Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[silverspawn]

hm, right. the boiling point of water is of course not an elementary law but a consequence of lower laws, but nonetheless, not clear why trheshold-y effects couldn't similarly result on the C. side.

Maybe 'strong emergence' is the better framing for this reason. I do think the "counterfactuals can't count" step still goes through?

[faust]

Anyway, among the things that are not ok is caring counter-factuals. Say you have a system that can be modeled as a causal graph with nodes A-L. The right part of the graph is something like J -> L <- K or something (e.g., and AND gate), involving only the last 3 nodes. Currently, only J and K in the system are on, and together they cause L to be on due to the connection. In describing the C. of this system, you're not allowed to look at connections that didn't matter, like an edge from node A to node B. Since consciousness has causal effect but the connection from A to B didn't play a causal role in what just happened, in can't have affected the system's consciousness.

More generally, only the things that are happening right now are allowed to matter. If a brain reacts to input x, you can't look at [how it would have reacted to input y] to figure out what its consciousness is like. You can only look at the computation that actually took place.

I'm not quite convinced. As an analogy, consider a program that takes the values 2 and 5 and outputs 7. I cannot conclude from this that the system performs addition. Indeed I cannot conclude that no matter how many tests I run. The only way to know for sure is to do some sort of formal verification which analyzes the entire system.

How is "X performs addition" a different property from "X has consciousness"?

[silverspawn]

The "X has consciousness" property is different because the precise type of consciousness always has causal effect. If you change a system such that its qualia (=current state of consciousnesses) changes from A to B, this must be accompanied by a corresponding material change, which has causal effect on the future.

This is not true for addition. You can have System1 with behavior "if input=(4,4) perform addition, otherwise perform addition" and System2 with behavior "if input=(4,4) return 0, otherwise perform addition". Both systems behave identically for all inputs except (4,4). (They don't just output the same thing, they perform the same steps, hence the if condition for System1.)

I think this exact example works as a proof for why counter-factuals cannot matter. Say you have System1. It does addition. Each instance of it doing addition exhibits some kind of qualia. Now it receives input (2,5). Right before this input is processed, I go inside and change it into System2. If this did matter, then it would now exhibit different qualia despite doing exactly the same thing. This is a contradiction; changes to consciousness that have 0 change on causality are not allowed.

[faust]

I think I don't get what your point is. If I change System1 to System2, then surely this will have causal effect on the future, at least on a possible future in which I give the input (4,4).

[silverspawn]

The change of qualia didn't have a causal effect. You can view an action as a tuple (m,q) with the casual material effect m and the qualia q. You can't have two tupels (m,q), (m',q') such that m=m' but q≠q', and this would happen in the above example.

You can also imagine changing it back into system1 after the input where it didn't matter.

[faust]

The change of qualia didn't have a causal effect. You can view an action as a tuple (m,q) with the casual material effect m and the qualia q. You can't have two tupels (m,q), (m',q') such that m=m' but q≠q', and this would happen in the above example.

You can also imagine changing it back into system1 after the input where it didn't matter.

So the claim is: two systems with different qualia can never cause the same material effect? This still seems very unintuitive to me.

[silverspawn]

That was the implicit premise going into the argument, although it may in fact be the main sticking point

(That is, if you take the statement to be about one specific time -- two systems with different qualia in one instance -- which probably is what you meant.)

[silverspawn]

Will spell out reasoning for that, but I think that'll wait till tomorrow

[silverspawn]

So Anatomy of a scandal, which is the show I am still watching, is either about a false rape accusation or about, well, a true rape accusation. I legitimately am still not sure, though I'd take a bet at even odds.

But it made me realize that I'm not actually familiar with any real incident of a confirmed false rape accusation. Ofc I don't watch the news so...

[silverspawn]

I think part of what makes this show good is that the accused feels like a person with integrity. So whether the accusation is false or correct, it's more morally complex than what you typically get in film

[silverspawn]

The Buddha also said, whatever we think a lot will become the inclination of our mind

Which is the more general version of the point yesterday. The unconscious part of the brain is using a bunch of deeply flawed algorithms to decide which things to present to consciousness, and a major parameter is how often they've appeared recently. So briefly thinking about x isn't just a time loss, it's also a training input to the unconscious part of your brain, and if you do it too much, your brain will then helpfully supply x when you're trying to focus on something else.

