The amount of suffering by animals in the wild is vast. What can be done? My conclusion (see "Activists Should Focus on Public Outreach" at the end) is that it's impractical to undertake direct efforts to curb wild-animal suffering at the moment, but that there is enormous expected value in creating a movement of concern for the problem, both to ensure that our technologically advanced descendants will want to address it and to guarantee that they at least think twice before vastly multiplying the number of wild animals in the cosmos.

How, then, can we best encourage people to care about brutality in nature? I propose some ideas at the end of the above piece, like

* posting on animal-rights forums and writing blog comments;
* writing to influential philosophers, activists, animal-welfare scientists, and other thought leaders to ask for their opinions on wild-animal suffering (Are they aware of people working to address the issue? Would they consider talking about it explicitly in their own work?);
* participating in animal-rights meetups and events and asking attendees what they think (Might they consider diverting some of their time toward outreach related to wild animals?);
* writing conference papers, journal articles, or books on the topic, perhaps co-authored with ecologists, ethologists, or other scientists, to ensure that the work is not entirely armchair philosophy.

This wiki entry highlights much of the same.

I'm curious to hear thoughts from others: What do you think would be the best ways to change the idea that wild-animal suffering matters from a fringe claim into a social value that's taken as obvious? Feel free to contribute thoughts here or on the Reducing-Suffering Wiki.

Suffering in the rest of Nature is indeed horific. Just a quick note on how the question of wild animal suffering should be framed.

If you ask people if it's morally acceptable to feed live rodents to snakes in our zoos, most respondents will say "no". On the other hand, if you ask most people if we should intervene to prevent snakes in the wild devouring live rodents, they'll say "no" too. The contrast is presumably that in the former case we are seen as complicit in cruelty and suffering, whereas in the case of wild animals, any sense of our complicity in their suffering is lacking.

So I think focusing on complicity is important. Massive habitat destruction is increasingly leading to the creation of "wildlife parks" for the purposes of species conservation. Thus the distinction between captive animals in zoos (whose welfare is acknowledged as our responsibility) and wild animals in Nature (whose welfare is usually assumed to be no one's responsibility) is starting to break down. In fact most large terrestrial mammals will probably be extinct outside our designated conservation areas later this century. Either way, canvassing opinion - from both the public and policy-makers - on what level of killing, cruelty and suffering we want in "our wildlife parks" may elicit a different response from questions about suffering in "wild animals".

I say a bit more in
http://www.abolitionist.com/reprogramming/

After looking at your papers I tenatively agree that the question of wild animal suffering seem to dominate that of farm animal suffering, and thus also other more human concerns (my degree of belief is only limited by fact that so few rigorous studies have been done). And as you've pointed out, the question of insect suffering might dominate the question of suffering among other animals. When my planning allows it I hope to be able to read through and comment all your papers carefully.

First, I think it would be wise to wait and gather information before taking action. Of course, your post here (and your previous workon the topic) is an excellent start for such a project. The reason for caution is that our intuitions, personal theories and anecdotal experiences on what may convince people may be unreliably unless supported by evidence from studies. Of course, on a topic as important as this, in the absence of relevant studies, anecdotal experiences etc are far better than nothing. My point is just that action on this cause (as opposed to discussions) should perhaps wait until we have evidential support for the effectiveness of a certain line of action.

Perhaps such caution would be less important when targeting philosophers and scientists, since they might be more easily swayed by pure rational arguments, however.

Second, I think it would be useful to distinguish between different time spans for this goal. Different methods may spread concern for wild animal suffering in the short and long run. For instance, there may be some low hanging fruit among certain animal-welfare supporters that won't affect many more people than a select group in the longer run. At the same time, targeting certain scientists and intellectuals could lead to few converts in the short run but many more in the long.

It seems like the important time-span would roughly be "singularity time". The goal should be that the concern is maximally spread by that time among the public (those people who will be affluent enough to consume the new technology). For the group of people developing the technology (or controlling the development) a somewhat earlier time for maximization might be needed, so that they can have concern for wild animals in mind when doing the development.

Furthermore, aiming at making wild animal welfare an obvious social value may be a too ambitious (short- or medium term) goal. Perhaps it would be wiser to aim at convincing certain target groups, such as people that may likely affect the development of a "singularity".

Third, I think this issue highlights the importance of a more general question: promoting utilitarianism in the very long term. There is a plausible case to made for that the question of existential risk dominates all other more traditional utilitarian concerns. Classic examples of risks are things that have the consequence that humanity (including posthumanity) becomes extinct, or fails to pass a certain level of development. However, for a utilitarian it is not enough to avoid the standard risks to secure an optimal outcome. A non-utilitarian pothumanity may do other things than maximizing pleasure and might even cause cosmic suffering (as you have pointed out).

This shows a danger in the utilitarian focus on (standard) existential risks. If all utilitarian thought and resources would be used to minimize such risks but none to promote utilitarianism (or important utilitarian conerns, such as wild animal suffering) the expected outcome could be as bad or even worse as if the utilitarian resources were wasted on something else than existential risk. Because of this I think a strong case can be made for the promoting of utilitarianism (in the right time-span and groups) may be a question of rival importance to that of reducing (standard) existential risk.

