Hume asks, "what is the nature of that evidence" which can assure us of matters of fact, "beyond the present testimony of our senses, or the records of our memory?" And he answers that all of these reasonings "seem to be founded on the relation of cause and effect." Hume is here recognizing that much of our reasoning does not prove itself by direct perception or impression, nor by the intuitive or demonstrative analysis of the relations of ideas. These reasonings must presuppose the general idea of cause and effect in order to infer such propositions from experience. The presupposition of cause/effect is necessary in the proof or validity of those statements of truth. And since this is necessary in the justification of these matter of fact statements, Hume begins to question the validity of this presupposition.
Is there an empirical justification of cause and effect? If a take a directly perceived fact, such that X has occurred, then how am I justified in asserting that Y is the cause, or that Z will be the predicted effect? I am inferring Y and/or Z from known X, so I am supposing some kind of necessary connection between these events. But, can I really know this necessary connection? Our reason, unassisted by experience, can never draw this inference. It is not knowledge `a priori' or innate, because the statement that an event has a cause and/or an effect is not just a necessary definitional relation of ideas. It can be contradicted hypothetically, and it is not something that we just immediately assume before any experience whatsoever (though Kant would disagree here). The inference is not intuitive and neither is it demonstrative. So, "the knowledge of this relation is not, in any instance, attained by reasons `a priori'; but arises entirely from experience, when we find, that any particular objects are constantly conjoined with each other."
Somehow, the idea of cause and effect comes by way of experience, though we have no actual experience of it. Causation, as an event, is never experienced in any one moment, and "every effect is a distinct event from its cause"... so could not be "discovered in the cause." It is only because of our experiences of things conjoined in some temporal sequence, that is, we experience impressions in a temporal sequence or priority, that we infer the cause/effect relation or the power of things to move other things in the world. We cannot actually have impressions of causes or powers.
These inferences from experience presuppose some similarity between known events/objects and predicted events/objects, and they presuppose a uniformity of nature such that the future resembles the past and similar events/objects will behave uniformly. Experience will be found useless in these inferences if nature may change, or if the past (recorded in memory or on notes) cannot show something about the future. Therefore; we must prove the necessary resemblance of the past to the future, or the necessary uniformity of nature. But Hume points out that it is logically impossible to experientially prove this resemblance or uniformity, "since all these arguments are founded on the supposition of that resemblance (Pt.II). Since these inferences require [constant and uniform] experience over time, we cannot use experience to justify them.
It would seem though, to me, that a good amount of experiments or experiences would give us adequate reason to infer a general law of causality and uniformity of nature. I can agree with Hume that a person suddenly brought into the world could only observe a succession of objects next to each other in time, but could not observe any actual powers of these objects to affect the others so could not without doubt assert necessary causal connections, nor could assert a uniform law of causation in nature. But over a larger period of time, experience would gain validity in experimental inferences, such that one could say that given all of this experiential data it is reasonable and probable to believe in the uniformity of nature and the law of cause/effect.
The cause/effect relation is an
inference, based upon how the imagination makes such an association of
relation, according to
the contiguity and priority of impression-based ideas in the mind. * [I
don't believe that Hume meant that contiguios objects had to be butted
up to each other, but that they occupy the same area of space or the
same strech of time, or are `near by' although there would be concern
as to how large of space or how long of time; and
this is not resolved empirically or rationally, but only in the
imagination, or what mind considers as contiguious - is this
wrong? ]
Now, we must question the validity of such an inference. We have grounds for doubting the reasonableness of an inductive inference of a law of cause/effect based upon particular experiences of conjoined objects. We might skeptically say that at best all we can have is probable knowledge of causation, or of the uniformity of nature presupposed by such an inference of causation. And yet, Hume recognizes that, for all practical purposes, it is reasonable to believe in this regularity and uniformity in nature and expect it with great assurance, as long as "... no instance has ever yet been found of any failure or irregularity (ECHU VI)."
As long as our experience of nature continues to support our belief in uniformity and our belief in a necessary causal chain involving certain events, that is, our impreessionable experience does not contradict these beliefs (or expectations), then, our belief is justified as well as possible, assuming that these beliefs do not contradict other idea-beliefs, but this justification is not an epistemological certainty.
