Causal Explanations Considered Harmful
superbowl.substack.com
superbowl.substack.com
This observation, if we take it seriously, will encourage us (humans in general trying to make sense of the world) to be humble about the conclusions we reach and mildly skeptical about the conclusions others reach, even when those conclusions feel really incontrovertible.
It's another thing to have a FUNCTIONAL MODEL of HOW AND WHY humans make certain categories of mistake.
Such a thing can help us sort through proposed answers to complex questions with better discernment than just the basic realization that "this was arrived at by a fallible human, therefore it could be wrong, so I won't hold it too dogmatically."
Causal graphs are only circular if you are sloppy with time! We do not know if LSD in 1968 caused hippies in 1969, but we can rule out that hippies in 1969 caused LSD in 1968. And it may seem trivial, but it's not. That things don't happen before their causes is the main tool we have to break endless debates about what caused what, but both the article writer, Scott A. and the author he reviews seem to thrive a little too well in the world of muddy explanations.
I think you're missing most of the point of those diagrams, and the concept of feedback loops in general.
Yes, the cause cannot happen before the effect. But there was no singular "LSD" event in 1968 and no singular "hippies" event in 1969 - the two phenomena were feeding on each other over time. 1968 and 1969 are just years in which people now agree the phenomena crossed some notability thresholds.
Both here and in the general case, in such feedback loops you often won't know what the original cause was - the starting point has been lost to history, and the earliest records you can find are already from when the feedback loop already started. Did some random person ingest some LSD and, because of the experience, became the very first OG hippe? Or did the OG hippie do some LSD with their friends, and the shared experience made those friends more susceptible to the hippie ideas? Who knows?
Then there's an issue of boundaries. Even if you can point at some concrete events in time, like e.g. date of founding a company, it's probably not what you're looking for anyway. That decision itself is a discrete outcome of a different feedback system that's been running for some time, and that feedback system - not the singular signing of a document - provided the influences you're really looking for.
One hippie in 1986 could have ~dreams of a massive population of newly minted hippies in 1969 (in this case: virtual hippies), and that dream could motivate them to manufacture and distribute LSD in pursuit of that goal.
Lots of causality is due to false beliefs, and at the cognitive level distortion of time may be just another distortion. What time "is" is to some degree a bit hand wavy in the first place, isn't it?
Perhaps better not to think too much on this though.
The famous counterexample metaphor is farmyard turkeys. For months, for their entire lives, the turkeys observe the diligent farmer bringing food and water, providing for their every need so that they don't have to do anything. On the day before Thanksgiving, a turkey philosopher confidently declares that, based on this prior evidence, we can conclude that the farmer, their great benefactor, exists only to serve and further the interests of turkeydom.
Aristotle, of course, used a broader ontology of causes than is commonly used today, but there should be no confusion if we are careful to define what we mean. In particular, final causes are more commonly regarded as psychological attitudes, whereby if someone acts now to bring about a certain circumstance in the future, the circumstance is caused by what they were thinking prior to the action.
There is, likewise, no logically necessary reason to assume that that causal process of metabolism will continue to work tomorrow as it always has in the past, though it is indeed reasonable to prepare for that outcome as the most likely outcome.
If someone buys something to eat at breakfast tomorrow, but dies before then, there was no future event to cause the purchase.
I think time may be similar, in that while we seem to have an innate conscious experience/perception of ~absolute time and absolute reality, the way the subconscious mind implements the portion of reality it is the source of may be substantially different.
This is a logical inconsistency. Tomorrow's breakfast does not yet exist, only your "knowing" about it. Your thought has already happened today, therefore "knowing" you will be hungry is the true cause, not the potential for breakfast itself.
On top of that, merging their effects (in policy setting for example) sort of is a barrier to attributing causality. It forces a 2 dimensional signal to go to 1 dimension. Neural networks can backpropagate attribution through that, but only with the convention that this attribution is the marginal effect. And that's a complex problem: who would be responsible for a policy change when there's a population contributing long-term for 48%, but a short term population adds 4%. Which population is the real cause? the 48% or the 4%? Is there a right answer to this question?
