Causation – a Philosophical Question

The cat died

The cat died

What do we mean when we say something like ‘the death of my cat caused me great sorrow’? How should we understand such a statement? And what sort of truth-maker makes such a statement true?

The first question is a question about semantics, about the meaning of a causal statement. The second is a question about metaphysics, about what, in the world, makes the causal statement true.

Such questions are related of course. We would usually think that someone who understands such a statement would know how to determine its truth or falsity, and would know, therefore, what sort of situation in the world makes it true or false. Meaning and true go together.

But such questions are much easier to ask than to answer. When posed as philosophical questions they are very difficult to answer. When a philosopher asks such questions they are not, for a start, usually posed as questions about a particular causal statements, but about all causal statements. We are asking in effect what ‘A caused B’ means, and we are asking what sort of situation makes ‘A caused B’ true. We want to know about the semantics and metaphysics of causal statements in general.

We might start by saying that ‘A caused B’ seems to be about a relation, that of causation, between two relata A, and B. But if so, what is this relation, what sort of properties does it have, what sort of things does it relate?



All relations have properties. Some relations are, for example, transitive. The relation ‘taller than’ is transitive. If A is taller than B, and B is taller than C, then A is taller than C. Knowing that a relation is transitive tells us something about the logic of statements referring to that relation, something about the entailments to which one making the statement is committed.

So if we are assuming that causation is a relation, we can ask ‘is causation transitive?’ If A causes B, and B causes C, then does A cause C?

The difficulty is that some causal statements appear to be transitive and some don’t.

Horseplay by the river

Horseplay by the river

If Sam’s pushing Fred caused Fred to fall into the river, and Fred’s falling into the river caused Fred to drown, we would usually think that Sam’s push caused Fred’s drowning. Here transitivity seems to reign.

But here is a set of causal statements for which transitivity does not seem correct:

If Churchill had been German he would have been a Nazi

The Nazi Swastika

The Nazi Swastika

If Churchill had been a Nazi he would have been a traitor

If Churchill had been born German he would have been a traitor.

So is causation transitive? Or isn’t it?


Causation: A Very Short Introduction:

An interview with the authors of ‘A Very short Introduction to Causation’:


About Marianne

Marianne is Director of Studies in Philosophy at Oxford University's Department for Continuing Education
This entry was posted in Blogs, Monthly Conundrums and tagged , , , , , . Bookmark the permalink.

29 Responses to Causation – a Philosophical Question

  1. John-Brian Vyncent says:

    Ms. Talbot,

    Fascinating topic as usual! I aspire to one day write as eloquently and invitingly provocative as you do.

    I am lost over one aspect of the “Churchill transitivity example”: How am I to interpret Churchill’s traitorous ways? If I understand your premise correctly, if Churchill were a Nazi (holding true to Nazi ideology) then he would not be a traitor to the United Kingdom as he was born German. If the proposed “Nazi Churchill” were still as principled an individual as we remember him to be, then we delve into the philosophical arena on the prospect of “a good Nazi.” Could it be that someone could be a Nazi, not agree with the aspects of concentration camps, mass murder, suppression of minority rights, etc and still not be a traitor (such as the participants of Operation Valkyrie)?

    If he were a UK national born on German soil then I could see the traitor argument but then that is a matter of perspective. But “perspective” does not match the drowning analogy as drowning is an outcome that is not dependent on someone’s perspective.

    What am I missing?

    • Marianne says:

      Thank you for your kind comments, John-Brian. I think you might be reading my example incorrectly. I use it only to show that from the first two premises, the third doesn’t follow. It is a counterexample to the idea that causation is transitive. But I do agree there are different ways of reading the counterfactuals involved (perhaps that is the problem!?)

      • Ghazi says:

        Wittgenstein’s thought is very suitable here, “relation between thought and reality is to be found in the grammar of language”, “perception follows a vital role” but here in this case churchill’s example was just to show independence of the statements and to show how transitivity is invalid right?

      • Marianne says:

        Wittgenstein’s thought is often suitable!Yes, I wanted to give an argument for, and an argument against the claim that the causal relation is transitive.

