Like many submissions to the Cluelessness Competition, this essay wasn't featured in the entries list for some criteria reason, so I am asking for potential feedback here. I think a summary of this post is sufficient (paste into Claude), because it is long, and you might expect the argument I am trying to make.
Critiquing premise three (from the summary post):
P3 from the announcement post:
--
Warning: Contains references to self-ending.
--
I want to position this challenge by posing a solipsistic agent. When I reference “agent” (as a fluid “her” or “them” or “him”) in this critique, I am talking about all reality, or the only universe known to exist (the agent’s reality). Solipsism is the idea that one’s mind is the only thing known to exist. This definition comes short because it assumes there is “one” in comparison to "others," and not just the totality of existence, which I will refer to as the agent’s reality.
There is (1) epistemic solipsism and (2) metaphysical solipsism. For our purposes what exists is all that is possible to know. The key move here is that when the agent interprets cause X or acts on cause X, they are acting on all available knowledge. I will explore the upshots of this and clarify here the implications of holding this view for cluelessness.
Beginning scramble
There is something here in this critique. I know there is. Let’s flesh it out:
Someone can be solipsistic and disregard others — they are more likely to be depressed psychologically as a general psychological statement, and reality here will end as far as we are concerned if this person self-ends by making choices without a moral framework toward others, or without meaning. Doing good makes us feel good and thus supplies the reality that we need to maintain our existence. The point of acting with "good" in mind heuristically, or with EVs, means that we are sufficiently interested in acting to return our reality and so doing good is a reality-supplying thing. Reality then is all of the causes we can make out.
Cluelessness depends on a hypothetical world where our actions influence a web of causes that could "wash out" as bad. Solipsism counters in that these causes aren't sure to exist for the agent, so we cannot even propose a hypothetical that the long-term future could wash-out (or be bad) because these causes are not sure to be in existence until at one point in time they are encountered. When it is encountered, it exists. All we have is what supplies our reality, and if doing good more readily supplies it, then the causes and goodness that we affect are genuinely good in the “known” (tautology) universe.
As far as I am concerned, one issue could be that believing “one” is the only reality to exist can also be psychologically damaging; however it is less so, I suppose, if we commit to doing good in the world. Similarly, believing this is a necessary truth, since there is no proof that other minds exist. It is also a necessary truth that we, to some extent, must believe or act in a way in which there are other minds.
But I don’t need to prove these things or iron out solipsism insofar as doing good is possible (whether through EV or heuristics) for the agent’s causal chain (reality). P3 roughly argues that if the universe can wash out any attempt at doing good, we are unaware of what causes, actions, or effects are good/bad. The universe, however, is the totality of the causes our mind “encounters,” so the solipsistic agent will know by definition what is good/bad.
So, we can propose hypothetically that a cause can come in at some point in time and wash out the effects of a good action. While the agent can be a hypothetical reasoner (e.g., I donated to a the Malaria Consortium and I feel good, but I can suppose that this action will be washed out in the long term by another cause after I am gone), this hypothetical is not what they can observe but only what they can hypothetically conjure. This is a level of unreality that we cannot see and supplies the actions we could take. Importantly, this is unreality even while the agent is alive — if it is not perceived, it does not exist. Reasoning exists in the domain of the hypothetical (e.g., counterfactual worlds where I do X instead of Y), so it seems that we have a problem, especially when we can hypothetically imagine a universe can be bad regardless of what we do. It is important to note that this universe is hypothetical and exist in belief, but let’s tread lightly:
Reasoning is an action in and of itself, given we have limited time as an agent. It is a fact that the agent cannot perceive all (hypothetical and theoretical) causes and act on them. It is also a fact that all causes the agent perceives = reality. It is also the case that all the hypothesizing is bound in time and is both (1) an action of the agent itself and (2) beliefs that the agent chooses to endorse. It is also important to note that theorizing (via known facts, predicts action) is distinct from hypothesizing (for our purposes: Proposing a reality in a way that suggests guessing).
The agent has limited time to weigh moral reasons, others’ perspectives (since to be an agent and to do good, even in a universe where the agent is the only one to exist, we must accept others as entities we can perceive), so the causes that they examine as being potentially wash-outable with indeterminism exists as time-dependent reasoning to inform time-dependent actions. This is the ultimate triage. The unreality that the agent sets via hypotheticals exists in the unreal (to be or to not be endorsed via belief). The action that is finally carried through might be a more real reality, but the agent is always in triage with the ability to act on the hypothetical reasons they encounter with their limited time. Of course, the agent’s time and space form the whole of the universe, known reality, and the causal field. I will iron this out several times in this essay, and hopefully it will be clear why this matters.