It seems reasonable as David says that focusing on complicity will probably be effective. However, there is the danger that this is more effective in short run than in the long run. As long as people draw an act/omission distinction we are faced with the danger that they will chose not to intervene in cases where they could have prevented massive suffering. On the other hand, some sort of complicity can probably be invoked for most future dangers. For instance, the creation of new universes containing lots of suffering by an agent would presumably trigger feelings of complicity (for instance look at the problem of evil and the complicity mythological gods are often accused of).

* posting on animal-rights forums and writing blog comments;
* writing to influential philosophers, activists, animal-welfare scientists, and other thought leaders to ask for their opinions on wild-animal suffering (Are they aware of people working to address the issue? Would they consider talking about it explicitly in their own work?);
* participating in animal-rights meetups and events and asking attendees what they think (Might they consider diverting some of their time toward outreach related to wild animals?);
* writing conference papers, journal articles, or books on the topic, perhaps co-authored with ecologists, ethologists, or other scientists, to ensure that the work is not entirely armchair philosophy.

1. I'm unsure about the use of posting on animal-rights forums and writing blog comments. From my experience forum discussion is usually very time consuming and seems to have limited results (Felicifia would be an exception due to the much higher level of discussion.) But a smaller initial investment in making some posts and watching the reactions would probably be justified, for purposes of gaining further knowledge on the usefulness of this type of intervention. In case you have tried that, what were the results?

2. Your second point seems very interesting. I feel that it could be useful to get the issue known in the philosophical community, and hopefully make it a topic raied at ethics and environmental ethics classes. Probably people who take such classes are more likely to be persuaded by argumentation on a topic like this. This should also go in a higher degree for the thought leaders themselves. The cost to do the intervention seems very small also. Have you written to anyone yet, any feedback?

3. Again, your third point seems more time consuming, but could be initially useful to gain more information on it's own effectiveness.

4. Also costsome, but could potentially have big consequences. It seems that 2 should preceed the implementing of 4 to get information about the receptiveness to the idea in the fields in question. Do you think non-armchair philosophy is more influential? In which circles?

Hi, I agree with all that Alan writes in his post.

I would like to comment a part of the Jesper's answer:

"1. I'm unsure about the use of posting on animal-rights forums and writing blog comments. From my experience forum discussion is usually very time consuming and seems to have limited results (Felicifia would be an exception due to the much higher level of discussion.) But a smaller initial investment in making some posts and watching the reactions would probably be justified, for purposes of gaining further knowledge on the usefulness of this type of intervention. In case you have tried that, what were the results?"

My background is in the animal rights movement (not in the utilitarianism), and I write about this topic in forums, my blog (sorry, in Spanish!), Facebook... I think it's cost effective, because there are some vegans that can to accept this issue.

Writing some words about the topic (perhaps Twitter is very cost-effective) you can to know interested people. After that, you can send articles to them.

Anyway, animal rights activists can be the target of people like me (as an animal rights activists), and utilitarians can be the target of people like you.


"3. Again, your third point seems more time consuming, but could be initially useful to gain more information on it's own effectiveness."

Well, I think it's cost effective if you are well-know in the AR movement, but it's less cost effective if you are a "strange utilitarian guy".


So, you utilitarians, go for the other utilitarians; we AR activists, go for the ARists.

Hi Daniel,

That seems like a useful division of labor. Great to hear from an AR-activist who is engaged in this question .

How would you estimate the sentiments in the animal rights movements you have experience of when it comes to suffering among wild animals (by "natural" causes)? Are you alone, or are you many like-minded? Is there resistance to the idea of caring about wild animals suffering from non-human causes or is it more that people haven't thought about the problem?

Hi Jesper,

Intervention in the wild could be accepted for die-hard anti-speciesists, but the AR movement is usually very linked with the rights ethical theory. Most anti-speciesists are rightists, and utilitarianism is usually more accepted for speciesist welfarists. And although someone can be rightist and still are against predation (for the same reason than a rightist can be against kill humans), that's very uncommon.

Steve F. Sapontzis is an animal rights philosopher that wrote supporting to avoid the predation when it was possible, in an article ("Predation"), that was published after as a chapter of his book Morals, Reason, and Animals. There is a French journal, with an utilitarian approch and some articles about the horrors of nature. And I know a few Spanish and South-American activists interested.

Most people in the AR movement dislike this idea, but things can change in the future. A problem is a lot of people think we are ecologists, and even there are animal rights advocates that defend an alliance with ecologists. So it's a priority to direct the AR movement to a rationalist anti-speciesist, non-ecologist approach.

Thanks for the great discussion, everyone!

davidpearce, I like your point about differing intuitions over wildlife parks vs. pristine nature. Still, I'm not convinced that most wildlife (especially small animals like fish, frogs, and insects) will live in wildlife parks at any point. Even in cities, there are plenty of "wild" cockroaches and pigeons! And as you often point out, a large fraction of wildlife live in the oceans, where parks are unlikely any time soon.... Nonetheless, even as just an intuition pump, the point is a great one.

Jesper,

My point is just that action on this cause (as opposed to discussions) should perhaps wait until we have evidential support for the effectiveness of a certain line of action.

