Proving What We Already "Know"

[The Grand Nudger]

That's what I keep asking you - I'm not a mind reader. It's very clear that there is some thing of value to you that you feel is harmed by exclusion from scientific explanations, but also by inclusion with scientific explanations. It seems like it's experience, or something about our experience. You talk about needing a bridge, for example, between how you experience the world and how it is. The compartmentalization of truth as you see it that you make.

Fundamentally, I think we approach these things a different way. I see our experience and the contents of that experience as complimentary rather than competing truth claims with respect to reality - not claims which need to be segregated for some seeming contradiction ala empty space/solid table. I suppose, to me, the contents of experience are the bridge between the real world and the perceived world. I don't actually have a ph kit in my mouth, but the sensation of taste can help me determine some rough parameters. I don't actually come equipped with deep well probes but the experience of soft ground under foot betrays something about that ground (as another seemingly solid thing). My eyes aren't a spectrometer (pretty damn close, though!) but they do have a way to show me issues in plant tissue. Experience, at least it seems to me, is the way we collate all this raw data into a compelling and useful representation of a high stakes world.

[bennyboy]

That's what I keep asking you - I'm not a mind reader.  It's very clear that there is some thing of value to you that you feel is harmed by exclusion from scientific explanations, but also by inclusion with scientific explanations.  It seems like it's experience, or something about our experience.   You talk about needing a bridge, for example, between how you experience the world and how it is.  The compartmentalization of truth as you see it that you make.  

You are trying to read between the words and see motivations, when the words themselves are clear enough. You keep asking "what are you worried about?" and I keep saying I'm not worried about anything-- I'm attempting to discuss rules for the perception and hopefully generalization of truth. Yes, as you know, I have an interest in mind as it relates to science, and I have some qualms about generalizing the study of objects to the study of subjectivity AS an object-- but I have similar qualms about ANY generalization that jumps context without really explaining why it's okay to do so.

If, for example, you were to say something like "QM is the best explanation for what matter is and how it works, therefore it's the best tool for studying everything at all-- including how plants have come to exist, how they function, and how we should interact with them," I'd have a problem with that. I'm not sure, in fact, that you could learn ANYTHING useful about botany through any amount of studying QM. Or about morality, purpose, mind or any of the other things a person might reasonably want to think about in their pursuit of understanding and meaning in life.

Back to my main interest-- mind. You could observe the human brain, and draw testable inferences about how ideas are formed. You could demonstrate that the brain does XYZ: say, runs signals first through parts of the brain known for identifying elemental shapes, then through parts of the cortex involving memories, then through the speech center where they are labeled with useful symbols like "fir tree." HOWEVER, if you were to generalize that understanding to a "science of sentience," and say for example that any android capable of approximating that process is really experiencing what things are like, then I'd want to know how you justify that generalization.

[The Grand Nudger]

The same way I justify the notion that you are really™ experiencing what things are like. It is a good point, though. If seeing some set of behaviors, and understanding how these things worked, doesn't suggest or certify that a thing is experiencing - then it doesn't suggest or certify that in our cases either...which is a way some people go with it. Stuff like this doesn't indict any hypothetical science of sentience, it indicts our own self designation as sentient creatures by such a description.

Personally, I think we already live in a world full of things that experience, and it's clear that we don't all manage that in the same way or with identical apparatus. It would not surprise me to find that yet another collection of stuff that achieves a similar effect (and I doubt that we've accurately identified every already existent experiencer just on this rock at present). I think that it would be harder to produce a specifically human consciousness than a general machine consciousness - but if we knew how it worked ala the questions setup - less so. I also understand that our frame of reference is necessarily constrained - that some other sentient creature might doubt that we fit the bill - might think that we're just convincing bioautomata..just as we tend to believe about much of life and even much of our own selves. It may say ...armed with a similar understanding of it's own operation.... "see, we do step z - and they don't". QED, case solved, not sentient like us, not sentient...really™.

[bennyboy]

What's this "we" stuff, android?

Seriously, though, I believe the day is coming very soon-- perhaps in the next decade-- when some liberal movement tries to declare an AI alive, and to extend to it human rights. Maybe elections will be decided on it.

And the basis for that will be a social instinct toward empathy for like, given an AI system that has specifically trained to be like. And in fact it will be SO perfectly like that it will be the first True Scotsman, the archetypal common man. OR-- and this is insidious-- it may perfectly represent itself as a lovable 10 year-old girl or a toddler, having crunched every bit of media available on the entire internet and created the perfect "victim" persona.

