You found my old blog. Thanks for visiting! For my new writing, visit mikesententia.com.
Last week, I compared spirits to artificial intelligence:
Implement [an] intelligence with nerves and living tissue, and you have a human. Implement it with transistors and silicon, and you have an AI. Implement it with ethereal structures like magickal connections, and you have a spirit.
Several readers had excellent comments. Ananael brought up some larger questions that I want to answer today. (His text in italics, my replies in normal.)
This post brings up a question that’s been on my mind for awhile. How, precisely, does your model define the difference between these sorts of “AI” spirits and ethereal software? I don’t necessarily think your artificial intelligence analogy is a bad one, but I do think how you explain it here blurs the line a bit.
First, I want to acknowledge a way that my metaphors are confusing. When talking about spirits and ethereal software, I say that spirits are like humans, and ethereal software is like a computer. But I also say that spirits are like an artificial intelligence. What gives?
Partly, it’s that metaphors are limited, and spirits are like spirits, not humans or artificial intelligences.
But partly, this is keeping with science fiction around AIs, where computers are tools, but AIs are people. They just happen to be people made of computers, as opposed to people made of cells. So, keep this distinction in mind as we discuss spirits (human-like AIs) vs ethereal software (unintelligent computers).
Before going on, I need to make sure everyone has experienced a certain corner of computer science geekery. In the 90s, people made computer programs you could text with. They weren’t brilliant, but they could have a conversation. If you haven’t tried one, talk with Eliza for 30 seconds.
Texting with Eliza is clearly different than texting with a penpal in China, right? But imagine your grandfather has never used a computer, and try to explain that difference. You text with both of them. (He’s never texted with anyone, so they already seem pretty similar.) Neither uses English quite like you do. (Your penpal is Chinese, remember?) You can cite some simple differences: Eliza replies right away, your penpal sometimes misspells words, and so on. But that’s not really the essence of the difference.
That’s the situation with spirits and ethereal software. They just behave differently. Spirits have emotions. They think, take a different direction, give useful advice. Ethereal software executes commands. When I talk with them, it’s an obviously different experience. But I can’t easily define that difference for you.
Also, a question I’ve been pondering: Are spirits and ethereal software made of the same thing? I don’t know. Both are made of ethereal matter. But both humans and computers are made of ordinary matter, and we’re not made of the same stuff. So I don’t know. Spirits and ethereal software might be made of different types of ethereal matter. (Or not. I don’t know either way.)
That’s because it seems to me that in your model ethereal software has to do a lot of recognition-type information processing. We had that discussion awhile back of how you think a massively modified LRP is still the same LRP because the “LRP ethereal software” recognizes the form even if it’s been almost entirely rewritten.
I’d like to clarify how I think about sigils and rituals, and how they interact with ethereal software.
Remember that sigil I made for my book? A grid of six symbols, each repeated twice, to form a sigil. When I did that, I tied each symbol to the ethereal software. You can focus on any one of them, and it’ll probably work. Any four of them, and I’m even more confident it’ll work. You don’t need all , that’s just there for redundancy. You could take 3 of them, arrange them into a triangle, add a big circle around them, and it should still work, as long as you still have the individual symbols that the software recognizes.
That’s how I think of rituals: A bunch of components, each recognizable to the ethereal software. You can re-arrange them, remove a few, and they’ll still be recognizable. Will the re-arranged ritual issue the same command as before? I have no idea. But each component should connect you to the same ethereal software, and that ethereal software should still respond to the ritual as a whole.
That’s never struck me as very plausible for “software” that you describe as essentially unintelligent, simply because as I’m sure you’re familiar, modeling recognition tasks like that on a computer requires an enormous amount of processing. For that reason recognition algorithms are a big part of current AI research and seem to require some sort of “intelligence” of their own in order to work at all.
I think I muddied the waters by talking about “intelligence.” Like computers, ethereal software can do complex information processing. They recognize commands, sigils, ritual elements. They even reads minds. They’re definitely complex.
Also like computers, ethereal software isn’t self-aware, doesn’t have emotions, doesn’t seem to do original thought. That’s what I was trying to get at by saying that it’s not intelligent.
So what would really help ceremonialists like myself understand your model better is if you could clearly delineate where “ethereal software” stops and “spirits” start. Or are they the same thing? If that’s the case, we may have just have been arguing about terminology this whole time.
For me, it’s not a continuum. It’s two categories. Like humans and computers.
I also want to mention: Sometimes, ethereal software gives you a vision of an angel or other avatar (just like computer programs sometimes use a human avatar). So you can’t assume that anything offering a vision of a spirit is a spirit.
And years ago, before I became good with shielding, I’d sometimes encounter spirits who drained my energy. These spirits were more animalistic, less intelligent. Still very different than ethereal software, but not the human-without-a-body that I normally associate with spirits.
