## Was Hume right about causation?

• 42
Hume said that it was possible for events to not have causes. But what does that even mean? When we say that an event is uncaused, do we mean that, before the event in question occurred, it did not have a high probability of occurring? For instance, if a flame flickers to the left (in the absence of any breeze or external movement) perhaps we’d say that it had no cause for doing so. Because the flame could have just as easily flickered to the right or in any other direction. But wouldn’t it be more accurate to say that the flame’s leftward flickering was caused by the tendency of fire to flicker unpredictably? Maybe Hume would deny that events necessarily have probabilities of occurring before they occur. But if they do occur, then it follows that they are possible. And either all possible events are equally possible, or some have higher probabilities of occurring than others (anywhere between 0=impossible and 1=necessary). It seems obviously false that each possible event is equally likely to occur at all times, with each event having an instantaneous probability of 1/n, n being the total number of possible events. But even if it were true, we could still say that the events in this chaotic hypothetical world were caused by the universal principle of equal probability.
• 152
Inferring to probabilities also proves that we do not really have justifiable evidence of possibility. If everything is so unpredictable, that even potentials of incomprehensibly small chances can occur or reoccur, then inductive reasoning is illogical.

From A comes B, but they are conjoined together in mutual inclusiveness, they're constantly conjoined together. Whenever we encounter A, B must be present. Likewise, the statement is in equal value to saying 'when we find B, A must be present'. From there, we are fumbled with a weak form of necessity. Consequently, inductive reason becomes illogical but obviously that is not the case - however, if things are beyond our finite grasp; like the universe, infinity or God, inductive reason cannot define a logical stance towards these sorts of things.

So things that are beyond our contingent, finite grasp are measured by possibility; or epistemic justification. They cannot be measured by inductive reasoning.
• 3.8k
Correct me if I'm wrong but I don't believe Hume said that it was possible that events can be uncaused. Rather, he said that it was impossible to establish such a thing as causality. His point was epistemological - about the nature of our knowledge - and not ontological - about the nature of the world. To put it awkwardly, we lack the ability to 'know causes'. Whether or not there are 'in fact' uncaused events is not the kind of question Hume engaged.
• 1.5k
Correct, Hume never stated that acausal events were possible. The OP is conflating acausality with Hume's actual Problem, viz., that there is no logical a priori justification to assume that a sequence of events will occur (under ceteris paribus conditions) simply because they occurred countless times before (Hume offers an example of billiard balls being struck), as there is no a priori justification establishing a causal connection. Which is, to your point, a conflation of ontology with epistemology.
• 8k
Remember Hume was 'enquiring into the foundations of human understanding'. He accepted that it is possible to know deductive proofs with certainty, that a - a = 0, or a priori truths, that bachelors are unmarried men, because the conclusion is directly dependent on the premisses.

But when it comes to knowing the relationship between a cause and its effect, no such logical necessity applies, as what we really observe is only the habitual or recurring conjunction of cause and effect. We customarily call that 'causal' but it doesn't have the same binding necessity as deductive proofs. So how can we claim to know that there is such a relationship? That is his challenge.

It was just this statement which Kant said 'awoke him from his dogmatic slumbers', so have a look at Kant's response, in which he claims that he

rescues the a priori origin of the pure concepts of the understanding and the validity of the general laws of nature as laws of the understanding, in such a way that their use is limited only to experience, because their possibility has its ground merely in the relation of the understanding to experience, however, not in such a way that they are derived from experience, but that experience is derived from them, a completely reversed kind of connection which never occurred to Hume.

(my bolds).

In other words, the validity of such principles as 'the relation of cause and effect', is said to be one of the foundational elements of knowledge itself. So where Hume, the empiricist, presumes that all knowledge is derived from experience only, Kant is showing that, in order for experience to be intelligible at all, such fundamental categories as the relation of cause and effect must obtain. This is closely related to Kant's famous dictum 'Percepts without concepts are empty; concepts without percepts are blind' and his so-called 'copernican revolution in philosophy'.

