Several ideas presented by David Hume never clicked for me until I read this great piece by Darren Allen on consciousness, most notably the metaphor of the fourth dimension, and how it helps us make sense of a lot of problems in philosophy, related to the irreconciliable perspectives that emerge from dualism.
One of the ideas that I actually rederived for myself before even knowing who Hume was, though I didn't examine the consequences very much, was the idea that causality cannot be observed in the world out there, only correlations can, and then our mind jumps in and adds in commentary to make sense of it all. We can call this the problem of causation, making sure to distinguish it from the problem of induction which I will discuss later here. Another way to phrase it is that observations can only bring you correlations, whereas causality is brought in by a mental model of Reality.
I thought about this because my classes on statistics always focused on correlations between the variables observed, but never on causality, and I thought to myself about what would be required to establish that two variables are causally linked with one another, and I realized that causality is a function of explanation, not of something which is undeniably 'out there'.
What a scientific experiment does is give you increasing levels that you can trust the model, by having more data, and by isolating variables and seeing if the results are affected, but there is never any fundamental certainty that the model is true, because models can at most be useful at making predictions (within a certain rage of parameters), never true in the absolute sense.
How do you know for instance that the future will not magically be bent out of shape by some external metaphysical entity, that the laws of physics do not change over time, or in general, that the future will behave like the past? Those questions, as fanciful as they may seem, point to the more general case of the last question, a very relevant problem presented by David Hume which is called the problem of induction, and which deserves to be taken very seriously, because it reveals some the weaknesses at the foundation of modern science.
The problem of induction thus asks us: how do we know that the future will behave like the past, or in general, how can we be sure that our generalization from observed to unobserved cases is justified? The immediate problem is that it is impossible to give 'proof' of our trust without running into an obvious circularity in reasoning.
For instance, if you say that the distant past and the recent past both behaved in the same ways—in the context which is relevant to what you are observing, for instance, the sun setting and rising—then you are ... using the past, in two different forms, to make the claim that the future behaves like past. You haven't proven anything, you have implicitly assumed that the future of the present behaves like the future of the past, which is still the past, which means that what you wanted to 'prove' has been assumed to be true to begin with.
In general, if you take the assumption of uniformity of Reality, that up until now, our models that assume that they can generalize the observed data to the unobserved cases have worked, then you are ... using your observed data (of the models that have worked) to make a generalization about an unobserved case, (of the future models you intend on building), a circular reasoning.
The most practical and intellectually honest resolution to the problem of induction is to simply say that science, and inductive reasoning in general, do not give us the absolute truth with an ironclad certainty, but is instead an iterative process to arrive at more and more useful models of Reality, increasingly practical and general explanations of what goes on around us.
This is the view of science that I subscribe to, and it would certainly be better for the sake of intellectual honesty if scientists subscribed to this view too, but unfortunately, due to their reputation and the power that they hold in our world—a reputation which is undeniably dwindling as the faith in institutions is sharply declining—scientists can't help but claim that their methods and principles are not merely useful heuristics, but actually are the truth, that building a metaphysics on causality and on the idea that the future will behave like the past is totally justified, when it simply isn't.
Scientists can tell us a good deal of things when they are within their specialized field, but as soon as they step into the domain of metaphysics, what Reality actually is, they are often more ignorant than high school students. You can of course build models from those assumptions, and scientists do and are successful at predicting phenomena from those, but explaining the whole of Reality with those? Certainly not. Finite methods and ideas can go far, but they can never grasp the Infinite.
What are the problems with those assumptions, that we can take causality as a given, and that we can generalize the unobserved from the observed (in particular, that the future behaves like the past)? Well for one, if you take causality and the uniformity of Reality as a given, you must try to find some explanation about the origin of Reality. In our regular circumstances, it would be strange to say that something happened, but that there wasn't any cause for it, that it simply happened out of thin air, just like that. Scientists are certainly not the type of people who would be satisfied with such an explanation, they are the people who will constantly find more general causes to understand a given phenomenon. This habit of mind is known as the principle of sufficient reason (PSR): that everything must have a reason, a cause for being the way it is.
PSR sounds very sensible when you apply it to our usual experiences, but when you extend it to metaphysical realities, such as the origin of Reality, then it suddenly stops being so obvious. 1 If everything must have a reason in the past, it either leads to an infinite regress, that the past stretches back forever, which goes against the finitist intuition of most scientists, or if we assume that the time since the beginning of Reality is finite, we get a theory such as that of the Big Bang. People have heard about the term 'Big Bang' so much that they are no longer shocked at the utter absurdity of the statement: that Reality itself, which includes time and space, somehow started ... in a finite time from now! Scientifically inclined people love at those who explain everything in terms of God, yet that claim is nothing short of a miracle: that the beginning of time happened ... in time!
What are the alternatives to the Big Bang then? Well, claiming that the Universe is infinite could be one explanation, though it certainly doesn't explain ... how the infinite Universe came about in the first place; again the explanation that it started from God actually sounds more believable.