Indeed, you're very right about this. Hard statistics on the kinds of approaches that historically have led to the greatest long-term change would be ideal. I would just point out that there's always a tradeoff between "do more planning / research" and "find more supporters." If we first undertake suboptimal efforts to identify a few more people who care about the cause, they can help with the more systematic research phase. This is the type of resource-allocation tradeoff that any organization faces. Indeed, it applies to armies and empires, as well: Do you build up your weapons and manpower with your current resources, or do you go conquer more territory first to capture even more raw materials?

Furthermore, aiming at making wild animal welfare an obvious social value may be a too ambitious (short- or medium term) goal. Perhaps it would be wiser to aim at convincing certain target groups, such as people that may likely affect the development of a "singularity".

Again you make an excellent point. Since singularity-type change, if it happens, would be vastly more important than any non-singularity type change that humans might bring about, the expected value of focusing on singularity-influencers does seem high.

Because of this I think a strong case can be made for the promoting of utilitarianism (in the right time-span and groups) may be a question of rival importance to that of reducing (standard) existential risk.

Exactly right. It's the tradeoff I point out in this post on SIAI: Should we focus on promoting general knowledge (or preserving humanity), or is it more important to promote the right memes so that smarter / non-extinct humans eventually do make the right decisions? The space of goals that an advanced civilization might advance is indeed vast (including, as you suggest, the goal of promoting suffering in some way), and it's not obvious to me that post-humanity-minus-utilitarian-memes will end up in a net favorable memetic state.

But a smaller initial investment in making some posts and watching the reactions would probably be justified, for purposes of gaining further knowledge on the usefulness of this type of intervention. In case you have tried that, what were the results?

Just a few examples:

* Vegan Society of Austria
* Veggieboards
* Scienceblogs.com
* RichardDawkins.net
* Accelerating Future

The cost to do the intervention seems very small also. Have you written to anyone yet, any feedback?

Not yet, but I agree the approach sounds promising. I have written to some professors who study ecology or animal welfare to see what research has been done on the subject of "wild-animal welfare," applying the same sorts of standard techniques that are used to assess the welfare of lab and farm animals. In most of the replies I've been told that they know of little such research being done....

It seems that 2 should preceed the implementing of 4 to get information about the receptiveness to the idea in the fields in question. Do you think non-armchair philosophy is more influential? In which circles?

Again probably right. Non-armchair philosophy is more influential in at least public-policy circles (e.g., advocacy organizations on Capitol Hill, Wall Street Journal readers, etc.), I would imagine. The importance of influencing them depends on how much their support matters relative to the philosophical fringes of society. This raises the fundamental question, I guess: Who will be the big players in post-human meme shaping, anyway?

Finally, Daniel, thanks for your insights. A division of labor does seem to make sense. I wonder, how much influence might the AR movement have on future human values relative to utilitarians? I wouldn't be surprised if it's much larger, though of course the fraction of the movement that could be persuaded to change its mind is probably much smaller than the fraction of utilitarians.

Hi Alan. Yes, I think the AR movement will be more influential than utilitarians on future human values. Influential for good or for bad.

I think a charity with a message and work acceptable for both utilitarians and ARists makes sense. Perhaps ARists wouldn't fall in a die-hard rights ethics (forgetting the suffering of wild animals), and utilitarians wouldn't fall in speciesism (forgetting the suffering of all the non-human animals).

Although a charity could have a message acceptable for both utilitarians and animal rights advocates, perhaps this "no man's land" feature causes than only a few utilitarians and animal rights advocates would support it because they saw the project as too soft. If we want a world with less suffering, perhaps a better strategy would be to have two types of charities:

1. Utilitarian charities that take present and future farmed and wild animal suffering into account.
2. Animal rights charities that take present and future wild animal suffering into account.

I would like to know the opinion of the utilitarians here. Do you think a lot of utilitarians would support a type 1 charity? Do you think a lot of less utilitarians would support a charity with a more anti-speciesist explicit message?

broadly labelled social movements
If you asked a historian - although I'm not one - I think they'd say you need to label your social movement broadly. The title needs to absorb people from all kinds of backgrounds. Ideally, it can even coopt some of those you are opposed to on most other topics. For example, secularism is a banner that theists can unite under. That prevents alienation. Then, religious people can be persuaded by religious secularists. atheists can be persuaded by atheist secularists, and so on. I suppose that does little do address the topic of how the organisation should be structured though.

Utilitarian animal suffering charity vs. communication of utilitarianism itself
I don't know whether utilitarians would support a utilitarian wild animal charity. My own personal view is that bioengineering wild animals and abolitionism are charities that we should not be donating to yet. suppose bioengineering of wild animals is 1% likely to ever become viable. Should we try to convert wild suffering abolitionists or utilitarians? Well utilitarians are moderately likely to become wild-suffering abolitionists on their own. In the other 99% of possible future scenarios, including completely unforseeable ones, the utilitarians will do us proud.

So I think that speculative futuristic philanthropy is outweighed by the promotion of utilitarianism. More specifically though, I think that the promotion of utilitarianism should exclude wild animal charity, SIAI, abolitionism, existential risks and so on. Ethical treatment of animals, climate change and help for the poor is more than a hard enough sell already. Something like transhumanism would perhaps be worth selling, perhaps not. You could regard it as the borderline. Those fields that are more speculative and unintuitive than it need not be marketed.

