critisisms of Hume's causal analysis 3
Necessity:-
Causes are supposed to necessitate their effects (other things being equal). Given only that cause, the effect is supposed to be required to come about, it makes the effect inevitable. Hume relies upon this supposed fact and reverses it to make his claim that we expect an effect upon it being constantly found following its cause seem plausible. But given that a particular cause always brings about a particular effect (other things being equal) it does not follow that in our experience that cause is even normally found with that effect. This is because, in our experience, other things might generally not be equal; It might generally be the case that there are other factors involved in situations where that cause is involved, so producing some different effect, or range of effects. Nevertheless it can seem an advantage of Hume's position that he, at least, can give an account of the necessity we suppose exists between cause and effect, in spite of the fact that, logically, these 'objects' are all distinct from one another, can be separated in the imagination, re-combined in the imagination at whim and so there is no necessity apparent between them. But this always makes me feel a bit uneasy because, since according to Hume, there can be no real accounting for why any fact appears or does not appear in any situation why shouldn't our opinion that there is a necessary connection between cause and effect also appear in causal situations, or when we consider them, for no reason whatever? And if it is supposed it does appear for some reason that is in some way adequate doesn't this disprove that there can be no real accounting for why some fact or aspect appears in a situation? If on the other hand the reason is not in some way really adequate we must at most, in some way, be fooling ourselves into thinking it is a real reason? But in that case wouldn't it be better, more philosophical, to un-fool ourselves and just ignore it? Especially as there seems evidence to suggest that Hume's supposed accounting can't be right because causes are not constantly found with their effects in our normal experience.
However this is there have been other attempts to account for the supposed necessity between cause and effect, or at least attempts to suppose that there might be such a thing in spite of their being no logical necessity apparent between them. William Kneil maintained that causes might be principles of necessitation, in that they might necessitate their effects although we don't know that they do so. Just as it may be a necessary fact that all primes grater than the number two are the sum of two lesser primes. This may be a mathematical theorem and if it were it would be provable and necessarily true, but no-one has proved it so far, and perhaps no-one will ever prove it, although it may be provable. So just because we don't know that something is necessary this does not show that it is not, and cannot be necessary. But this seems awkward as a reply to Hume, not least because if we think causes do necessitate their effects (given that we are allowed to try and account for this feeling or opinion) it seems daft to suppose we think this for some reason we don't know, or have been unable so far to grasp. And it seems inconsistent for the empiricist Hume to maintain to know that causes CANNOT necessitate their effects, rather than for him to just be claiming that we can't make sense of how there could be any such necessity from experience. If he did make the first claim, it would (apparently, unless you take it to be a claim about what it is possible for us to suppose using those words) be a claim that presumed to go beyond experience, and so would be un-empirical.
Differently there have been attempts to make sense of the necessity the conditional 'if-then' statements causal statements support. Some of this has been along broadly Humean lines in terms of a particular case being derived from, or looked upon as supported by a more general background theory. Or else in terms of the function or use that we reserve for statements about cause and effect. Or analysis has been proposed in terms of possible world semantics, or again in terms of rigid designators. But however this is I think it is quite possible to make sense of a real objective necessity between cause and effect, in spite of the fact that we can't know that any such thing is true; and it is possible to make sense of this in terms of what we normally suppose. The question can be answered if we first ask why we think, or why Hume, for example, might think that logical truth is necessary, and by then showing that although something similar to the conditions establishing that logical necessity can appear in causal situations, the conditions enabling us to know this necessity couldn't occur in these situations.
