David Hume philosophy is best known for its empiricist approach to investigating human nature. His philosophical account of causation is certainly revolutionary for its time, in that it assumes a mind-dependent perspective of how reality is manifested. He investigates the empirical foundations of what it means to have knowledge about objects, as well as what the nature of objects themselves are. In presenting a detailed explanation rationalizing our mind’s movement to immediately conceiving connections and beginning to draw causal inferences, he relies on the notion of constant conjunction and its significant role in causation. Hume’s analysis of this causation begins with the fundamental discussion of “relations” between objects we perceive. He posits that there are many different kinds of relations, particularly seven, which he terms as “philosophical relations”. This extensive list is interesting because when we take a closer look at what he says about these seven philosophical relations, we discover that they can be divided into two kinds: relations between ideas and relations between matters of fact. In this exposition, which I will explain in detail in the following paragraphs, Hume attempts to explain the connection between knowledge and causation; or in other words, how causation can give us knowledge of objects and things we perceive through our outer experiences, or sensations.
As a starting point, Hume begins with explaining relations between matters of fact. He presupposes that these sort of relations are independent from ideas themselves. To clarify this concept, take an example of this sort of relation: box and table. The table, which occupies an extension in space, is a relation of place. The box, on the other hand, is located above (on top of) the space taken by the table itself. According to Hume, this relation between box and table does not depend on the idea that one holds about the box itself, or the table itself. Most importantly, is that this relation can change without any changes in the particular way one represents either component (box or table). Now, an interesting phenomenon occurs; when someone grabs hold of the box, lifts it, and moves it from off of the table. In this situation, the box is now located to the left of the table. It seems at this point that the relations have changed, without changing the idea of either the table or the box as independent entities. Like the relation of place, relations of identity work in this same regard as well. Relations of identities consist of the numerical identity of objects across different and un-interrupted perceptions of them. Identities, like place, are therefore another instance of a relation between matter of fact. They are not, however relations between ideas because relations of these sorts are dependent on each other.
It appears, through the above analysis, that those relations that can change between objects without any change of the ideas we have of these objects are relations between matter of fact. Due to the fact that these objects, or the ideas that we have of these objects, do not change; they appear to be non-contingent on each other. Moreover, these relations of matter of fact are not analytical relations because they are not part of the ideas that we hold of them. According to Hume, causal connections are of this sort: they are relations between matter of fact. This explanation is incredibly important to remember because it allows the reader to discover that Hume, when he analyzes causation; that part of what “causal relations” are is what he comes to call necessary connection. Necessary connection, as we will soon discover, is part of causal connection. It is important to keep in mind, however, that when utilizing the term “necessary”, Hume does not mean to imply necessary in the sense of relations between ideas. Rather, he suggests that necessary connection is always contingent.
The discussion surrounding causal relations is important, according to Hume, because it allows one to arrive at probable beliefs (probabilities) concerning the numerical identity of objects, as well as of time and place. Hume suggests that in none of these instances, the mind can go beyond what is provided by the senses, or sense perceptions. These relations, therefore, cannot be made use of reasoning insofar as they have a causal relation. Instead, one comes to have probable knowledge about the numerical identity of objects, or concerning particular relations of time and place, by first making predictions. The importance of predictions is grounded in the mere fact that they involve inferences; and in turn, these inferences play a significant role in reasoning. As such, for one to have beliefs about objects requires more than sense perceptions; this process is dependent on inferences and reasoning. More specifically, the inferences that Hume refers to in this account of obtaining knowledge are particularly causal inferences. For Hume, of all these relations between matters of fact; when it comes to having probabilistic relations, the causal inferences are the most important component. From this point, in order to have justified and perfect understanding, one must trace the idea of causation back to its origin—beginning with impression, the first mental content, which bestows the idea in our mind—. From this account, Hume notes that any idea, in order for it to be respectable and clear, is an idea that must be traced back to the impressions from which it all began. This applies to the idea of cause and effect, as well as the idea of causal relations, just as much as any relation in the mind. It is this particular process, of mental back-tracking, which Hume regards as his methodology for epistemic discovery.
