The L-Index

Using a term like nonlinear science is like referring to zoology as the study of non-elephant animals.
Stanislaw Ulam
Last updated on

January 31, 2023

Lampposts in fundamental physics


In quantum gravity, there is a commonly used phrase, “looking for your keys under the lamppost”. It is the old tale:

A policeman sees a drunk man looking for something under a lamppost and asks him “what have you lost?” “My keys” replies the drunk, and the policeman gets on his knees and looks for the keys too.  After a while, the policeman asks “are you sure you lost them here?”  “No” replies the drunk, “I lost them in the park.” “Then why are you looking for them here?” asks the policeman. "Because this is where the light is” says the drunk, indignantly.


The story goes back to Hodja, and nowadays has its own name as a cognitive bias, the [streetlight effect]. It is an extreme problem in quantum gravity because there are hardly any lampposts, so little data to go on. Maaaybe in that case it is excused? But I’ve seen it popping up in many other places where there is no excuse.


To stay in physics, one such place is observational cosmology. The physics of the early universe, inflation, etc, is based on quantum field theory. QFT applies when the system is in equilibrium and the early universe most certainly was not in equilibrium. It is very much a dynamical, non-linear system.


A few people from non-linear systems tried to work in that space using actual non-linear physics but these were aliens from a different planet of science and the cosmologists did not even notice them. [There is a lot of work adding non-linear corrections to the linear system but that is not at all the same as a full non-linear system. Just like putting your hand in a bathtub full of water is not at all the same as the water in a river. ] I asked the cosmologists about that and the answer always is “sure, it is a highly non-linear system, but we only have tools to describe a linear system”. Aka, the “shut up and calculate” principle of the scientific method.


That is incorrect. We do have tools to describe non-linear systems. They can appear weaker than the tools of non-linear physics because their predictability is inherently less, the system itself is not predicable in the sense of being able to tell where it will be in the future. But there is a different kind of questions you can ask and get good answers: what type of system it is, which can tell you the possible types of futures it can have, and most importantly what kind of disturbances are likely to affect it a lot or a little. You can also make models describing that type of system and run simulations that will tell you things about its behavior.

SIPs. Economics and what kind of answers you can get.

Lampposts everywhere


However, it is always “rough” compared to the far more “polished” description of a linear system which in some cases can even be as simple as a closed equation.


Science is a bit like the joke about the drunk who is looking under a lamppost for a key that he has lost on the other side of the street, because that’s where the light is. It has no other choice. - Noam Chomsky


I was puzzled back then, I didn’t understand  the streetlight effect is not contained in fundamental physics, or even science, it is _everywhere_. It is may be the defining characteristic of our era. . I thought it was something special to fundamental physics, perhaps because the questions are too big and the data too sparse, which made it too hard to respond correctly to the pressures of professionalized science.
The ten years since I last lived within the science walls have been a constant lesson that the streetlight effect is not contained in fundamental physics, or even science, it is _everywhere_. It is may be the defining characteristic of our era.

Lamposts in Neuroscience


The neuroscientists tell us that there are no “knobs” in the brain that control a specific thing in the person. There are multiple parts that interact with each other and the environment to produce a behavior, a pattern. It is a dynamical system. There are huge variations in the patterns between different people. It is not even just in the brain, the body and the environment (the context) are crucial. The brain as a whole is not a knob, it is one part of a dynamical system.

Executive Function - quote Barkley

Lampposts in Mental Health


The DSM is a list of knobs too. A disorder is a particular knob that has gone awry and needs retightening. Everyone knows this is not true, the BAP says
>”....”.
We understand that in reality how we deal with this is to have the person spend time with another person, the clinical psychologist, and through that interaction something changes, hopefully in the way the person desires. This is bringing one dynamical system in contact with another, in ways that experience says can change the overall dynamics in a desirable way. This is something humankind has a lot of experience with. Before clinical psychologists there were tribe elders, a priest, or your grandmother. That’s a lot of data on what works and what doesn’t, what works in the short term and what works in the long term. The knobs of the DSM are very vaguely correlated with what happens in the room, the BAP says:
> some good quote

