David Hume is for me the earliest philosopher that can be read in a total modern, contemporaneous, no nonsense, relatable, atheistic key. These are notes on what I got out of his first Enquiry (written approx. 1748), not a scholastic study of what he exactly meant.

Core idea

The core idea is that human reasoning is at best inductive. This has profound implications about our knowledge of the world.

Humans are animals. There are pre-build body structures. Sensors like eyes, ears etc. are linked to the brain and give us perceptions: information about what is happening in the environment now. However the pre-build structures go deeper e.g. to the wiring of the brain. You can see some of this in animal instincts such as a hen that hatches eggs. The pre-wiring in the brain gives us things from basic ones such as memory to the propensity to learn and speak a language.

The brain can thus store, recall and recombine ideas that originated in perceptions. Some are simpler like associating by resemblance or by space or time relationship, but the crown jewel is causation, that thing that makes us provide justifications: this is caused by that.

Hume made the observation that causation is really two things: a body of experience that A is followed by B and a prediction that next time A will also be followed by B. Based on the body of experience some probability is involved. We’re pretty certain of some things, like we’re all pretty likely to die. In December, in England the weather is likely to be cold, but not a certainty.

Then the claim is that this is all that it is about human reasoning. There is no more profound reasoning mechanism than the causation mechanism described above.

As an example of an application, we say that a piece of lead falls to the ground because of gravity. We might go deeper and say that the piece of lead is formed out of atoms with nucleus of the atom holding most of the mass etc. But in the end there is no pure reasoning explanation decoupled of experience, there is no ultimate reason to the question “why?”. It all boils down to: that’s what we’ve seen happening in the past, we’ve got a theory which involves the gravitation word that seems to predict that quite well. When the theory predictions stop working we’ll try to come with some other theory.

Historical context

Philosophy is largely influenced by the ancient Greeks. Thorough the accident that the emperors of the Roman Empire decided to become Christians, a lot of the diversity of various schools of philosophy were inhibited/prosecuted leaving largely Aristotle due to the his view of the world kind of matching the Christian doctrine. For Aristotle the world is structured in concentric spheres: the Earth, water, air, Moon, planets, Sun and finally the immutable stars, fitting the Christian theory that the hell is down and heaven is up. For Aristotle stones fall to the ground because they are made of “earth” and they (actively) aim to go to the “earth sphere”.

Through various other historical accidents such as the Ottomans’ effective use of gun powder to topple the last remains of the Roman Empire and conquer Constantinople alternative views of the world reach again the Western Europe. Eventually we reach the likes of Galileo Galilei that challenged systematically the Aristotelian view: - the observation of a supernova questions the “immutable stars”

the Moon has mountains, whose height can be measured based on the shadows they make, so it’s more like the earth and yet does not fall
Jupiter has moons, rotating around it, not around the Earth
Venus has phases, crescents that indicate that it rotates the Sun, not the Earth

This led to ideas that maybe a stone is passive rather than active, it will just continue it’s movement until some external force acts on it. This eventually leads to Newtonian mechanics.

When I learned philosophy in school, Hume was given but a short entry under “Locke, Berkley and Hume, the British empiricists”, that is opposed to rationalists such as Descartes. How convenient, but also how misguided in it’s simplicity which delayed for so long me reading Hume. You see, Descartes in his Method book starts quite well. He claims that he’ll only accept things supported by reason, with the famous “cogito ergo sum”: I think therefore I exist. This is still a bit circular because it uses a language that has a subject which already identified an entity, but not too bad, it’s an interesting argument. But then he shortly convinces himself “rationally” that God exists and, being a good God, would not let Descartes delude himself. That sounds mystical to me rather than rational. For me Hume is much more rational. Descartes was a good mathematician though.