Following this suggestion, Hume then suggests several techniques that one may utilize in order to find the initial impression, which is critical to understanding. He proposes the method of looking for a particular quality of what is supposed to be the “cause” and some particular quality of a thing that appears to an “effect”. By quality, Hume specifically refers to “non-relational properties”, or more commonly known as “modes”. However, Hume immediately dismisses this account on the grounds that it would be impossible to find a non-relational property that makes all causes, causes; and all effects, effects. Moreover, no matter what properties one may discover, it is not guaranteed that these properties will be able to answer the questions “what is a cause” or “what is an effect”. From this stance, Hume contends that causation must be derived from some sort of relation. It seems now that the goal is to discover one or more relations that are supposed to be always present when something is cause, and always present when something is an effect. He identifies these two relations—those relations which are always present when something is a cause and when something is an effect—as “spatial contiguity” and “temporal priority”. These two relations are certainly relations between matters of fact because it turns out to be a relation of place (space) as well as time. According to the analysis above, Hume has already explained that these relations of time and place are understood as matter of fact relations. From this point, Hume arrives as the understanding that in all cases of causal connections:
X causes y if and only if:
- x is temporally prior to y
- x and y are spatially proximate
Unlike other philosophers of his time, Hume discounts the popular notion that cause can be simultaneous with its effect. Rather, these two relations between matters of fact are universally present whenever there is case of cause and effect. Hume implies that nothing would be a cause of some effect if it were neither spatially proximate, nor spatially continuous. In order to gain a deeper understanding of such relations between matter of fact, it is important to clarify what Hume really means by “spatial contiguity”. Common assumptions readily hold that this term has a literal meaning, implying that two things actually (physically) have to touch each other in order to have “spatial contiguity”. However, this is not the case for the same reason that the sun may cause the elliptical orbit of the earth, but certainly does not touch the earth. It seems, from this example that the causation can occur without any physical contact between objects. If causation required physical touching, or if the sun had to touch the earth to cause its elliptical orbit, then the earth would surely overheat and there would be no earth at all. As such, it seems to Hume at this point that these two relations are not everything that is required for causation.
Spatial contiguity and temporal priority are among the sort of relations, as such that we can discover them to be invariably present when we are having sense perceptions, or perceptions of objects. However, while spatial contiguity and temporal priority certainly remain as necessary conditions, they are not sufficient in order to discover cause/effect relationships, which lend to the discovery over particular impressions. A relation that is in fact superior to the previously mentioned ones is that of necessary connection. Again, necessary connection is not a relation between ideas, and still remains as a relation between matter of fact. These three relations, Hume contends, are the necessary components of what makes up “causation”. Accordingly, the following conditions are required:
X causes y if and only if:
- x is temporally prior to y
- x and y are spatially proximate
- x is necessarily connected to y
The big question remains, however: what is necessary connection? As suggested previously, the idea of necessary connections now exists in the mind. But, according to the appropriate methodology, it is essential to trace this idea back to causation. It seems, at this point of the analysis that we have arrived to a dead end; it appears to be a difficult challenge to find other impressions of relations that can provide us with a source of the origin of “necessary connections”. If this feat is too difficult, the idea of necessary connection is no good at all; being that, without causation, we cannot assume knowledge of objects. Yet, Hume maintains that the discovery of the source is essential to understand causal connections; otherwise, it may result in a horrendous kind of skepticism.
Hume admits that the idea of necessary connection cannot be traced back to any impression of sensations. Impressions of sensation—or of perceptual experiences—gave us the idea of the other two relations, but seem unqualified to give us the idea of necessary connection. However, Hume is confident that this difficulty does not result in a skeptical argument and claims instead that there is another, very different, set of impressions. These impressions are termed “impressions of reflection”. It is at this point where the reader is beginning to see the distinction between impressions of sensation, and impressions of reflection as two very different operations. The idea of necessary connection, therefore, is going to trace back to some idea of impression by reflection. In contrast, the other two relations—spatial contiguity and temporal priority—are dependent on impressions of sensation. Hume explains this phenomenon by first suggesting that impression of reflection is in itself an idea of reflection. Some of what is in our inner experience (reflection) is going to provide us with the right source of impressions. From such impressions, we will be able to understand the idea of necessary connections. As such, Hume posits that the idea of necessary connection is comprised of the following; it is in two parts a complex idea of sensation (outer experience) and in one part a complex idea of reflection (inner experience). These two sorts of experiences are essential because experiences allow us to make inferences, and ultimately devise complex ideas of necessary connection. Hume states that this explanation provides us with the insight of how the idea of necessary connection arises. The answer, he suggests, relies on the key notion of “constant conjunction”.
