, in part because he couldn’t describe a mechanism with the power to shuffle the continents around the globe. But in the end, Wegener was right.

The true power of a theory is its ability to make testable predictions. One obvious prediction of the theory that obesity is caused by a contaminant in our environment is that we should also expect to see paradoxical reactions to that contaminant.

Predicting Paradoxical Reactions

Sometimes drugs have what’s called a paradoxical reaction, where the drug does the opposite of the thing it normally does. For example, amphetamines are usually a stimulant, but in a small percent of cases, they make people drowsy instead. Antidepressants usually make people less suicidal, but sometimes they make people more suicidal.

Normally when we talk about paradoxical reactions, we’re talking about the intended effect of the drug, not the side effects. But from the drug’s point of view, there’s no such thing as side effects — all effects are just effects. As a result, we should expect to sometimes see paradoxical reactions in side effects as well.

And in fact, we do. A common side effect of the sedative alprazolam is rapid weight gain. But another common side effect is rapid weight loss. Clinical trials show both side effects regularly. One trial of 1,388 people found that 27% of patients experienced weight gain and 23% of patients experienced weight loss. In those who do lose weight, weight loss is correlated with the dose (r = .35, p = .006).

Severe weight gain is a common side effect of psychiatric drug clozapine. People can and do regularly gain ten or twenty pounds on this drug. But some people actually lose weight on clozapine instead.

Lithium increases leptin levels in most patients, and this is presumably part of the mechanism that causes people to gain weight on lithium. But in some patients, lithium reduces leptin levels instead.

Normally the weight loss from these paradoxical reactions is pretty limited. But occasionally people lose huge amounts. People can gain 4 lbs (1.8 kg) over only 17 days on alprazolam. In comparison, anecdotal reports from admitted abusers suggest that high doses of alprazolam can lead you to eventually lose 10 or even 40 lbs.

On clozapine, people usually gain 10-15 lbs. But some people lose huge amounts of weight instead, up to 50% (!!!) of their body weight. One patient, a woman in her 30s, went from about 148 lbs (67 kg) to about 75 lbs (34 kg) on clozapine.

AGRP neurons are a population of neurons closely related to feeding. One of the ways researchers established this connection was by showing that activating these neurons in mice led to “voracious feeding within minutes.” Another way they showed this connection was by destroying these neurons, a process called ablation. “AGRP neuron ablation in adult mice,” reviews one paper, “leads to anorexia.”

If weight gain is the main effect of a drug, the paradoxical reaction is weight loss. If the obesity epidemic is caused by one or more contaminants that cause weight gain, we should expect that there will be some level of paradoxical reaction as well. If obesity is the condition, the paradoxical condition would be anorexia.

If it’s possible to turn the lipostat up, leading to serious weight gain, it’s certainly possible to turn the lipostat down as well, leading to serious weight loss. For most people, these environmental contaminants cause weight gain. But just like with other drugs, in some people there’s a paradoxical reaction instead.

This is biologically plausible. People with anorexia have extremely low leptin levels, and some reports suggest that leptin levels are correlated with symptoms other than just BMI. Anorexia risk is genetically heritable and some of the genes involved have already been identified. The authors of one genetic analysis close by saying,

Low BMI has traditionally been viewed as a consequence of the psychological features of anorexia nervosa (that is, drive for thinness and body dissatisfaction). This perspective has failed to yield interventions that reliably lead to sustained weight gain and psychological recovery. Fundamental metabolic dysregulation may contribute to the exceptional difficulty that individuals with anorexia nervosa have in maintaining a healthy BMI (even after therapeutic renourishment). Our results encourage consideration of both metabolic and psychological drivers of anorexia nervosa when exploring new avenues for treating this frequently lethal illness.

Brain lesions alone can cause anorexia nervosa, complete with the characteristic psychopathologies like fear of fatness, drive for thinness, and body image disturbance. Many cases present as “typical” anorexia nervosa, complete with weight and shape preoccupations. When tumors are surgically removed, these symptoms go away and the patients return to a healthy weight.