(However it might be noted that in recent analytic philosophy [e.g. here], Kant's distinction between analytic and synthetic propositions has been criticized, however I think it's important to understand it first.)
• 737
Hume's beliefs about causation are antiquated. He didn't consider that there might actually exist natural law. Modern physicalist philosophers (e.g. Armstrong, Tooley, and Sosa) are "law-realists". They suggest it is reasonable to believe there actually are inviolable laws of nature.

The existence of natural law does not imply uncaused, contingent things can't exist. Adolf Grünbaum makes the case here.
• 3.4k
I think this is a case of non-overlapping magesteria. Deduction is distinct from induction. Necessity is a feature of the former but not the latter and causality is an inductive inference.

Perhaps Hume means to say that causation isn't a necessary relationship but wouldn't that be repeating the obvious, afterall isn't causality induction?
• 792
Hume's beliefs about causation are antiquated. He didn't consider that there might actually exist natural law.

On the contrary, Hume appeared to support the notion of a necessary connection between events in nature. All that he said about uncertainty and probability had to do with our knowledge of those connections, as others have pointed out.
• 5.8k

Wayfarer has it right...
• 42
I think this is a case of non-overlapping magesteria. Deduction is distinct from induction. Necessity is a feature of the former but not the latter and causality is an inductive inference.

Perhaps Hume means to say that causation isn't a necessary relationship but wouldn't that be repeating the obvious, afterall isn't causality induction?

I think you need induction to determine which causal relations exist and how they occur, but the fact that all events have causes is deductive.

The existence of natural law does not imply uncaused, contingent things can't exist. Adolf Grünbaum makes the case here.

Grunbaum's argument against the principle of sufficient reason only works if we take PSR to mean that every event has an explanation that not only suffices explain it but also precludes the possibility that any other event could have occurred in its place. I think this version of PSR goes too far. We can claim that everything needs an explanation without being determinists.

But when it comes to knowing the relationship between a cause and its effect, no such logical necessity applies, as what we really observe is only the habitual or recurring conjunction of cause and effect. We customarily call that 'causal' but it doesn't have the same binding necessity as deductive proofs. So how can we claim to know that there is such a relationship? That is his challenge.

It's true that we don't perceive causal necessitation directly through our senses. But if we observe that various types of events are constantly conjoined, can't we use a priori mathematical and modal concepts to demonstrate that it is extremely improbable that these constant conjunctions occur by random chance?

Kant thought that causation was an a priori concept through which all sensuous experience had to be filtered. I think he believed that without the concept of cause, we'd be unable to perceive time, as there would be no order to its moments. For instance, you wouldn't be able to process a basic phenomena like walking if you didn't understand that each step caused you to move forward and your position one second ago was therefore ahead of your position two seconds ago. Thus, by his logic, the only reason we understand the world as ordered by causality is that if we didn't, intelligible experience would be impossible. But I think the causal order we observe goes above and beyond what would be necessary just to allow us to differentiate between moments in time. We've discovered highly intricate principles of cause and effect both at the microscopic and macroscopic level. These principles aren't prerequisites of experience. We've been experiencing things for thousands of years without knowing about microbiology or general relativity. So why are we capable of making scientific discoveries and using them to great effect (e.g. modern technology)? To me, it seems like the best answer is that causal relations we infer into our observations actually exist in reality.

Correct me if I'm wrong but I don't believe Hume said that it was possible that events can be uncaused.

I'm pretty sure you are wrong, although most of what you said is correct. You're right that Hume's main point was to show that we cannot establish that necessary causal relations exist. But in order to make this point, he argued that there's no reason to believe that events couldn't occur without causes.

Which is, to your point, a conflation of ontology with epistemology.Maw

Yes, Hume was trying to show that we can't have knowledge of causes. But if he's correct, then it follows that causes don't necessarily exist in reality, at least as far as we know.
• 3.8k
But in order to make this point, he argued that there's no reason to believe that events couldn't occur without causes.

Produce the quote. We can take it from there. As far as I know, 'to make this point', he denied the principle of the uniformity of nature; or at least, denied that such a principle could be upheld. Not that 'there's no reason to believe that events couldn't occur without causes.'
• 42
Produce the quote. We can take it from there.