  2. Ghazi says:

    This problem of causation seems equally as perplexing as mind and body problem. To me it definitely doesn’t seem like a transitive property, there are many cases where transitive theory fails explaining causation. I believe, Causation will only be followed when two different cases are affecting each other, directly, for example if we developed an equation that describes the heat flow and the lets say that variable in that function is temperature difference, cross section area and other factors then one can say that changing the area/temp/other factors involved in the equation will change the heat flow. But what if surrounding is playing a role in the heat content of the body (latent heat). Transitive theory fails there. Thus there are way more cases where transitive theory fails and if the question persists what sort of truth maker statement makes it true then eventually it all seems to conclude at assumptions. Recently whilst writing a paper on logic i came up with this problem, what justifies syllogism itself? what truth maker brings consistency with the applied logic? how do we know if the logic applied in the deduction is true? what is the reason behind logic? Ahh!! a little messed up.

    • Marianne says:

      Hi Ghazi, the problem of causation IS as perplexing as the mind-body problem. Yes, I agree there are many different counterexamples I might have give that suggest causation isn’t transitive. but many people think it is. The question of what justifies deduction is a very interesting one. I argue in my book that t i as difficult and as intractable as the question of what justifies induction.

      • ghaziphil says:

        I just came up with a logic to show how Causation is NOT transitive. If all A is B and if all B is C then only all A is C (though reverse might not be true) now i can take another step to show how it fails at transitive theory. Causation was transitive if the reverse of the above example was true (hope you understand me). Causation can only be justified by transitive theory if the reverse follows same relation as the forward flow does. Then only interdependence can be shown. Talking about others, those who support causation by transitivity are the ones who have not seen the vice-versa and i believe their examples are limited to a certain periphery, to prove something, a statement must be Necessary NOT contingent isn’t it?

        In which book have you stated about justification of the deduction, i would love to read 🙂
        and to me Logic is the grammar of thoughts, no thought can be true unless right grammar is used and still i am trying to track down the origin of logic, would hate to put it on the shoulders of instinct haha 🙂

        and thank you for such an excellent topic. 🙂

      • Marianne says:

        I am very sorry but I am not following your argument. The causal relation is transitive if whenever:

        A causes B
        B causes C
        A causes C

        It is not transitive if it could be that A causes B, B causes C and A does not cause C.

        I tried to give examples of each case in my blog.

        The book in which I talk about the justification of deduction (actually I say it can no more be justified than induction can) is my Critical Reasoning: A Romp Through the Foothills of Logic. It features on this website.


  3. Safwan says:

    I think a refreshing perspective about ‘causation’ is possible. If event A leads to B – then, first of all we must accept that A had a potential {P} or tendencies – to be in a relationship with B. The inner power within A, its “inner potential” to react (according to its own intellect) – this element of potentiality (or the possibilities-within) – invites for viewing Cause and Effect as a mutual agreement between possibilities – or potentials – acting in both A and B.

    The perspective of acknowleging A as containing potentials : A(p) which emerge naturally from its intellectual properties) – this perspective allows for great flexibility. The highest possibilities win but not all possibilities {pi} can materialise because of interconnectedndess of A with other conditions.

    In science, Cause and Effect expresses consistency: laws, principles, specific numbers, certain patterns. But the mental realm is not deterministic, as it allows for “potentials”. Potentials include varieties of scenarios.

  4. ghaziphil says:

    thank you Marianne, for the recommendation 🙂 i shall go through that book. I still am coming up with examples that shows how causation is Not transitive. If causation was transitive, wouldn’t we be able to eliminate some of the steps? I believe transitivity is sufficient to explain Causation but NOT Necessary. One more example i would like to take from mathematics, i wish at the moment i had some philosophical argument but why not math.
    Example: If there are exceptions in the steps then transitivity doesn’t hold valid any longer,
    A is differentiable IF and ONLY IF A is continuous,
    so only continuity can lead to differentiability
    BUT functions which have sharp turn in their graphs , even if they are continuous they can not be differentiated.