Brain lesions are not the only purely biological issue that can cause anorexia. In some cases, it appears to be closely related to the gut microbiome. In one case study a patient with anorexia had a BMI of only 15 even after undergoing cognitive-behavioral therapy, medication, and short-term force feeding, and despite maintaining a diet of 2,500 calories per day. Physicians gave her a fecal microbiota transplant from an unrelated donor with a BMI of 25. Following the transplant she gained 6.3 kg (13.9 lbs) over the next 36 weeks, despite not increasing her calorie intake at all. This is only one case, but the authors indicate that they are planning to conduct a randomized controlled trial to investigate the effects of fecal transplants in individuals suffering with anorexia. To the best of our knowledge this next study has not yet been published, but we look forward to seeing the results.

Eating and maintaining weight is a central cognitive problem. “The lipostat does much more than simply regulate appetite,” says Stephan Guyenet, “It’s so deeply rooted in the brain that it has the ability to hijack a broad swath of brain functions, including emotions and cognition.”

Remember those children we mentioned in Part II, who were born without the ability to produce leptin? Unlike normal teenagers, they aren’t interested in dating, films, or music. All they want to talk about is food. “Everything they do, think about, talk about, has to do with food,” says one of the lead researchers in the field. A popular topic of conversation among these teens is recipes.

These teenagers have a serious genetic disorder. But if you put average people in a similar situation, they behave the same way. The Minnesota Starvation Experiment put conscientious objectors on a diet of 1,560 calories per day. Naturally, these volunteers became very hungry, and soon found themselves unable to socialize, think clearly, or open heavy doors.

As they lost weight, these men developed a remarkable obsession with food. The researchers came to call this “semistarvation neurosis”. Volunteers’ thoughts, conversations, dreams, and fantasies all centered on food. They became fascinated by the paraphernalia of eating. “We not only cleaned our plates, we licked them,” recalled one volunteer. “Talked about food, thought about it. Some people collected as many as 25 or 30 cookbooks” (one such collection is pictured below). Others collected cooking utensils. “What we enjoyed doing was to see other people eat,” he continued. “We would go into a restaurant and order just a cup of coffee and sit and watch other people eat.”

Subjects became overwhelmingly preoccupied with food, and some collected dozens of cookbooks, like the collection shown above. 

These are the neuroses of people whose bodies believe that they are dangerously thin, either correctly (as in the starvation experiment) or incorrectly (as in the teenagers with leptin deficiency). The same thing happens when your mind, correctly or incorrectly, believes that you are dangerously fat. You become obsessed with food and eating, only in this case, you become obsessed with avoiding both. A classic symptom of anorexia is “preoccupations and rituals concerning food”. If that doesn’t describe the behavior above, I’m not sure what would.

But avoiding food and collecting cookbooks isn’t the lipostat’s only method for controlling body weight. It has a number of other tricks up its sleeve.

Many people burn off extra calories through a behavior called “non-exercise activity thermogenesis” (NEAT). This is basically a fancy term for fidgeting. When a person has consumed more calories than they need, their lipostat can boost calorie expenditure by making them fidget, make small movements, and change posture frequently. It’s largely involuntary, and most people aren’t aware that they’re burning off extra calories in this way. Even so, NEAT can burn off nearly 700 calories per day.

When most people eat less than they need, they become sluggish and fatigued, like the volunteers in the Minnesota Starvation Experiment. But people with anorexia fidget like crazy. A classic symptom of anorexia is excessive physical activity, even in the most severe stages of the illness. When one group measured fidgeting with a highly accurate shoe-based accelerometer, they found that anorexics fidget almost twice as much as healthy controls.

This kind of fidgeting is the classic response in people whose bodies are fatter than they want to be. In studies where people were overfed until they were 10% heavier than their baseline, NEAT increased dramatically. All of this is strong evidence that people with anorexia have lipostats that mistakenly think they desperately need to lose weight.

Of course, this does sound a little far-fetched. If anorexia were really a paradoxical reaction to the same contaminants that cause obesity, then in the past we would see almost no anorexia in the population, up to a sharp spike around 1980…

While there’s not as much historical data as we would like, the pattern we observe is just about that (see figure below). Cases were quite low until about 1970, when prevalence suddenly shot up. When we look at specific sections of historical data, finding evidence of an increasing trend (often only in young women) is pretty common.

