Inducing the Law of Causality

Metaphysics and Epistemology
1

Hi, over winter break I’ve been listening (or re-listening, as in the present case) to some lecture courses, most recently Leonard Peikoff’s Objectivism Through Induction. In the first lecture of the course Dr. Peikoff makes a point which, as I recall, I never really understood or agreed with when I originally listened to the course, and which I still find quite confusing. He essentially states that the law of causality does not follow deductively from the axiom of identity, but rather that it (causality) must be induced independently through the observation of a wide range of concretes.

My understanding of Peikoff’s argument is as follows: The validation for the law of cause and effect is not its connection to identity (which is only grasped later in the process of integration) but the fact that, time and time again, entities are observed to act in a lawful fashion. We look out into reality and observe that when dropped, balloons float into the sky and rocks fall to the ground. When shaken, rattles make noise and pillows remain silent. (These are some of the examples given in the lecture). Since there is no proper distinction between necessary and contingent facts, between “it is” and “it must be”, to discover that entities do act lawfully is to discover that those entities must act lawfully. Hence after observing a huge number of entities across a broad spectrum of existence (enough to exhaust all of the signifiant kinds of entities known within our present context), all of which have in the past acted in certain specific ways, we are justified in forming an inductive principle: “Entities of a certain kind act in a certain fashion, and only in that fashion.”

My problem is that I do not understand how one can validate the inductive process itself (i.e. solve the so-called “problem of induction”) without first assuming that causality is true. In other words, the law of causality seems to be at the base of induction, in the same way as existence, identity, and consciousness are at the base of the hierarchy of knowledge. Trying to validate something at the base of induction by using the process of induction would then be hopelessly circular.

I found a few past threads dealing with the problem of induction, most notably this one, where several knowledgeable Objectivists take the position that a given induction is valid if and only if one can establish a causal link between an observed phenomenon and the identities of the entities involved, and from what thinking I’ve done on the issue so far, I agree. Take the rock versus balloon example again. Say we observe 1000 times that the rock falls and the balloon floats. Without already having grasped that an entity’s action at a given moment is caused by its identity, how can we ever be sure that on the 1001st time the rock will not fly off into space and the balloon will not shoot down to the ground? Or, without identifying causality as a corollary of the law of identity, how can we be sure that cause and effect applies to previously undiscovered entities, such as quantum particles?

I know that Dr. Peikoff has a second course on the topic of induction, Induction in Physics and Philosophy, where he may discuss these issues further, so if anyone has listened to this latter course, could you explain a bit further his theory and how the law of causality can be induced (or if this has been discussed before in these terms, which is probably the case, point me to some relevant past posts)?

Thanks!

2

Update: I found this short, anonymous post in the comments section of Diana Hsieh’s NoodleFood blog. Can anyone confirm/deny, and if true, could you give a brief summary of what, exactly, Peikoff said?

Comment ID: #23 (link)

Name: Axioms and Induction

In Objectivism Through Induction LP said that the law of causality was induced. (I don’t think he ever said that the axiom of identity was). He explicitly reversed this position in a Q&A to Induction In Physics and Philosophy.

3

I recall this being true. I believe (from my memory of a few months ago) that Dr. Peikoff said that one cannot really induce causality. Causality is simply identity applied to action.

4

Yes, he revises that position in Induction in Physics and Philosphy.

5

An interesting thing to consider is that one can only establish knowledge of the identities of things through causality, because all knowledge, being an aspect of consciousness depends on the perception of actions unto the senses. For example: when one sees a red ball, one is really only percieving the action of light reflecting off the surface of the ball and only those wavelengths of electromagnetic waves in the spectrum of red. When one feels a wall, one has to touch it, and have it touch you back, firing off the receptors in your fingertips. All consciousness is a result of such action. Without light one can not see. Without sound waves, one can not hear, and sound waves and light waves are a result of actions; candles burning, or friction. Causality then too is at the base of knowledge, because implicit in every identification, in every “it is this” one is saying “it is this type of entity which acts on my consciousness this way, and only this way, thus far.”

The process of induction is the process of identification, using the actions of those entities being identified. This is the only way to identify anything. Those things which do not act, can not be seen, or heard, or felt, or tasted, or smelled.