As Hume says, "we give the preference [of belief] to that which has been found to be most usual (VI)," but not to close out any consideration of alternative beliefs or overlook contrary evidence. So, what stands as reasonable beliefs concerning cause and effect are those which are most supported by frequent or regular conjoinment in experience, that is, supported by some degree of contiguity, priority, and repetition in ideas gotten from spatial and temporal experience.
Hume attempts to show that there is no real empirical evidence of cause and effect, or the power of something to move something else. He says, "we are never able, in a single instance, to discover any power or necessary connection; any quality, which binds the effect to the cause, and renders the one an infallible consequence of the other. We only find, that the one does actually, in fact, follow the other (VII,Pt.1)." Nothing, in any single instance of experience, can adequately suggest the idea of power or necessary connection. The power of one thing that might causally move other things is completely concealed from us.
Our true observation, according to Hume, is that "all events seem entirely loose and separate. One event follows another; but we can never observe any tie between them. They seem conjoined, but never connected (VII,Pt.2)." The only connection is in our thought and not in experience. And the connection in thought is merely a habit of assuming the future to be like the past, a kind of unjustified habit of the mind which gives rise to the inference of causality. This inferred causality is merely a customary explanation of what is really only a common habit of thought.
The true definition of cause, therefore, should be a psychological definition, a habit of one thought being conveyed by another thought, due to the continual conjunction of the two together in experience, such that the appearance of one automatically conveys the thought of the other. Causality is thus a mental anticipation of the occurrences which usually precede or follow a certain event, or similar kinds of events. And, I would assume, that Hume would also believe that a general notion or idea of causality could be assumed as a law of nature due to many kinds of conjunctive experience.
It would seem, then, that these acquired beliefs are
inductive inferences of the causes or effects of certain events, due to
the experimental evidence of constant conjunction between the events
we call `causes' and the events we call `effects'. In other words, if
we regularly observe two events constantly conjoined together,
that is, one is never without the other, then, at least within the
limits of our experience, we are moved to believe in some necessary
connection between them. But, Hume says that the reason for this belief
is psychological, not reason or observation. The belief
cannot be justified according to just reason or just empirical
evidence.
It is not a logical or analytic necessity that necessary connections
exist, and nobody ever observes these connections. All that is observed
is the regularity of constant conjunction. But somehow,
out of these observations a belief in necessary connections arises. But
Hume thinks this belief is justified from a practical standpoint, even
if unjustified in the traditional philosophical sense. And not only is
the belief useful in everyday practice, it
is virtually necessary psychologically in order to maintain some
sense of uniformity and coherence in our everyday thinking about the
world.
Another possibility for validity is in a combination of the two
traditional types of validity. If one type of validity is
concerned with coherency and non-contradiction in the relation of
ideas,
while the other type is concerned with regularity of idea with actual
experience, then we might be justified in our beliefs if they satisfied
both criteria, that is, coherency in the relation of
ideas and regularity in the experience of conjoined events/objects. Of
course, even both of these justifications together do not secure
that the belief is infallible, or that future experience or developed
ideas will not contradict and invalidate the belief at some time. But,
what can be expected of any not-just analytic beliefs, except
probability?
For Hume all real knowledge of fact must come from impressions of the senses or of reflection. He grounds knowledge, then, in the same empiricism as Locke, for out of these simple impressions we can construct complex ideas. Ideas find their validity in this empirical source. So, knowledge can either be impressions or ideas "copied" from impressions, and "all ideas, especially abstract ones, are naturally faint and obscure,"(H 319). in comparison with the more vivid impressions. Most knowledge is a complex arrangement of ideas linked together in thought according to certain psychological principles of association using memory and imagination. Hume lists three main types of association: resemblence, whereby we can classify things according to their common image or properties; contiguity, which is being reminded of things and events near to the same time or place; and, cause/effect, which is actually the experience of conjoined objects or events having occured in temporal sequence and appearing to be of the general place.
By making these associative arrangements of ideas, based upon sensory or reflective empirical impressions, we are using memory and imagination to reason about truth, and when we reason we form propositions or statements of truth, justifyable or not. Hume then makes a distinction between types of propositional knowledge, "divided into two kinds, to wit, relations of ideas, amd matters of fact" (H 322).
The first type is a priori to experience, and their validity is found within the definition itself, within the relations of ideas. Kant calls these propositions `analytic' because the predicate meaning of the sentence can be found within an analysis of the subject meaning, so that these analytic statemnets cannot be denied without contradiction. The predicate is contained within the subject, even though the terms are not the same. As Hume says, "Propositions of this kind are discoverable by the mere operation of thought, without dependence on what is anywhere existent in the universe" (H 322).