It's also not obvious how valuable academic research into futurism-esque fields actually is, since a) it's partially self-referential, b) it doesn't exactly have a proud history and c) it's often almost pure philosophy, hence unfalsifiable and difficult to judge its practitioners' merits.

With two or three immediate, virtually unprecedented global risks facing us in the very near future, it doesn't seem like addressing them directly would go that far wrong, whereas research into supposed alternative uses of our money might provide answers so speculative that it amounts to throwing good money after bad.

It's also not obvious how valuable academic research into futurism-esque fields actually is, since a) it's partially self-referential, b) it doesn't exactly have a proud history and c) it's often almost pure philosophy, hence unfalsifiable and difficult to judge its practitioners' merits.

This will need a more thorough discussion at a suitable time, but for now a couple of quick ideas. a) Could you expand on this? The problem that the futurism-acitivity itself might affect the actual outcome or something else? b) This is a useful criticism. It seems "intuitively" correct. But I am unaware of any rigorous studies on the issue. c1) In what way are complaints about unfalsifiability themselves falsifiable? Are they not based on pure philosophy? c2) The merits of the people or of the work produced?

With two or three immediate, virtually unprecedented global risks facing us in the very near future, it doesn't seem like addressing them directly would go that far wrong, whereas research into supposed alternative uses of our money might provide answers so speculative that it amounts to throwing good money after bad.

A. Which risks do you include among these three? B. This seems a bit like loss-aversion. I also sometimes feel the pull of wanting to invest money in saving lives here and now, the wish to do something that will surely cause good rather than risking to throw it all away. It can even seem silly to invest money But by utilitarian standards this reasoning seems irrational. As has been discussed many times, the expected utility of something with a probability even as low as 1% or even much less is justified if the stakes are high enough.

Now, I don't advocate donating money to SIAI at present. But they seem as a possible candidate for being an optimal target for donation. I think more information on the topic discussed is needed before that. Because of the high stakes such information should be very valuable.

Dani, I'm glad you like the idea of an anti-speciesist SIAI research project -- I obviously do as well. As far as your question, I think it's actually the case that most utilitarians tend to support the factory-farming charity over a future-wild-animal-suffering charity. Look at Peter Singer, for example, or all the other utilitarians (including on this forum) who support New Harvest and Vegan Outreach.

As far as, Do you think a lot of less utilitarians would support a charity with a more anti-speciesist explicit message? I think RyanCarey's comment provides good insight: Many utilitarians would probably regard a more anti-speciesist message as unnecessary, because anti-speciesism is already built in to utilitarianism (for the most part -- with the exception of, say, most welfare economists!).

I personally would support a more anti-speciesist message, though, because I think the goal of making people full utilitarians is probably too difficult for mainstream audiences. It is possible to convince a lot of people to care a little more about animals, but the number of people you'll find who will adopt utilitarianism full-hog is very small. Of course, we can focus on both types of efforts: Get issue-based support from larger numbers and then full ideological support from smaller numbers. That's the best we can hope for, I think. This is, of course, the classic tradeoff that all movements have to make. For instance, should communists try to spread their entire ideology, or should they settle for smaller issue-based victories that are approved of by the ideological core of the movement?

RyanCarey, I agree we shouldn't support on-the-ground bioengineering at the moment. My idea of a wild-animal charity is one that helps to spread the meme that wild-animal suffering matters. The question of whether to focus on this meme or go for the entire utilitarian meme is the tradeoff question I mentioned in the previous paragraph -- it's not obvious to me either way.

As far as the point about 1% chance of bioengineering being viable, I would suggest that perhaps the main benefit of promoting concern for wild animals is to make people think twice about massively increasing the number of wild animals through terraforming, directed panspermia, lab universes, etc. Of course, these are all themselves rather speculative, but considering just the probabilities here isn't sufficient: We have to compute the entire expected value of working on wild animals vs. working on general utilitarianism. The calculation isn't obvious either way to me.

Why not promote utilitarianism proper, if concern for wild animals is a special case? I guess I'm concerned that many ideological utilitarians actually end up being less utilitarian in practice than many people who subscribe to other ideologies. My impression is that many utilitarians today tend to focus on rather conventional human causes like Oxfam that aren't orders of magnitude more cost-effective than many of the things that non-utilitarian do-gooders support. (Most Oxfam donors aren't utilitarians.) Other ideological utilitarians are interested in philosophy but never get around to doing much at a practical level. Finally, full-hog ideological utilitarianism carries a lot of philosophical baggage that tends to encumber it more than improve it: For instance, theoretically endorsing the possibility of torturing someone if watching the spectacle is sufficiently pleasurable for a large enough group of others.