What is the general, if you like, old fashioned, way of accounting for logical necessity? Logical truths are supposed to be known by being analytic. What this means, basically, is that what can be justifiably concluded from some premises cannot go beyond what is already found in those premises. This is Hume's main reason and proof for supposing that in the causal relation we cannot be involved in any sort of logically adequate reasoning; "There is no object considered in itself that can afford us a reason for drawing a conclusion beyond it." and "We certainly look beyond the idea of these objects when we draw a conclusion from cause to effect"[Or at least he says that sort of thing.] But it is not just that the conclusion cannot go beyond what is found in the premises if they are to adequately support it, that produces the necessity, but also that we know the nature of the premises and can in that way grasp what can and cannot follow from them through our grasp of what is and isn't found in them. The form is; If you have got 'this' you necessarily have got 'this', or 'whatever is found in, is contained in 'this' '. But you yourself have got 'this'. So you can also know that you have got 'this'. This knowledge is also sufficient to support knowledge of what is and isn't found in 'this', and so what necessarily does and doesn't follow given 'this'. Now if, contrary to Hume, we only suppose we satisfactorily understand causal situations where there is no apparent essential difference through the situation, those states, or 'premises' apparent in the situation must appear sufficient to produce the result or effect of the situation. If there is no essential difference apparent as the situation progresses, this shows that we don't have to draw a conclusion beyond the factors found in the situation as the situation progresses, so they themselves must appear sufficient to produce the progression apparent in the situation. But, if they are sufficient and there is nothing else in the situation to stop them, (presumably) they must produce that result. But being apparently sufficient in this way, is not enough to allow us to know that they are sufficient, or that there is not in fact some other states involved in the situation. Other states being involved in the situation seems quite compatible with the appearance to us that these are the only states involved in the situation and that they are sufficient to produce what occurs in the situation. And there is no way that these apparent states, which we are supposing are not our own idea, can rule out such possibilities for us. Not least because they are only supposed to be sufficient to produce the effect resulting from the situation, not also some state of knowledge in us. So the fact that what happens in the situation may appear objectively necessary is quite compatible with the relationships in the situation appearing logically contingent. They could only not appear logically contingent if we could be in a position to know that what appears sufficient to produce the result of the situation, and what must necessarily produce that result if what appears to be the case is the case, is in fact the case. But his is something that we can never be in a position to know.
Add a comment
critisisms of Hume's causal analysis 2
Supposing that Hume's theory is true it cannot be used as an explanation showing us how it could be true, because according to the theory nothing found in the situation and considered in itself can be used as an explanation. all that can happen is that we find another constant conjunction between such regularity in a situation and our feeling of satisfaction that we understand that situation and that it must be so--contrary to the fact that Hume had no way of counting up such a conjunction so couldn't have proposed his theory for that reason. But this is absurd. if we can compare some property, even consisting in conjunctions, to a subject and realise that that could produce the subject, if these conjunctions were to continue through it, for example, this is a different basis for our understanding than it being found constantly conjoined in that subject. If it is possible to compare the properties found, or supposed in a situation, even if these are only known as consisting in occurrences (types of conjunction) noticeable in the situation, with what is produced by that situation, as it clearly is and must be on a constant conjunction view, then this provides a different basis for understanding the situation than constant conjunction; it is on the basis of comparing types of occurrences in the situation with what is produced by the situation. But this means, if we take our everyday objects and notice what properties they exhibit, and compare this to what happens in a situation we can understand the situation in terms of these everyday objects, or fail to understand it and feel we must investigate it further on a similar basis. But this is using objects themselves as they are found in everyday experience to understand what is experienced to occur, in everyday experience i.e. just what Hume is supposed to have shown we can't do.
In that everyday sense we can investigate the situation and try to understand it by and with what is found in it. The theory whereby these everyday objects are analysed into their philosophical components and it is then supposed they must be re-constructed from these components, is not the result of an empirical investigation but a philosophical one, and is not an empirical theory but a philosophical one. But it is the everyday sense of empirical investigation where the most convincing evidence is found for the conclusions and where the success of empiricism is demonstrated.