Kind of amazing that I needed this spelled out. Observing it a million times wasn't enough.

[silverspawn]

So the two system thing. Let's call it the qualia uniqueness problem. Given two processes (m,q) and (m',q') where the first element describes the full causal effect and the second the qualia, does q=q' imply m=m'? If so, this is qualia uniqueness. Conversely, if qualia uniqueness doesn't hold, there exist cases with q=q' but m≠m'.

I thought qualia uniqueness was immediate from Model 3; unfortunately thinking about it more made me realize that it's not. This should also answer how proponents of IIT (which agrees with Model 3) would respond to yesterday's argument; they'd deny qualia uniqueness. In fact, I'm now realizing that the IIT paper had a small part disputing qualia uniqueness explicitly.

Here are the two attempts that I thought would work but don't

- Model 3 says that C. and matter are two aspects of the same thing. That means there's a function f that maps qualia to matter. Most people (at least in my target audience) will assume f is injective, so two different states of qualia will yield two different material descriptions
    - But, you could simply say that the material description of a system includes non-causally-relevant things, and then you can have the above while denying qualia uniqueness. So you have uniqueness for the entire system, but not for processes.
- At least in humans, everything we perceive seems automatically causally relevant because we can report on it and reporting is itself a causal effect. If I notice that something feels different, a causal effect has already occurred. So this seems to imply qualia uniqueness.
    - But, you could argue that this only happens if we explicitly pay attention to the thing that's changed (attention is different from perception), and if we don't pay attention, in fact qualia may not have causal effect, so again qualia uniqueness doesn't hold.

Alas, the entire line of attack looks less useful than I'd hoped.

[silverspawn]

In reality, all of this is bunk because consciousness is a state rather than a process; qualia uniqueness is true. but alas.

[silverspawn]

Apropos to nothing here, but the entire thing where people make fun of safe spaces and trigger warnings is kind of hilarious. Safe spaces and trigger warnings are some of the most fundamental assets of society. Do you value not being exposed to pornography out of context? That is a safe space. would you like to be warned before a movie or speaker discusses mass murder is detail? That is a trigger warning.

Like it's literally about the ability to select what kind of things you engage and don't engage in. It's a very basic form of customization.

Obviously one can be for marginally more or less of those things, but being categorically anti safe space is an admission that you haven't realized what spectrum it lives on.

[faust]

Apropos to nothing here, but the entire thing where people make fun of safe spaces and trigger warnings is kind of hilarious. Safe spaces and trigger warnings are some of the most fundamental assets of society. Do you value not being exposed to pornography out of context? That is a safe space. would you like to be warned before a movie or speaker discusses mass murder is detail? That is a trigger warning.

Like it's literally about the ability to select what kind of things you engage and don't engage in. It's a very basic form of customization.

Obviously one can be for marginally more or less of those things, but being categorically anti safe space is an admission that you haven't realized what spectrum it lives on.

I think these people just don't care about having a consistent belief system. "Yes I am categorically against trigger warnings. Also, you should put 'explicit content' warnings on music that uses swear words."

[Awaclus]

Apropos to nothing here, but the entire thing where people make fun of safe spaces and trigger warnings is kind of hilarious. Safe spaces and trigger warnings are some of the most fundamental assets of society. Do you value not being exposed to pornography out of context? That is a safe space. would you like to be warned before a movie or speaker discusses mass murder is detail? That is a trigger warning.

No as long as there is no risk of someone else seeing me and thinking I'm looking at porn on purpose where it's considered "inappropriate" (e.g. work), and no, respectively. I have had panic attacks from being unexpectedly reminded of a certain mass murder that killed a lot of people I actually cared about even though I didn't know them personally, so I can speak from experience and say that it's very far from being as simple as it would have to be for trigger warnings to be useful.

For example, hearing someone say something like "2013 was such a different time" in a completely mundane context triggered a panic attack for me once because my brain immediately went "YEAH THAT WAS BEFORE THE ARSON ATTACK" and that was too much for me to handle in that moment because it took me by complete surprise. A lot of the other times it was likewise something mundane that just randomly came up and then my brain drew some kind of a connection between it and e.g. some details about the attack I had seen on the news. A few times I have actually been triggered by a trigger warning, even if it's for something else (e.g. rape), because it reminds me I have a risk of getting panic attacks, which of course I can't remember without also remembering the reason why that is.