Arepo, we've discussed previously the tradeoff between academic research vs. concrete and immediate concerns. I'll add one further point. What we're talking about as far as academic research is mainly exploration of issues related to how best to reduce suffering, or promote happiness, or promote positive technological developments. In other words, it's basically doing exactly what we're engaged in right now on this forum: Spending resources (time and money) not concretely acting to help others but just talking about how we can best do so. If it's worthwhile for us to have these sorts of discussions on Felicifia, why might it not also be worthwhile to engage in more comprehensive analysis in the form of academic papers? Perhaps your intuition suggests that the amount of planning that goes on through these forum discussions is sufficient and beyond what we talk about here, the marginal payoff of further analysis drops below the marginal payoff of concrete action. I'm simply someone who views the marginal-payoff curves as intersecting at a slightly different point.

Jesper, I'm curious: What organizations besides SIAI do you think might have higher value? Have you come up with a list of candidates? I would be very interested to see it!

Jesper Östman wrote:

It's also not obvious how valuable academic research into futurism-esque fields actually is, since a) it's partially self-referential, b) it doesn't exactly have a proud history and c) it's often almost pure philosophy, hence unfalsifiable and difficult to judge its practitioners' merits.

This will need a more thorough discussion at a suitable time, but for now a couple of quick ideas. a) Could you expand on this? The problem that the futurism-acitivity itself might affect the actual outcome or something else?

Annoyingly I can't actually remember what I had in mind when I wrote this - somehow I'd assumed it was self-evident. I think I probably meant that given that they claim to merit any funding at all, such organisations clearly have to assess themselves as a potential recipient of funding. Aside from their inevitable bias (which could amount to massive overestimation of the amount of finance they should receive, given that they're theoretically wielding most of the global budget), this seems like an impossible task in philosophical research: how can you reach any reliable estimate of the value of the (non-data-dependent) conclusions you're going to reach without actually having reached them? In other words, insofar as such research is reliable, it's obsolete, since the organisation has already performed its most important task - insofar as it's not obsolete then it's basically guesswork.

b) This is a useful criticism. It seems "intuitively" correct. But I am unaware of any rigorous studies on the issue.

No, me neither. But the media seem prone if anything to overreporting successful predictions, so I doubt we massively underestimate their history.

c1) In what way are complaints about unfalsifiability themselves falsifiable? Are they not based on pure philosophy?

Well, 'falsifiable' is usually an oversimplification, since we're concerned with the strength of reason it might give us for acting. But an example of the kind of thing that frustrates me is Bostrom's simulation argument. We have no way of knowing how likely any of his possibilities are - and if the conclusion is true, everything we think we know about the world is seriously undermined - so it really just seems to me like a restatement of the brain in the vat concept. If we think it's likely, we might want to change our behaviour radically, but we cannot think it either likely or unlikely and we cannot know what to change our behaviour *to*. So for the direct cost of funding Bostrom to produce the paper, and the opportunity cost of him doing something else with his time (or someone else having been educated instead of him), what have we gained?

c2) The merits of the people or of the work produced?

I was thinking mainly of the people, actually, since I think argument from authority is seriously underrated as a potential source of guidance.

A. Which risks do you include among these three?

Peak resources, peak oil in particular, meteor impact, climate change, and more specifically the strain they put on international relations leading to us using existing or imminent technologies to wipe ourselves out (turns out this is a risk even for meteors). Or leading us to destroy civilisation and it never recovering.

B. This seems a bit like loss-aversion.

I disagree. I think we maximise our impact by dealing with problems that we expect soon and have quite a lot of info about now, rather than looking for data to deal with future problems. Even if we don't make the difference between extinction and continuation, we might well improve the global economy enough that we can afford to make up the research difference in other areas later on. One exception might be studies that examine some aspect of the present that future studies couldn't replicate - happiness measurements, for eg. But such studies are relatively cheap, don't require specialised organisations, and don't seem to be the focus of groups like SIAI.

Alan wrote:I guess I'm concerned that many ideological utilitarians actually end up being less utilitarian in practice than many people who subscribe to other ideologies. My impression is that many utilitarians today tend to focus on rather conventional human causes like Oxfam that aren't orders of magnitude more cost-effective than many of the things that non-utilitarian do-gooders support. (Most Oxfam donors aren't utilitarians.)

I think you're being unfair here, Alan. The few utilitarians (and people who often think like utilitarians but might not accept the label) I know in person prefer such causes precisely because they judge they have the highest expected return.

Arepo wrote:We have no way of knowing how likely any of his possibilities are - and if the conclusion is true, everything we think we know about the world is seriously undermined - so it really just seems to me like a restatement of the brain in the vat concept. If we think it's likely, we might want to change our behaviour radically, but we cannot think it either likely or unlikely and we cannot know what to change our behaviour *to*.

What we've gained is a recognition of some underlying logical (or at least probabilistic) connections between ideas, which we might not have recognized before. In particular, we've realized that if post-humans don't go extinct and have motivation to run lots of ancestor simulations, then we're almost certainly one of them. So either our probability of post-human ancestor-simulation running has to change, or our probability of being a "brain in a vat" has to change, or both. If your probabilities were already fully consistent before hearing the argument, that's great! But the point is that for many people, the two probabilities are inconsistent and so they have to revise one or the other.

Arepo wrote:I think you're being unfair here, Alan. The few utilitarians (and people who often think like utilitarians but might not accept the label) I know in person prefer such causes precisely because they judge they have the highest expected return.