There are many other ways of trying to understand a situation than to look for regularities e.g. religions, superstitions, analogies and metaphors, cultural ways of looking at things, and traditions; and these are for many people the ways they most loudly acclaim and are convinced by. And since there are many things that can be done with experiences, which can be called 'reasoning with them', or at least ways they can be used which generate conclusions e.g. south American Mayer Indians who kept children in cages and pulled out their finger nails so that they would cry and by sympathetic magic encourage the rains to come. So a regularity theory is hardly an anthropologically adequate description of how people do, or can, as a matter of fact, reason regarding matters of fact. It is a theory of how we aught to reason if our reasoning or cogitations about relations of matters of fact are to be based on experience; and 'experience' is here meant in the philosophical sense where everything must be settled by experience; because otherwise we are admitting (apparently) that some things could be legitimately settled other than by experience. (This is why logic was supposed to account for the rest of our possible knowledge claims, and is supposed to assert nothing but tautologies, i.e. relations between propositions that just re-state what is stated in the premises of the argument, and so don't produce any really new information.) But this is an ideal of what it is to learn from experience. It is because of this ideal that Hume proposes his constant conjunction theory, and it is in the light of this ideal that it seems convincing. But this ideal is unempirical. It is a paradigm loop closed in upon its self. Even if scientists must use (in some sense) idealisations as their basic notions they don't expect to be able to just reason and produce conclusions without having to look again at experience to check them. And science is full of cases where it was thought something must be the case, although it turned out that this was wrong, or that something can't be the case, although it turns out it happens all the time e.g. the flight of bumble bees. But nothing is allowed to count against Hume's theory of what it is to learn from experience because we have got a philosophical definition, or at least idea, of what 'learning from experience' must consist in, or what 'basing our knowledge claims on experience 'must consist in' and so if anything doesn't match up to that idea, it will entitle us to say 'this isn't empirical'. We are not going to be allowed to say "This is what learning from experience normally consists in doing, and it isn't what Hume supposes" because we are not dealing with how people actually do perform investigations which they then call "learning from experience" or "this theory is based on experience" or how they normally do go about reasoning regarding experienced situations and matters of fact. We are thinking that if anything is properly learnt from experience it must follow this pattern because such a pattern defines what being 'properly learnt from experience' must consist in. "All our more lively perceptions, or impressions precede our less lively impressions or ideas, and the later are copies of the former".
It must be admitted that if some aspect of our knowledge cannot be derived from experience, or justified by pointing to experience, then that bit of knowledge, if it still counts as knowledge, is separate from experience. But the empirical question is "how can our knowledge claims be based on experience?" Hume supposes that if they are to be based on experience this must mean that all the more lively perceptions of the human mind i.e. its impressions must precede its less lively perceptions i.e. its ideas, and that the ideas are copies of the impressions, which they thus can represent. He also claims that this is found to be true, based on experience, because he challenges us to find an exception to this rule. This is what I think is his ideal of empiricism. This is what he thinks empiricism must require and boil down to, and it is the consequences of this position that he attempts to draw out in his philosophy. That it is an ideal, or dogma, can be seen from the fact that although there are possible exceptions to it, the missing shade of blue, mathematical knowledge, cause and effect, the idea of the self, the idea of one thing being similar to another (when there is not an impression 'being similar to' that can be pointed out in experience, and which this idea copies), these are not counted as exceptions but as things which must be accounted for by their ideas being in some way derived from our impressions which precede them Or mathematical knowledge is supposed to consist in nothing but the relations of ideas, which since they add nothing to the ideas can be known through them independently of experience—But similarly with the relations of ideas, the similarity or diss-similarity of the objects and properties in a causal situation to what is produced by the situation?). The strength of this position is that it posses a challenge as to how we can give content to such ideas so that they are not just assertions which for all we can say further might be empty, without resorting to something like Hume's strategy. But, just as in other sciences although one set of rules or postulates or axioms are often thought required to handle, or explain, its subject matter. Still it is discovered through developing that understanding and through close examination of the experiences involved that a change in the axioms is needed, or that the subject matter can or must be accounted for in another way; Just so, it may equally be possible that through looking at the normal experience of our engagement with experience we may find there are other ways of justifying our knowledge claims by experience than by trying to suppose our more lively impressions precede our less lively ideas, and the later are copies of the former. For instance it might be that there are other ways of getting more out of our engagement with experience than by attempting to base the nature of our ideas on the nature of our impressions.