Thus, these ideas of relation, though having their foundation in impressions since every idea is an impression, are not exactly empirical, since they cannot even be denied by future experience, and they do not give us additional information as to matters of fact. Hume, then, questions what could give us ideas of "matters of fact", other than our direct sensory impressions in the moment (or in memory), or the intuitively or demonstratively known "relations of ideas". He asks "what is the nature of that evidence" which can assure us of matters of fact, "beyond the present testimony of our senses, or the records of our memory" (H 323). And he answers that all of these reasonings "seem to be founded on the relation of cause and effect" (H 323).
Hume is here recognizing that much of our reasoning does not prove itself by direct perception or impression, nor by the intuitive or demonstrative analysis of the relations of ideas. These reasonings must presuppose the general idea of cause and effect in order to infer such propositions from experience. The presupposition of cause/effect is necessary in the proof or validity of those statements of truth. And since this is necessary in the justification of these matter of fact statemnets, Hume begins to question the validity of this presupposition. Is there an empirical justification of cause and effect?
If I take a directly perceived fact, such that X has occured, then how am I justified in asserting that Y is the cause, or that Z will be the predicted effect? I am inferring Y and/or Z from known X, so I am supposing some kind of neccesary connection between these events. But, can I really know this neccesary connection? Our reason, unassisted by experience, cannever draw this inferrence. It is not knowledge a priori or innate, because the statement that an event has a cause and/or an effect is not just a neccesary definitional relation of ideas since "every effect is a distinct event from its cause"... so could not be "discovered in the cause" (H 326), and it can be contradicted hypothetically, and it is not something that we just immediately assume before any experience whatsoever (though Kant would disagree here).
The inference is not intuitive and neither is it demonstrative. So,
"the
knowledge of this relation is not, in any instance, attained by reasons
a priori; but arises
entirely from
experience, when we find, that any particular objects are constantly
comjoined with each other" (H 324).
Somehow, the idea of cause and effect comes by way of
experience, though we have no actual experience of it. It is only
because
of our
experiences of things conjoined in some temporal sequence that we infer
the
cause/effect relation or the power of things to move other things in the
world. All inferences from experience presuppose some similarity between
known events/objects and predicted events/objects, and presuppose a
uniformity of nature such that the future resembles the past and similar
events/objects will behave uniformily.
Experience is found useless in inferences if nature may change, or if
the
past (recorded in memory or
on notes)
cannot show something about the future. Therefore; we must prove the
neccesary resemblance of the past to the future or the necessary
uniformity
of nature. But Hume points out that it is logically impossible to
experientially prove
this resemblance or uniformity, "since all these arguments are founded
on
the supposition of that resemblance" (H 332).
It would seem, though to
me,
that a good amount of experiments or experiences would give us adequate
reason to infer a general law of causality and uniformity of nature. I
can
agree with Hume that a person suddenly brought into the world could only
observe a succession of objects next to each other in time, but could
not
observe any actual powers of these objects to affect the others so
could not
without doubt assert necceary causal connections, nor could assert a
uniform
law of causation in nature. But over a larger period of time, experience
would gain validity in experimental inferrences, such that one could say
that given all of this experiential data it is reasonable and probable
to
believe in the uniformity of nature and the law of cause/effect.
Hume
disagrees. Or, at least he does not believe that these inferences are
justifyable only by experience or by a priori reasoning. Since these
inferrences require [constant and uniform] experience over time, we
cannot
use experience to justify them?
First, we have concluded that we do make some kind of inferrence about
cause/effect. Now, we must question the validity of such an inferrence.
We might then have grounds for doubting the reasonableness of an
inductive
inferrence of a law of cause/effect based upon particular experiences of
conjoined objects. We might skeptically say that at best all we can
have is
probable knowledge of causation, or of the uniformity of nature
presupposed
by such an inferrence of causation. And yet, Hume recognizes that, for
all
practical purposes, it is reasonable to believe in this regularity and
uniformity in nature and expect it with great assurance, as long as
"... no
instance has ever yet been found of any failure or irregularity" (H
347).
