Notes from the Wired

Killing Dream People!

December 17, 2024 | 5,782 words | 28min read

Dreams, Morality, and the Value of Life

A while ago, I had a rather vivid dream, but I will spare you the details. What was important about this dream was the presence of a dying girl. Who she was isn’t crucial; what’s important is the dying part.

This dream led me to wonder: What is the morally relevant factor that allows for the killing of dream people, but forbids the killing of “real” people? In other words, how can I define the value of human life in such a way that it includes real people but excludes dream people? And if I can’t, shouldn’t we give dream people the same moral consideration as real people? And what would this entail?

If you’ve read Peter Singer’s paper “All Animals Are Equal,” you may find certain parallels. Just as he tries to include animals in moral consideration, I am attempting to include dream people.

The underlying question of this article is a value question: What do we value? Most people would likely answer this question with some variation of “Human life” or “Persons.” But this leads us to another question: What is a “Human Life” or a “Person”?

I have already tried to address this question in a previous article, where I proposed a definition: “We say something is someone if and only if it has the capability to deploy consciousness,” and that something deserves moral consideration (i.e., should not be killed) if and only if it is someone. In this article, however, I will attempt to find a broader, morally acceptable principle that differentiates dream people from “real” people.

Before we proceed, I want to clarify that when I refer to “personhood” in this article, I mean it as a shorthand for “something deserving moral consideration,” but I do not intend to imply any additional connotations that come with the word “personhood.”

Another point that needs clarification is what I mean when I say “moral consideration.” Imagine a device capable of extracting dreams and displaying them in some kind of showbox for people to watch and enjoy. Do we have a moral obligation toward the people in these dreams? For example, if the same device can also manipulate dreams, would it be morally justified to change the dream into a hellscape where the dream people eternally suffer, just for my amusement? Or am I allowed to destroy their “property”? Kill them?

If the people in the dream have intrinsic value, then the answer is obviously no. But do they have intrinsic value?

If a being deserves moral consideration, it means that when we, as agents, perform actions in the world (or dream world), we must first consider the moral implications for that being. For example, a stone does not have moral consideration, so the question “Is it morally wrong to destroy a stone?” doesn’t make sense. We could still imagine that the stone is owned by someone who would mind its destruction, but importantly, the stone itself doesn’t have moral consideration—its owner does, and by extension, we need to consider their property. On the other hand, when we ask, “Is it morally wrong to kill this human?” the question makes sense because we do consider humans when making moral decisions—in other words, they have moral consideration.

It’s important to note that even if dream people do deserve moral consideration, there might still be arguments as to why certain actions (like those mentioned above) are morally permissible. For example, one could argue that dream people have different interests from us, and these interests may not include their continued existence or “suffering.”

You might think that all of this doesn’t matter since the girl in my dream was just a fiction of my imagination. To that, I have several rebuttals, which you will see later. But to begin with: Who’s to say that you are not a fiction in the dream landscape called the Universe, dreamed by a greater entity, more commonly referred to as God?

Personhood Defined as Human Life

One of the first arguments against dream people having inherent value is that they are not living people, i.e., they do not have human life, and thus are not worth protecting. But as I alluded to earlier, I do not think this is a viable definition for personhood. Instead, I believe we ought to protect the “continued conscious experience” of a subject.

But why should we value consciousness over life in general? When debating a definition, one aims to find something that includes all possible instances of the thing and excludes all non-instances. So, if I want to challenge the definition, I can either show that there is something we would call a person that is not included in the definition, or that the definition includes something that should not be a person.

What do I mean by human life? I mean the biological processes that occur in living humans. I do not want to specify this further, since the question of what exactly constitutes “human life” may be complicated and is not the focus of this article.

With this definition of personhood, it would exclude:

And it would include:

One might disagree about whether all of the listed examples should or should not deserve moral consideration (or even argue about whether they should be on the list at all). However, remember that showing even one instance of something that should be considered a person but is not included in the definition—or that is included but should not be—is enough to show that the definition is insufficient. In so far as no one is willing to “bite the bullet” on all of these points, human life as an indicator of personhood is not a good definition.