Registered yearly incidence of anorexia nervosa in mental healthcare in northern Europe in the 20th century

In general the data is pretty scattered and spotty. Rarely does a study look at rates in the same area for more than five years. When there are such comparisons, they are usually for periods before 1980. For example, van’t Hof and Nicolson, writing in 1996 and arguing that rates of anorexia are not increasing, at one point cite studies that showed no increase from 1935-1979, 1935-1940, 1975-1980, and 1955-1960. But data from the Global Health Data Exchange (GHDx) shows that rates of eating disorders have been increasing worldwide since 1990, from about 0.185% to 0.215%. This trend is small but reliable — 87.5% of countries saw their rates of eating disorders increase since 1990.

(If that’s not enough for you, we can mention that in 1985 the New York Times reported, “before the 1970’s, most people had never heard of anorexia nervosa.” Writing in the 1980s, presumably they would know.)

There are other ways to look at the relationship. For example, we can compare the most obese countries to the countries with the highest rates of eating disorders:

Share of Adults that are Obese, 2016. Reproduced from ourworldindata.org under the CC BY 4.0 license.
Share of Population with an Eating Disorder, 2016. Reproduced from ourworldindata.org under the CC BY 4.0 license. 

With the exception of a few notable outliers (genetically homogeneous South Korea and Japan), these match up really well. The fit isn’t perfect, but we shouldn’t expect it to be. There are large genetic differences and differences in healthcare practices between these countries. They may use different criteria to diagnose eating disorders. But even given these concerns, we still see pretty strong associations — Chile, Argentina, and Uruguay are the most obese countries in South America, and they also have the highest rates of eating disorders.

We can go one step further. Looking at the data, we see that these are statistically related. In 2016, rates of eating disorders were correlated with obesity in the 185 countries where we have measures for both, r = .33, p < .001. If we remove the five tiny island nations with abnormally high (> 45%) obesity (Kiribati, Marshall Islands, Micronesia, Samoa, and Tonga), all of them with populations of less than 200,000 people, the correlation is r = .46:

Prevalence of eating disorders and obesity, 2016. Kiribati, Marshall Islands, Micronesia, Samoa, and Tonga not shown.

We see the same correlation between rates of obesity and rates of eating disorders when we look at the data from 1990, r = .37, p < .001.

Perhaps most compelling, we find that the rate of change in obesity between 1990 and 2016 is correlated with the rate of change in eating disorders between 1990 and 2016. The correlation is r = .26, p = .0004, and it’s r = .30 if we kick out Equatorial Guinea, a country where the rates of eating disorders tripled between 1990 and 2016, when none of the other countries even had their rates double. You can see those data (minus Equatorial Guinea) below:

 Increase in the prevalence of eating disorders and obesity, 1990-2016. Equatorial Guinea not shown.

That’s no joke. The countries that are becoming more obese are also having higher and higher rates of eating disorders.

We even see signs of a paradoxical reaction in some of the contaminants we reviewed earlier. You’ll remember that when mice are exposed to low doses of PFOA in-utero, they are fatter as adults — but when mice are exposed to high doses as adults, they lose weight instead. The dose and the stage of development at exposure seems to matter, at least in mice. It’s notable that anorexia most often occurs in teenagers and young adults, especially young women. Are young women being exposed to large doses all of a sudden, just as they start going through puberty? Where would these huge doses come from? It may not be that much of a stretch — PFAS are included in many cosmetics.

In one study of 3M employees, higher PFOS levels led to a higher average BMI, but also to a wider range in general. The lightest people in the study had some of the highest levels of PFOS in their blood. The quartile with the least PFOS in their blood had an average BMI of 25.8 and a range of BMIs from 19.2 to 40. The quartile with the most PFOS in their blood had an average BMI of 27.2 and a range of BMIs from 17.8 to 45.5. Remember, a BMI of below 18.5 is considered underweight.

In the study of newborn deliveries in Baltimore that we mentioned earlier, researchers found that obese mothers had babies with higher levels of PFOS than mothers of a healthy weight. But underweight mothers also had babies with higher levels of PFOS. In fact, babies from underweight mothers had the highest levels of PFOS exposure, 5.9 ng/mL, compared to 5.4 ng/mL in obese mothers, and 4.8 ng/mL in mothers of normal weight. “The finding that levels were higher among obese and underweight mothers is interesting,” they say, “but does not have an obvious explanation.” Knowing what we know now, the obvious explanation is that PFOS usually causes weight gain, but like all drugs, it sometimes has a paradoxical reaction, resulting in weight loss instead.