    By the above example i just wanted to say that, transitivity is sufficient for causation but its Not a Must.

    • Marianne says:

      Dear Ghazi, there is no suggestion whatsoever that transitivity is an EXPLANATION of causation. The question is does it CHARACTERISE causation? I.e. if causation is a relation, is it a transitive relation? Any examples of causal relations that are non-transitive or intransitive (including those from maths) suggest accusation isn’t characterised by transitivity.

      Many thanks for your comments.

  5. ghaziphil says:

    Talking about induction and deduction, I am more inclined towards induction because it is just impossible to know something unknown from unknown, we can only know or take an attempt to know unknown from the known variables, which somewhere defines that induction tends relatively more towards the truth compared to the deduction (deductions seems more dependent up on the premise compared to induction).

    • Marianne says:

      One doesn’t have to choose between induction and deduction. We need both. They are simply different forms of arguments, with different characteristics. Both argument types are dependent on their premises. But ‘following from’ in each case is different.

  6. anonymousG says:

    By way of introductory remarks, please pardon my lack of knowledge of the language of philosophy, or the more precise terms often employed by philosophers. When I consider the question of the existence of transitive causation, I would like to answer with a qualified yes.
    Hume’s theory of causation serves as the qualification starting point for my reply, with full acknowledgment that other philosophical approaches to the question might open up alternative points of view.

    Arguing in the affirmative begins with a brief (and admittedly rough) sketch of Hume’s claim that his theory of natural causation is well suited to building causal connections between any natural or political (perhaps a bit of an overstatement) phenomena that takes the form A–>B.

    Hume’s causal theory, consisting of three basic elements: contiguity, priority in time and necessary connection, have been, and will continue to outlined in a variety of ways. A simple three step framework is used to describe them here.
    Principle of the Uniformity of Nature: Best considered Hume’s metaphysical starting point for his theory of causality, Hume asserts that empirical experience serves as the the theory’s building block. It’s a relatively easy metaphysical principle to explain. One sees the sun historically rise in the morning and set at night, and after repeated experiences, one makes an inference that tomorrow will be pretty much like today. The sun will rise in the morning and set at night.
    Temporal Contiguity: The only thing we really know about causation is the existence of temporal contiguity of two actions, A and B. Hume’s famous billiard ball example . Ball A strikes Ball B, and Ball B moves. A then B, A then B. B does not move without A striking it.
    Necessary Connection: When making an inductive inference about the cause of billiard ball movement, the human mind differentiates between matters of fact (physical properties of the billiard balls) and relations of ideas, the mental construct that serves as the causal explanation of billiard ball movement)

    Describing Hume’s theory of causation in such a manner provides a path for building a similar theory of political causality. Rather than starting with the principle of the uniformity of nature, a theory of political causality would start with a principle of the uniformity of human nature.

    Hume understood the importance of institutions, social or cultural (described in the essay, Of National Characters) and formal legal (described in That Politics May Be Reduced to a Science) as the key to promoting regularity in human behavior. His conversations regarding customs and culture as regulating human behavior suggest that he viewed the science of politics in the same way he view the science of nature.

    So great is the force of laws, and of particular forms of government, and so little dependence have they on the humours and tempers of men, that consequences almost as general and certain may sometimes be deduced from them, as any which the mathematical sciences afford us.

    In other words, while Hume assumed a difference between a science of the natural world and a science of the human world, he further assumed a transitory theory of causation, applicable to both worlds. I’m still unsure of the degree of Hume’s further thinking on the topic comparing the differences in causation between the natural world and human world. My first thought is to assume that natural world causal links transcend time and space, while human world causal links, due to both culture and laws, tend to be more time and space dependent.

    At first glance, it appears that many examples of natural world transitive causation exist. I hit Billiard Ball A, Billiard Ball A hits, Billiard ball B, Billiard Ball B hits billiard ball C. A–>B, B–C and therefore A–>C

    The case of transitive causation appears to weaken a bit when moving from the natural to the human world. Perhaps some typology might be established as an organizing tool. Physical acts of humans might fit into a transitory causation schema. A trips and falls on B. causing B to fall on C. Therefore, from a physical POV, A–>C.