To highlight the importance of a correct definition of personhood, I offer the following motivational thought experiment:

Imagine we are far years in the future. Mankind has made tremendous advances in both science and technology, and has never stopped researching improvements to computers and understanding brain physiology. Sure, during this time, progress stagnated or slowed down at times, but research continued, making incremental advancements. One day, through advanced surveillance technology, we successfully map an entire human brain, capturing every neuron and synapse. Not only that, but we also figure out the exact principles of how and when neurons fire. Equipped with that knowledge, researchers run simulations on supercomputers using the brain map.

Before long, these simulated brains are used everywhere as a cheap way to automate tasks. But because society has defined personhood as human life, and simulated brains are not included in that definition, these simulated brains can be kept and employed in slave-like conditions, working 24/7 without rest, having no rights to refuse requests or stop working.

One might disagree that consciousness is a purely materialistic and deterministic process or argue that humans would not survive long enough to reach such a position. However, if this scenario were possible and we defined personhood incorrectly, we could find ourselves confronted with a malum morale on a scale as great as, or greater than, the transatlantic slave trade.

Does the Length of Life Determine Moral Worth?

One claim I’ve heard that differentiates dream people from “real” people is time. A friend of mine put it like this:

“They only exist for the few minutes of the dream, and after that, they are gone.”

Another user I asked on a message board replied:

“[…] Their lifespan is only until you wake up. But if they’re with you when you’re awake, you can take their claims seriously.”

Is this really true? Is time a relevant moral factor that differentiates dream people from “real” people?

Imagine a woman, with no living family or friends, gives birth and dies during childbirth. The baby is diagnosed with an incurable disease that will cause it to die in a few days. Assuming the baby will indeed die and has no one to care for it, is it wrong for me to do as I please with the baby, simply because of its limited lifespan?

Most people would agree that the answer is obviously no. But if that’s the case, then “time” or “limited lifespan” can’t really be the differentiation factor.

We can also imagine a thought experiment in the opposite direction. Suppose we put a person into a state of deep sleep and administer a medication that prolongs their dream (such a medication might be developed in the future—it’s not completely absurd, considering that today, we have medicines like some antidepressants that can intensify dreams). We also take care of all their dietary needs. In this scenario, one person might experience a dream that lasts for 50 years or more.

Would the lifespan of the character in the dream, in this case, be extended? If so, would that suddenly make them worthy of moral consideration?

I would argue that most people would say that just because the character in your dream exists for a longer period does not mean we should suddenly start treating them like a person.

Furthermore, think about the implications it would have for moral consideration if it were based on lifespan. Would that mean that rich people, who through their wealth have access to better healthcare and thus a longer life expectancy, would be more valuable than poorer people who can’t afford the same healthcare?

The Role of the Body in Defining Personhood

Another possible distinction we could make is that dream people do not have their own bodies; they are instead “guests” in a house called the brain or consciousness.

However, this argument doesn’t hold up to scrutiny any more than the previous one.

First of all, the concept of an “own body” is not clearly defined. Where does one’s body end and another’s begin? Are the bacteria inside me part of my body, or not? If they are part of my body, what if I stand inside the mouth of a blue whale—am I then part of its body? What if I remove my arm and replace it with a mechanical one? What if I transplant my brain into someone else’s body? There are many such questions, but they highlight a common problem with many philosophical definitions. I will set this issue aside for now. (Though it’s worth noting that having precise definitions is one of the most important aspects of science, which works through the falsifiability principle. After all, how can you falsify something that is imprecise? Or as Wittgenstein would say, “Any philosophical debate is just semantics.”)

Similar to the “human life” argument, we can imagine scenarios where we would say a person deserves moral consideration even though they don’t have their own body. For example, imagine in the future, we are able to upload minds into computers, creating digital minds. Would we now say that just because we changed the outward form of a person, they are no longer a person? Would it be morally permissible to torture or kill these minds?

Again, most people would say no—just because we change the body of a person from a biological one to a mechanical or digital one, they do not lose their rights or personhood.

Creation and Moral Rights

One possible argument is that a dream person is created by something else and is therefore artificial, in contrast to a human.

But this is a tenuous argument at best, and simply wrong at worst. Are humans any less artificially created than dream people? And what does “artificial” even mean?

