I just finished Johann Hari’s Lost Connections.

Its argument, in brief, is that the “chemical imbalance” theory of mental illness was always ill-founded, that its been pushed by the pharmaceutical companies, that depression and anxiety are complex and caused by a tangle of biological, psychological and social factors which can be abstracted as missing “connections”: to nature, to each other, to meaningful work, etc. These are not just individual ailments: they are societal issues that we need to resolve, and he points to legalizing psilocybin and creating a Universal Basic Income as reasonable first steps.

So far so good. I agree with the overall story here but have some serious points of departure.

Sex and exercise

Even on his own terms, there are some major omissions in the list of lost connections: Hari completely ignores the body. There’s nothing in this book about exercise or physical health, which from what I understand is an extremely important part of alleviating these conditions. He also misses love and sex. He talks about meaningful connections to other people, but love and sex are their own complex part of human connection.

Bad science

He is also far too uncritical of the psychological and sociological research that he cites. I suspect many of the citations would not hold up in replications, and many of the effect sizes are so large as to be completely implausible: for example, a single primary care visit where a doctor asks about childhood trauma is alleged to reduce health expenditures over the following year by 35%. No way. I doubt this book would hold up to a Guzey-on-Walker or Banana-on-Thaler style takedown.

In that same vein, there are serious issues with causation in many of these studies. They often have the shape “people with bad life experience/condition X have higher rates of depression, and X precedes depression, therefore X is a cause of depression.” For X substitute poverty, bad health, childhood trauma, having an unfulfilling job etc. This is an extremely unsatisfying story. First, most bad things are correlated (in today’s society). Poverty, bad physical and mental health, being exposed to crime, and so forth all travel together and it takes a particular kind of blindness to pick two elements out of the maelstrom and decide that one causes the other. Much more careful methods are necessary. And second, just because depression is diagnosed after these other conditions manifest doesn’t mean that it wasn’t there, latent, before a diagnosis.


Finally, he doesn’t mention genetics. According to this random page I just found, “Heritability is probably 40-50%, and might be higher for severe depression.” This is an extremely important part of the story even if that genetic package didn’t lead to depression under other circumstances, or if it interacts with the environment in complex ways (as it likely does). For example, let’s say that one day the king decides to enslave everyone taller than 6 feet tall. The king is the proximate cause of the enslavement, but to fully understand its societal effects we would want to understand who becomes tall and why.

Style issues

I also found the style of this book quite annoying. He doesn’t just say “Cacioppo’s research describes X and Y.” Instead he’ll tell a story about visiting Cacioppo in his office, being told about the initial discovery of X. Then he’ll talk about how he went digging into the research further, and found out Y too! And he’ll refer to the researcher as “John” every time. This gets old fast.

Overall: pretty good

This reinforced some beliefs I already had and provided a fairly lucid framework for tying them together. I don’t think I learned all that much but it was a decent read, and good to get some practice thinking critically about bad psych/social science papers. However this might be extremely helpful for someone who hasn’t been exposed to these ideas, and thinks of mental health as a purely chemical condition, especially since it details the author’s own struggles and might be relatable for people in a similar boat.