28 thoughts on “A Chemical Hunger – Part VIII: Paradoxical Reactions

  1. This is the most compelling twist yet!

    A strong competitor to the environmental chemical hypothesis is the overly palatable food hypothesis. It was, in my opinion, quite feasible that food crossed some threshold and became a superstimulus causing people to systematically overeat, which then caused metabolic dysregulation in other ways (diabetes) reinforcing the problem.

    BUT the mere existence of bags of cheetos seems incredibly unlikely to explain a correlated rise in anorexia, and that too synchronized across 30 years (negating the impact of new media types) as you’re describing.

    Great series!

    I wonder if there’s a way to look for populations that have been given an intervention which artificially reduces PFOS in their blood. E.g. patients who regularly get dialysis for non-diabetes related reasons (assuming dialysis can actually reduce PFOS more effectively than kidneys).


  2. MasteringTheClassics says:

    IMO the anorexia correlations with obesity are easily explained by fatter populations being more into dieting, and anorexia just being excessive dieting. This doesn’t obviate this entire post, but it does a good chunk of it in.


      1. MasteringTheClassics says:

        It would not change my mind on the cause of most (not all) human anorexia, but it would provide strong evidence for your overall thesis.


      2. MasteringTheClassics says:

        And in defense of my intransigence, note the existence of atypical anorexia, wherein someone is engaging in anorexic behaviors but is not alarmingly thin. This phenomenon actually provides evidence that your overall thesis is correct (something has gone badly wrong with these people’s set points), but it’s anorexia without any apparent paradoxical reaction.


        1. Interesting point. We’ll think about this more — there might be an explanation based on the fact that the lipostat is really multiple different control mechanisms, which might “disagree” about the appropriate body weight.


  3. Mr A says:

    I love reading all these! So fascinating and well written, hopefully you guys are the spark that will solve overweightness/obesity in our modern world, look forward to the next parts !


  4. I’m confused by the anorexia leptin connection you’re making. Anorexics have low leptin levels, which generally increases hunger, so are you suggesting anorexics have a paradoxical relationship to leptin? Because otherwise I’m not clear why low leptin levels would be an explanation for anorexia rather than a consequence of it. Or am I misunderstanding the connection you’re positing?

    Wouldn’t the most straightforward explanation of what links the leptin insensitive teens, the Minnesota starvation experiment subjects, and anorexics be that in all cases their bodies believe them to be starving (correctly, in the latter two cases)? Why would the assumption be that the anorexics’ food obsession comes from their lipostat sensing they’re too fat, rather than from their lipostat correctly sensing that they’re starving? I can see the above average levels of NEAT as one evidence point, but doesn’t that seem to point the opposite way to the low leptin levels (which you’d expect to correlate with the lipostat sensing the body is underweight)?

    (Btw I do agree that there’s good evidence that anorexia is not just a social phenomena, particularly the animal studies that you’re planning to go over next. I just don’t understand the leptin point.)


    1. So, these are good questions, let us try to explain. The leptin and lithium paper in particular is a weird case. The paper suggests that some of the patients lost weight when taking lithium — “The change in leptin levels correlated with the change in BMI (r = 0.58, p < 0.05) and that in the weight (r = 0.57, p < 0.05)" and "The leptin levels were decreased in 6 patients" — but never outright says that some people lost weight. So we tried to describe that without saying anything that isn't actually in the paper.

      The other paragraph on leptin is also a little messy. The part we were trying to emphasize was that "leptin levels are correlated with symptoms other than just BMI" — that is to say, not only do these people with low body weights have low levels of leptin, their leptin levels are also correlated with the "psychological" symptoms that most people would expect to be social, but the correlation with leptin suggests they are biological instead. It would have been better to quote sections like "weight gain in AN patients can induce relative hyperleptinaemia in comparison to controls matched for body mass index" and "leptin levels in patients at target weight have been found to be disproportionately high in comparison to a healthy control group upon adjustment for BMI and % body fat."

