I have read a lot about theories of causation lately. None of them define causation satisfactorily. Either you end up with a definition that comes very close to common sense, but at the same time is prone to rather technical counter-arguments (e.g. common causes, preemption, etc.) or you end up with a definition that makes the whole configuration of the universe a cause for a certain effect and therefore is far from our common understanding of causation.

Now, my question, as stated in the title, is: Isn't the reason for this inability to define causation sufficiently, that we use that notion in a diffuse way in everyday life?

  • 2
    The reason is that one cannot serve two masters. The everyday use of causation singles out "the cause" from the background based on what the users are interested in and can affect. That cannot be a basis for a detached objective concept that science can study. So if the "perfect definition" is supposed to do both there is no point looking for it.
    – Conifold
    Aug 6, 2021 at 0:05
  • @Conifold, thanks a lot, this pretty much answered it for me. I guess I was just wondering why pundits seem to put that much effort into it if they must all know it's completely futile. A weird objection in the realm of philosophy, I am almost ashamed of it. Aug 6, 2021 at 0:25
  • 1
    What's wrong with directed acyclic graphs within the background of spacetime? And just realizing everyday life simplifies everything? cran.r-project.org/web/packages/ggdag/vignettes/…
    – J Kusin
    Aug 6, 2021 at 1:16
  • Yeah, basically, we have some way of deriving later states of the universe from earlier states - fundamentally the laws of physics - and we say a piece of an earlier state caused a piece of a later state, if we can derive the piece of the later state from knowledge of the piece of the earlier state. Everything else is quibbling over technicalities, within the same basic concept.
    – causative
    Aug 6, 2021 at 1:31
  • 1
    I think it's more likely that the idea of causation is an intuition that lets us deal with the normal world as we experience it, but those definitions are trying to turn it into metaphysics, and there is no coherent definition of it that has the universality and precision they are trying to get. Aug 6, 2021 at 2:37

2 Answers 2


Isn't the reason for this inability to define causation sufficiently, that we use that notion in a diffuse way in everyday life?

Isn't the reality of everyday life that all the notions that we use we use them in a "diffuse way"? This doesn't seem to have prevented progress to be made in mathematics, science, engineering. What you say implies that there is a lot of theories of causation, which means that many bright minds have tried.

For one thing, maybe we should pay attention to the everyday notion of cause. This is clearly not what many people do, as demonstrated by one answer here. I quote:

So if we take as a premise some situation A, and we can use our rules to derive as a consequence a second situation B, with the consequence B happening later than all of our premises, then we say that A causes B.

I would describe this view of causation as typically scientific. Yet, it is apparent that this is not the ordinary notion that issues from everyday life. So, formal theories of causation fail to provide any perfect definition of causation probably because they don't even try to model the everyday notion, not because the everyday notion is "diffuse".

The reasons that they don't try may be trivial. The history of humanity is replete with false theories because it will always be much easier to produce at least some of the many false theories which are conceivable than to produce the only one which is true. Isn't that enough of a good reason to fail?

Another good reason to fail seems to be that many if not most bright minds tend to pooh-pooh everything which is not formally articulated, and the everyday notion of causation is certainly not formally articulated. And if you pooh-pooh something, you are not going to be motivated enough to want to dignify it with a proper formal model of it.

Yes, that's about it.

  • "Yet, it is apparent that this is not the ordinary notion that issues from everyday life." How do you justify this claim? Common sense reasoning can be understood as cognitive rules that allow us to infer what will happen from an initial situation. Thus the definition you quoted coincides well with the ordinary notion.
    – causative
    Aug 7, 2021 at 1:04
  • Yeah i agree. The definition is way looser than implied. And close to folk causality. (The only truly scientific notion of causation is that this state of the universe, plus the laws of physics and evolution of time and resolution of quantum uncertainty, created this later state of the universe. There is never an A and B alone.)
    – Al Brown
    Aug 7, 2021 at 1:42
  • @Causative 1. "How do you justify this claim?" As Conifold observes in his comment on the question, "The everyday use of causation singles out "the cause" from the background". 2. "Common sense reasoning can be understood as cognitive rules" And the everyday notion of cause is not anything formalised, so reasoning is irrelevant here. Further, common sense and reasoning are antinomic notions, so your "common sense reasoning" is a very ambiguous expression. Is that reasoning based on common sense premises, or reasoning which is common or commonly acceptable? Aug 7, 2021 at 10:27
  • @Speakpigeon 1. But by singling out one cause for the effect, we are invoking a rule, that says the effect "normally" does follow the singled-out cause. 2. "Reasoning" is a broad term not meant to imply just formal deduction, but includes informal or inductive inference as well. en.wikipedia.org/wiki/Commonsense_reasoning "Common sense reasoning" means reasoning about topics that most people would call "common sense." ("most animals move forwards more rapidly than they move backwards" would be an example of a common-sense proposition.) It is an informal process.
    – causative
    Aug 7, 2021 at 10:44
  • @causative 1. "we are invoking a rule" We are not. It may be in scientific notion of causation, but not in everyday life where it refers mostly to one-off contingent connections between events. - 2. The Wikipedia article doesn't say that we do common sense reasoning. It say it is a notion in AI. Reasoning and sense are antinomic. We may reason from common sense premises. The process followed is formal or it is not reasoning. There are no rules, certainly not any that we know of. Our inferences are based on our logical sense. Aug 7, 2021 at 17:22