"We can never demonstrate the necessity of a cause to every new existence, without shewing at the same time the impossibility there is, that anything can ever begin to exist without some productive principle" (Treatise of Human Nature, Book1, Part 3, Section 3)

I was trying to prove him wrong on this point in the OP. You are right about epistemology vs. ontology. He is talking about what we can prove and not what exists. But I think that if we can prove something to exist, then it does exist.
• 3.8k
Hmm, fair enough. The triple negative(!) is tricky - we cannot show that it is impossible that effects exist without causes. But yes, the point bears on what can - or rather cannot - be 'shown', moreso than what is or is not the case.
• 1.6k
the fact that all events have causes is deductive.

Not axiomatic? :chin:
• 629
n other words, the validity of such principles as 'the relation of cause and effect', is said to be one of the foundational elements of knowledge itself. So where Hume, the empiricist, presumes that all knowledge is derived from experience only, Kant is showing that, in order for experience to be intelligible at all, such fundamental categories as the relation of cause and effect must obtain.

That's true, but I also have to agree with other posters that only deduction can be known for certain, and that Hume was right to say, in cases of inference, there can be no absolute certainty that causality exists.
• 1.8k

We need to remember Hume is talking about the necessity of a cause. The problem here isn't a question of proving existence. We do that every time we encounter something. Each time we encounter a concurrence of events, we've observed/proven they existed together (as fast as that goes).

The necessity of a cause asks for more. It's not just a proof of existence, but a demonstration that a presence of one state can only lead to another.

In other words, it is not about what exists, but rather what doesn't exist . If I'm dealing with necessary causality, I'm trying to the presence of something resulted in something happening rather than not. My goal is to say every time this thing occurs, it will mean this other state MUST occur rather than not.

This is the impossibility, to have or show this necessary connection. Since any causal relationship is defined by its effect, none can be stated as necessary in terms of a cause. Any cause we might consider is caught without the existence to make it case.

Will I wake up tomorrow? The fact causes might exist doesn't tell me. It's not yet defined what those causes, if any, lead to. We cannot yet say whether my heart has caused me to live tomorrow or not. I might be alive (heart working as it is right now). Or I might be dead (heart stopped).

The nature of any cause has to wait for how I exist in the future. A necessity from cause alone is impossible.
• 8k
I think the causal order we observe goes above and beyond what would be necessary just to allow us to differentiate between moments in time. We've discovered highly intricate principles of cause and effect both at the microscopic and macroscopic level. These principles aren't prerequisites of experience. We've been experiencing things for thousands of years without knowing about microbiology or general relativity.

But we still validate all of those discoveries against experience. That is what experiment and observation does, after all; you predict what will happen, and then you observe what happens, to either confirm or falsify your prediction. Experience has a fundamental role in all that. Besides, one of Kant's basic questions is how is it possible to arrive at synthetic a priori judgements, which surely are fundamental to all mathematical physics.

So why are we capable of making scientific discoveries and using them to great effect (e.g. modern technology)? To me, it seems like the best answer is that causal relations we infer into our observations actually exist in reality.

Yes, but - we have had very many contentious debates on this forum about the philosophical implications of physics. I don't want to divert this thread down that rabbit-hole, except to note that it's precisely the reality of the so-called fundamental particles which has been called into question by physics (hence the profusion of science titles about quantum mechanics with sub-titles about 'the quest for reality'.)

So, science makes predictions that work - no contest there. But that can be accomodated in terms of instrumentalism, without presuming anything about what 'actually exists in reality'. So, I'd just be mindful of the implications of presuming 'what exists' on that basis. There are still many open questions.

I also have to agree with other posters that only deduction can be known for certain, and that Hume was right to say, in cases of inference, there can be no absolute certainty that causality exists.

