When I see, for instance, a Billiard-ball moving in a straight line towards another; even suppose motion in the second ball should by accident be suggested to me, as the result of their contact or impulse; may I not conceive, that a hundred different events might as well follow from that cause? May not both these balls remain at absolute rest? May not the first ball return in a straight line, or leap off from the second in any line or direction? All these suppositions are consistent and conceivable. Why then should we give preference to one, which is no more consistent or conceivable than the rest? All our reasonings [sic] a priori will never be able to shew [sic] us any foundation for this preference. - David Hume, 'An Enquiry Concerning Human Understanding 1748
A philosophical skeptic, as evident in the quotation above, Hume (disingenuously I’d wager) didn’t believe in cause and effect. Generally when we are trying to explain a phenomenon, we search for its cause. If we find the cause, we have then explained why the phenomenon has come to be, or why it is the case. This seems simple enough. Science, at the very least, is based on a belief in cause and effect, or so common sense would have us believe. Hume argued that we are actually mistaken in our belief about cause and effect. If we observe that event A is followed by event B, we may feel that there must be some property of A which causes B, but all we actually see is the contiguity and succession of the events.
We have no other notion of cause and effect, but that of certain objects, which have been always conjoin'd [sic] together, and which in all past instances have been found inseparable. We cannot penetrate into the reason of the conjunction. We only observe the thing itself, and always find that from the constant conjunction the objects acquire an union in the imagination. - David Hume, A Treatise of Human Nature 1740.
We cannot actually say that one event caused another. All we know for sure is that one event is correlated to another. For this Hume coined the term 'constant conjunction'. That is, when we see that one event alway 'causes' another, what we are really seeing is that one event has always been 'constantly conjoined' to the other. As a consequence, we have no reason to believe that one caused the other, or that they will continue to be 'constantly conjoined' in the future. The reason we do believe in cause and effect is not because cause and effect are the actual way of nature; we believe because of the psychological habits of human nature.
(On first hearing this, I asked my lecturer what he thought Hume would do if I pointed a loaded gun at Hume’s head. He didn’t answer.)
This is quite a profound philosophical problem. Personally, I can't stand it. It's amazingly unsatisfying. Literally giving up. If someone asks you why something happens, all you can do is shrug your shoulders and say: "I dunno?" I call this the "sh*t happens" account (pardon my descent into scatology). There is no way to explain why something occurred, or to predict when it will occur again. "Sh*t just happens man!"
One of the best solutions to Hume’s radical skepticism comes from Wesley Salmon. The following (admittedly simplistic summary) is from the wikipedia entry on Salmon to which I contributed. (Can one plagiarise oneself?)
A major feature of Salmon’s work consists of providing a philosophically sound basis for scientific explanation. Salmon claimed that a scientific explanation is the state of affairs of something fitting into or being a part of a pattern in the world, where the pattern is constituted by at least one causal process. A process is the real physical connection between cause and effect. For example, the heat from the flame of a gas stove excites the molecules in the water via the iron atoms in the bottom of the pan. This is a process where every step from the cause, the flame, can be traced to the effect, boiling water. This view has come to be known as the ‘ontic’ account of explanation. It is thoroughly realist and rests on an ontology that is designed to answer the question: "Just what is a causal process?"
Salmon argued that events are intersections of two or more causal processes. This interpretation leads to depicting the world as a network with lines (causal processes) and nodes (intersections of causal processes - events). Thus events, causes and effects are reduced to causal processes, where causal processes are real connections between events. According to Salmon, causal processes transmit 'structure', or energy and momentum or 'information' from one spatio-temporal location to another. There are two ways, in principle, by which it is possible to demarcate causal processes from 'pseudo-processes' – how causal processes are transmitted through space-time and what they transmit. As for how causal processes are transmitted, the theory holds that the transmissions must be continuous, with no discontinuities or 'jumps' in space-time. Unfortunately this is not enough to be sure that it is a genuinely causal process, as 'pseudo-processes' can also be continuous.
Demarcation requires that it is necessary to examine what causal processes transmit. Salmon argued causal processes transmit information, and as such we should be able to 'mark' a process by modifying it, to see if the modification is transmitted. This 'marking principal' serves to demarcate causal process from pseudo-processes, as the latter cannot be marked. Further to this, 'things' are 'causal agents' if they are originators of causal chains and not merely 'passers on' of causes. They are 'uncaused causers'. The 'uncaused causer' is the first cause in a new causal chain. It is the initial cause in the chain of causal events we are interested in explaining. The uncaused causer obviously had a cause of its own, but in terms of the phenomenon that we are attempting to explain, the uncaused causer's cause is irrelevant
Salmon's ontology of causal processes seems to fit scientific explanations very well. For example, genes are very much 'causal agents'. Geneticists have inserted the 'antifreeze' genes from flounders into the genetic code of tomatoes. This then protects the tomatoes from frost damage. A process (normal reproduction) has been marked, such that the information that is transmitted (the DNA) is modified (the addition of the flounder gene). Genes, therefore, are a part of a genuine causal process. Specifically, genes are made up of DNA. Inside the cell nucleus, a particular nitrogen-base sequence of DNA controls precisely what proteins are formed in the cytoplasm. By controlling the synthesis of proteins, DNA determines what chemical reactions take place in the cell. The chemical reactions of cells affect the chemical reactions of the body. A small chemical change to the way a particular molecule forms can produce a considerable effect on the phenotype.