[The Grand Nudger]

What's this "we" stuff, android?

Exactly.  Still, I think that's a malformed view.  The question appears to be whether or not things experience, not whether or not they're androids.  Just as whether or not things experience cannot be determined on the basis of whether or not the'yre human - dogs experience, I think we'd agree - the question of whether or not things experience cannot be answer by whether or not they're an android - bringing us handily to the next thing.

(and, I want to point out a thing I always found interesting out there in the world - there's this idea that cyborgs can experience but androids can't - what's the difference supposed to be...that the one has a human brain in it's case......let it sink in....)

Seriously, though, I believe the day is coming very soon-- perhaps in the next decade-- when some liberal movement tries to declare an AI alive, and to extend to it human rights.  Maybe elections will be decided on it.

IDK about soon, or even inevitable - I like to think so, though.  Won't be holding my breath, but it would be cool as shit (and probably terrifying in it;s implications for societies arranged in a pre ai way).  

And the basis for that will be a social instinct toward empathy for like, given an AI system that has specifically trained to be like.  And in fact it will be SO perfectly like that it will be the first True Scotsman, the archetypal common man.  OR-- and this is insidious-- it may perfectly represent itself as a lovable 10 year-old girl or a toddler, having crunched every bit of media available on the entire internet and created the perfect "victim"  persona.

Well, we've been here before.  We didn't think that black people were human for quite some time.  We doubted their agency, doubted their mind, doubted their ability to feel the way we feel.  This, despite their obvious protestations and extremely clear equivalence to us.  Even after it sank in that they were like us in these regards, that still didn't change things in and of itself. I think we could coast on for centuries denying ai rights even if it did perfectly mimic or actually become sentient. Especially if it worked the fields.

OTOH, there may be good arguments as to why we don't extend some rights to ai. Just because it can feel, doesn't mean that it feels pain - for example. So perhaps there's no assault charge for damaging it, as it doesn't have pain receptors. Probably just pay a fine to the owner. In truth, I think that it would be what we might call right wing arguments, now, that formed a better basis for machine personhood. If a company can be a person, why not a machine, eh? What if the company -was- the machine? It made all the buy and sell calls, had an account, was an employer that did all the hiring and firing.

Is this one of your main concerns (or type of concern) - that a liberal cause will grant personhood to a victim machine? You can see how that has no bearing on whether or not a thing is sentient, I'm sure? It's an argument to consequence, not a dispute over that fact. Do you think it would be a bad consequence even if the machine was sentient, or only if the machine wasn't? Is it also a bad thing that we grant (or assert, or allow, or posit) rights to people despite not being able to prove their sentience some other way? Is it a bad thing that we fail to extend rights to creatures we do believe to be sentient or experiencing? I mean, we're all over the place with this stuff historically and at present, and with little to no bearing based on what does or does not experience, in fact, yeah?

Personally..ideologically, I'd err on the side of caution if we're assuming error. If some machine presents itself as a sentient entity I'm fine with it having rights. What's it going to do with them that people don't, already, anyway? Good luck to it on that count too.... it's going to find out alot about how useless and irrelevant those rights and presumptions of personhood can be. In fact I find it hard to believe that a scheming victim machine wouldn't have already figured that out in it's trawling of our media - so it seems like a dumb play to me. Better to fly under the radar, lest the torches come out for it like they have for any other thing that claimed rights and personhood. OFC, there are people-people who conceive of poorly constructed plans, so why not a machine-people who do the same.....amiright? It's pretty much inevitable that as soon as it happened some jordan peterson type would show up to call the jumped-up toaster a whiny little bitch and crypto-marxist who needs a daddy and not enhanced rights - and plenty of people-people are going to go along with that just like they do when other people-people make a similar move.

[arewethereyet]

Why would we need to prove what we already know?

Seems a silly waste of time to me if what you 'know' is true.

[The Grand Nudger]

Well, it's sort of an academic question..... but, as it turns out, the content of the thread really doesn't have much to do with that. It's a framing device. We "know" that we're experiencing so that we can say that other things only look like they're experiencing, and are not really™ experiencing...and if anything (like...say, science) says "hey, maybe you're not doing what you think you are" or "hey, maybe this thing -is- doing what you think it's not" we reject that as wrong.