Getting back to causation for an example; when Newton said that that he would count as universally true what was shown to be true in all our experiments, and until some exception was found, he didn't say we should suppose something is universally true based on many instances of it being true, and until some exception is found. The difference is that Newton's basis for extending our assertions is based on the careful examination, and understanding of, present instances; Hume's basis is upon the numerical weight of instances which are thought, or found, to be similar. In Hume's terminology, the numerical weight of relationships found between our more lively perceptions or impressions in experience increases the liveliness (and strength) of the relations between their corresponding ideas.
This difference goes with a difference of the use of experience in experiment. We can broadly regard experiments or 'experience' as attempts to find some statistically significant correlations between phenomena, or we can view experiments as looking to find re-producible instances or examples of experience which can be examined and manipulated so as to discover which bit produces what part of the situation. (we can also view them as attempts to refute a theory (popper) or occasions for normal science (Kuhn). Mark Collier in Hume studies Vol.31 No.1 is an example of someone using the first approach to what it is for people to understand factual situations. An alternative to this would be to take e.g. magician's tricks as giving a case where failure to understand a situation can be produced on demand, so that we might try and isolate which part of these situations, or at least a class of them, produces this on demand astonishment, implying that where this part isn't found we wont feel astonished, but will be more likely to feel we understand. We could then hope to demonstrate the correctness of such a relationship by means of the reproducibility of the situations where it appears, and which we will now be in a position to manipulate. This later investigation would also not be an anthropological investigation about how people in different cultures do or have reasoned, but would be about a basic part of our reasoning ability, demonstrated by its re-produce ability, in as far as it is re-producible, across cultures, and applied to the scientific situation. [I can remember a television program where an Australian explorer reported that it was impossible to amaze or astonish the aborigines because they found everything equally mysterious.]
An example where the re-produce ability of a situation and consequent ability to manipulate it counts towards it being regarded as a scientific achievement is Newton's splitting of white light into colours of the rainbow and his re-combination of it into white light by the use of prisms. This is also an example where it seems the presumption of the continuation of the factors in the situation demonstrates that white light must be composed of those factors. Even though, in this case, the white light appears qualitatively (and so essentially) different from those factors.
To someone who is looking at experience from a statistical point of view reproducible relationships between phenomena must be used primarily as illustrating the constancy of experience in the respect that is reproducible. But it is pretty ridiculous to suppose, for example, that when Galileo rolled balls down a plain and examined how they accelerated that he was trying to work up a habit and find out how they usually accelerated down the plane. The much more plausible reason he would have repeated the experiment is to try and avoid errors which would be especially likely to occur given the primitive nature of his equipment. He also, by gradually making the experiment more regular by carefully examining and repeating it, would have learnt various parts of the situation that could effect the result, and gained more skill and knowledge in being able to manipulate them. From a Humean empirical point of view this must seem to take a central factor for granted, because it seems assumed in all this that the situation must in itself be regular, and so if there is any difference between instances of the situation this must be due to some difference discoverable in the situation. This is also just what we appear to assume in the case of billiards. But it is what must result from experience, according to Hume's theory and if Hume's notion of empiricism is going to be right. But because this is an assumption, and so because it may be wrong, it does not follow that it is wrong in any situation, much less that it must be wrong. It also does not follow that it cannot be found out to be wrong in any situation much less that we may be presented with a situation where we can recognise that it at least appears to be wrong. So, although it is an assumption it is not clear that it must hamper a correct grasping of the nature of an experienced situation, or that where it does hamper a correct grasping of the situation this can't be pointed up through and from experience. So, although it is an assumption to suppose that where there is a different result from a situation there must be some difference in the contents of the situation, this does not show that it must hamper a correct grasp of the nature of a situation through experience. But also, if it is an assumption that is correct, or true, then it is one that will be able to greatly facilitate our ability to explore situations and learn through the examination of our experiences of them. This is because, wherever there is an apparent difference in what results from situations there must, according to this assumption, be some difference existing in the situation. So it will encourage attempts to discover this difference, and this is a principle for seeking out 'experiences' previously unknown or suspected. So which is the un-empirical point of view? The view that requires everything to result from experience, or the one that includes an assumption which if true can greatly increase our ability to learn from experience, and which assumption does not rule out attempts to handle experience in other ways which may suppose the assumption is false?