      It's also weird that leptin levels are lower in AN patients than in matched BMI controls. We agree that this is the opposite of what we might expect from how we understand leptin, but it's certainly in line with the idea that something biological is going on!

      So we don't want to pretend we know exactly what is going on with leptin in anorexia, but we do think this all strongly points to a biological cause over a social/psychological cause.


      1. Thanks for the response! I have a few thoughts but they’re all very speculative. Sharing in case they help anyone think things through.

        > The leptin and lithium paper in particular is a weird case. The paper suggests that some of the patients lost weight when taking lithium — “The change in leptin levels correlated with the change in BMI (r = 0.58, p < 0.05) and that in the weight (r = 0.57, p The part we were trying to emphasize was that “leptin levels are correlated with symptoms other than just BMI” — that is to say, not only do these people with low body weights have low levels of leptin, their leptin levels are also correlated with the “psychological” symptoms that most people would expect to be social, but the correlation with leptin suggests they are biological instead.

        So my main thought on this is (and I give the explanation for this in more detail below): the weird leptin levels in anorexics can be explained as a response to their food intake, rather than as an explanation for it. In this theory, their psychological symptoms correlate with leptin levels because the leptin levels are an indicator of starvation (which is impacted by recent caloric intake, not just weight, hence most of the differences vs controls). So the ‘psychological’ symptoms are, as you say, a biological response, not a social one. But that biological response is a starvation response (or, in the weight restored anorexics, a response to recent intensive refeeding), rather than anything more complex than that.

        So to flesh that out a bit more (sorry):

        > It would have been better to quote sections like “weight gain in AN patients can induce relative hyperleptinaemia in comparison to controls matched for body mass index” and “leptin levels in patients at target weight have been found to be disproportionately high in comparison to a healthy control group upon adjustment for BMI and % body fat.”

        I think this is potentially explained by this, from your linked study (https://www.nature.com/articles/4001909/): “In healthy individuals, short-term overeating preceding any substantial weight gain induces an upregulation of leptin secretion”

        Recently weight restored anorexics have basically ‘overeaten’ to restore weight. And as the study says: “If however a patient maintains her target weight for a period of several weeks, leptin levels initially fluctuate substantially to then drop into the normal range.” (i.e. when they’re eating at maintenance levels consistently, leptin levels are normal.)

        > It’s also weird that leptin levels are lower in AN patients than in matched BMI controls.

        I think this is more easily explained by a) the fact underweight anorexics are presumably actively restricting caloric intake (which we know reduces leptin: https://bit.ly/3kxwkJM), and b) the fact that they tend to have lower fat mass relative to controls. As they say in the study: “In studies that have compared leptin levels of AN patients with those of healthy underweight females (Table 1), both per cent fat mass and leptin levels were higher in the healthy underweight, despite overlapping of BMI; it appears that the higher fat mass in the healthy underweight largely explains this finding; a contributing role of acute dietary restriction in the patients appears possible.”

        The factor that isn’t explained by my theory is the NEAT point. That feels like the most interesting thing to focus on, from my perspective. Sorry if that’s an obnoxiously long and nitpicky response! I’m just a bit obsessed with this topic, as a former anorexia patient.


    2. [Adding to clarify your reply]: I’m also curious about this part: ‘Lithium increases leptin levels in most patients, and this is presumably part of the mechanism that causes people to gain weight on lithium’. Would we not expect the increased leptin to be a consequence of the weight gain, rather than a cause of it (if that’s what you’re suggesting here)? Isn’t increased leptin what we’d expect in most cases where someone has gained fat?


  5. Lyla says:

    I’ve been wondering about the decrease in smoking cigarettes as we see the rise in obesity, particularly in the US. Nicotine causes weight loss, loss of appetite, and a more demanding metabolism. Are the studies and samples across decades controlled or disaggregated for smokers? If not, why do you think this is a non-issue? My thought was that perhaps we should be heavier with the modern western lifestyle, but because people smoked more when we established base measurements (early-mid 1900s) this was obscured.