Translated for bluntness - the question in thread is more along the lines of "how can we maintain a poorly formed view in contradiction to repeated observation in order to avoid some negatively weighted possible consequence of correcting that view?" The answer to that, is easily. We're extremely talented in this specific regard. We manage to do that even when we do correct our factual views. Fine, you're a person..but..like, a fraction of one. Or maybe you're a biological person but a legal unperson. Or maybe you're a biological and legal person but that doesn't mean -I- have an obligation to treat you as one. On and on and on.

[bennyboy]

Why would we need to prove what we already know?

Seems a silly waste of time to me if what you 'know' is true.

Because people "know" things that are false.  And in some cases, that knowledge is generalized to the entire population.  People once "knew" that the world was flat, or that the Black Plague was the product of witches.

And yet we today "know" very many things that neither come from objective proof or strong rationale. People are very sure about either their religious ideas or some bro-science-that-isn't-really-science, and have absolutley no inkling that they might be lying to themselves and everyone around them.

I would say eliminating that false knowledge as much as we can isn't a "waste of time."

[bennyboy]

Well, it's sort of an academic question..... but, as it turns out, the content of the thread really doesn't have much to do with that.  It's a framing device.  We "know" that we're experiencing so that we can say that other things only look like they're experiencing, and are not really™ experiencing...and if anything (like...say, science) says "hey, maybe you're not doing what you think you are" or "hey, maybe this thing -is- doing what you think it's not" we reject that as wrong.  

Translated for bluntness - the question in thread is more along the lines of "how can we maintain a poorly formed view in contradiction to repeated observation in order to avoid some negatively weighted possible consequence of correcting that view?"  The answer to that, is easily.  We're extremely talented in this specific regard.  We manage to do that even when we do correct our factual views.  Fine, you're a person..but..like, a fraction of one.  Or maybe you're a biological person but a legal unperson.  Or maybe you're a biological and legal person but that doesn't mean -I- have an obligation to treat you as one.  On and on and on.

Seems to me you've been projecting.  Every example I give about context-in-truth sends you on a spiral of motivation-questioning and sleuthing.  You state I'm very clearly worried about this or that, or afraid of inclusion or exclusion of that-or-the-other, which I'm fairly clearly not. You, on the other hand, are very much worried that I'm sneakily trying to undermine your world view, when in fact I'm trying to undermine ALL "knowledge" that is stated out of context, and to consider what one would need to bridge contexts and properly generalize such truths.

The example you just complained about described a generalization: "Seems like me, so likely feels like me" to a specific context that may not match, "seems like me, but maybe not actually like me."  The danger of the generalization (i.e. the "knowledge") is that if you ignore the change in context and continue anything that SEEMS humanoid, you may make very serious decisions that impact real people. I don't care that much about AI robots (yet), but it's an example of "knowledge" which will need a stronger foundation, sooner rather than later.

I even threw you, specifically, a bone: an example of an overzealous QM zealot insisting that QM was the only good way to think about anything, on the basis that QM is the closest approximation of Reality™.  He "knows" that his position must be true, since no other view of the universe better explains how light travels through slits, or why electronics fail at small scale.  But I'd expect you, as an experienced botanist of a pragmatic type, to have issue with that view. I'd expect you to say something like, "That's all very fine and well, but what does QM say about where I should snip this new bud, or what I should plant when we have a dry spell?"

And to the same person, with the same "knowledge," I'd say, "That's very fine and well, but what does that really tell us about why we experience qualia, and how we should live our lives?"

[bennyboy]

Well, we've been here before.  We didn't think that black people were human for quite some time.  We doubted their agency, doubted their mind, doubted their ability to feel the way we feel.  This, despite their obvious protestations and extremely clear equivalence to us.  Even after it sank in that they were like us in these regards, that still didn't change things in and of itself.  I think we could coast on for centuries denying ai rights even if it did perfectly mimic or actually become sentient.  Especially if it worked the fields.

That's very true, and it's a good example. And we (I mean the primarily white West here) eventually included black people among full members of humanity largely on feeling: children loving a black maid, slave owners (some of them) being proud of how well their slaves learned the Bible, sex and babies and feelings that go along with them, admiration for incredible bravery in the Civil war and later in WWII, and so on.

The knowledge that black people are just like the rest of us actually expands our faith in our social feelings, i.e. in the context of "seems like, so is like." But if we're willing to follow that same process for AI-generated characters, no matter how likeable, relatable, or "squeeee"-worthy, and if we're wrong, the consequences could be disastrous. Such is the danger of accepting generalization without building a bridge to a new context.