My point was that if we want to encourage donations to Oxfam, there are much easier ways to do it than by promoting utilitarianism. As I mentioned, lots of people resist utilitarianism on intuitive grounds (e.g., the torture-can-be-good point), and utilitarianism is a far more demanding philosophy than what's required to get people to donate even significant amounts to charities like Oxfam. Ask the Oxfam fundraising department yourself: What's their recommended strategy for increasing their donation base? I'm almost certain they won't tell you that it involves promoting the utilitarian philosophy!

"such organisations clearly have to assess themselves as a potential recipient of funding. Aside from their inevitable bias (which could amount to massive overestimation of the amount of finance they should receive, given that they're theoretically wielding most of the global budget"

It seems very unlikely that futuristic organisation will have control over most of the global budged in the near future. But perhaps you mean of resources that could be extra usefully used? And yes, overestimation of one's own work seems likely (although this is a problem anyone will have). For a fun criticism of much philosophy from that perspective, see this. This is a good reason for caution in one's estimations of funding for oneself or those similar to oneself. But I can only see it as a reason for the importance of more careful and bias-aware work on Existential Risks, not as a reason for why E-risk research isn't important.

"how can you reach any reliable estimate of the value of the (non-data-dependent) conclusions you're going to reach without actually having reached them? In other words, if such research is reliable, it's obsolete, since you've performed your most important task - if it's not obsolete then it's not reliable."

It seems a similar problem can be generalized to most cases of eg grant writing. How can any researcher know in advance what conclusions they'll reach? One way of solving it would be to do an induction on relevantly similar work. I can predict that the next paper David Chalmers writes will be something a certain group of people will consider insightful (making an important contribution to the issue) with eg probability >40%.

It wouldn't be that hard to look at a representative sample of what SIAI has produced, judge it's importance, make predictions about further work, check the new work etc, see to what extent they are fulfilled, etc.

"No, me neither. But the media seem prone if anything to overreporting successful predictions, so I doubt we massively underestimate their history."

It's also great fun to read about the foolishness of predictions. Probably it would be most useful to compare a class of people using similar methods for prediction.

"But an example of the kind of thing that frustrates me is Bostrom's simulation argument. We have no way of knowing how likely any of his possibilities are"

I agree with you in that this is a useful criticism. It is easy to get frustrated at examples like this. I've been a bit skeptical about SIAI myself since I've thought there's a risk they focus too much of their energy on interesting but ultimately not very relevant paradoxes and technical problems which easily catches the attention of philosophers/logicians/mathematicians. For instance, if we look at what SIAI officially plan on doing, most of it seems less arcane and more useful from a utilitarian perspective http://singinst.org/challenge#grantproposals. In particular, very little of the total time and resources of SIAI/FHI/Bostrom that are spent on the simulation argument.

Furthermore, it is always dangerous to dismiss issues that many persons one consider at least somewhat authoritative are concerned about (as should follow from what you point out below). In particular, a reason for caution is that the mechanisms that make one frustrated over questions like these may be the same as make most people think most plans discussed on Felicifia are "absurd".

"I was thinking mainly of the people, actually, since I think argument from authority is seriously underrated as a potential source of guidance."

I agree very much.

"Peak resources, peak oil in particular, meteor impact, climate change, and more specifically the strain they put on international relations leading to us using existing or imminent technologies to wipe ourselves out (turns out this is a risk even for meteors). Or leading us to destroy civilisation and it never recovering."

Agreed. This is important, and as I've mentioned before there's a risk that futuristically minded people don't give them enough attention since they seem too mundane and boring.

"I disagree. I think we maximise our impact by dealing with problems that we expect soon and have quite a lot of info about now, rather than looking for data to deal with future problems. Even if we don't make the difference between extinction and continuation, we might well improve the global economy enough that we can afford to make up the research difference in other areas later on. One exception might be studies that examine some aspect of the present that future studies couldn't replicate - happiness measurements, for eg. But such studies are relatively cheap, don't require specialised organisations, and don't seem to be the focus of groups like SIAI."

1. And I think more reasearch is needed on precisely this issue, comparing long term (futurism) and short term risks. Disagree? 2. The problem with spending resources at increasing economic growth is that it itself might be increasing existential risk.

Perhaps it couldn't be justified spending a huge part of humanity's resources on investigating futuristic questions. My main point is that barely any resources at all are spent. SIAI for instance is extremely cheap, even compared to something like happiness studies.

Jesper Östman wrote:I agree with you in that this is a useful criticism. [...] In particular, very little of the total time and resources of SIAI/FHI/Bostrom that are spent on the simulation argument.

Hmm, it's not clear to me that the simulation argument is a waste of time. Indeed, I think it's one of the most important questions we can think about. Whether we're simulations (and at what level in a potential simulation hierarchy -- basement level, first-order simulation, second-order simulation, etc.) makes a huge difference to our assessment of how much computational power we're likely to have in the future and so on.

Jesper Östman wrote:Perhaps it couldn't be justified spending a huge part of humanity's resources on investigating futuristic questions. My main point is that barely any resources at all are spent. SIAI for instance is extremely cheap, even compared to something like happiness studies.