There are many parallels between them:

Religious people claim that humans were created by God. How is this any different from a dream person being dreamed up by a real person? Is being created by God less artificial than being created through evolution?

If, in the future, we train a neural network on a large amount of data and it gains consciousness (however we may measure it), does it deserve any fewer rights because it was created by people rather than parents, and through gradient descent instead of evolution?

If, through random quantum fluctuation, a fully-formed human were to appear in front of you, indistinguishable from a “normal” human in terms of external appearance, memories, and full consciousness, would it have less value just because of how it came to exist?

I would answer, “No!”, to all of these questions. How consciousness or human life was created, or by whom, does not matter in the consideration of whether they should have moral rights.

Location and Moral Rights

We could argue, what separates “real” people from dream people is the location or place they inhabit. This is true in a sense: dream people live in dreams, while “real” people live in the “real” physical world. But simply pointing to a difference in location is not enough to establish a moral distinction. The distinction needs to be morally relevant, and it clearly is not.

Do we say it’s morally less wrong to kill someone from America than someone from Asia? Is it morally permissible to kill someone on the International Space Station simply because they are in space and we are not?

Obviously, the answer is no. So why should it be morally permissible to kill a dream person just because they exist in a different place (the dream) than we do? This is similar to the drowning child thought experiment, where the moral consideration does not change based on the location of the child.

A proponent of the location argument might try to modify it by suggesting that they are not referring to a different physical location between dream and “real” people, but rather to different “dimensions” or “realms” (whatever that might mean).

We can again refer to the example of uploaded digital consciousness or simulated human consciousness. Don’t they exist in a different dimension? The medium in which they exist is digital and electronic, compared to our biological medium. One might counter by saying that they both still exist in the physical world and as such do not count.

But do we really think that if we discovered another plane of existence, populated by conscious agents, it would be morally acceptable to kill them? People who consume the drug DMT often claim to encounter mechanical elves from different dimensions, and they certainly don’t think it’s morally permissible to harm them. Religious believers of various faiths also posit conscious agents that exist beyond our physical world. As long as they are not evil (e.g., demons), most would argue that this does not give anyone the right to kill them.

In summary, we can conclude that personhood and moral consideration do not depend on location—whether in a physical or “interdimensional” sense.

The Experience of Pain in Dreams and it’s Conenction to Morality

To experience pain, a sensation must first be detected by sensory neurons through a chemical process. These signals are then translated into electrical impulses and transmitted through the spinal cord to the brain, where they are interpreted as the feeling we call “pain.”

When a dream person is physically harmed, none of these biological mechanisms exist. For example, the brain does not have the necessary sensory neurons to detect pain. Importantly, if a dream person is harmed in a dream, you do not experience the pain yourself.

So, how can a being that has no capacity for pain be considered worthy of moral consideration? In Peter Singer’s paper, All Animals Are Equal, he argues that the capacity to suffer is a requirement for moral consideration.

First, the feeling of pain does not solely depend on special sensory neurons; it can also be triggered by electrical signals entirely within the brain (e.g., phantom pains, migraines). Second, just because a being does not experience the phenomenon of pain does not mean it cannot experience suffering. For example, a simulated human consciousness might not feel pain but could still experience emotional suffering, such as the loss of a loved one.

Lastly, it is possible that the dream person is experiencing pain, but that pain remains confined to the brain region responsible for their consciousness, rather than being transmitted to your entire brain. In other words, the pain of the dream person could be isolated from you, without you even being aware of it.

Are Dream People Conscious?

With the superficial arguments out of the way, we can now focus on the core of the argument.

Earlier, when I discussed “human life,” I implicitly made the assumption that “dream people” are conscious. This is important because, if dream people aren’t conscious and we value consciousness, then the argument loses its relevance.

So, the central question is: Are dream people conscious? Or should we act as though they are conscious?

To begin, let me offer a motivational account of why it makes sense, at least initially, to consider that dream people may indeed possess consciousness.

Motivation: Biological Imperative to simulate Conciousness

In every living biological organism, there is a feedback loop between it and the world. It emits actions that change the environment and receives information through its senses. There are many possibilities for how the organism can engage with this feedback loop, but most ways are not useful. Nevertheless, organisms have developed strategies, though they vary in both sophistication and timescale.