    1. People didn’t smoke much around 1900. We have some data from around then and people back then were even leaner than they were in the 1960s. Many smokers today are obese, so something has still changed. Cigarette consumption dropped in the US since 1980, but hasn’t changed much in some countries (e.g. Spain) and has gone up in some countries (e.g. Norway and China as far as we can tell). Despite these differences, rates of obesity are on the rise everywhere. So declining smoking rates might make a small difference in the US but can’t be making a big difference.


  6. HoldMyGin says:

    A very strong piece of supporting evidence for this theory would be if we find the same inverse correlation between altitude and anorexia as between altitude and obesity. A quick search isn’t turning up county-level anorexia data for me, unfortunately.


  7. Those results are not, in fact, paradoxical. The setting point of fat is determined by balance of different amino acids, peptides and so on which are governed by population balance equations, like Lotka–Volterra, but less chaotic (otherwise we would explode). The intervention shifts stable orbit of those equations (actually something like attractor or stable orbit) with averages of attractor being important for the setting point of fat percentage. Sometimes intervention shifts to wrong (i.e. higher averages) attractor, it seems there are many of them.

    Liked by 1 person

  8. Mks Mary says:

    I keep expecting that you’re going to do an “interlude” on endocrine disruptors. I don’t know much about them, but there was sure a furor about BPA when I was shopping for baby bottles for my kids, and I remember reading that the other chemicals which replaced it in “BPA free” plastics may also be endocrine disruptors. There was also a furor a while ago about low sperm counts worldwide, and speculation about whether that was related to endocrine disruptors (though I understand the evidence for declining sperm counts is kinda weak?) Anyway, when your series of posts here was shared on Metafilter, at least one commenter complained forcefully about the fact that it didn’t discuss the havoc that plastics may or may not be wreaking with our hormones. I don’t know if that’s warranted, but I’d certainly be interested in at least a brief discussion of it.

    ( https://www.metafilter.com/192025/A-riddle-sauted-in-a-mystery-deglazed-with-an-enigma if you want to see the rest of Metafilter’s reactions to this series.)

    Liked by 1 person

  9. Chris says:

    You mention antidepressants and antipsychotics a few times. Our bodies they fully break then down and many are dumped. The first antidepressants hit the market in the early 50s and the first SSRIs in the early 80s. Overall use of them has been spiking too. We know that crustaceans exposed to antidepressants spend more time in sunlight, and fish in a lake exposed to antianxiety meds explored more and took greater risks. A lot of these meds cause weight changes and obviously changes in the brain. Seems like a potential contributor as well.

    Liked by 1 person

  10. This was a great read, both interesting but very informative. Based on what I picked up, if I wanted to lose weight (fat) I should avoid consuming pesticides and antidepressants as well as avoid certain plastics. I am genuinely interested in how to put this into practice. Should I distill my own drinking water? Do I need to get rid of all of my plastic plates and cups? It seems like avoiding meat is ideal, but how do I know that the vegetables I eat are low in pesticides, or is there an effective method to clean my produce? Is there any other lifestyle changes you would recommend?


    1. Yeah, all good questions. It probably depends on which contaminant(s) you think are most likely. If you think it’s lithium, don’t work closely with auto grease. If you think it’s PFAS, replace your carpets with hardwood floors, don’t volunteer as a firefighter, etc. Distilled water is probably a fine idea but also most people probably don’t get most of their exposure from drinking water so it may not make a big difference. We’re going to look into it and try to make some recommendations soon. Surest bet might be to move to Colorado (or Thailand if you want to go more extreme)?


  11. Todd Cole says:

    “Lithium increases leptin levels in most patients, and this is presumably part of the mechanism that causes people to gain weight on lithium.”
    That sounds backwards. Wouldn’t increasing leptin make people eat less, and thus lose weight?


  12. BZ says:

    Overall I find this series compelling, but this post seems like the weakest of the bunch. To me the phrase “paradoxical reaction” appears as a special pleading to recast evidence against your theory into evidence for it. The positive evidence for your theory is just evidence, the neutral evidence is below statistical significance and can safely be ignored, and the negative evidence is a paradoxical reaction. With this kind of reasoning, you can convince yourself of anything.

    If you had a really clear hypothesis for the mechanism of the paradoxical reaction, ideally with details about when it should occur and how strong it should be, I’d feel more comfortable considering it to be evidence in favor of your theory.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s