Tuesday, December 31, 2019

The Missing Arrows

How to read a model chart



What is the problem with this picture?

First some translation. "HÃ¥llbarhet" means Sustainability. The other circles say, in clockwise order, "economical", "technological", "cultural", and "social". The boxes say, in clockwise order, "values", "knowledge", "skills", "ecological method".

The graph is a mess. Can we be specific about why it is a mess? First, there is a feeling of confusion. How is this to be interpreted? Are the "values" only related to "economy" and "technology", and not to "sociality" and "culture"? That seems like an odd point to make. And "skills" is not linked to "technology". My dominant hypothesis is that the words in the boxes have been thrown in at random.

Second, the what do we learn from this? We learn that Sustainability affects the Economy, and that the Economy affects Sustainability. Okay? And not only that: Sustainability affects all the other circles, and is in turn affected by all of them. All the arrows are also the same size, so all the interaction are given equal importance. What does it even mean to say that Sustainability affects Culture? The graph leaves this as an exercise to the reader. The cooperative reader will certainly come up with some interaction and be content. What has the cooperative reader learnt? To fantasize.

The problem with the graph is that everything affects everything. Any experience or observation that the reader has about an interaction between two of these concepts will fit into the model. The reader will not reject the graph based on experiment: it permits everything.

When reading a chart, pay attention to which arrows are missing. The missing arrows are what make the actual claims. In science, a model with many nodes and few arrows eliminate a large part of the hypotheses, which make for more powerful predictions. In engineering, a design with few interactions between different parts makes for a system that is easier to maintain and modify.

An example of a good model with only a couple of clear interactions is the Central Dogma of molecular biology:

The central dogma of molecular biology
What claim is made here? That DNA does not produce proteins directly, but that the information goes via RNA first. It also claims that information is never added to DNA, except at replication. Another claim is that RNA and proteins do not replicate.

The extended central dogma.
This model was falsified by experiment. It turns out that RNA sometimes does replicate by itself. And also that RNA can sometimes write to DNA. Perhaps not the way a human engineer would have done it, but we want to describe nature well, so the arrows have to be added.

How to start the scientific revolution

The enlightenment was about striking arrows, not adding them. Striking the arrow from morality to disease was the foundation of medicine. Striking the arrow from planetary movements and theology to the outcome of games of chance was the foundation of the mathematics of probability. Striking the "transmutation" arrows from elements to other elements was the foundation of chemistry. Striking the arrows between human relationships and financial transactions was the foundation of economics, a less successful example.

How to use your time with an expert efficiently


Geology - always lower = older?
(suppose we chance upon a geologist)

Astronomy - assume that stars will move in a strictly periodic pattern? This is what tells planets apart.

Demography - humans are born, grow up, have children, and die - in a quite predictable pattern!

Ecology - what species do not interact directly?


How to ask questions in order to find out these hidden reductionist assumptions in different fields? [1] First of all, we should find out whether this field and this person is actually committed to being consistent. If the field makes no solid predictions and the person is agreeable, then we should be able to make them agree to pretty much any absurd hypothesis. If the field makes no predictions and the person is disagreeable, then we are dealing with a game of Mao. Mao is a card game where the rules are being made up on the fly by the game leader, the chairman. The chairman punishes transgressions of rules that have never been spoken. The disagreeable but scientifically empty field is just politics and fashion. The fiercest players win, instead of the most correct hypotheses. So as a way of finding out the truth, it serves no purpose.

Suppose now that we have good reason for believing that our expert will not agree to anything, but will also not discard our arguments simply on the basis of coming from an outsider. So the expert will treat our arguments skeptically but charitably. Now we have the basis for actually exchanging knowledge. How should we proceed to find the reductionist assumptions in the field? The problem is that these assumptions may not be so conscious for the expert. It does also not make for interesting conversation to simply ask directly "what are the hidden reductionist assumptions in your field?". One thereby misses a chance to prove to the expert that one is worthy of their best response. A clever person finds a way to find out the hidden assumptions with questions that seem to be leading a different way at first.


[1] Why even ask a person at all? Why can't we google it? This is the type of thing that is typically hard to find in written sources. Another problem is that before knowing a subject's hidden reductionist assumptions, basically nothing one reads within the subject makes deep sense.

No comments:

Post a Comment