That is the reason why Hume is categorised with the sceptics. It is why Russell says in HWP that Hume essentially called into question the notion of scientific certainty.
• 7.9k
Hume said that we do not "observe" causes, but only constant conjunctions of kinds of events and that from that, out of "mere habit", we infer causation. A criticism of this is that Hume is privileging sight over other forms of perception. Of course we don't hear, smell, taste or touch causation either. But it might be reasonable to say that we feel causation happening to, in and with our bodies.

Conceptually the difference between correlation and causation is that causation involves some kind of force, which can be conceptualized as an exchange of energy, and in fact that notion is absolutely central to the sciences. Where do we derive our idea of force or energy? Don't we feel those in our own bodies? We feel ourselves pushing and being pushed by other things. We feel the force and resistance of gravity, the cooling of the wind and the burning of fire. We know directly, as directly as we can know anything at all, that these natural forces act on our bodies.

Another point is that there is a difference between accepting the idea that events are caused on account of in a sense knowing (not with deductive certainty of course, but by a familiarity of which there can be no genuine doubt) that causal forces are at work in our world and knowing precisely what those causes and forces are. If things are globally interconnected then we might be talking, not about exclusive causes, but merely about more or less proximal causes. And further, there is Aristotle's four categories of causes to be considered. As Kant pointed out in response to Hume, the question of causation has to do with the synthetic a priori logic of our common experience.
• 850

Judgements of experience, or judgements with empirical content, are synthetic, yes, but they have nothing to do with experience if their content be a priori, or contain an absolutely necessary relation, re: mathematics.

Would you mind, and how would I be wrong, if I re-wrote your statement as: Kant pointed out to Hume, the question of causation has to do with the synthetic a priori logic of our pure understanding, and in no way our common experience, which had thus far been described by mere habit.

“....But to synthetical judgements a priori, such aid is entirely wanting. (I do not here have the advantage of looking around in the field of experience)....” (A9, B13)

Not sure what you meant by a priori logic. Perhaps you intended it to stand for the categories, which holds cause and effect in the category of relation. While these apply to cognitions with respect to experience, they are derived a priori from conceptions alone.

Bottom line.....Kant was saying Hume misunderstood the power and function of the faculty of understanding.

Just a thought.........
• 10.6k
Hume said that it was possible for events to not have causes.

Did he? I thought he was famed for his scepticism.
• 7.9k
Would you mind, and how would I be wrong, if I re-wrote your statement as: Kant pointed out to Hume, the question of causation has to do with the synthetic a priori logic of our pure understanding, and in no way our common experience, which had thus far been described by mere habit.Mww

My understanding is that the synthetic a priori conditions of and for possible experience requires prior experience in order to establish just what are the necessary general conditions of and for experience. Once it is established (by experience and the phenomenological examination of experience) just what is essential to all coherent experience (that it be spatial, temporal, causal etc) is a priori in the sense that further actual experiences do not need to be examined in order to know that these conditions will necessarily apply.
• 850
Once it is established (by experience and the phenomenological examination of experience) just what is essential to all coherent experience (that it be spatial, temporal, causal etc) is a priori in the sense that further actual experiences do not need to be examined in order to know that these conditions will necessarily apply.

Problem is......no experience or possible experience, from which that kind of knowledge arises or may arise, will ever connect the concept of cause with the concept of effect essentially, or which is the same thing...necessarily. Put another way: immediate knowledge of cause is not given by knowledge of effect.

Once logically established, yes, but such establishment is the purview of reason, not common experience.
• 5.8k
Once logically established, yes, but such establishment is the purview of reason, not common experience.Mww

And yet a human without language - and thus without reason - can and does learn that touching fire causes pain. The attribution and/or recognition of causality does not require reason. It only requires common experience. It only requires drawing a correlation between the behaviour and the subsequent onset of pain at the location of the body that actually touched the fire.

That is to think about the fire and to believe that touching it causes pain.

It only requires pre-linguistic thought/belief.

No language required.

Hume's edifice is inherently inadequate for taking proper account of the attribution/recognition of causal relationships/causality.