So we can say that in solipsism: Thoughts of an agent = moral reasoning in time, actions themselves, but also more unreal compared to causes perceived and (moral) actions taken. Thoughts of an agent are actions, too, but are more unreal because it supposes theoretical worlds or hypothetical guesses that inform the eventual actions the agent will take in, for example, moral calculations.
An example here would be helpful: The solipsistic agent considers hypothetical paths A and B. She acts on path A after a certain amount of time weighing all the causes she can theoretically conjure. She can always assume (1) that if she spends more time theorizing, she may clarify her plan of action, (2) spending more time yields decreased returns after a certain point, (3) one point in time to act might be counterfactually (in the hypothetical space) better than another point in time, (4) she can always hypothetically conjure that causes are always able to be washed out, and therefore (5) she can hypothetically conjure that unawareness persists. She concludes that this hypothetical universe extends beyond what can happen in theory (the totality of facts, causes, reasons and predictable effects she has to work with). She does this because she has no evidence for a hypothetical, infinite universe besides what she can causally see and theoretically predict based on these causes. She chooses to go with path A after organizing all theoretical pro-causes and con-causes, assuming she is a thoughtful agent. She uses both EV and heuristics to build her claim and act on path A after reasoning (an action itself) at an amount she believes is necessary. She sees the effects play out well as evidence of a good action.
We can critique her decision making but as long as I have established that cluelessness cannot hurt provable reality (solipsism), I win.
Clarifying the view
That was a lot to track. Let’s see if I can clarify:
The causes that the agent perceives are the only things in the "theoretical domain” to exist, since the agent supplies reality. Theorizing before actions (e.g., we have facts XYZ which inform that I should do path A over path B) and hypothesizing what will happen to this action (guesses, more with an essence of unreality) are all actions that the agent chooses to make in a universal triage of time.
Do we have reasons to reject solipsism? I do not think we can. Proving that others exist, or that a tree falls when the agent is not watching, is impossible. If we are to reject something like cluelessness, I believe we must strip down all assumptions and engage with the most basic first principles possible.
Believing that others exist is merely a belief that was somehow made in a causal chain at some point in time, and the agent (or reality, all of existence) stuck with it. This is probably because he has no choice, since others are indispensable to our living, and form the feedback of morality that ensures we do not self-end, or end reality. But the agent does have a choice to believe the universe is an infinite mess in which acting in it (which the agent is, again, forced to do) will lead to indeterminacy. Believing that hypothetically the good things that we choose to do will fail in an infinite universe relies on things the agent (1) cannot know and must propose on behalf of the agent’s time/causality (or the universe), which is no evidence for this, (2) will more likely lead to self-ending since the belief is depressing, and (3) denies the present and unavoidable results that you see from your actions (ideally good ones).
This is not wagering because we can conclude that this hypothetical belief (in unawareness, an infinite universe of washing out) is false based on the evidence (solipsism). Nevertheless, it is a choice here to believe in this falsity. I will cover this further on.
Additionally, what we “cannot know” is only a hypothetical belief that the agent chooses to endorse or spend time on as an action in reality. So saying that we cannot know that our action is good or bad, if the agent sees the action as good, is a psychologically-damaging and self-denying belief, which may end reality if the agent is that much harmed. (Ending reality is bad and the agent exists to avoid this). This is possibly stepping into a critique to the normative premise tied with the third: The agent does what is good, which supplies reality, and so every action that the agent takes is ideally done with faith that they are either heuristically doing good or using EV to calculate the good for (supposedly) more important or isolated actions. Either way, a hypothetical-posing of an unawareness is a belief that the agent has no evidence to believe unless they encounter causes that serve as evidence for wash out.
And even if this were to be the case, the (near-infinite) amount of actions that the agent has taken in reality have been oriented toward the good as a self-sustaining process (agent is good-striving, psychologically). The very nature of reality is oriented toward a continuous good based on what it means to sustain this reality.
So you are admitting hypothetically the agent can supply credence to the idea that there are wash-out causes?
Yes. The agent can believe whatever they want to. That does not mean it is true, or justified.
If the agent doesn’t see her causes wash out, then she can assume that her action had the attended (good) effect. The agent makes many actions and perhaps a near-infinite amount of decisions in reality. If this is an agent with good faith, the reality she builds and actions she chooses to take will be net-good despite any hypothetical-beliefs. This net-good also depends on the necessary-beliefs that the agent holds, like the belief that other people form the basis of acting morally, and the source of reality is fulfilling “good actions” toward others. This is achieved, then, with a knowing of which actions, causes, and effects bring about goodness. The agent is in full knowing as the crux of reality.
The agent can confidently discount hypothetical unawareness scenarios where the universe flushes any attempt at doing good. That does not mean that she must not believe this, even though she probably should (and should just focus on the theoretical-belief realm: Given I know ABC for path E and XYZ for path D, I will make a maximal-bracketing decision).