Exactly. Changing your mind through new insights offers the possibility of increasing your cost-effectiveness by many orders of magnitude; in contrast, the best efficiency gains you can get with more concrete, short-term projects are a few orders of magnitude, usually much less. I think Gaverick put it well here:

The cost-effectiveness of projects varies widely. Some human welfare projects cost between $1 and $10,000 per life-year. But many of these gains in human welfare are offset by losses in animal welfare, as longer human lives and higher incomes typically lead to a net decrease in the number of animals who exist and a net increase in suffering per animal. Compared to human welfare projects, animal welfare projects are probably more efficient, with costs as low as $0.02 per life-year.

The value of all these life-years is probably small, however. Humans and other animals probably evolved the minimum hedonic capacities needed to motivate fitness. Anything more would waste resources that animals could apply to growth and reproduction. Increasing hedonic capacities is an engineering problem that could be solved with advances in neuroscience. 150 years ago, few believed that anesthesia was possible. It's now possible to undergo invasive surgery not only painlessly, but even euphorically. An entirely euphoric life is no longer implausible. Once we understand the underlying biology of happiness in the brain, people could choose to experience much higher levels of welfare than they do now.

If so, future life-years would have higher values than current life-years. Projects that ensure the existence of future lives could then be most cost-effective. In fact, these projects are cost-effective even if we assume no improvement in welfare. For instance, even if one expects no improvement in welfare, a short human future, and a constant population, projects to deflect extinction-level asteroids cost $0.20 per future human life-year. Nonhumans would also benefit from such projects, as there is an annual 1/billion probability of an asteroid impact that extinguishes all animal life. If one assumes there are now a trillion sentient nonhumans on Earth, the cost of these projects is under $0.00000001 per life-year. And asteroids seem like a small risk compared to others.

Given the range of $0.00000001 to $10,000 among plausible projects, donating an additional 10% of one's income seems much less important than choosing the most cost-effective projects.

By the way, I think Gaverick himself approves of donating to research organizations rather than direct on-the-ground action-oriented charities.

Hmm, it's not clear to me that the simulation argument is a waste of time. Indeed, I think it's one of the most important questions we can think about. Whether we're simulations (and at what level in a potential simulation hierarchy -- basement level, first-order simulation, second-order simulation, etc.) makes a huge difference to our assessment of how much computational power we're likely to have in the future and so on.

Ironically, I've actually held a much higher degree of belief in the simulation hypothesis than Bostrom himself. However, I've been a bit dismissive about considering it as an ER since I've made the (perhaps hasty) judgment that such concerns were based on non-utilitarian concerns about the survival of particular humans or the human race (including posthumans perhaps) as opposed to maximizing utility (note for instance that shutting the simulation down might be a good thing if suffering is more common than equal amounts of pleasure in the universe as a whole - for a negative utilitarian it will be the best outcome).

A related consideration is that a simulator may have good utilitarian reason to shut down the simulation, if it does so. What reasons do we have for believing that we are more effective utilitarians than a simulator? (Ironically, this reasoning reminds me of Bishop Butler's idea of God as the ultimate utilitarian). Fighting against this may thus have negative consequences.

On the other hand, there is of course some risk of a "bad" simulator (similar to a Pascalian God). And perhaps it could be argued that this constitutes an existential risk of similar magnitude to more standard risks.

The "useful criticism" part was meant to qualify the criticism that it is hard to give evidence based estimations of the probabilities necessary for the simulation argument. As opposed to saying something about unfalsifiability

Exactly. Changing your mind through new insights offers the possibility of increasing your cost-effectiveness by many orders of magnitude; in contrast, the best efficiency gains you can get with more concrete, short-term projects are a few orders of magnitude, usually much less. I think Gaverick put it well here

I very much agree

Jesper, I don't dispute many of your points about the simulation hypothesis. However, in order to reach all those conclusions, you had to first consider the simulation hypothesis, work through various scenarios, and evaluate the utilitarian implications. All of that is "research work" of precisely the type I'm suggesting is valuable.

(Maybe in hindsight you could regret spending the time on it. But how would you have known that beforehand?)

Jesper Östman wrote:What reasons do we have for believing that we are more effective utilitarians than a simulator? (Ironically, this reasoning reminds me of Bishop Butler's idea of God as the ultimate utilitarian).

One reason is the problem of evil: Why would a good god create so much suffering?

I. Yes. Still, one has to make choices. Even beforehand. One needs to distribute resources to problems in some suitable way. The hard question is how. How much resources should go into research on sim vs AI vs bio/nano etc?

II. Perhaps that is a reason for that there isn't much suffering? It could be a solipsistic simulation. (Cmp Descartes argument for why animals aren't sentient)

Yup. Number I. is basically the multi-armed-bandit problem.

Interesting. Yes.

Alan Dawrst wrote:Whether we're simulations (and at what level in a potential simulation hierarchy -- basement level, first-order simulation, second-order simulation, etc.) makes a huge difference to our assessment of how much computational power we're likely to have in the future and so on.

It's not the significance of the argument being true that I think it's salient. It's that, like Descarte's demon, we have no idea what the chance of it being true is, and, like the demon, if it is true we don't know anything about the real universe. Its effect on our action should be infinitessimal.