6

IAmMetaphysical: That is an interesting observation. My first inclination was to disagree with your point that causality is at the (hierarchical) base of all identification, but having thought it through a bit more I agree. The reason for my hesitance was that although it’s true that our consciousness has a certain identity such that entities must interact with it in a certain way to be perceived, I think that is a relatively much later discovery, whereas the data given to us by perception is the lowest-level foundation on which all other knowledge is based. The directly given, the starting point in knowledge, is the entities we percieve, not the light or sound waves that compose our sensations. However, I see that implicit in grasping those entities is the fact that we grasp them through some means, which, obviously, must follow causal laws. So yes, although we may not explicitly grasp that fact until much later, the validation of perception definitely depends on the law of causality.

7

Nice formulation. :stuck_out_tongue:

This is basically Kant’s problem with perception, that it must use a means.

8

The explicit, conceptual form of any axiom is derivative, a product of induction over a vast number of observations. But the implicit, perceptual form of any axiom is self-evident in every act of consciousness, including every act of induction. The explicit philosophical defense of induction requires the explicit, conceptual forms of the axioms. But the method of induction is the primary means of gaining knowledge, even before the conceptual forms of the axioms are known.

9

All the axioms of objectivism are obviously reached through induction. But this brings about the question of how we came to trust induction in the first place. By what method do we reach the conclusions, including the concept of induction and the laws of causality? Here I think the constructivists as well as the phenomenologists have some useful insights. We are thrown into the world without knowledge and without concepts and from this stream of experiences we try to solidify a base of concepts. The major mistake of the constructivists however was to assume that we cannot properly ground this stream in solid concepts.

The grounding occurs according to what I would call the axiomatic method, and is very similar to the similarly named process of mathematics. Mathematics consists of identifying certain basic premises – axioms – and then reifying these into propositions through the process of logic. What is not obvious, however, is the fact that every single new proposition in the axiomatic system is an opportunity to falsify the axioms. All that is needed to do this is to prove a statement A and ~A deduced from the same axioms. Thus, every time we deduce a new non-contradictory proposition in the axiomatic system we actually strengthen the validity of the axioms. At some point the truth of the axioms is so well-established that the occurence of a contradiction is unthinkable.

This same process is at work with the metaphysical axioms as well, even though the main process here is induction rather than deduction. Induction here is the same process of holding all propositions together forming a non-contradictory whole. Like with mathematical concepts there is a golden opportunity to bring bad axioms down by new observations, and by the axiomatic method some axioms ARE genuinely bad and will fail in the face of observations. But those axioms that survive any attempt at finding a contradiction are the ones we call self-evident truths. These are of such a nature that they 1) reveal themselves in all our observations and 2) cannot thinkably be contradicted.

At the base of the axiomatic method is indeed the principle of non-contradiction. That is, even before we conceptualize it we use coherence as an epistemological evaluation.

10

how we came to trust induction in the first place

You induced and trust in the concepts of “trust” and “induction”.

Induction is the primary method of gaining conceptual knowledge from perceptual observation. Just as the formation of a statement denying the axiom of existence implicitly upholds and assumes the axiom of existence, any question of trust in induction implicitly assumes the trustworthiness of it. You formed the concept “induction” by induction, by means of observing and integrating many examples of the operation of your own mind and that of others. Denying the trustworthiness of induction in effect denies the trustworthiness of one’s own denial. (Note that I don’t say induction is axiomatic, since it’s not a perceptually self-evident primary.)

11

I don’t agree that the axioms of Objectivism are reached through induction, and neither does Leonard Peikoff, at least in the OTI course. By my understanding, axiomatic concepts are reached through induction - i.e. we reach the concept of existence, identity, etc. by observing countless concrete instances of such things and abstracting away the specific concretes. However, the axioms themselves consist of an identification in propositional form of the facts at the root of these concepts, facts which once identified are self-evidently true. So induction is integral in the process of coming to grasp the axioms and understanding their real meaning in the world, but the validation of the axioms is their self-evidency, not the fact that we see them expressed in reality over and over again. New experiences neither strengthen the axioms nor provide an opportunity to falsify them. As Ayn Rand states in Introduction to Objectivist Epistemology, “After the first discriminated sensation (or percept), man’s subsequent knowledge adds nothing to the basic facts designated by the terms ‘existence,’ ‘identity,’ ‘consciousness’ – these facts are contained in any single state of awareness…” (p. 55 in the 1990 edition).