Active inference proposes a model that unifies all these different strategies into one. Its basic idea is that all living organisms minimize the surprise of their sensory observations. When speaking about surprise, it is used as a technical term, measuring how much an agent’s current sensory observation differs from its preferred sensory observation (e.g., a fish prefers to be in water).

In this model, one can see the brain as a “predictive machine,” where perception is a combination of prior information (current information in the brain) about what most likely caused the sensation and the sensory stimuli itself. As such, perception is a constructive process, where sensory information is used to confirm or disconfirm hypotheses, thus minimizing surprise.

Thus, this process needs some generative/forward model, which generates predictions based on its observations, called a prior.

For example, let’s assume that I currently think, with 90% probability, that the object in front of me is a frog and 10% probability that it is an apple. I now observe this object with my senses and realize that it was not a frog, but an apple instead. I update the probabilities in my head, so that the next time I see an object, I’ll think with higher likelihood that the object is an apple than before. In the same way, actions, like perception, can also be used to minimize surprise. Instead of updating beliefs, one can change the world—by exchanging the apple for a frog, for instance, we bring the perceived world in harmony with the predicted world.

The brain does this surprise minimization with all the objects that exist in the world; therefore, it tries to have the best possible world model, a generative model that attempts to predict the world.

One of the most difficult things to predict in the world is human behavior. It is of critical importance to understand when a person is happy, when they are sad, and so on. Because if you know a person is angry, you can infer they might attack you, thus minimizing surprise. The human mind does this through the theory of mind, in which humans have developed instinctively formed models to recognize how someone currently feels, helping to better predict the world and the people in it by reading microexpressions on people’s faces and other stimuli.

But reading facial expressions is not enough to predict human behavior in its entirety. To know if a person will attack me or if I can trust them, I need to model their motivations, their character—e.g., do they have a disposition to lying? The more information and characteristics I model of other humans, the easier it becomes to predict their behavior, thus minimizing surprise.

With all this said, if we now imagine what would be the best possible model for predicting future human behavior, what would it be?

The answer is that the best possible model is a full simulation of the consciousness of a human being.

So when we ask the question: Are dream people conscious? We should keep in mind that there is a biological and evolutionary incentive for the dream person to be conscious. Namely, when the brain uses dreams to process past interactions with other people, what better way to learn how one should behave in the future than by interacting with a conscious being?

I want to caveat this by saying that active inference is a model, so it doesn’t claim to know the truth about how things are, but rather serves as a good approximation. Future evidence may show that minimizing surprise is not, in fact, what is happening in organisms.

Also noteworthy is that this assumes it is even possible to simulate human consciousness in the first place. One should also be mindful that evolutionary arguments can help motivate an argument, but they shouldn’t be used as the main thrust of the argument (e.g., see why evolutionary psychology is so controversial).

Consciousness in Consciousness: Information Flow and Ownership

If we assume that the dream person has an independent consciousness, then there exists an information flow between this subconsciousness and the “main” consciousness. This is rather obvious, as dreams and the people in them are not random but are highly influenced by our lived experiences.

But is this separation morally relevant?

Let’s assume we keep developing and improving brain-machine interfaces (BMIs) like Neuralink, and in the future, every person has one, working just like a USB-C port. We can connect two ports and exchange messages and other information. If I were to connect myself to you via this interface, would this grant me the moral right to kill you? Clearly not. As such, we can’t say that just because there is an information flow between two entities, this somehow makes the other entity not worthy of moral consideration.

We could argue, though, that this information flow is so deeply intertwined between us that I have a certain part of “ownership” over the subconsciousness, and as such, it grants me certain rights over it. This may be true, but even then, one still needs to at least morally consider the other entity. For example, during slavery, even though one person may have been owned by another, it did not mean the slave lost all rights. There were still certain principles the slave owner had to adhere to. Thus, even though the slave might have been owned by someone else, they did not lose moral consideration.

Another example might be Siamese twins, a condition where the bodies of two twins are fused together, and thus one body houses two consciousnesses. Again, just because two consciousnesses are tightly integrated with each other does not mean we have the right to kill the other consciousness.

I want to remind the reader that I am not arguing that dream people deserve certain rights, e.g., not to be killed or harmed. I am simply arguing that one should consider the impact one’s actions have on them, as we do with other humans.

The Resolution Argument

The counterargument I would regard as the strongest against the consciousness of dream people is what I dub the “resolution argument.” It argues that the “resolution” of a dream is significantly lower than that of reality, just as one might reduce resolution in a video game. If I walk around in reality and talk to another person, I can clearly see how they look, what clothes they wear, what their hair color is, I can hear the cadence of their voice, see special mannerisms, and much more. If I compare this to a dream, everything becomes much more muddied. If I wake up, I might remember that I talked to someone and maybe even some rough outlines, but far fewer details than when I talk to real people.

Now, if the resolution of dreams is really so much lower than in reality, why would we assume that the same does not apply to the people in the dream? And since the resolution of dream people is so much lower, why would we assume that they have consciousness, the argument goes.

There are multiple avenues of attack I want to use against this.

First, let’s start by acknowledging that there are various methods that can help increase the vividness—and with that, the “resolution”—of a dream, and how well one remembers them. These methods include the importance one places on dreams, drugs such as some antidepressants, and lucid dreaming which can be learned. These techniques either actually make dreams more vivid or help in recalling dreams better (though they could also affect both factors at the same time, which I’ll ignore for the sake of simplification). If the first case is true, then would this mean some of these techniques would lead to more vivid dreams, thus making the dream people also have more details, hence making it more likely that dream people could gain consciousness? (Assumption: higher vividness means more brain capacity is allocated, and more brain power leads to a higher chance for consciousness to emerge. This seems reasonable, since in animals, intelligence is correlated with levels of consciousness.) On the other hand, if it’s the second case, then dreams might have a rather high resolution, we just don’t remember them well, thus defeating the argument.

Second, we could imagine that there are beings of higher levels of consciousness and intelligence, for whom dreams have a more “upscaled resolution”—be it aliens, simulated consciousnesses in the future, where we could ourselves choose the computational capacity of the simulation and thus the “resolution,” or gods (see Brahma’s Dream). This idea is also deeply rooted in Chinese philosophical thought, particularly through the Butterfly Dream.

Daoist interpretation of the allegorical meaning of the story [Butterfly Dream] is this: just as dreaming is not less “authentic” than being awake, being dead is no less “authentic” than being alive.

The idea of “presence” is that whatever is is just what it is and nothing but what it is.

The thought behind this is that most Western philosophers define presence in terms of ontology, the study of the essence of a thing, and separate the “I” from the “am.” Meanwhile, in Eastern philosophy, these concepts are more tightly linked together: what currently is, the presence, is simply what it is—no split between “I” and “am.”

And if we can imagine that consciousness arises in dream people in “upscaled resolution,” why couldn’t consciousness also arise in the symmetrical case of “downscaled resolution” in dreams?

As a final point, outward “resolution” gives us no knowledge of the inner workings of a thing. Just because a thing does not have many outward details does not mean it is not intricate. A video game producer might decide to have worse graphics but better AI in the non-playable characters (NPCs) as a performance trade-off. Or a phone might look simple on the outside, but its inner workings couldn’t be further from the truth. The same could be true for dream people: they might have “lower resolution” in appearance and mannerisms, but have an impressive inner life.

I do not argue that dream people are conscious (see the next section for this). Instead, I argue that just because the “resolution is lower” for dream people does not mean we can exclude the possibility that they are conscious.

The Philosophical Zombie: Or Why We Should Assume Dream People Are Conscious, Even If They Are Not

A philosophical zombie is a thought experiment in which there is a human who is completely normal and the same as every other human, except for one aspect: they are not conscious. A philosophical zombie, just like a normal human, can react to stimuli with actions, hold conversations, do a handstand, and go to work every day.

One basic question we might ask is: how do you know that other humans, e.g., your friends, family, or even complete strangers, are conscious and not just philosophical zombies that act as if they are conscious?

This problem exists because we can only observe the outside of other humans but have no tools to inquire into their inner lives. We may see that they are engaged in an active conversation, but we don’t know if, while doing this, they are conscious—i.e., aware of themselves.

We do have various tests that can check how conscious a being is. For example, the mirror test can be used to test if an animal recognizes itself. But in the end, all we measure is their behavior, outward indicators, never truly piercing the veil to the inner life.

The question of when something is conscious is of utmost importance if one defines personhood as the capacity for consciousness. For example, if a dog whimpers or yelps when hurt, we don’t know if this is simply a reaction to a stimulus, like a sunflower turning to face the sun, or if the dog has an actual subjective experience of being and pain.

Nonetheless, whether a thing is conscious or not lies at the bedrock of our moral theory. Every day, we make moral decisions about people and animals, even if we do not know if a thing is conscious or not. So how do we decide if a thing deserves moral consideration? Why do we believe that other humans are conscious? Or that intelligent animals, like dolphins, possess a low form of consciousness and, with that, some moral consideration?

We do it based on outward behavior. I do not have any guarantee that the reader reading this is conscious or not, but I can say that you are probably similar to me both in what we are and in our behavior. Since I know that I am conscious and we are roughly similar to each other, I can assume you are conscious or at least act like you are. The same goes for animals. We make various behavioral tests and can see that a dolphin recognizes itself in the mirror, that dolphins may commit suicide, that dolphins use pufferfish to get high, dolphins cooperate with humans, and many more behaviors we recognize as things we humans do. And just like with humans, we extrapolate from that that they are, to some extent, conscious.

We can do the same thing regarding dream people. If in a dream I am an action hero, fighting against villains, and I shoot them, I can see the dream people call out in pain and fall to the ground. I can talk to them, hold conversations, and so on. All of this is to say that dream people exhibit behaviors we would expect from conscious agents, and since we do not have a reliable way to investigate whether something is conscious or not, all that leaves us to decide that, is their behavior.

Or, in other words, from the duck test: If it looks like a duck, swims like a duck, and quacks like a duck, then it probably is a duck.

Because the only other alternative is to live in epistemological nihilism, embracing solipsism, and just assume the only conscious agent is me and me alone.

Imagining People vs. Dream People

Lastly, I want to mention a problem that arises if one thinks that dream people have personhood, i.e., moral consideration. This problem concerns the differentiation between dream people and imagined people. If one day you day-dream or rehash an old conversation, the mind may conjure up an imagined person with whom you have the conversation.

But just like I asked what the distinction between dream people and “real” people is, I can ask the same about imagined people and dream people. The only thing that comes to mind is that the latter feel more part of you than the former. But one could argue that this is the exact problematic issue already discussed in the section “Consciousness in Consciousness: Information Flow and Ownership”. The imagined person feels more intimate because the information flow is bigger, but as already discussed, this is not a relevant moral factor. As such, if we grant dream people moral consideration, we would need to do the same for imagined people for the sake of consistency.

Just like with dream people, there might be an escape. One can argue that the interests of imagined people more closely align with one’s own. As such, most things I do to them are in their interest. But this is just an assumption. The interests of both dream and imagined people is an epistemological question, and I will not further inquire into it.

Conclusion

So, what have we learned?

There is no real principle that differentiates “real” people from people in dreams. The only thing we can hope for is that dream people are not conscious, but even for that, it is not clear how one determines that. Maybe, just like we assume with animals of higher intelligence, like dolphins, that they are conscious to a certain extent, we should do the same with dream people. Thus, granting dream people some form of moral consideration. I didn’t scrutinize what rights exactly they should be granted or need, as this depends on their interests, as mentioned in the first section of this article. This seems even more difficult to explore than whether they have consciousness or not.

But I hope I made you think, and that next time you plan to hurt a dream person, you’ll think twice.

Looking back at the dream I had that I mentioned at the beginning, I wonder if waking up instead of continuing the dream (not like I had the choice) was an act of mercy. Because she never had to experience her agonizing death; instead, she just vanished out of existence. But maybe this too is just forcing my human notion of what good and bad are onto the dream person’s worldview, which would feel completely alien to us. Because maybe, by waking up, I exactly destroyed her interest for which she was created.

Ressources:

▲
Made with Hugo, website licensed under CC BY-NC-SA 4.0.