Infinity and finite time is confusing
We can suppose that the agent does near-infinite actions, since the totality of existence does not suggest that things are finite (all actions = all thoughts = all of reality). But I would be hesitant to call the totality of existence “infinite” since there is limited time and limited causes to perceive in time, not because the universe is necessarily infinite.
Further, the only way we can assume that the agent exists in triage, in limited time, is to compare what could happen if the agent chose one action over another, suggesting counterfactuals.
The agent chose path A after t amount of reasoning which precluded her from doing path B since path B could have only been done in that point in time, too. The agent is thus time-dependent since limited causes can be perceived, although this is misleading since these limited causes are all what reality is, so the agent, actually, is not limited unless we again believe in the hypothetical unawareness. The agent is limited in triage only by comparison to what the agent could have done differently, e.g., acting on path B.
How is morality possible?
Solipsism is usually defined as bad because it assumes that “we” are aiding our imagination. But solipsism is the truth and the only known reality to exist. Morality is the term, perhaps, that ensures that reality is supplied; when “we” act and do good in the world, reality persists, and if “we” do not, “we” defeat ourselves. The agent defeats herself.
The agent is the sole anchor of reality. Causing good within the only world that exists is, by definition, objective, genuine good (considering the agent avoids things like circular preferences or money pumps, which raises interesting questions about an agent going back on themselves in time; but this is out of the scope of this essay). It cannot be "fake" or "meaningless" because there is no alternative, higher reality to measure it against.
The agent can act with impartial-altruism in this way. Sure, it may be egoism. But as long as the reality that persists (premise three) is possible not in a (choice of) belief in an infinite universe with unawareness but in the agent’s reality, causes that they can perceive, we can compare actions and a genuine consequentialism.
Isn’t this just wagering?
A construction: “we can't know whether the optimistic or pessimistic scenario about the future is true, so we should choose to believe/act on the optimistic one because it's more livable, prevents psychological collapse, and honors what we can presently observe.”
First, strip out the “we.” I am going to drive this down as a redundancy: The agent can believe in a hypothetical (infinite universe’s unawareness) as a choice, but the agent is forced to believe and act in a way that supplies their own reality for certain, necessary facts. Doing good and being oriented to the good is what is in the nature of the agent, so it is not wagering, but supplying reality. Making heuristics or EV calculations falls squarely here because the agent is weighing causes against one another in this reality, since again, the only causes the agent can perceive are the only reality that necessarily exists. It’s important to note that the agent is perhaps not “believing” in reality, though, since beliefs (in hypotheticals) are not forced. Believing in solipsism is forced so it really isn’t a belief that may imply choice.
Beliefs can either work for this reality or work against it.
Can the agent be bad?
The real risk that I run here is that I am writing on behalf of myself in a meta reflection. I believe that what I am doing is good and therefore I am still alive, or at least, “my” reality is still supplied. The agent will be drawn to goodness and meaning. Since what constitutes the agent is good and meaning-driven, the agent cannot be “wholly bad,” because this would result in a self-defeating appraisal of the agent’s own self which would have more tendency to end the reality for the agent (e.g, depression). Ending the reality for the agent is ending the universe and all causality.
So, the agent can definitionally not assume that they are bad, or this would end reality. I am unsure that answering this question, though, would preclude the solipsistic system as providing grounding (reality) to make, for example, heuristics or EV calculations.
Moral luck
This idea came up when I considered: Can we suppose there are degrees of good? Sure: As long as the agent picks up the causes in the theoretical space. Acting on path A vs. path B and weighing the two would assume that doing so causes more degrees of good.
Does the agent have to get lucky to perceive and act on the space of high-value causes? Moral luck relies on (necessarily) believing in others’ existence and can judge consequences, and it assumes that there are other ways to make actions (counterfactuals) which would only exist in some way if the agent conjured this. For example: I was lucky that other “people” found my action to be good, since luckily I was able to find the set of causes that led to this and thus improved the good in the universe (agent’s reality).
Again, this assumes the hypothetical that there is unawareness out there in an infinite universe. So getting “lucky,” then, is an illusion that is self-denying, since the causes that the agent “intercepts” is reality itself. There is not a pool of causes “other there” to “intercept,” but the totality of causes that the agent can see as being a product of all of reality.
The agent notices a wash out. What happens?
This must be contrasted with the near-infinite amount of actions the agent makes in reality. It might be important to note that reality begins and ends, it is a totality. So if the agent finds evidence in causality that an action = wash out, it is important to note that none of this is made in isolation, distinct from the general orientation of the agent as one that does good, that supplies reality. The agent is doing good as part of reaffirming existence; so if one action washes out, others are compensating.