On the other hand, I can discuss mass murders on Twitter for days and be completely fine, and even bring up the Kyoto Animation arson attack as evidence for my arguments and still be completely fine. I watched the Buffalo shooting video and I was completely fine. I get very emotional (sad or angry) about it sometimes especially if people are suggesting I don't care about victims just because I don't believe banning pistol grips and other cosmetic features on rifles is useful, but emotions aren't harmful, and if anything, I feel like it's helping me that I get to process those emotions sometimes.

The key thing here is that when the topic of mass murders comes up in a conversation or a movie or something, it's almost always predictable without any trigger warnings anyway because there are things that lead up to it, and when it's predictable, it's not going to take me by surprise and I'm probably going to be fine, or if I think there's a chance I might not be, I can stop watching or take a break from it or try to redirect the conversation to a different topic or something. When I open an article and the first thing it says is a trigger warning, that's completely unpredictable (most of the time I'm still going to be fine though because trigger warnings are common enough that I don't usually pay too much attention to them, but there is certainly a substantially elevated risk there, which as far as I can tell is not that unusual of an experience based on what conversations I have had with other people).

I also don't think you should put 'explicit content' warnings on music that uses swear words, and I only do so when the platform forces me. I am in favor of flashing lights warnings, however, and I put them in all of my (and Birds of Necama's) videos if there's a need, which I check with an analysis software.

Safe spaces are whatever as far as I'm concerned, people can have them if they want and I don't see a need for Making Fun of them for it, as long as they aren't trying to make a normal space used by other people into their own safe space.

[silverspawn]

I mean, the things that annoy, harm, etc. different people are extremely varied and can seem ridiculous from the outset. I get legitimate enjoyment out of people viciously insulting me online, but when I see someone write gg after they win, my blood boils. Similarly (I'd say), you have extreme reactions to ostensibly harmless things and no averse reaction to conventionally extreme things.

But this just means that the problem is hard, and you can't solve it perfectly. There's no practical way of avoiding hurting some people sometimes. I don't think "and therefore you shouldn't try" is the conclusion to draw. Maybe in your case, trigger warnings are legitimately so ineffective as to be worse than useless, but there are plenty of people who are scared by conventionally scary things. It remains pretty obvious to me that they're part of a reasonable society. It just makes sense, you're about to show thing x that's hurtful to an unusually large number of people, so you briefly announce that.

[silverspawn]

The focus on swear words in particular does seem completely ridiculous to me, but I don't take that as an indictment of the concept, only one thing that we probably don't need to censor.

[silverspawn]

when the topic of mass murders comes up in a conversation or a movie or something, it's almost always predictable without any trigger warnings anyway because there are things that lead up to it

maybe, but it's not apparent at the start of the movie

[Awaclus]

I mean, the things that annoy, harm, etc. different people are extremely varied and can seem ridiculous from the outset. I get legitimate enjoyment out of people viciously insulting me online, but when I see someone write gg after they win, my blood boils. Similarly (I'd say), you have extreme reactions to ostensibly harmless things and no averse reaction to conventionally extreme things.

But this just means that the problem is hard, and you can't solve it perfectly. There's no practical way of avoiding hurting some people sometimes. I don't think "and therefore you shouldn't try" is the conclusion to draw. Maybe in your case, trigger warnings are legitimately so ineffective as to be worse than useless, but there are plenty of people who are scared by conventionally scary things. It remains pretty obvious to me that they're part of a reasonable society. It just makes sense, you're about to show thing x that's hurtful to an unusually large number of people, so you briefly announce that.

Is a mass murder of people whose creations made my life substantially more tolerable an ostensibly harmless thing?

[silverspawn]

No, but you described how you were reminded of that by ostensibly harmless things

[silverspawn]

My point was that the things that cause the bad reaction are hard to predict

[Awaclus]

No, but you described how you were reminded of that by ostensibly harmless things

I don't think this is very uncommon either. Brains tend to create associations between things.

[silverspawn]

But do you really think that it's so common that trigger warnings shouldn't be used at all?

[Awaclus]

But do you really think that it's so common that trigger warnings shouldn't be used at all?

I really think that trigger warnings are, on average, more triggering than the content they are warning about, and especially in a way that reinforces the trauma as opposed to helping the person heal.