At the center of causality is the concept of a source for a process to come into being,

a generating a condition underlying an observation.

We often talk about primordial causes as first principles.

We also think in terms of what we call a mechanistic model, a model that can be followed from cause to effect

like in an algorithm.

A whimsical and rather famous picture of what came to be known later as an example of a Rube Goldberg machine, after

its creator (a cartoonist), is the perfect illustration of a mechanistic model because in such a rather silly

convoluted mechanism, to every process corresponds an action and a consequence all tied in a long chain of

causes and effects.

In this cartoon, a so-called

self-operating napkin is activated by the action of the person using a spoon to eat soup.

It is comical because the mechanism to activate the napkin performs a series of sophisticated, if not very silly,

tasks only to achieve a rather mundane final action, which is to move the napkin.

A Rube Goldberg machine, or just a Goldberg machine for short, illustrates another important aspect in the challenge of causality discovery.

And in particular, to the area of modeling,

which is that if we were not able to watch the processes leading to the movement of the napkin,

one would not be able to rule out a Goldberg machine as a generating mechanism.

The funniest thing about a Goldberg machine is that it is very transparent as we are seeing its inner workings.

But if we were only able to look at the napkin, you wouldn't be able to know if it is the hand of a person moving

the napkin or a series of sophisticated or silly actions behind it.

In other words, a Goldberg machine also illustrates that, in contrast to the cartoon,

It is impossible in general to tell what is behind.

In some way,

I guess we would like to say that a Rube Goldberg machine is among the least likely explanations for what happens in the world.

But we truly don't know and we have to find arguments in favor or against those explanations.

Perhaps what appears to be funny in this cartoon is not that far from what actually happens for example inside the human mind,

or at the molecular level even if it is not exactly happening with gears, pulleys, and levers. For many centuries,

the geocentric model was based on what we called epicycles.

This model was used as a model of our universe to explain the movement of the stars and planets in the sky,

keeping the earth at the center of the universe.

In this model, planet were following an additional movement around a small circle called an

epicycle, which in turn moves along a larger circle called a different.

Both circles rotated clockwise and were roughly parallel to the plane of the sun's orbit.

It explained the apparent retrograde motion of the five planets known at the time,

but one can see how complex running the epicycles model was.

Even if the epicycles model of the universe was followed, it was able to explain and describe with virtually the same

accuracy and modern models, such as the heliocentric model, the movements of the objects of the sky.

One of the most difficult movements to explain was that of a planet because planets appeared to move in one direction

and suddenly change to the opposite direction depending on the relative position of Earth and the planet their regular

revolution paths around the Sun.

This retrograde movement was difficult to explain without placing the Earth outside the center of the solar system,

so the epicycles were needed for the early eccentric model to be causal and mechanistic.

Moreover, both the epicycles and later the heliocentric models have about the same predictive power for certain features,

meaning that it was possible to run the models into the future to simulate the movements of the Sun,

planets, and stars just as they appeared moving in the sky to make predictions.

For example, it explained changes in the apparent distances of the planets from the Earth.

So both the geocentric and heliocentric

interpretations allowed mechanistic models with the epicycles corresponding to the mechanistic geocentric model and,

for example,

Kepler's laws of planetary motion corresponding to the mechanistic model of the Copernican geocentric model.

Clearly, the effectiveness of both the geocentric and heliocentric models are an indication that generating mechanisms and

predictions are not necessarily related to each other.

Nevertheless,

mechanistic models are very important because they are not passive descriptions of an active process.

They can actually be built, run, and followed step by step. This is why they are called mechanistic.

A mechanistic model may reveal more information by virtue of being

mechanistic as opposed to be, for example, only

descriptive. For example, in the heliocentric model, if the Sun cycle requires two gears

twenty times larger than the gear for the Earth's moon,

the mechanical model would suggest that the diameter of the Sun is 2 multiplied by 20 times larger

than that of the moon and would also shed light on the moon distance to the Sun, especially

under the observation of Sun eclipses where the Moon and Sun appeared to be of about the same size.

Geocentric models would have required more convoluted explanations to reveal just the right distances of at all times and

this makes mechanistic models much more important than descriptive.