    Transitory causation tends to get a bit more muddy when moving from physical acts to moral or legal considerations. Does A tripping and Causing harm to C make A legally or morally liable for C’s injuries? Again, without access to information about the specific historical circumstances of the physical activity, along with the specific legal and moral theories that might be used to justify or refute transitory legal or moral causation in the circumstances, it’s difficult to answer the question. I assume that some legal and moral theories do posit transitory causation under specified circumstances.

    • Marianne says:

      Far too long for me to read this now, never mind comment. Will try to do so asap.

    • Marianne says:

      Dear Anonymous, don’t worry in the slightest about your lack of philosophical knowledge. I hope this website will attract complete beginners as well as people who have a bit more understanding. But I DO urge you to be more concise: I have a full time job and can only do this round the edges of that!

      Philosophers generally agree that causation is transitive. But there ARE problems with that view as I outlined in my blog.

      Your account of Hume’s account of causation is basically correct except for one thing. Hume denied there was such a thing as necessary connection. He admitted that it is part of our concept of causation, but also that it arose from habit – our experience of constant conjunction prompts us to expect B to follow from A when we have always experienced Bs following from As. It is, says Hume, this expectation that we mistake for a necessary connection.

      The application of the Principle of the Uniformity of Nature to the human world is, I think, limited. Here we need (in addition to the PUN) the Principle of Charity. The view that the appearance of irrationality (rather than irregularity) is evidence for error. In the same way (and almost certainly for the same reason, causation, as applied to the mental, seems to be quite different. Human beings act for reasons, and if reasons ARE causes (and most, but not all, philosophers think they are), they are causes of a very special kind. I do not ground my claim that Jane is crossing the street because she wants an ice cream on the basis of my experience of a constant conjunction between Jane’s wanting an ice cream and her crossing the street!

      Thank you for taking the trouble to post!


  7. anonymousG says:

    Hi Marianne,

    I appreciate your pointing out the problem with my representation of Hume’s concept of necessary connection.

    I also agree with your point on the importance of the principle of charity for interpreting or understanding human reason and its role in human behavior. In my own language, I often consider them importance of reflecting on ppl as subjects in and of themselves, rather than placing them in what often turns out to be a stereotypical behavioralist subject-object relationship.

    Thank-you and best wishes

  8. Scarlatti says:

    It’s fascinating that what you wrote caused several peculiar misreadings. If some of the misreading lead to the writing of books about problems you never intended, would you deserve credit for the books in any way? Or might you, in all fairness, deny all responsibility?

    This matter seems a question of reliability. We call an action a cause properly when it reliably leads to something. If someone wins the lottery it is not strictly correct to say buying the ticket caused them to win, it was simply one necessary circumstance among others, including the existence of the universe as a whole, that lead to that jackpot.

    PS: I don’t find the drowning causally convincing. That’s just a matter of legal blame. The man might not have drowned, or was he so feeble that it was almost inevitable? And on the other hand, Churchill, given his fanatical energy and cruelty to teetotaler women, would likely have made a good buddy to Goring.

    The main thing is, the transitive bit is misleading. Even the push does not lead to the fall (without the rest). There must be also a man who is weak enough to be pushed over, gravity, the existence of the water and the past existence of the entire world. But, then, someone must tell us, as well, how cause is intelligible at all (how cause is caused as a possible thought), since one state of the universe leading to another is nonsense, they would both be the same universe. And so the so-called cause would be the same thing as the so-called thing caused. Apparently this is necessary, since without it being the same one couldn’t speak of the change. It is all entirely unintelligible gibberish, the talk of causes.

    • Marianne says:

      Is it what I wrote that caused the misreadings, or were the misreadings imposed on what I wrote? What does ‘reliably leads to’ mean except ’causes’?

      Why do you say that one state of the universe leading to another is nonsense because they’d be the same universe. This seems ti me to be just false, but you must have a reason for saying it, so maybe I am misunderstanding you….?