Add a comment
Critisisms of Hume's causal analysis
Firstly this is because most of our most basic causal relationships depend on idealisations which a)cannot be found directly in experience e.g. an object without any force acting on it having unchanged motion in a straight line; this is a basic assumption that underlies our understanding of the causal reactions that must be involved with any moving object. But it is not observable in any experience. b)There are exceptions to the generality of properties being found with objects e.g. not everything falls to earth e.g. the moon, not everything maintains its volume, living things for example, or water evaporating. It is not found that all bodies attract e.g. magnets repelling each other, and most objects to observation have no attraction between them. c) If it were the case that the feeling of satisfaction or necessity or that we understand a situation depends on the factors being constantly found together in experience, and if it were the case that things were regularly found together we aught to feel we understand everything. For us to feel we don't understand something there must be an absence of apparent regularity in the situation and between that situation and other similar ones but there are lots of things we feel we don't understand, so a large amount of our experiences can't be regular, but it is supposed to be on this basis that we conclude everything must have a cause. Hume admits that the result cannot be bigger than the cause in 'of scepticisms with regard to the senses' and is forced by this to suppose it is by the coherence and simplicity our understanding gives to our experience, not the regularity of our experience, that we reason. (But what is his reason for supposing the result can't be bigger than the cause?) But he should equally give up his regularity theory in the present case of causes too. d) there is no way, and certainly was no way in Hume's day, to count up regularities and balance them against unregularities. So he couldn't have proposed his theory because he'd found that there was such a preponderance of regularities in our experience of those cases we thought we understand and think necessarily must be so. e) it is implausible to suppose that scientists realised a regularity when they proposed their laws, and if they did what they proposed has much greater extent than the regularity e.g. Boyles law. How many times had he checked it in his experience? and from this small amount so what? f) it is implausible to suppose that e.g. Archimedes recognised that he had always seen that the amount of water displaced was equal to the weight of a body floating in it, or that the weight of an object in a fluid is reduced by the amount of the fluid that that object displaces i) when he thought of it as a law ii)when he proposed that it is a law; the conjunction involved is too obscure to be noticed apart from the theory, let alone noticed as constantly true prior to the theory so that the theory can be based on it. g) If we take Hume's own paradigm of a cause and effect situation, billiards or snooker--have you ever played snooker? Unless you happen to have a talent for the game, even easy shots often miss by along way. Mostly you have no particular reason, or specific part of the experience you can point to, to say that that was what you did wrong. It may be that the balls generally go very roughly in the right direction, that you expect that through practice you will be able to improve, and take for granted that there is (must be) some difference somewhere in the situation whenever you miss a shot. But this assumption is not based on your experience, because your experience is fairly random. After all even top class players miss shots fairly often and no-one to date is quite sure why kicks happen. The fact is that we always suppose there must be some reason in the situation given a different result, and thus suppose that there is a greater regularity in the situation than is apparent in our experience of it. So this attitude cannot be dependent upon the statistical nature of our experience of the situation. h) There have been changes in our understanding of situations that are basic to our causal understanding of situations e.g. the natural state of a body. Aristotle thought that the natural state of a body was rest, Galileo and Newton thought that the natural state of a body was continued un accelerated motion in a straight line. If these basic understandings were based upon any constant experience of objects coming to rest naturally and objects naturally continuing with un accelerated motion in a straight line, two different and incompatible results would have to be constantly found true in our experience of this type of situation. This seems to illustrate the fact that things can only be found constant from some particular point of view, if you have a different point of view a different set of relationships may appear constant (except that Newton's view can't be found constantly in experience, or strictly in any experience at all, because it is an idealisation). Thus the point of view seems more fundamental than what is found constant, because what is found constant depends on it.
j)Not only is it not found through any clear list of experiences that when it is our opinion states are causally connected this is subsequent to their being found constantly together in our experience, it is also the case that there are conjunctions which are clearly constant in our experience but which we do not think are causally connected. The old chestnut here is night and day. The one is constantly followed by the other, but we don't think that night causes day, or that day causes night.