Another set of explanations consist from sidestepping PSR to begin with. This is where we get the Kantian tradition, though reading Kant himself is hardly recommended to anyone who values their sanity (and their time). The basic insight we get from Kant, or at least the tradition of philosophy that sprung from his fundamental ideas, 2 is that many of the principles of Reality that we get from our day-to-day perception of life, and thus take for granted, are accurate within the confines of our regular life, but are not fundamentally true, and cannot be extended to metaphysical realities.
The main principles being firstly, the principle of separation (or what Darren Allen calls facticity), that I am a subject here experiencing a separate object there, and secondly, the principle of causality, that separate things cause one another to happen in time. Yes that's right, we go back to one of the problems of David Hume, that causality cannot be taken for granted, because Kant's main work was a direct critique of the empiricist philosophy that Hume tried to build, (and ironically ended up unearthing many of the weapons that would be used against it) hence Kant writing The critique of pure reason.
Within the Kantian view, there is no 'origin' of Reality to begin with, because 'origin' is a temporal-causal concept, and Reality is fundamentally beyond those. We can apply our temporal-causal intuition to our day to day experiences, to explain for instance why the loaf of bread didn't come out right, but not to metaphysical realities. This type of dualistic split, between the finite world of day-to-day perceptions and intuitions, and the metaphysical domain of Reality, finds an echo in the past in the form of Platonism and Gnosticism, which respectively view the reality of our senses as being a mere reflection of a Higher World of Ideal Forms in the clase of Platonism, or which views our common human reality as being 'inside' the realm of the Demiurge, a false god, which itself is 'inside' the True Reality of Divine Creation, in the case of Gnosticism. In both cases, we are invited to treat our regular senses, knowledge and intuitions with more skepticism than how we would usually understand them, which is why Gnostics use the fancy term 'gnosis' to distinguish it from regular knowledge, to talk about a direct relevation of truth, unmediated by our faulty senses.
We should be careful however with taking this dualistic split too literally. Reality, being One, United, is not literally split into 'regular human reality' and 'metaphysical True Reality'. This itself is a metaphor that points at a truth deeper than what the words literally reveal, it's a metaphor of transcendence. It is useful for the self to understand this dualistic split, to inform the mind of its limitations, but for consciousness, reality and Reality are One. (see part 1 of Darren Allen's series on consciousness to get more details from the man)
Other notable ideas that Hume expanded upon are the is-ought problem, that facts by themselves can never reveal any moral conclusions, which strikes me as hardly debatable (it simply means that there is more to life than objective facts), or his argument that miracles shouldn't be supported due to their lack of empirical evidence, which is both a circular reasoning, and an idea which will deny you from the most fascinating experiences by definition (see also Taleb's idea of the Black Swan), or his skepticism on the self because it is a 'bundle of thoughts and impressions', or his discussions on the problem of evil—if God is so good and powerful, why does he allow evil to exist?—which I wrote about in a previous post.
I can't say I have read Hume directly, but nonetheless, the important questions that he asks are utterly fundamental to the metaphysics of science, and the fact that many scientists nowadays are not even aware of those is a testament to the philosophical poverty of our times. The idea that it is possible to model Reality without understanding it deeply is both very naive on a practical sense—how do you know if you haven't checked?—but also an absurd claim to make at the level of metaphysics, how can you study something that you don't even know the nature of?
These types of philosophical questions all sound like navel gazing, and to be fair, a lot of contemporary philosophy is essentially intellectual masturbation, but what they point to are the foundations for our worldview, and therefore what we consider to be solid to begin with. If you build your life on a worldview with shaky foundations, you risk pouring your energy into projects that are simply misaligned with Reality.
This is becoming more and more of a reality in our times, now that technology allows people to get more and more of what they want, or at least, what their self wants, and realizing the limitations of how that self operates, mainly, the fact that it converts intangible, ineffable experiences, the mystery of being itself, into effable, useful, totally unmysterious things it can then own, manipulate, and advertise to the world.
This is the useful part of philosophy in my opinion: understanding the nature of the self, and the fundamental problems and limitations it runs to. But understanding the self also requires some understanding of metaphysics, because self converts (again, ineffable) Reality into effable spatio-temporal reality, where things are separate from one another, and caused by other things. Taken to its extreme, a life of pure separation and causality is quite literally hell, as Darren Allen's work shows.
1 Similar to how the Axiom of Choice (AoC) in mathematics is obvious for a finite collection, such that there is no need for AoC to pick an element from a finite number of non-empty sets, but becomes a necessity for infinite collections of sets, especially the uncountable ones, and leads to all sorts of strange consequences. What is intuitive in our day to day experiences might not be so obvious when we are dealing with metaphysical foundations, or in the case of mathematics, with infinity.
2 Namely Schopenhauer, upon whose philosophy Darren Allen bases his metaphysics, though he cleaned up the confusion of what Schopenhauer calls 'will' and instead put it its place consciousness, the knowledge of the thing-in-itself by being it.
Go back to the list of blog posts
Science Hume Rationality Causality Induction Philosophy
2026-06-22