According to Hume, we discover what is really occurring when we come to think that there are causal connections somewhere. In the mind, we begin to realize that there have been, in the past, many cases where things or objects of the same kind (x) have (always) been found to be in the relation of spatial contiguity and temporal priority. By constant conjunction, Hume means that there are many cases of x-like objects, or perceptions of x-like objects; where what is perceived is temporally prior to a particular instance of perception of a y-like object. As a result, the mind receives a myriad of x’s and these are similar in some ways—similar within one particular kind—and we get a bunch of y’s that are similar in different ways—similar within its kind, but of a different type than x’s—. From here, the mind now thinks back through experiences of the past, and sees that there have been repeated instances of x-like things which are spatially continuous to, and temporally prior to, many y-like instances. This is constant conjunction, a repetition of associate pairings of x’s and y’s, and the mind slowly becomes aware of this observable fact. As such:
X1 — (is paired to) — Y1
X2 — (is paired to) — Y2
… and so on, and so forth.
When the mind becomes aware of these past constant conjunctions (pairs between different kinds) it then, at some particular time, t, (present moment) becomes “impressed” by the parings. In other words, the mind realizes this pattern and the conjunctions become impressed within our inner experience. This is how impression of reflection arises, being that it is dependent on inner experiences. This impression can be further described in a number of different ways. In particular, the mind forms the impression of expectation and therefore begins to demand more of the same pairings. The mind comes to believe that the future will continue to display the same sorts of pairings, being that the mind presently has become aware of parings that were experienced in the past. The future works in the same way as the past; nature is uniform, and it will continue in the same manner, according to demands and expectations. This appears to be almost like an emotional reaction that the mind exhibits in response to the past experiences, which will eventually give us the idea of necessary connection. Most importantly, this is an idea of reflection, and it is by means of this reflection that the mind is able to form expectations of such pairings.
Similar to his contemporaries, Hume agrees with Descartes, Locke, and even Berkeley, that knowledge about objects requires inferences. Having an outer sense perception is nothing but simply perceiving all that comes before us as the particular content. We do not, strictly speaking, by means of our senses alone, represent objects in any way. Instead, to have knowledge of them, Hume highlights the importance of inference, or synthesis, because our beliefs about objects are always complex and therefore require this analysis. As such, Hume attributes causal connections as an essential mechanism by which we can eventually form empirical beliefs, and therefore maintain knowledge about objects. This is important because we have to engage in “causal inference”, but to explain this causal inference we must explain the nature of causality in general. As such, his reasoning allowed him to derive the idea that there are supposed to be three parts of the complex idea of causal connection. The third part, necessary connection, is the most important—not because it suggests “logical” necessity by any means, but rather that it is a relation between “matters of fact” (not ideas). These complex ideas come before the mind when the mind turns to inner experience; this inner mind is the only possible place, according to Hume, that we can find the source or origin of the idea of necessary connection. As such, the idea of necessary connection resides within some impressions that arise before inner sense.
More specifically, the sorts of impressions or reflections Hume refers to are those that experience feeling, or emotion. As suggested, the mind is reacting emotionally to its own mental contents by means of expressing a desire for “more”, a determination. Furthermore, these mental contents are those of constant conjunction, and the mind comes to insist that constant conjunctions continue will persist in the same pattern from the past into the future. These constant conjunctions are important because they are paralyzed connections between things that are of a particular kind, or that are similar to one another in a certain way due to the regularity, repetition, relations, “contiguity”, and “spatial priority”. In each of these cases (X1—Y2, etc), some X1 is spatially contiguous to Y1, and the pattern perseveres in the same form. Hume explains this constant conjunction in more formal terms, as the “invariable repetition or regularity of some pair-wise arrangement” (Hume, 192). In sum, it appears that we obtain these constant conjunctions in the past, and then Hume believes that the mind emotionally responds to the present awareness of past conjunctions. This response is unique, and is of a particular sort.
