The Nature of Causation

THE NATURE OF CAUSATION

Empiricism: All knowledge is derived from the senses. Anything we know is either directly observed or else deduced from observation.

Hume is an empiricist. He reasons:

· In order to analyze causation, we must first understand what we mean by “causation”.

· Since knowledge is derived from observation, it follows that:

· To understand (and therefore analyze) causation we must look to the sensory sources of that idea.

Hume on causation

Consider any two events that you wish to call “cause” and “effect”. What do you observe that gives you this knowledge?

Suggestion: We notice that one has a certain property that renders it a cause; the other has a property that makes it an effect.

Problem: Any kind of event can be an effect and any kind can be a cause.

· There is nothing that all causes (effects) have in common.

Conclusion: Causation is a relation between two events.

· But what kind of relation do we observe when we perceive cause and effect?

Observing causation

1. Contiguity: Causes and effects are proximate.

· Something can affect something distant only by a chain of intermediate, contiguous causation.

2. Succession: The cause precedes the effect.

Against simultaneous causation:

· If causes and effects are simultaneous, nothing could happen that would take time—everything would occur at once!

What else is there?

Hume: If I examine my experiences of causation, I can find only contiguity and succession:

First suggestion: A caused B iff A and B are contiguous and A precedes B.

An objection

E.g.: I tap a light bulb: it turns on.

· There is contiguity and succession, but did I cause the bulb to illuminate?

Imagine that, unknown to me, someone flicked the switch just as I tapped the bulb:

· So, in my case, we have succession and contiguity but no causation.

What’s missing?

· Anti-Humean: Causes must necessitate their effects.

· That’s how we distinguish causation from coincidence.

Observing necessitation

Hume: We can’t observe necessitation!

· We don’t see, hear, touch, smell or taste it.

Consider the person who flicked the switch:

· Did she observe “necessitation”?

Hume: No. She only observes contiguity and succession?

Okay, but: how do we learn that flicking light switches causes bulbs to illuminate?

· What distinguishes causation from coincidence?

Repeated instances

Hume: We do not learn of causation by observing a single instance.

· If I tapped the bulb once and it illuminated, I wouldn’t know whether I’d caused it to light up—coincidence?

But we do get the idea that A caused B if we observe A and B to be constantly conjoined.

· If I tap the bulb on subsequent occasions and it doesn’t illuminate, then there is no causation.

· But if it does illuminate every time it is tapped, or the switch is thrown, then there is causation here.

Final definition:

A caused B iff: A-events are constantly conjoined with B-events.

Constant conjunction vs. necessitation

Hume: If, we can’t observe necessitation when we observe B following A once:

· We can’t observe it by seeing B follow A many times.

· Repeated viewings don’t show us anything new.

Connection of ideas

After observing the constant conjunction of A and B, the mind forms a certain habit: it expects B upon observing A.

· I.e., we feel an impulse towards B when A is observed.

· It is for this reason that we feel A “necessitates” B.

Since the idea of necessitation doesn’t come from things we observe:

· It must be derived from the mind; from the propensity we have to think of B whenever we see A.

Necessity is in the mind, not in the world. The world only contains constant conjunction.

Summary of Hume’s view

A causes B if and only if A’s and B’s are constantly conjoined, i.e. if and only if:

· A is contiguous with B

· A precedes B

· A events are always followed by B events (anything similar to A is followed by something similar to B).

· The only “necessitation” between A and B is the habit of mind that their constant conjunction produces in us.

Objection (?):

· Day and night are constantly conjoined.

· But day doesn’t cause night.

Must Hume admit that day causes night? How might he respond?
Reid’s objection

Reid: Hume bases all possible knowledge of causation on the observation of external events.

· But if we reflect inwardly, we can directly observe necessitation:

When I move my body as the result of an act of will, I am directly aware of my own power, my ability to necessitate something.

· Our knowledge of causal power comes from voluntary actions.

The general concept of causation comes by extending this one:

· A is a cause of B if the relation between A and B is the same as that between a voluntary act of will and an action.

Consequences

If this is right, then we can directly observe necessitation. But there are problems here:

Problem 1: If causation is the relation between an act of will and the result:

· It follows that only conscious beings can cause anything.

· The world of matter, plants, etc. is non-causal (passive).

· But surely causation governs the material world!

Problem 2: Hume can challenge:

· How do you know you observe “power”?

· What do you observe other than constant conjunction between acts of will and actions?

Against Causation

Russell: Hume is right; our everyday understanding of causation is:

1. A is always followed by B

2. A was necessarily followed by B

The last condition rules out day causing night.

Problem 1: What we normally take to be causes and effects are never constantly conjoined.

· E.g.: Taking arsenic causes death.

· We take this to be true, but there are cases (rare) where someone survives swallowing arsenic.

· Also, one can be shot before the arsenic takes effect.

Response: Okay, so causation is nearly constant conjunction.

· E.g.: Taking arsenic is usually followed by death.

Russell: A whistle almost always precedes workers going for lunch.

· Does the whistle cause them to leave?

Yes: nearly constant conjunction.

· But, many other whistles in factories nearby are also nearly constantly conjoined with the lunch break.

· They satisfy the condition as well, so how do we rule them out?

Hume: contiguity!

· Only one whistle is contiguous with the workers (via a chain of smaller contiguities).

Infinitesimals

Russell: Okay, but let’s see where this leads.

· Arsenic can’t cause death because taking arsenic is not really contiguous with death.

After all, one doesn’t die immediately; there are processes that occur in between the two events.

· So, we must “shorten” the events considerably.

This makes sense anyways:

· We don’t care what happened before he swallowed or what happens after he died.

Russell: Once we start down this path, we see that we have to “shorten” the events indefinitely.

· Until we have infinitesimals.

· These are the only “events” that are truly contiguous.

But how do we describe what occurs at the infinitesimal level?

Using differential equations!

· These tell us how something’s motion is changing at a moment (acceleration).

· They don’t tell us where that something will be at a later time.

What they do not tell us:

· That one event is always followed by another

· That one event necessitates another

· They describe processes, not relations between individual events.

No events

Moreover, whatever these infinitesimals are, they don’t look like ordinary events.

They are not such things as:

· Gun shots.

· Arsenic swallowing.

· Explosions.

· Etc.

So at the fundamental, physical level, there is nothing even approximating our idea of causation.

Conclusion: The scientific view of the world leaves no room for anything resembling our everyday understanding of causation.

So, when does this ordinary idea come from?

Empirical generalizations

“Arsenic causes death” is a rough, empirical generalization.

Advantages:

· Observable/confirmable.

· Nearly constant conjunction

· Useful, pragmatic guide to living

But they are not fundamental.

At the basic physical level of reality, there is nothing that comes close to our idea of causation.

· So, causation is not physically real.