You could take the view that the knobs are there to get health insurance and we can ignore them. But we don’t, they create problems. They of course become part of the dynamics. The person will grapple with how they fit the description. It will affect how they think of themselves. It will affect how the psychologist interacts with them, since they’ve been taught to stay aware of whether the person fits the description or not. At the very least, it adds extra work for everyone. There are good intentions there: the psychologist keeps the DSM description in their mind because they’ve been taught that if it turns out to be wrong the person will be better served by someone else. The person pays attention because they are open to whatever will make their lives better. The system sticks to it because they want to help as many people as possible, as effectively as possible.


Explain how this becomes lamppost: the change in the dynamics is to focus attention on the aspects of the person on the DSM description, not the entirety of the person. It really comes down to how much the psychologist and the person can tap into their own wisdom and experience and deviate from the instructions.
Clinical studies look at a narrow problem designed to remove the non-linearities. That’s totally how it works in the lab, almost all the time. examples: depression, again, and something else. All with good intentions.
Digital mental health wants to help people and takes this system and automates it, making it data driven. But what data? What you can capture.

Digital mental health wants to help people and takes this system and automates it, making it data driven. But what data? What you can capture.

Data and metrics


Data-driven has put the lamppost bias on steroids.


You can’t improve what you cannot measure.


That is correct and would be a great guideline **if you can get data on what matters.**
But Technology has limitations of what data it can gather, so it becomes “improve only what you can measure”.
VCs add to this with good intentions too, they want to know that you are data-driven and have metrics.
It is not that we couldn’t get better data. It’s that

  1. The problem is not well-understood because of the gap between the practitioners and the implementers
  2. The correct data will not be as prescriptive as what we imagine is required. We mistakenly think this is a problem with the data or the method instead of realizing that it is an inherent property of the system: a human in a society.
  3. The timelines and money-lines do not match the problem. The ecosystem is one of straightforward “problem- solution”. What is the problem, what is your solution, best answered if you say you have a better knob than what is currently used.
  4. Media is also best suited for simple stories. The more sophisticated publications, like the NYT, are full of “do x to fix y problem”.


We now have a new feedback loop.

All of the above have created a new story for how people think of themselves. They expect knobs, they can only see knobs. They see themselves and each other as a collection of knobs.
The stories and narratives of dynamical systems have withered to nothing. People might say things like “everything in the universe is alive” or “tap into the oneness”, which is a reflection of their correct instinct that it is a dynamical system. The language though is woo woo, wrong for our times. If you were to replace “alive” with “dynamical” it would immediately make sense but we do not do that.
Since we do not put any effort in the alternative, and all solutions are of the linear dynamics sort, we do not get to see the advantages of an alternative. It does not have a chance to get new stories, appropriate to our times.

The depression story: 

Antidepressants are about serotonin but there is no serotonin. Tell the story of where that came from (maybe separate post)
Clinical studies look at a narrow problem designed to remove the non-linearities. That’s totally how it works in the lab, almost all the time. examples: depression, again, and something else. All with good intentions.

Optimization:

Add a paragraph from the old report and add a mini-post eventually. Need mini categories!

Lamppost storytelling

All of the above have created a new story for how people think of themselves. They expect knobs, they can only see knobs. They see themselves and each other as a collection of knobs.
The stories and narratives of dynamical systems have withered to nothing. People might say things like “everything in the universe is alive” or “tap into the oneness”, which is a reflection of their correct instinct that it is a dynamical system. The language though is woo woo, wrong for our times. If you were to replace “alive” with “dynamical” it would immediately make sense but we do not do that.
Since we do not put any effort in the alternative, and all solutions are of the linear dynamics sort, we do not get to see the advantages of an alternative. It does not have a chance to get new stories, appropriate to our times.

Links & Tendrils...

Turn into good intentions story? Combine with the market? Has it made it accelerate?

Maybe it's move fast and break things, the friction in the academic system may be like democracy. Can also be upgraded some day.

See also

more