As long as our experience of nature continues to support our belief in uniformity, or of the object/events thought to be linked together in a necessary causal chain do not contradict our theory; then, our belief is justified, not as epistemological certainty but for all practical purposes, assuming that a contrary belief would be less practical. As Hume says, "we give the preference [of belief] to that which has been found to be most usual" (H 348), but not to close out any consideration of alternative beliefs or overlook contrary evidence. So, what stands as reasonable beliefs concerning cause and effect are those which are most supported by frequent or regular conjoinment in experience.
These acquired beliefs, then, are inductive inferrences of the causes
or effects of certain events, due to the experimental evidence of
constant
conjunction between the events we call `causes' and the events we call
`effects'. In other words, if we regularily observe two events
constantly conjoined
together, that is, one is never without the other, then, at least
within the
limits of our experience, we are moved to believe in some necessary
connection between them.
Hume says that the reason for this belief is
psychological, not reason or observation. The belief cannot be justified
according to just reason or just empirical evidence. It is not a logical
or analytic necessity that necessary connections exist, and
nobody ever observes these connections. All that is observed is the
regularity of constant conjuction. But somehow we make an inferrence
from
these observations to believe in necessary connections. How? And is
this valid?
These are two separate questions. Hume maintains that this
is
not an empirically valid move (but if this is not a valid induction,
then
what the heck is?), nor an anaytical move. It is [instead] a
psychological
move or a tendency of the mind to infer truths based only on constant
conjunction of experience (could we not then say that all inductive
reason is a
mere tendency of the mind -- that would sound kind of redundant). For
Hume
the belief in causality is a psychological tendency of practical man.
It is
a habit or custom of the mind because of the regularity of such constant
conjunctions. This does explain the psychological mechanism for
cause/effect
beliefs, which is still based upon the empirical model of acquiring
ideas,
and does not invoke innate ideas nor require reasoning
by proofs or demonstration. Is the belief then valid, or justified? Not
if
we demand justification by either a priori reasoning or empirical
perceptions of these supposed connections.
Hume has shown here that
neither
statements of "relations of ideas" nor statements of "matters of fact"
can
validly apply to such beliefs of necessary connection. Each kind of
justification fails, and we are left without a category to place such
inductive beliefs. Kant later comes along and places this within a
category
of knowledge, which is synthetic a priori knowledge, but this would not
appeal to Hume because he finds that beliefs of necessary connection are
acquired through some temporal process, while Kant finds this belief to
be
necessarily [a priori] presupposed
in any kind of science or thinking in general.
So, we might then stop here and conclude that beliefs in necessary connection are not justified. But Hume thinks it is justified from a practical standpoint, even if unjustified in the traditional philosophical sense. And not only is the belief useful in everyday practice, it is virtually necessary psychologically in order to maintain some sense of uniformity and coherance in our everyday thinking about the world. Another possibility for validity is in a combination of the two traditional types of validity. If one type of validity is concerned with coherancy and non-contradiction in the relation of ideas, while the other type is concerned with regularity of idea with actual experience, then we might be justified in our beliefs if they satisfied both criteria, that is, coherancy in the relation of ideas and regularity in the experience of conjoined events/objects. Of course, even both of these justifications together do not secure that the belief is infallible, or that future experience or developed ideas will not contradict and invalidate the belief at some time. But, what can be expected of any not-just analytic beliefs, except probability?
All that Hume should be concerned about is the correspondence between such complex ideas as causation and the impressions from actual events in the world. What we should be validating, I think, is that ideas or beliefs have their foundation in actual impressions or experience.
Hume sees the absurdity of having an idea of connection or power, because it "never appears to our outward sense", but this is only reasonless because Hume has already defined [true] ideas as necessary copies of impressions. So, as I see it, "these words [the proposition of necessary connection] are absolutely without any meaning" (H 360) because of his definition of ideas.
The mind has not actually observed a necessary connection between events; instead, the mind has inferred the connection because of regular past perceptions of constant conjunction. The mind has come to expect the one event upon the appearance of the other, and the mind has acquired the custom of believing in causality. We do not see a connection and then make an inferrence about future events, and we do not see uniform connections in nature and then inductively reason causality in nature; instead, we imagine a connection because of a habit of mind to suppose that future experiences of constant conjunction will be similar to past experiences. Yet, I think that whether or not this is a reasonable assumption or merely a habit of mind, whether it is justified by reason or by psychological mechanism, is but a quibble over definition, for "reasonable assumption" and "habit of mind" are essentially the same since both are based upon regularity of experience and coherancy of ideas.