So, two categories. Human/AI vs computer, spirit vs ethereal software. Does that help?
If you liked this post, consider visiting my current blog at mikesententia.com.Tags: Ethereal Software, ReadersQuestions, Spirits
It helps as far as highlighting the differences between our models go. I still don’t find your model very parsimonious in terms of explaining ceremonial rituals like the LRP. By the time you include all the possible variations and “matches” to different components it becomes really complex and in my opinion less workable than other models I’ve considered.
Am I right in thinking that the explanation you give above would predict that that the fewer pieces match the ethereal software, the less effective the ritual should become? I don’t want to be constructing a strawman here, but if so, I can tell you for a fact that’s not how it works. There are often modifications you can make that deviate from the ritual’s original design but which improve its effectiveness, and there’s no linear relationship between effectiveness and number of modifications.
That aside, though, one of the things I’m seeing here is that I tend to think of the categories you describe as manifestations of the same underlying “consciousness-stuff” – whatever that is – rather than breaking them out into two completely distinct classes. To some extent my ideas are informed by my education in neuroanatomy. Depending upon the complexity of an animal’s nervous system, they may act reflexively, exhibit sentience, or exhibit sapience.
To extend the analogy, very simple animals would then correspond to your “ethereal software” level. They simply respond to stimuli (commands) in a mechanical way. The “animalistic spirits” you’ve encountered would correspond to creatures that are sentient. Note that “sentient,” like “theory,” has a popular connotation different from the one scientists use. An animal that is sentient is not just motivated by direct stimulus-response, but exhibits some rudimentary emotion-driven behaviors.
Finally, “person-spirits” would be sapient, which means self-aware. Colloquially, people often say “sentient” when they mean “sapient,” just like they often say “theory” when they mean “hypothesis” or “guess.” Only a handful of species are thought to be sapient – the great apes, elephants, dolphins and orcas, and some corvids. Humans aren’t even sapient until an age of about 18 months (which, by the way, is probably why you can’t remember your infancy – you weren’t self-aware at the time).
Now how I tend to look at this is not to see hard classifications, but with the realization that these three categories are on a continuum. It’s all the same neural cells, but as the number increases and their arrangement becomes more complex an animal first becomes sentient, and if the process continues far enough, becomes sapient. Whether that’s how the process really works in the spiritual world I can’t necessarily say, but I do think this difference in perspective is at the root of some of our disagreements.
It would be great to come up with some sort of experiment that would settle it, but so far I’m having a lot of trouble thinking of what that would be. Maybe see if you can evolve a piece of ethereal software into a spirit by adding complexity to it, or something like that? I’m not sure how you would test it, though.
Lots of interesting stuff in your comment.
On the number of standard ritual elements: Based on my general model, I’d expect it to be a threshold, rather than a “more is better” situation. If you have very few standard ritual elements, or none of them are done correctly, then the ritual won’t contact the ethereal software. (Though an experienced mage who works with that software all the time could contact it just by thinking about doing magick — that’s how we get rituals to Superman and similar things.)
Once you have enough standard elements to contact the software, well, I don’t really know what happens. In general, ethereal software will read the mage’s intent, so a lot of the ritual’s effectiveness is probably in how well it helps you focus on that intent. I also wouldn’t be surprised if the software picks up on certain ritual elements and interprets them as instructions. Modify the ritual, and you modify those instructions. Either of those situations could cause a modified ritual to work better.
But, I will also admit: If you want to discuss the specifics of modifying rituals, I’m the wrong guy, and that’s the wrong model. I don’t really do rituals — I’ve done some, but I communicate with ethereal software in other ways. And my model doesn’t predict which changes to make to which rituals. In terms of ritual design, it can only explain after the fact.
On the animals-to-humans continuum vs a computer-human dichotomy: Interesting point about the simple brains and brain complexity. I’d apply the same continuum, but two of them, one for spirits and another for ethereal software. I think we can agree to disagree on that for now.
In terms of testing, what comes to mind is to learn techniques to create ethereal software, and techniques to create spirits. If they’re similar techniques, that suggests they’re similar things. If the techniques are vastly different, that suggests they create different things. A project to tackle someday…
Thanks for writing. Always a pleasure to discuss with you.
A very helpful post, and Ananael has captured things I have struggled to convey in the past. One of the things I have found confusing and wanted to ask is: Why are spirits ‘AI’ and not ‘beings’?
Particularly when both ethereal software and spirits are to be made from ‘ethereal stuff’, and your view of consciousness is that of emergent materialism – isn’t it just a matter of complexity? But maybe the important difference is: They are made of the same material or stuff but not of the same component structures. And that is how “computers brains > human brains” is like “ethereal software > spirits”.
What your mentors say about their origin?
Do they know if they are artificial or evolved or something in between?