The mind comes to form a definite and strong belief (insistence) that the constant conjunctions—which are instances of the same kind of constant conjunctions in the past—will continue in the future. Hume posits that this is the “principle of the uniformity of nature”, which denotes a uniform correlation and suggests that the future is like the past and thus, continuous. The mind develops “customary expectations” which infers that the events of the past will continue. From this process, the mind forms customs and habits, which are ultimately responsible for our beliefs or expectations that this process will continue. Most importantly, Hume thinks that the “necessity” (necessary connection), which we now ascribe to these correlations, has to be understood and explained according to the custom and habit that now exists in the mind. This concept is certainly a revolutionary idea, which many of Hume’s contemporaries would criticize. Ordinarily, one would think that necessary connections (matter of fact relations) are somewhere out there in the world; that they are completely mind independent, and it is up to the human mind to discover that there are such necessities in the world, as well as investigate which particular kinds of necessities there are. An example of what I mean here is the kind of necessity required to make gravity work, among other observable and natural phenomena. However, Hume turns this idea on its head and suggests a different account; the kind of necessity that he refers to is to be explained by the mind’s reaction, or customary expectation, for these correlations to re-occur in the future. If the mind were not to become impressed by its present awareness, then there would be no necessary connections in the world. Connection, therefore, is to be completely analyzed and understood by the mind’s habitual expectations. In short, Hume contends that necessary connection is mind dependent—specifically explained according to the mind’s emotional reactions. This idea of necessary connection, therefore, is primarily dependent on the idea of reflection altogether.
This whole account starts with outer sensation in order to arise at temporal priority and spatial contiguity. Then the story takes an inward turn to account for the third idea of necessary connection by means of reflection, or inner experience. As such, Hume holds both experiences (inner and outer) to a high regard, believing that they are equally important in enabling us to arrive at knowledge. At this point, one might raise the objection of whether or not the mind is a reliable structure at all. However, for Hume, the concern is not about “reliability”; for there are these necessary connections in the world, and they are independent of how the mind works. It is the mind’s job to discover these connections; and he admits that sometimes the mind falters in the investigation. Hume would say that this is the wrong picture to maintain; he is trying to convince the reader that there is nothing in the mind-independent world to be discovered, but rather everything is mind-dependent. As such, all the manifestations are a function of how the mind works. Therefore, necessary connections are, for their existence, dependent on the mind’s functioning. Hume notes that all minds, for the most part (if they are working properly) will be working in the same sort of way as each other. All properly functioning minds will be able to form the same expectations; those that do not, are in some way or another, “defective” in their nature. He admits that there are deviations from the standard norm of universal human nature, but attributes this to something wrong with the particular mind itself.
Hume’s account of causation attaches a special importance to the nature of human experience. We remember to have had instances of one type of object, and we remember that these objects have had related cause and effect relations. As such, we call one “cause” and the other “effect”. Yet, in all cases where we reason—draw inferences—concerning them, there is only one that is perceived or remembered and the other is inferred. This definitely happens over time, as we come to form expectations. Interestingly enough, this parallels Galileo’s predictions; that the beliefs formed are rooted in inferences based on expectations. In other terms, this is “inference of prediction” which depends on the union of ideas—or the association of ideas that allow our mind to always put the x’s together with the y’s in an orderly, paired, fashion. This habit of the mind has developed to entertain the idea of Y1 when it has the idea of X1; where the belief in Y1 is based on the inference, and the belief in X1 is based on the mind’s customary expectation.
What Hume continues to emphasize in his account is that reason alone cannot demonstrate proof of causal principles. Rather, the internal impression by reflection allows the idea of necessity to arise, which is a key component of causation. If we were to lack an idea of necessity, we would be hindered from drawing causal inferences. In this case, we cannot justify beliefs about objects, we would have no knowledge, and we would ultimately surrender to skepticism. This (uncontrollable) necessity is nothing but that determination of the thought to pass from cause to effect, or effect back to cause, in accordance to their union. On page 197, Hume elaborates on this idea of cause and effect, and explains concretely what he takes causal relations to be. His second definition is more complete, and he contends the following: a cause is an object precedent and contiguous to another; and soon united with it in the imagination (the mind things about them, obtains them in memory, thinking about past experiences and how things will be in the future), that the idea of one determines the mind to form the idea of the other. It seems here that Hume’s analysis of causation, by constant conjunction, is essential in order to describe what empirical knowledge of objects are. A relevant question now arises, of what objects are; and Hume’s analysis of causation certainly has consequences for his analysis of objects. However, this is a separate topic that Hume delves into next and one that is not particularly important to this particular discussion.