Causality is prior to our account of it. The subsequent attribution and/or misattribution of causality is as well. The former can be wrong. That is, we can misattribute causality. Causality cannot be wrong. Rather it is simply a part of the ever-changing universe, and thus not only an integral part of what actually happened but necessary for anything at all to happen.

Causality existed long before we become aware of it.
• 5.8k
My understanding is that the synthetic a priori conditions of and for possible experience requires prior experience in order to establish just what are the necessary general conditions of and for experience. Once it is established (by experience and the phenomenological examination of experience) just what is essential to all coherent experience (that it be spatial, temporal, causal etc) is a priori in the sense that further actual experiences do not need to be examined in order to know that these conditions will necessarily apply.

All reason is existentially dependent upon rudimentary thought/belief. All experience is as well. There is no reason completely devoid of experience and/or emotion.

Any such notion is akin to a brain in vat; such notions are based upon logical possibility(coherent use of definition) alone. There are no brains in vats prior to language use for "brains in vats" is a name for a particular kind of thought experiment. That experiment is itself existentially dependent upon thinking about thought/belief. As such, it cannot glean knowledge of that which existed in it's entirety prior to language.

Causality. Meaning. Truth. Thought. Belief.

All of those things existed in their entirety prior to account of them. Any account to the contrary is wrong.

The problem with some logical possibility is that they stand in direct contradiction to the way things actually are. All talk about apriori is bunk. All talk about reason without emotion is bunk.

A priori reason is equivalent to the most delicate of apple pies without apples. It is a brain in a vat. A brain cannot do anything by itself. Sensory stimulation/interaction is necessary in the sense of existential dependency. A brain in a vat cannot think for the same reason that an apple is not enough for an apple pie. A priori reason is itself existentially dependent upon experience. There is no reason absent and/or devoid of experience. Reasoning in the manner set out by Hume and Kant is a kind of experience that requires thinking about thought/belief.

I'm not approaching you in a combative stance, by the way... My approach sucks. The above reflects some of our past discourse. I think we agree much more than we disagree.
• 5.8k
Establishing causality in great detail is reporting upon that which already existed in it's entirety prior to our account of it. Doing so requires language use. Establishing Hume's position does as well.

Causality does not.

We can learn about that which existed in it's entirety prior to our account of it. Language can be used to impede our efforts or help facilitate our success in doing so.
• 7.9k
I was thinking more along the lines that experience, to be coherently understood at all, presupposes causality. And we know that from experience, because examination of our judgements about events shows us that they can only be understood in terms of causation.
• 7.9k
I'm not approaching you in a combative stance, by the way...

No problems; I didn't read this response as combative. My definition of "a priori" is simply that which we know that experience must look like in order to be experience at all. Of course this knowledge requires prior experience and phenomenological examination of that experience in order to discover what the general common attributes of all experiences are.

I think we agree much more than we disagree.

And here's an example: I agree!
• 792
Hume said that it was possible for events to not have causes.

"We can never demonstrate the necessity of a cause to every new existence, without shewing at the same time the impossibility there is, that anything can ever begin to exist without some productive principle" (Treatise of Human Nature, Book1, Part 3, Section 3)

As you can see, Hume didn't say anything of the sort. His thesis is that we can never demonstrate the existence of a cause. Again, Hume wrote about human understanding - how we come to know, whether we can know - not about the nature of things. He didn't actually have much to say about metaphysics and ontology, he was mainly concerned with epistemology.
• 1.3k
As you can see, Hume didn't say anything of the sort. His thesis is that we can never demonstrate the existence of a cause. Again, Hume wrote about human understanding - how we come to know, whether we can know - not about the nature of things. He didn't actually have much to say about metaphysics and ontology, he was mainly concerned with epistemology.

In a way, his epistemology was his metaphysics - what is known is identical to what is.
• 792
In a way, his epistemology was his metaphysics - what is known is identical to what is.

I am by no means an expert on Hume, but I don't think this is true. Where does he say this?
• 1.3k

He doesn't say anything like that. This is simply one way I interpret his philosophy as a whole, especially as taken in its historical context.
bold
italic
underline
strike
code
quote
ulist
image
url
mention
reveal