  9. 8formsofkinglymadness says:

    You are too unjust and dishonest to become a philosopher, putting aside the other parts of your failings.

    • Marianne says:

      What a strange thing to post. You don’t even explain why you are saying what you are saying. Still, I have approved it because I’d hate people to think I only approve complimentary posts!

  10. Pingback: Remarks on Causation, Free Will and Grammar | Naturalistic Philosophy

  11. JC says:

    In chemistry one pours ‘x’ into ‘y’… causing ‘z’. It makes sense here to say: “pouring ‘x’ into ‘y’ causes ‘z’”. If a physics student were to hear this explanation, however, she might say: “what a strange way to look at it … the cause is fermions, bosons and their relations!”

    She might hear me say… “I choose___!” and respond: “How odd… each atom follows a specific course, and you are made up of only atoms! How can you choose anything?” This is the same mistake!

    In the first example, transitivity applies because we accept the causal relations between those events. We blame the Sam’s of the world. “Language is a form of life” means it is a way of experiencing the world. When I follow the rules of mathematics, a triangle becomes a tool and I see it as such. In a different game it might be a mountain… a sail, or nothing at all.

    If Churchill had been a German… what would it mean for Churchill to be German? ( Compare with the modal values of belief and certainty. .)

    Does this mean of the same construction, atom for atom? What of his history?

    Consider Wittgenstein’s…

    “The case would be like the following — certain kinds of plants multiply by seed, so that a seed always produces a plant of the same kind as that from which it was produced — but nothing in the seed corresponds to the plant which comes from it; so that it is impossible to infer the properties or structure of the plant from those of the seed that comes out of it — this can only be done from the history of the seed. So an organism might come into being even out of something quite amorphous, as it were causelessly. – Zettel”

    …and next…

    “One is tempted to say, “A contradiction not only doesnt work it cant work.” One wants to say, “Cant you see? I cant sit and not sit at the same time” as one says, “I cant talk and eat at the same time.” The temptation is to think that if a man is told to sit and not sit, he is asked to something which he quite obviously cant do.” – Lectures on the Foundations of Mathematics

    These elucidate the grammatical confusions which lay within the propositions of the second argument.

    Context is the theme of the former. Showing a particular aspect of commands – the way we attempt to follow them – is the aim of the latter.

    One shows the senselessness of the argument:

    1. {C, G (¬C)} ⇒ N,

    2. {C, N} ⇒ T,

    3. {C, G (¬C)} ⇒ T.

    And the other… that the correct response to the observation: in one situation causation seems transitive, and in another it doesn’t is… “so what?”

    We are tempted to think logic as we do maths…

    In some ways the metaphor “the mind is a machine” is helpful. It draws up a picture of cogs and pistons, which serves to say: “the mind is predictable; it moves in certain ways.”

    When such a description is applied to understanding however… problems arise. One is likely to imagine inputting the code “2 + 2 =” into the mental program ‘mathematics’ whereupon a computation occurs and then: “4!” whirls out. – Setting the Stage for Plato

    This is the problem!

    “But surely “2 + 2 = 4” is more than a rule… it fits the world and so must apply whatever the circumstance.”

    Let us see.

    How can the mathematical proposition

    “2 + 2 = 4″ describe reality, as it does?”

    Well, how do these cubes fit together? In many senses they do not; in others they do. One might call the latter the application of imagination. (The bottom cube is not larger, but closer to me.) This is how ‘2 + 2= 4’ fits the picture of reality. There are many different ways it can (perspective)… but any with use must necessarily describe a part of the world; as it is only in the world a rule has a point.

    It is here the meaning of the proposition comes together with reality (context). – Clearing Plato’s Confusion

    In this language game we speak of causation in this way… in that one that way.

    The correct question: “how does this way cause us to act… and that way?”

    Not only which move is permissible using this piece on this square.

    But which is not.

    (originally posted here:

    • Marianne says:

      Goodness, I am sorry JC, but I have so little time, this is a very long answer, and your other answer refers me to your page. I am happy to put your answer up, and to refer others to your page, but will have to wait until i have time to read it myself. But thank you for replying.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s