Hume’s “An Enquiry Concerning Human Understanding” was published in 1748 (the version I read was based on the 1777 edition). It’s called Hume’s first Enquiry to distinguish it from a later (also similarly named) “An Enquiry Concerning the Principles of Morals”. The first Enquiry is a revision of ideas from his similarly named Treatise published (anonymously) earlier in 1739. Hume brings lots of ideas to the table, though the Enquiry gains by being shorter and more focused than the Treatise.

An example

This is a personal story that I remembered while trying to figure out how many repeated experiences do we need to form a causation in our brains.

My dad took me and my not even two years old daughter to the cemetery where his mum, my grandma, is buried and he lit a candle. And I saw something in my daughter’s eye that made me shout at her: “DO NOT TOUCH THAT CANDLE!”. As we departed, and carefully stepped on the narrow path between the tombs I hear my daughter shouting and crying. She touched the candle: “It jumped at me”. I checked she was fine: it probably hurt, hence the shouting, but there was no burn damage. “No, it did not jump, you touched it, I told you not to”.

You can see how from her point of view, at that age, you can touch anything, it’s fine, even glowy things like laptop monitors and tablets and phone screens, they are shiny and it’s fine to touch. The candle is a bit different, but surely it should be the same. But she learned that the candle flames are an exception from the rule.

Now on the surface, one experience was enough, but it’s more complicated, we need some repetition for the induction:

She already had a large body of experience of how things look like: that enabled her to notice that the flame of the candle is a bit different.
I told her not to touch it. She benefited from my experience. We learn not just from direct experience, but from the experience of others e.g. by reading.
The heat sensing nerves and pain sensation are pre-wired structures. There is an evolutionary reasons for such structures to emerge and learn quickly that fire is dangerous.

The story continues as some months later I was baking bread in a Dutch oven (basically a large-ish cast iron pot with a lid). As I took it out of the oven I saw that very same spark in her eyes that she was going to go and touch it. My dilemma was that I had to put the pot down on the table as the heat was already coming though the kitchen gloves, but my hands were on the pot handle and could not stop her touching it, so I used my elbows and shouting the best I could.

But you can see again how from her point of view you can touch anything, but not the candle flame, but the pot looks nothing like the candle flame and yet we need another exception from the rules for the (hot) pot.

The examples are personal and somehow trivial, but they show Hume’s causation at work, and that causation is enabled by mechanisms that even children employ.

Where is Hume wrong?

He seems to imply that mathematical reasoning (Euclid’s Elements style) and Newtonian physics are something different to the causation mechanism that involves induction, in his words they are demonstrative, concerning relations of ideas rather than matters of fact.

This is to be expected as the understanding of non-Euclidian geometries, of formal systems, of the limitations of the axiomatic systems, they all came later in time. On the physics side relativity, quantum mechanics also came later in time. The argument is a bit more complex for mathematics, but for physics it’s clearer that the inductive method is at work. We build a body of evidence, then a theory that gives predictions. If the predictions continue to match evidence, then fine. Else build a new theory: relativity for high speeds, quantum mechanics for atomic scale etc.

Consequences

We can’t extend the inference to infer something different from what’s observed. The causation works by noticing that A is followed by B and making the prediction that A will be followed by B, which we condense into “A causes B”. But it’s not valid to say “Look how much structure there is in the world, therefore a God must have created the world”, we can only infer something “Look how much structure we found in the world, therefore we expect we’ll find more structure in the world”. This generalises into paying attention if, in an argument, the things that follow “therefore” are actually related to the “experience” preceding it.

In practice we’ll have multiple sources of experience. Some will contradict each other. Many Christian churches will offer bread and wine and say that they are the body and blood of Christ. On the other side we have the direct experience that they are bread and wine. We know that other people either have an interest to delude (the priest would loose their job if they admit they are wrong) and we should not let that overtake more direct and simple experience. This generalises into paying attention to the source and strength of the body of experience that is used as the basis of the reasoning.