Of course, this is not a rebuttal that impartial-altruistic actions are impossible and unrealizable. The agent has seen an action as being bad, and is noticing how future actions can be improved. This is not unawareness.
Further, it is merely a picture of an unfortunate circumstance. The solution to this is not reinstating cluelessness but the agent choosing to “do good better” in this isolated calculus, finding what went wrong, or doing better heuristics/EVs in the future.
Again, what supplies reality is a tether to the good; so the agent has to be acting here with good faith, and can admit mistakes, but not in a way that would self-defeat or self-end, because that would go against the nature of the agent’s reality. The agent makes good and meaningful decisions all the time by virtue of being the agent that supplies reality.
So you can ‘find out’ that your action is bad?
Yes. That is a cause not coming into view, but one isolated causal action in the totality of the agent’s reality that the agent has focused on as being bad or not intended in their consequence.
The agent could find out that their action was bad, years in the future. It is not finding out as in a cause coming into a “limited view.” (I am not sure we can distinguish the agent as a “view”). But for our purposes, if the agent can find out an action is bad, he can find out that impartial-altruistic actions are good and possible, that evidence exists to see these take hold, and the agent’s evidence is all that there is. This doesn’t preclude someone from ranking, seeing evidence for a good action, rather, it just means that something turned out bad, in which the other, vast majority of near-infinite actions that the agent takes over the course of reality are good as a matter of self-sustaining, especially if they were done methodically. The agent is not perfect, though, and seeing evidence for the vast majority of good actions that the agent fulfills as a part of their numerous (heuristic or EV) moral calculations fits into the ultimate, self-sustaining essence of reality.
Merely that badness is possible does not shape unawareness.
Finally, if the agent was acting in good faith, and if the agent and his projects turned out to be very bad, this would certainly be psychologically damaging but we must assume that this wouldn’t end reality. The agent uses the evidence available in reality to make these decisions, reinforcing and learning what is good. This is because every action that the agent makes is oriented to the good, so there will be naturally (nature of the agent/reality) more affected good than bad, which is not unawareness.
Perhaps there are causes out there waiting to be perceived
If my claim is to say that the bad effect was real all along, that the agent just didn't perceive it yet, then I have abandoned the "only perceived effects exist" claim, because now unperceived effects clearly did exist and were doing real causal work the whole time.
Key word: Now. These causes only exist at the moment the agent perceives them.
A good faith agent will search for evidence of their actions because to do so they would be affirming their tether to the truth which is psychologically reaffirming. Again, when the agent pushes themselves in good faith to see the evidences in their actions, they are not pushing “into” untapped causality, or an infinite, messy universe. “Good faith” can mean the agent’s orientation to the world is the mechanism for sustaining reality, so far as this reality doesn’t end. The agent sources meaning from doing good and believing the good is sourced from duty toward others.
Circular reasoning?
I can see circular reasoning probably being a big issue. However, I cannot come to see in what way this can best be formulated.
Why can we maximize actions and weigh consequences? Because all the evidence that the agent perceives can be weighed by the agent in a theoretical space. What about evidence that the agent does not perceive? That does not strictly exist because the agent is the only “being” necessarily to exist, the causes that the agent sees = all reality. But I thought other beings necessarily exist? Other beings also necessarily exist in the agent’s reality because we are forced to interact with them and ground our moral weights with them.
Ending reality is bad?
Perhaps a thorn in this work is that ending reality might not even be a bad thing. However, the agent wants to exist and this sustains reality. The agent does not want to self-end.
Conclusion
We are not unaware and we can weigh the causes. We can do decision theory, maximality, etc.
For example, the agent can do maximality simply because the causes that they perceive are limited, and the agent can rest easy with this since they are the limit themselves, the reality that the agent perceives is all that is to exist.
If I save fifty people from a fire or prevent a million shrimp from dying a painful death, and these people (or shrimp) do not become dictators and they and their loved ones thank me for having the joy to live another day for their rest of their lives, I have all the evidence I need that what I did was good, that what “the agent” did in reality was good, compared to not doing these things. We can draw out theoretical charts and comparisons between saving these people and shrimp, or not saving these people or shrimp. This evidence is direct while the hypothetical (that they are net-bad by virtue of living, by a hypothetical scenario that the agent can choose to endorse) relies on grand formulations, assumptions, and unreality.
The agent must pass up the hypothetical, harmful belief that saving these people and shrimp will then be net-negative because (1) this belief is a choice that the agent chooses to endorse and (2) you can weigh the theoretical causes and effects in this action and you can still come to the conclusion that what you did was good given all that “you” know, which is all of reality.
For these reasons premise three fails, and thus unawareness falls short of describing reality.
--
(Image by Yoga Digest, butterfly effect)
