What is more important: simplicity or induction?

Question

An argument by analogy is an inductive argument for the existence of other minds. An argument by analogy is enough to justify the belief in the existence of other minds.

But Occam's Razor offers a simpler explanation: solipsism.

Solipsism contains the smallest number of entities: only those human bodies and objects that I see at the moment.

Solipsism does not have billions of conscious people, no atoms and subatomic particles, no galaxies, etc.

The existence of other minds also requires the existence of billions of minds, a huge number of atoms and subatomic particles, a huge number of galaxies, etc.

Science and philosophy often prefer simpler explanations.

But it is obvious that no one believes in solipsism.

How was solipsism rejected?

Is induction more important than simplicity?

Thank you.

Answer 1

Occam's razor is the central principle of induction.

Theoretically, we seek the simplest explanation - in terms of the shortest and simplest formulas - that exactly matches and predicts all of the data under consideration. This is the Minimum Description Length principle.

"Simplicity" in this context is about the formulas, not the objects mentioned in the formulas. We can describe the motion of water by postulating very large numbers of H2O molecules bouncing off each other. The formulas required to describe this situation (Navier-Stokes equations) are short and sweet, even though the number of molecules invoked by the formulas is very large. So it is a good explanation.

Solipsism is not a good explanation. It doesn't explain anything about why things are the way they are. So, there's only you and a hallucination of the world - why is the hallucination the way it is? For solipsism to fully explain all the observed data, we would need to add on to it a complete explanation of why the hallucination is the way it is. This explanation would be just as complex as the laws of physics; you'd probably just have to invoke the laws of physics anyway. So, solipsism doesn't gain anything in terms of Occam's razor. In fact it comes out behind, because to explain everything it has to include the laws of physics plus it has to include the "self" as something separate. An explanation that unifies the self with the laws of physics would have fewer parts. This would be a realist explanation.

Answer 2

We could differentiate metaphysical simplicity from explanatory simplicity.

Metaphysical simplicity is what you're talking about: the number of entities existing in reality. This has issues: shouldn't we also consider the simplicity of each of those entities, how would we measure that, how do you delineate where one entity ends and another begins (in one sense, the universe is one entity, in another sense, a single person is billions upon billions of entities that are atoms), and why is this a useful metric. This also doesn't seem to consider the chain of causation that gets you to the current state of reality, which seems fairly important.

Explanatory simplicity is arguably more useful, and that's what Occam's razor tends to relate to. This relates to the explanatory power of claims. For this:

• We reject claims that don't explain any evidence. This would include things that have no effect on our observable reality.
• Let's say claim A can sufficient* explain 10 pieces of evidence and claim B can explain 2 of those pieces. We accept claim A for all 10 pieces, rather than accepting claim A for 8 pieces and claim B for 2 pieces.
• We favour claims with fewer surprising facts. If the claim is that some specific person loves you, it would be surprising if they make you suffer.
• We favour claims with fewer "just so" facts. If one says that only one's own mind exists, floating in a void somewhere, and this entire reality is a hallucination, this raises the question of why you're experiencing this reality instead of any other. Every part of this reality is "just so".
• This framing may still need some refinement, but we favour claims with a smaller "model space". If we compare this reality objectively existing versus us living in a simulation, one might say that the "metaphysical space" (the scope of what exists in reality) is similar. But the model space (the scope of models that exist as part of that claim) is much larger for a simulation, because in both cases you'd have a model that is this reality, but for a simulation you'd also have a model that is the "true" reality.
• Related to the points above, we'd reject claims that are interchangeable with other claims without affecting explanatory power much. One might posit that you're a mind floating in a void, or one might posit that you're in a simulation, or one might posit that you're in a simulation inside another simulation, and maybe the simulation was made by a dude named Steve, or maybe his name is Albert. Many of those explain the evidence similarly well across a number of factors, so it would be hard to pick one above the other. But we can also pick none of them, and just say that observed reality is objective reality.

There is a tradeoff between these factors: you could say that hallucination explains all testimony and therefore you don't need any other explanations for testimony, but that may create more "just so" facts, where you may not be able to explain why those specific hallucinations happened.

Explanatory simplicity seems to do a good job of leading to reliable beliefs, and we evaluate reliability through correct predictions and coherence.

Solipsism doesn't measure up well by this standard, so we'd reject it.

---

The argument from analogy for the existence of other minds can form part of the explanatory power of that claim.

Answer 3

Descartes and Solipsism

Solipsism follows inexorably from Descartes. A simple reductio ad absurdum (or contrapositive or modus tollens) suffices to get us out of that mess. i.e.
If
   Cartesianism ⇒ Solipsism
And if we don't like the conclusion, then
   ¬Cartesianism
is a natural conclusion.

Here is another less standard approach to reasonablized Descartes.

Occam and Simplicity

As for Occam and simplicity there is a spectrum of concerns
Description ↔ Narration ↔ Prediction ↔ Control.

Solipsism does splendidly on the left and becomes more and more pitifully inadequate as we move rightwards, ie. solipsism is a maximally parsimonious description of the phenomenal putative reality that presents itself to us. But is useless at prediction and control.

So for organizing one's own spiritual outlook on meditation, life, love and death solipsism is fine and good.

If you have an illness and need a doctor it's not!!

Idealism

On a more philosophical note if you recognize both that

• Solipsism needs to be rejected
• But so does crude materialism/naive naturalism

... then you may want to look into idealism.

Note: The word "idealism" in philosophy and in common usage have very little in common.

Observe how

• Original Idealism of Plato
• Subjective/Empirical Idealism of Berkeley
• Objective/Absolute Idealism of Hegel, Schelling et al
• Transcendental/Critical Idealism of Kant

it has remained a high point of (western) philosophy.

Today western philosophy has lost its way (in this space at least).
Vedanta, Taoism, (Zen) Buddhism are taking its place.

But there are interesting current western philosophers at the fringes eg. Analytic Idealism of Bernardo Kastrup.

Induction and Solipsism

Science is based on certain metaphysical unfalsifiable assumptions. Here is a typical listing of these assumptions. A meta-assumption behind all these is that the objective world — which includes other people — exists.

Induction is downstream from these assumptions.
Solipsism is an upstream alternative.

So if you are exploring the possible truth of solipsism you cannot also assume science principles which includes induction.

[BTW: You also cannot assume that this site and its respondents — including this one — exist observer-independently 😁]