Detailed reading notes

species of philosophy: easy vs abstruse
perceptions of the mind: impressions and ideas
3 kinds of connections between ideas: resemblance, contiguity, causation
every idea we have is a transformation of something we experienced as an impression or a sentiment
objects of human reason or enquiry: relations of ideas e.g. Pythagorean theorem and matters of fact e.g. the sun will raise tomorrow
reasonings are of two kinds: demonstrative, concerning relations of ideas OR probable reasonings, concerning matters of fact
reasonings on matter of fact based on cause and effect
causation based on experience, not reasoning or à priory
causation is based on inference between I’ve seen this object followed by this effect AND I foresee that other similar objects will be followed by similar effects
from the first appearance of an object, we never can conjecture what effect will result from it
the issue with unpalatable consequences of these ideas, e.g. with regards to religion and ethics, the answer is that you don’t asses truth based on convenience, also these particular ideas are not incompatible with ethical behaviour
on miracles: extraordinary claims contradicting experience need to be backed by extraordinary proofs e.g. claims of resurrection after death (a brave man Hume in 1770); Also good insight on how our curiosity e.g. about travels to exotic places provides incentives for people to exagerate and lie

Other notes

in Hume’s 18th century moral meant related to man e.g. moral philosophy is the study of man, not ethics
I read Descartes as a teenager: cogito ergo sum was impresive at the time, though the pure reasoning leading to the existence of god was meah, unconvincing. Hume is heads and shoulders above.
Similar ideas in Simon’s book: that we evolved in a slowly changing world which allow inference that things that happened in the past are likely to happen again. Another way to look at it is that for things that are not predictable we can’t establish causation.
Similar idea in Thinking Fast and Slow demonstrative reasoning might be system 2
moral facts the truth of which is determined by experiment similar to Popper’ falsifiability
For historical reasons it overstates the truth in mathematical reasoning or Newtonian physics and determinism
the human brain still has predispositions e.g to acquire language, to think in terms of agents
though one experience might be enough, not really: e.g. my daughter touching the flame of a candle
one can question the details, e.g. the delineation of the terms employed, but Hume is surprisingly right, especially around cause and effect being probabilistic thinking based on experience
the comparison with animals, precursor to Darwin? The observation leads me to question rationality as something unique to humans
the chapter on miracles is amazing, a precursor of Sagan’s extraordinary claims need extraordinary proofs
when discussing atheism he takes the politically prudent approach of putting his ideas in the mouth of an imaginary friend. Hume was wise to not engage in polemics (other than publish his books) during his lifetime.
end reference to books about “quantity and number” is really about the Euclid approach of using numbers and magnitudes, Hume was clearly familiar with Euclid
determinism: can’t know if world is deterministic, but when we can create areas which are deterministic, there are advantages
the implication on artificial intelligence: it uses the same probabilistic prediction based on training data, fundamentally humans don’t have a better reasoning mechanism (other than the body of evidence/training data)

Notes on other Hume writings

On the immortality of soul (separate from Enquiry)

the first and last paragraph that gives the appearance of religious orthodoxy: hume’s ability to navigate the historical context of his time in order to deliver his ideas
the paragraph on “women’s inferior capacity” is a reflection of the ideas of his time. Not everything he says is right. In particular though in this his argument that given this, it’s unfair to inflict the same threat of eternal damnation on wrong behaviour. That raises interesting questions on the equal treatment of genders in religions, which might be in selected Bible paragraphs, but clearly not reflected in historical practice e.g. male only popes

Letters

“There is very little more delicacy in telling a man he speaks nonsense by implication, than in saying so directly”

Closing remarks

I really enjoyed reading Hume. He provides a compelling model on how humans reason. This has wide ranging consequences and can help with better reasoning. He also brings a “no nonsense”, wise approach to life and situations. Although illustrated on religious circumstances, humans have an incentive to mislead and provide miracle solutions, magic bullets. This is part of the human nature and it’s useful to develop a nose for detecting such false reasonings and not letting us deluded by them.