The basic concept of causation is: we have some rules that allow us to take some initial events as a premise and infer what events will happen temporally after the initial events.

We assume that the premises and conclusions here are tagged with the time that the events happen, so that we can order them in time, with the conclusion occurring after all the premises.

  • The true, ultimate laws of physics are an example of such rules, though don't know exactly what they are.
  • More imperfect, limited laws of physics that only apply to certain situations are also examples of such rules. For instance, the Lagrangian of a Newtonian system allows us to derive future states of the system when given initial conditions.
  • Common sense reasoning is another example of such rules. A child knows that if you strike a match, it lights - the child has a cognitive system in place that takes as a premise the event, "the match is struck," and deduces as a consequence the later event, "the match lights."
  • For a simplified sandbox, Conway's game of life has three simple rules for deriving the next state of the grid from the current state of the grid. Causation can thus be tracked in Conway's game of life. You can run the simulation one way, then rewind it, flip one cell from white to black, and run it again to see what happens afterwards as a result of the flip.

So if we take as a premise some situation A, and we can use our rules to derive as a consequence a second situation B, with the consequence B happening later than all of our premises, then we say that A causes B.

That's the basic idea of causation. We can quibble over definitions - when we will technically say "A causes B" - but this isn't a problem in the notion of causation itself. It is just a problem of standardizing the terminology around the same basic idea. Of course, different people have different ideas of how the terminology should be standardized. This is like all the different and conflicting electric plug and voltage standards in different countries - a problem of coordination and agreement, not a problem with the fundamental idea of alternating current.

  • It is not technically right to speak of "the" cause of an event. There are any number of possible causes for a given event. Ultimately every event B is caused by the Big Bang, and there are many other sets of events A that we can say also are sufficient to cause B.
  • Our rules may not allow us to say that B happens with certainty; they may be probabilistic. For this we may say things like, "A probably causes B," if from A we deduce an increase in the probability of B. To be more confident of the causation, we may appeal to a better set of rules or a more complete set of initial conditions.
  • You mentioned preemption. If A "would have caused" B, if not for the interference of C, then really if we said, "A causes B," we were speaking loosely; A by itself was not sufficient to infer B. In truth it is, "A and the non-interference of C (and the non-interference of D, and E, etc) cause B." This is an appeal to a more complete set of initial conditions. It's not exactly wrong to say "A causes B" if that's what usually happens - it's just not exactly right either. We might call it a useful shorthand.
  • In conways game the resulting configuration also has innumerable other causes. (The player was born, Conway invented that. Player chose this square. The electrical power didnt fail. The big bang. Ad infinitum.) So whats the cause of the final configuration. I commented on the other answer and q too. Enjoy this
    – Al Brown
    Aug 7, 2021 at 1:45
  • @AlBrown Eh, that may apply to other things, but those are not causes within Conway's game of life itself. Those are causes within our universe, external to Conway's game of life. Conway's game of life doesn't actually have players. It is a self-contained universe.
    – causative
    Aug 7, 2021 at 1:50
  • Not causes within the game, just of it. And of it’s initial conditions. But no, I do understand thanks. Zero-player game i heard it called awhile ago. Many would say our universe is too. Self contained and zero player. If they get complex enough, the gof entities will say “Well no, agents do exist and claims here in our universe can be made. See we could set-up a correspondence between...”
    – Al Brown
    Aug 7, 2021 at 2:44
  • Also, is it self-contained if it is created and every game is set-up from the outside?
    – Al Brown
    Aug 7, 2021 at 8:21
  • @AlBrown as a formal system, yes, in the same way that mathematics is self-contained. The game of life in the abstract is a mathematical structure that is not the same as any particular instance of the game of life that a human sets up.
    – causative
    Aug 7, 2021 at 8:49

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .