Book Thoughts: The Mountain in the Sea by Ray Nayler

19 Dec, 2025

Following is the distillation of a long conversation thread I had with a friend, with the first email from August 24, 2024. In the process of distilling this, I updated a few sections and occasionally combined our wording together, though I also did leave some sections I quoted directly from my friend.

I listened to this book via audio while playing games. I didn't pay much close attention to the first half of the book, but ended up liking the second half a lot. For what it's worth, I still rated this like a 4.25/5 even though I have a billion critiques.

Audio book vs. Text

The person who recommended me this book actually gave me a text copy of it after I had already heard the whole thing through audio book. Flipping through it briefly since I didn't feel like rereading the whole thing, I realized that there were visual representations of the octopus symbols they were talking about in the actual text (and on the fore edge). I wished I could've seen these symbols when I was listening to this book, but I will say that the audio book is well narrated, especially when it came to the names I would've had no idea how to pronounce. Eunice Wong also clearly knows how to pronounce tonal languages (her website says that she has experience with the sounds of Cantonese, French, Mandarin, Russian, and Lakota), so that was much appreciated.

RE: The biggest theme — "To See and Be Seen"

The Mountain in the Sea argues that true connection between people, and species more broadly, is of utmost importance, and part of that connection is willing to risk rejection. Minervudottir-Chan is a scientist who refuses this type of connection in favor of control, going so far as to make sure that the android she created can still be exploited mentally to her own benefits. It is ultimately her refusal of connection that leads to her unceremonious and impersonal death at the hands (or rather, tentacles) of the octopi, while other characters who strove to make connections with each other are rewarded with life, from Ha and her team, to Rustem, to Eiko and Son.

And what is connection? It is to see and to be seen as you are, flaws and all, and appreciated; and to try to understand each other even when you can never fully do so. To see and be seen in equal measure is the greatest reward. Various parts of the book lead itself to the see/be seen theme, but I'm not going to list all of them here. Most notably, Ha's reward for interacting with octopuses and learning to see them through her attempt at understanding their symbols is that she is seen in return by the octopuses, culminating in a display where the storytelling octopus shows her back her own face to indicate that it recognizes her and that it will tell stories of her to the others to the point that they likely deify her.

From a construction standpoint, I appreciate that the book is very thematically consistent with itself. However, I and my friend took issue with the assumptions of universality in what is important to species and what is of inherent value: empathy, intelligence, beauty, individuality, and memory. Do other species have to value these things in order to be special? Must they have art and religion to be "evolved"? These things are actually inherently human biases that are related to how our own biology and societies work, and they create a hierarchy of importance in other species even when they claim not to. Other sci-fi have grappled more with difference, going so far as to even question what "death" would mean to other species, but The Mountain in the Sea has a bias towards a climbing ladder, and at the top lies perfection that, if grasped, can relieve Earth the burden of humanity.

The strangely uncomplicated ending of the book, which neatly ties everything up with the death of Minervudottir-Chan and the Buddhist Republic's possession of the island, feels like a cosmic blessing of Ha's ideas over Minervudottir-Chan's in a typical empathy=good/no empathy=bad way without actually contending with differing perspectives or multiple truths that can work in conjunction with each other.

Likewise, the fact that Ha was recognized by a singularly important octopus as a singularly important human felt very human-centric in its concept of "awareness". It was also tokenizing in a way that we would recognize more readily with other human peoples: For example, if the octopi were instead a group of Indigenous people, it would be as if the ending revolved around the singularly named and characterized Tribe Leader having a moment of recognition with the non-Indigenous protagonist.

Ultimately, the biggest trap The Mountain in the Sea falls into is of blaming humans as a species and acting as if humans are inherently inclined to environmental destruction (which is unintentionally anti-Indigenous in general), while still being largely human-centric in its thinking of individuals and leaders and "art" and "altars" and "empathy" as inherently crucial.

I'd argue, rather, that any individual's capacity for empathy is less important for decreasing exploitation around the world than recognizing that all life is interconnected, and understanding that exploitation always works its way back to your own life and the lives of the people around you. Even the richest people on this planet will not escape some form of loss if they continue to ruin our planet. Empathy does not guarantee channeling one's efforts into effective measures for reducing exploitation, and emotion does not equal action.

In the end, while I find the "to see/be seen" theme very compelling to my own biases and enjoy the idea of Ha's connections to the octopuses, it doesn't quite work when it's universalized like this. Something I always wonder when I read works that so obviously value performance in people and other species ("This species is intelligent, therefore it matters more; the capacity to see in a way that can be understood by humans is inherently more valuable") is what the authors think of plants. Or for that matter, people with severe intellectual disabilities with limited "awareness." And what of the languages that include rocks as living things, too? If we privilege function as earning more respect, we inevitably value some lives more than others. On an individual scale, maybe it doesn't matter so much, but if that's a worldview we're looking to push, it might do us better to reconsider it.

Things I liked about it:

1. This is the only time I have ever seen interspecies communication done well without the typical dumb human bias for symbols. Like usually it's wow, we have a translator, or wow, they also use [very culturally specific hand symbols like thumbs up or hearts or whatever that aren't even consistent between Earth peoples] and every time I have called bullshit. So this book not doing that was just an epic plus to me.

2. Liked the way it was organized with following multiple people to further its theme of all being interconnected. Setting aside each part with an excerpt from one of the two competing character theorists' books with points of view was satisfying from an organizational standpoint. More on the contents in the critique section.

3. Liked the book's argument for connection and for respect for all species on this planet. Or as my friend put it: "I also liked all the different points of view and how they both did and didn't directly intersect. I liked the ideas of consciousness and how the interplay of AI, the operator/drone bilateral thing, and arguments against indifference and apathy."

4. The book is very aware of global capitalist exploitation, and argues that we should try harder to understand each other and make sure our targets for violence are correct (dismantling the system rather than just the replaceable cogs). That understanding we're all interconnected means taking risks is necessary and sometimes that comes at the price of yourself, but also the greatest reward is being seen and seeing in equal measure.
   
   The whole thing of the ship had a really good payoff — the fact that the rebellion on the ship failed because they didn't actually strike at the system (the boat) itself and only ended up getting rid of replaceable cogs at the expense of their own mental well-being. Eiko would've become a new guard in that system, because it's a systemic problem. Violence is necessary, but it also has to be aimed right. And it's the real relations that will carry you through and are worth risking things for, even though there might not be a way to save everyone. To see/be seen. Also in a way, learning how to talk to the ship in its own language of the market to bring its demise.
   
   (Though, both of us agreed that it didn't ring true that the ship would've given them a lifeboat at the end.)

5. Thought the book's conceptualization of what makes beings was more interesting than just the Turing Test. The book insists that Evrim is alive just because they perceive themself to be. (More on this in the critique section)

6. Minervudottir-Chan being killed by octopuses just randomly because she was in their way. After being built up for so long as a possible antagonist, she basically squeaks off page. Because she's likewise not the real enemy in its entirety, much like the guards of the ship weren't. The real enemy was her corporation, which was also working and making its own decisions without her in a lot of places. That was a metaphor for systems of global control, though nation states like the Buddhist Republic don't exactly go critiqued. But this book defies the urge (that I felt at least) to have an epic ending for an epic antagonist, but the octopuses can't give a shit about that because they don't know anything about that. I think in presenting her as a complex character and then letting her just die like that, it also speaks to what we lose when we kill other beings (or would, if not for Evrim's perfect recall that I have gripes with). That other people and beings are full of complexities just like that, just like all the people Son lost at sea and never got to know.

7. I liked the concept of point-fives (basically realistic AI chatbots) and the reveal that Ha was talking to one, especially since I was like, "How the hell did they let this device clear security". But also when they first brought up point-fives, to some degree I was thinking like, yeah tbh what if all I want in life is someone who isn't a someone so I don't have to ever feel endangered? But I think knowing that it was basically catered to whatever I wanted to hear and couldn't help me materially would make me choose differently because of the inherent feeling that I'm just playing myself because I'm running from my real problems. But the book really did say that, actually. That interacting with real people is important even at the expense of rejection. To see/be seen.

Major Critiques:

1. It talks about writing as a form of human advancement, unfortunately lending itself to anti-Indigenous bias that suggests that Indigenous societies that don't rely on writing are less advanced/"evolved."

2. Places a lot of emphasis on empathy as necessary for caring and for humanity.
   Friend:

   I didn't like how the book presented Ha's Empathy Feelings Mind/Body Separation perspective as Right and the other scientist's Materialism as Wrong. Rather than showing different views of awareness or consciousness, the competing viewpoints went down the usual kind of "cold hard science has no feelings and is evil and bad and exploitative, while empathy and feelings are good and warm and compassionate". I would've better liked it if, say, different people with different philosophies all worked on the octopus language.

3. Prioritizes animals with sentience closer to humans, though that is also how most people think (and the characters would definitely think this). Still could've done better for this part I think. It tried with the thing with the turtles and cuttlefish and the like, but there was still the sense that the octopuses were more important.

4. My friend immediately clocked the writer as a man, though I didn't really notice as I don't read a lot of Sci-Fi to begin with. The roles given to the women are still typical of male authors who are trying to do better than blanket sexism with Strong Female Characters: the "female scientist" and "hardened soldier woman." The sexy parts in the book were basically all focused on women, but I didn't pay attention much to this. On my part, I found the writing really heterosexual, to the point that I kept accidentally reading Evrim as a man rather than as androgynous.

5. The way Evrim was described felt along the lines of androgyny fetishism; there was a constant focus on beauty while being sexless. The descriptions also basically fit "male-default with feminine traits of beauty," — slenderness, androgyny, wearing stuff like golden robes, long legs, looking like a god — and their voice lines were acted as being more masculine than Ha's. This paired with the fact Ha has zero hints being queer-aligned (but an obviously heterosexual past) and chooses to hug them at the end despite not being a naturally intimate person made me read their relationship as fulfilling romantic & heterosexual flags. This was to the point that I kept messing up Evrim's pronouns as "he/him" in these emails.
   
   ^{That said, being sexless is something that appeals to me extremely, as in, I want that for myself. But this relationship failed the QPR vibe for me, even though I would've much preferred to read it that way.}

6. The answer to Evrim being alive or not was answered by the fact that they perceive themself as being alive. This leads to the next question though: "Do they perceive themself to be or do they simply say that they do?" For example, we could program an AI to ask the same philosophical question and say that it believes that it's alive and feign distress. The book tries to fix this problem by saying that Rustem closed all of the portals where someone could hijack Evrim's processes, but that's just bad science (see #3 in the Wrong Science section.)
   
   As for the question of whether they're human and at what point do we recognize something as being human: the book indicates that rather than being "just human," they are "more than human." This is even despite this line, which is one of the latter lines in the book regarding Evrim's humanness:

   But it was not a human hand: it was Evrim's. No less, and no more.

   Rather, the stuff about perfect recall and the indication that Evrim will build a better species that will be able to live in harmony with all others indicates that the book treats Evrim as human+. From an earlier part in the book when Ha recognizes Evrim as human:

   What determines you are human is that you fully participate in human interaction and the human symbolic world. You live in the world humans created, perceiving that world as humans perceive it, processing information as humans process it. What more is there? Being human means perceiving the world in a human way. That's all. So, you are human. You may also be more than that, but you are certainly as human as most people I have known. More human than some.

   Evrim themself says that they know themself not to be human because they don't understand humans, to which Ha says that even humans can't fully understand other humans. And then they hold hands and we get the "no less, no more," line. The book can't seem to escape its own trap here though, because the only thing not human about Evrim is that they are simply better, and that better is enabled through ideas of perfection of memory that is what is supposed to lead them out of the mistakes of humanity.
   From one of the final lines in the book:

   As I said — I am all of Arnkatla. But I am also more than her. I am my own being.

   Otherwise, Ha's quote about being human through participation in human culture is nice, but I think ultimately whether you "are" anything is a doublefold question of: You need to claim it, and likewise, it needs to claim you. Ha claimed Evrim and so they are human to her (and more), and they claim being human[/being Arnkatla] (and more) as well, so the loop in the community is closed and therefore we can say they are indeed human (and super)... But that's just to the two of them anyway, as humanity at large isn't so keen on claiming Evrim (yet).
   
   Friend: "I wish that it had been more 'and also something else' rather than 'something more (superior)'. I don't like that them making more of them self is some kind of great accomplishment of starting a Species."

7. Presenting the Tibetan Buddhist Republic almost uncritically as a "win" for them taking over the island (of which they are also not native/local) was...off. While the book set up the Buddhist Republic to have a monopoly on advanced drone tech (or whatever they called them) and mentioned UN involvement to give them a fighting chance in terms of military might, and they also mentioned that they might not be able to protect the place forever, it begs the question: Why the Buddhist Republic and not Vietnam, when the island was right off the coast of Vietnam?
   
   Or more radically, why not whatever Indigenous peoples actually live near there, or who were originally on the island? Instead, there were basically sections in the book (IIRC months later as I'm writing this) about how people largely chose destruction to survive. This seems to fall into some weird idolization of Tibetan purity when it's presented without any indication that Buddhist Republic possession of the island could come with its own issues.
   
   Taken together with another of Nayler's works, Tusks of Extinction,¹ Nayler seems to just have a savior mentality in general rather than one of Indigenous self-determination, where the populations doing the "saving" are basically anyone who has highly specialized Western scientific knowledge.

8. The Mountain in the Sea also argues that as a species we're inclined to destruction in a way that blames basically all of humanity equally, when polluting structures and structures of war have actually been set up by a very small part of our global population. Its themes unfortunately has some sympathies with eco-fascism by basically suggesting that humanity as a species are irrevocably flawed and thus the only suggestion is to birth a new life form entirely (Evrim). Given that Evrim also has "perfect recall" (more critiques on the science of this later), the work betrays the idea that the reason why we hurt other people is because we aren't perfect enough — this is an individualistic lie that doesn't account for Elite Capture, and doesn't believe that humans can live alongside nature, even though Indigenous peoples were doing that before colonialism. Even if violence will always exist in human societies (and other animals), it doesn't necessarily mean destroying the planet. All of humanity is not responsible for nuclear weaponry.
   
   A more generous reading of this would say that Evrim is just a new species that will continue to exist alongside humans rather than replace them entirely, but kinda hard to fully believe when their name literally means "evolution," as Rustem reveals, and they are far more perfect than a human's brain.
   
   Friend: "He actually had a decent quote in the Tusks book. 'All fathers think they can pass the good on to their sons with none of the bad.' That's how I feel about the idea of having to birth a new species to be better than humanity. I think humanity can just git gud instead."

9. The best part of the book is figuring out octopus symbols and how to communicate, but there's not enough of it. Kind of feel like this was because the author wasn't confident in building a whole language.

Smaller Critiques:

• My friend mentioned that most of the interesting facts about octopuses was found readily on Wikipedia pages, so it's unclear how much research Nayler really did.

• The automatic translator doesn't make sense or follow patterns of real-world machine translation and envisions that they'd speak in broken English, which is not a real issue with machine translation. The jokes with it are primarily written to be funny in English — it's unbelievable that there would be something said in Mongolian that would translate to "fornicating" to "shitting rain."

• I didn't read the text version, but my friend said that the use of italics for foreign words is baffling. Lots of random foreign food is in italics but gaijin isn't. A lot of the cultural stuff also does come off as exoticizing.

• Friend: "It's pretty unbelievable that governments would actually make AI illegal. ChatGPT and original kinds of language models (shown off on TV like the 'therapists' or that bot that showed up on Jeopardy) weren't made illegal. More likely that people would dismiss it as not real or not having a soul or not being human or make Alexa jokes or being a hoax."

• Book plays into the myth that "disabled people have super senses" (as opposed to having to learn to rely more on other senses and adaptive devices) by bringing up echolocation. People shouldn't expect echolocation from blind people or assume everyone can use it and when it's brought up it's often to ignore the need for accommodations.

• The part where Ha smashes her point-five until it's unusable felt a little anti-technology "touch grass" type of stuff. Point-fives are like having a companion/emotional support pet. Would've preferred the message to be, "It can't be more than a companion animal, so don't let it replace real relationships with humans".

The Wrong Science according to my friend

I can't say I know better, but it sounds right to me. This section is just quotes.

1. "Meaning" and "Language" need physical architecture

"Meaning" does have a mass and charge. Saying it doesn't is often used as a "science can't explain feelings/empathy" rhetorical appeal, but information doesn't exist in a vacuum, devoid of any context. The concept of meaning having no mass or charge is false. Information only exists as long as something stores it, and information is only stored in the movement of particles and energy. Meaning cannot exist outside of the physical architecture that conducts it.

The universe as a whole can only store so much information in it. In fact, a model that would be able to accurately predict the universe would take so much storage that it couldn't exist inside the universe! There is an upwards information cap on how much the whole entire universe can contain within it. A multiverse could exist, but the point is that information and meaning have quantifiable masses and charges, and there are "finite" quantities of these. I find it incredibly cool and awesome and important that even stuff like facts and truth inherently requires "storage" and is fallible as anything else. Many people want some kind of metaphysical specialness to meaning, but there's nothing more special to me than the act of being alive, wetware teeming in electricity, permitting meaning to exist. Our architecture houses us. Much as I wish mind-body separation were real, and I could just copy myself to the Internet, the physicality of my electrical signals forms the being known to me as myself and my meaning.

Imagine if our question was whether piles of sand had mass and charge. I think that we would both agree that sand has mass and charge. So sand has mass and charge. Therefore, any pile of sand will also have mass and charge. However, how do we define a pile of sand? How many grains does it take to make up a pile of sand? When we're measuring the mass of a pile of sand, do we need to subtract any non-sand particles from the pile? Different people have different ideas of what a pile of sand is. Different piles of sand may have different quantities or compositions of sand and non-sand particles, and we'll have different masses and charges. So the answer to the question of does sand have mass and charge, is yes. However, the answer to the question of, how much mass and charge does this particular pile of sand have, can be more challenging to answer. A pile of sand is not a concretely specific thing. We define what a pile of sand is based on our subjective human (social) categorization. Even what we define as a grain of sand is ultimately a human social definition. So what this means is that, the way that we define the pile of sand, depends on our personal subjective definition in that moment. For example, I could define a pile of sand in a particular way and also decide that, as long as it has less than 5% impurity, I don't have to worry about subtracting non-sand things from the pile. Then I have to figure out how I'm going to measure the percent impurity and so on. However, none of this changes the fact that piles of sand, regardless of how we decide to define them, have mass and charge.

Okay, so what does this have to do with the question of ideas or meaning having mass and charge?

The difficulty of defining the total neural space in which everyone has ideas does not change the fact that the physical architecture required to house meaning and idea has mass and charge.

You're right that we can't ever fully express our internal meanings to another person. We may have similar meanings, but they will never be identical. I agree that most people want to perceive some special other about their personal internal meaning. However, their personal internal meaning is necessarily housed in physical architecture, which gives it a mass and charge. Theoretically, that mass and charge can be quantified and measured. Quantifying the mass and charge does require placing an arbitrary human definition on how to quantify mass and charge, just as we have to quantify and define what a pile of sand is. But the fact that we have to use an arbitrary definition doesn't mean that it doesn't have mass or charge. It just means that we have to define how we're going to measure that mass and charge and what we're going to include in that mass and charge, but anything that we're going to measure is going to have some component of mass and charge. The inability of language and other means of interpersonal communication to fully convey our internal meaning does not take away from the mass and charge of the architecture that houses the internal meaning.

A shared meaning by itself is also an arbitrarily defined category. How do we decide when meaning is shared? What is close enough to indicate that meaning has been shared? What degree of difference would we find acceptable to still say that the meeting is shared? For example, my perception of red differs from your perception of red, but we probably have similar overall perceptions of a red due to its societal significance in the cultures in which we grew up and immerse ourselves, such as passion or anger or whatever. What difference between my perception and your perception, and what difference between my internal meaning in your internal meaning, makes this perception a different meaning? These questions don't actually change the fact that my neurons and your neurons both take up mass. You are correct that defining the exact mass of a particular meaning in a particular population would require a lot of a priori definitions.

Nonetheless, the claim in the book was that meaning just doesn't take up Mass. That's false. any information in the entire universe or anywhere else has to take up mass. It is impossible to have information, store information, read information, access information, convey information without that information occurring in mass. This goes with the idea that there is no such thing as an objective truth. There is no true information floating out there free of human subjectivity (or whatever subjectivity we're talking about). So, the fact the information requires mass also means that many truths can exist, because no information exists outside of the universe that pertains to the information. There could be a multiverse or whatever that we just don't know about, but this fact would then just apply to the multiverse rather than the universe, and still mean that all of the information has mass and that no objective truth exists. Changing the size of the container does not change the inherent properties of information.

I think that many laypeople have this idea that information floats disconnected from stuff. I think this idea has to do with the desire to believe in a soul or in unique personalities (identity), and also the fact that we can't see information. It's interesting because we understand that computers have limited storage and that when we place information in that storage we fill up the storage, and yet many people struggle to apply that same idea to themselves or their brains. I do think that some people don't realize or don't understand this, and I think that others just don't want to see it that way.

And I'm not saying this (that information doesn't exist as a metaphysical other but exists in concrete matter) in a dehumanizing sense. I actually feel like I have more optimism for the human condition than some people who don't see it this way. But brains are essentially very elastic, very plastic, very adaptive computers and so have limited storage. That storage changes over the course of our lives.

What I mean by this is that our brains don't store the complete memory. Our brains store enough information that they can then reconstruct a plausible version of that memory when we call upon it. Our brains try to extract the most useful information to form a memory, and usefulness here is not an objective measure, but just whatever the brain has subjectively determined was useful in that moment and in subsequent moments where the memories are recalled. But just like how our senses and qualia are essentially mass hallucinations based on some stimuli from the outside environment, our memories are super made up jumbles based on enough identifying characteristics. It's why people can create false memories. Our brains literally make this shit up as they go. And this actually permits them to store a lot of information in a relatively small space. It's like how a video game that uses procedural generation can have a smaller total storage, because the individual rooms can be constructed from a much smaller library of assets, compared to a traditional game where every room has to have its own assets or whatever. This is a bad analogy because real games are also composed of many different specific assets and so on, but I hope this helps to illustrate the point that I'm trying to make.

But I do think that the average layperson, who recognizes the gap between their internal meaning and what they can externally convey, perceives that gap as existing outside of mass and charge. The reality is more that the internal meaning exists in mass and charge. And then the systems that we use to externally convey it also exists in mass, and the systems that we used to externally convey things just can't express an internal meaning on a one-to-one basis. What we externally convey is ultimately a model of what we internally mean, and models will always be simplified or otherwise inaccurate versions of whatever they're trying to model. We can only try to make the model as close to the original as possible.

Actually if we could read each other's minds, I think that we would have a really hard time because, as I said, our internal meetings are made up of fragments that are essentially constantly generated by our brains rather than some kind of infallible perfect record. In a sense, the external conveyance of words or whatever other medium is actually much better to store information in in a way that can be comprehended by other people, even if it doesn't feel that way because our internal meanings don't always match. Even if we could connect minds, that wouldn't permit us to completely understand in other person's internal meaning because their internal meaning would always come through our own subjectivity, which would inherently distort the internal meaning and thus change it, much like how the process of communicating internal meaning through language also changes it.

"How [does one make a language where people can't lie]?" I think it shares a misunderstanding of what language is. Even if we could read minds, people could still lie.

2. Somewhat related to the above ↑

The android's perfect recall had me rolling to be honest. Where are they storing all this data. What if their hard drives fail. Where physically is it? There's an upper limit on information density storage. Their little human-sized body cannot store all of that in perfect detail with sufficient redundancy. Their memory would be better than a human's but any storage media will fail with time.

3. Any system can be hijacked

Rustem closing the portal is really silly to me. Any system including human brains can be hijacked with the right tech. Closing the engineering backdoor can make it harder to hack it but not impossible. I wish the book had struggled with it more instead of giving them an easy answer. "Previously, you could have been controlled, but now you're free for real objectively!" is a much less interesting space to grapple in.

I mean, the idea that a system can't be broken into if the backdoor is destroyed is just wrong. Backdoors make it easier. It's funny because the author correctly used a cost of entry metaphor earlier in the book. The answer to the ending would've better been, "Now that the backdoor is destroyed, I think I've raised the cost of entry so much that almost no one will bother. It doesn't guarantee it, but you have a much lower chance of getting accessed." But I guess that isn't as feel-good.

Other stray personal feelings dump:

#books #fiction