Here we have a detailed account of the way genes cause phenotypic effects. DNA is information that tells the cell what kind of proteins to. This, in turn, governs the chemical reactions in the cell, which then produces a phenotypic effect. The gene is the 'uncaused causer' in the chain of causal events we are trying to explain.
Cop that Hume!
________
Tags:
Posts
3 comments:
I think what you say about Hume is very wrong: that he "didn’t believe in cause and effect", that he "argued that we are actually mistaken in our belief about cause and effect". He just thought that causal reasoning was done on the basis of psychological associations, rather than a priori reasoning or intellectual reasoning from experience. The view you're attributing to Hume is, I think, nowhere to be found in the Treatise or 1st Enquiry.
Off the top of my head (so errors may show up): First, Hume argued that we don't (and can't) learn about what causes what through abstract reasonings a priori (for Hume, "intuition" and "demonstration"). This is because any distinct things can easily be conceived apart from each other, and demonstrations/intuitions always close off alternate conceivability. So, he argued, we learn about what causes what through reasoning based on experience. What kind of reasoning? Well, intellectual reasoning would rely on the principle that the future resembles the past. But this principle cannot be demonstrated/intuited (we can easily conceive the contrary) and it obviously cannot be arrived at by causal reasoning (since that's what we're trying to explain). So instead our causal reasoning is based on experience in another, non-intellectual way: experienced conjunctions of things gradually train the mind to expect one when it sees the other. And the causal necessity we attribute to a cause-effect relation is a psychological projection of an internal feeling, akin to colors and sounds.
It's a psychological story about causal reasoning. That's why he talks about children and animals.
Maybe you'll believe wikipedia Cole?
Hume and Causation.
Thanks AF, but I actually contributed to that particular bit of the wikipedia entry, (based on my honours thesis - hence the similarity) so it's kind of circular to use that as a reference.
Cole, I thought the two citations I give from Enquiry and Treatise do exactly that? (I won't bother repeating them in their entirety.)
What else do I make of this sentence other than his disbelief of causation? "We only observe the thing itself, and always find that from the constant conjunction the objects acquire an union in the imagination."
A union in the imagination?
To be clear, by causation I go with Salmon's definition, a real physical connection between cause and effect. Hume explicitly rejects this.
You state: "And the causal necessity we attribute to a cause-effect relation is a psychological projection of an internal feeling, akin to colors and sounds."
To me a "psychological projection of an internal feeling" is an explicit rejection of a real physical connection.
A lot of this stems from the problem of induction, which itself stems from philosophers wanting/needing logical proofs as the basis for their epistemology, as opposed to a sound rational foundation. As you'll see in the wikipedia link, that's Hume's fault too. Though to be fair, he did make the distinction between logical and rational (in other terms however).
Here's a bit of Bertrand Russell's section on Hume in History of Western Philosophy (highly recommended to all):
Let us now ask ourselves what we are to think of Hume's doctrine. It has two parts, one objective, the other subjective. The objective part says: When we judge that A causes B, what has in fact happened, so far as A and B are concerned, is that they have been constantly conjoined, i.e. A has been immediately, or very quickly, followed by B; we have no right to say that A must be followed by B, or will be followed by B on future occasions. Nor have we any ground for supposing that, however often A is followed by B, any relation beyond the sequence is involved. In fact, causation is definable in terms of sequence, and is not an independent notion.
The subjective part of the doctrine says: The frequently observed conjunction of A and B causes the impression of A to cause the idea of B. But if we are to define 'cause' as is suggested in the objective part of the doctrine, we must reword the above. Substituting the definition of 'cause', the above becomes:
'It has been frequently observed that the frequently observed conjunction of two objects A and B has been frequently followed by occasions on which the impression of A was followed by the idea of B.'
This statement, we may admit, is true, but it is hardly the scope that Hume attributes to the subjective part of his doctrine. He contends, over and over again, that the frequent conjunction of A and B gives no reason for expecting them to be conjoined in the future, but is merely a cause of this expectation.
Russell further clarifies the objective part of Hume's doctrine:
This doctrine has two parts: (1) When we say 'A causes B', all that we have a right to say is that, in past experience, A and B have frequently appeared together or in rapid succession, and no instance has been observed of A not followed by or accompanied by B. (2) However many instances we may have observed of the conjunction of A and B, that gives no reason for expecting them to conjoined on a future occasion, though it is a cause of this expectation, i.e., it has been frequently observed to be conjoined with such an expectation. These two parts of the doctrine may be stated as follows: (1) in causation there is no indefinable relation except conjunction or succession; (2) induction by simple enumeration is not a valid argument.
I think I'll add all the above Bertrand Russell to the wikipedia entry.
As Russell points out, essentially Hume is making the correct point, that causation is not a logical necessity. Of course, this leads to a slight problem. Anytime someone says A caused B, by Hume's logic they are making the Correlation-Causation error. This all comes back to the difference between deductive and inductive reasoning. Deductive reasoning is always true (assuming one does it correctly), because the premises are accepted a priori. Inductive reasoning is not possible to "prove" in the same way. This in itself comes back to the difference between logical reasoning and rational reasoning. I wrote about this in Philosophy Now a few years back, so I'll post that essay in the near future.
Post a Comment