The problem of this causal explanation perhaps lies with the source material Scott Alexander cited without adequate critical analysis, namely David Brooks’ _BoBos in Paradise_.
“If Books Could Kill”, co-hosted by Michael Hobbes and Peter Shamshiri, analyzes and humorously takes down book-length transgressions against critical thinking. A recent episode adroitly and conclusively disassembles Brooks’ _BoBos in Paradise_ as quite harmful and comically wrongheaded. [0]
[0] https://podcasts.apple.com/us/podcast/if-books-could-kill/id...
An example from the book are the resources of recurring customers and staff. If you add marketing, you'll gain customers, and those customers will feedback and bring in more customers by word of mouth if your service is high quality (positive feedback), but you'll also lose customers with poor qualify service and that can be negative feedback. These feedbacks aren't always flowing in only one way. If marketing leads to a surge in customers, and you don't hire more staff quality will sink, and staff might leave due to being overworked . This feeds back into having less staff. The book grabs hold of the causality, you have to understand how your actions will affect the entire system and plan for it.
In terms of determining causality, it’s true that major events aren’t ever simply driven by a single cause, but philosophy has studied this to deep depths and that’s where I think the article leaves the discussion too early. The arrow of time is not discussed and neither are necessary sufficient or contributory causes. LSD and Harvard Admissions may be contributory to the events of the 1960s, but not sufficient or necessary to explain assassinations.
Also I've noticed that many substack articles are doubled up. I read the article, and then it repeats itself. What's up with that?
To do the reverse - to start with a full causal graph and trim it down to a DAG - requires you to explicitly, intentionally make your argument flawed. Doing this would be malicious. I don't think this happens often, because developing a full causal graph takes a lot of work, which you don't need if your goal is to just lie to people.
--
[0] - Or a DAG, but you rarely see people capable of conceptualizing DAGs on the Internet. Even the software industry is still mostly stuck glorifying trees, and UX experts keep telling us most people aren't capable of working with trees either!
[1] - Or interesting for other reasons. This is how you get motivated arguments which aren't technically wrong, just don't tell the whole story, but they happen to align with someone's interests.
In English, we structure sentences as either a question, answer, or command of causality. We have "independent clauses" but the moment a clause gains utility is when dependence is added to it.
The only methods we have to express noncausal ideas are additive. Like an expansion pack for the English language, we just slap on some extra grammar and jargon, then call it good. It's a lot of messy boilerplate that we have to use every time we need abstraction. It's no wonder that we have ended up with doctor-speak and programmer jargon.
This has me wondering: what if we did have first-class noncausal language? What would that even look like? How would we structure clauses? Is there even a way to do it, or is what I am proposing simply what one finds searching the antonym of causal itself: redundant, frivolous, irrelevant, or meaningless?
What if they said "sorry, we can't fix your car because it doesn't run"?
"To argue with a person who has renounced the use of reason is like
administering medicine to the dead."
Just asking for your thoughts...
I appreciate attempts to explore diagramming, and so I appreciate the article. I am also informed by past exposure to Markov chains.
I am aware of Prigogine, and so I accept that reasoning about a dynamic system relies on its dynamic state. A simple example would be:
opening the gates causes water to be released from the dam
A different example would be:
adding additional propellant makes it go faster
As for LSD, how did the author miss Tim Leary? Referring to a Leary personality diagnosis radar chart I can't find an obvious mapping in the dynamic of post vs comments. Best WAG is that neither is expressing dominance (both are seeking order and guidance), with the article more in the region of acceptance and the comments more in the range of suspicion and hostility.
https://statmodeling.stat.columbia.edu/2015/06/16/the-david-...
Obviously, there are many situations where we can't do them (e.g., randomizing people to smoking / not smoking). In those cases, the challenges the article discusses become relevant.
Dishonesty is considered harmful.
All this does nothing to undercut the validity of causality itself. I hope that's obvious, despite the clickbait title of the OP.