Re animal welfare, I tend to think it's important, but it's the one potentially utilitarian activity that seems like it can't reduce our risk of extinction (perhaps with the exception of in vitro meat), so it's the one thing I feel able to rule out.

Arepo wrote:It's not the significance of the argument being true that I think it's salient. It's that, like Descarte's demon, we have no idea what the chance of it being true is, and, like the demon, if it is true we don't know anything about the real universe. Its effect on our action should be infinitessimal.

I can't quite understand the source of your radical skepticism about the possibility of rationally assessing the merits of the simulation argument. Indeed, I was previously under the impression that you thought the simulation argument failed. But now you seem to be saying that we cannot know at all whether it fails or succeeds. In any case, such skepticism would seem to require having a 0.5 credence in the proposition that the simulation argument is sound, and this would demand taking the implications of that argument very seriously indeed.

Perhaps your position is that, even if we take the argument seriously, this would have no practical implications, since "if it is true we don't know anything about the real universe." I disagree both with the claim that we would lack all knowledge if we thought we were in a simulation, and with the claim that, if we indeed lacked knowledge under such circumstances, this would have no practical implications. Concerning the first claim, we would know nearly everything we now know if we thought we were living in a accurate simulation: knowledge about the simulated reality would then yield knowledge about the real world. Even if the simulation is not accurate, we could make certain inferences about the world that runs the simulation, given knowledge of both the simulated reality and the necessary properties that reality should have in order to allow for that kind of simulation. Concerning the second claim, even if we lacked empirical knowledge we would still have some knowledge. For example, we could know that one hypothesis is more likely true than another on grounds of simplicity.

If I may for a moment engage in some armchair psychologizing, I'd say that your reluctance to take the simulation argument seriously stems from a bias against hypotheses that appear to you to be "silly" in some sense. We are all prone to this bias, but it is imperative that as utilitarians we make a concerted effort to overcome it. Many of the hypotheses that are the most significant from a utilitarian perspective --that our species may become extinct within the next few centuries; that insects may collectively suffer more than the rest of terrestrial sentience combined; that we may accidentally create baby universes with infinite amounts of suffering-- are ignored or ridiculed precisely because they strike man peple as "silly", or unworthy of serious discussion. But lacking a persuasive argument for the view that the silliness heuristic reliably tracks truth, such dismissals are unjustified, and should be challenged.

Arepo wrote:It's not the significance of the argument being true that I think it's salient. It's that, like Descarte's demon, we have no idea what the chance of it being true is, and, like the demon, if it is true we don't know anything about the real universe. Its effect on our action should be infinitessimal.

Arepo, a quick analogy. How do you decide what probability to assign to, say, chaotic inflation theory as opposed to the standard model of inflation? It seems to me you start with prior probabilities for each hypothesis (based on complexity, so that more complex hypotheses start out less likely) and then you update based on evidence you observe.

How is the hypothesis "we are ancestor simulations of past humans by advanced post-humans" different from any other hypothesis? Presumably you have some prior probability for it. Then you can update that probability based on evidence. The hypothesis makes various predictions: For instance, it predicts that humans will not go extinct before achieving simulation capability, that they will eventually acquire large amounts of computing power, that they will eventually discover how to generate subjective experiences of human minds (or at least your mind) through simulations, and so on. Even the basic fact that "we are simulations" of any sort tells us a lot: That the (simulated) universe in which we live is computable, for one thing!

Does it predict that humans should achieve simulation capacity etc as opposed to something more complex, such that we should have a similar probability of achieving it as the simulators had? (Assuming that an ancestor-simulation faithful in this detail would be most probable)

Jesper Östman wrote:Does it predict that humans should achieve simulation capacity etc as opposed to something more complex, such that we should have a similar probability of achieving it as the simulators had?

It depends on exactly how you define the hypothesis; if you say that "ancestor simulations" means precise simulations of the history of the particular post-human simulators, then survival is guaranteed because those post-human simulators did survive. But you're right that arbitrary simulators could introduce randomness (and could even simulate things other than their ancestors).

Pablo Stafforini wrote:I can't quite understand the source of your radical skepticism about the possibility of rationally assessing the merits of the simulation argument. Indeed, I was previously under the impression that you thought the simulation argument failed.

Without wanting to get sidetracked into an epistemology discussion, I think knowledge is poorly defined, and is better thought of as the conclusions of seemingly useful assumptions. Given that, if an argument's conclusion is useless (ie gives us no motivation to change our action) then I find that equivalent to saying the argument fails.

But now you seem to be saying that we cannot know at all whether it fails or succeeds. In any case, such skepticism would seem to require having a 0.5 credence in the proposition that the simulation argument is sound, and this would demand taking the implications of that argument very seriously indeed.

I don't think not knowing the probability means I should have 0.5 credence in it - I have the same attitude towards a malicious demon, the Matrix, a benign god who's just allowing suffering for my own good and any number of such incompatible metaphysical propositions. I can't think them all 50% likely.

Concerning the first claim, we would know nearly everything we now know if we thought we were living in a accurate simulation: knowledge about the simulated reality would then yield knowledge about the real world.

This seems circular. If we don't know about our simulators, we don't know whether or how much their world resembles ours. If we assume things about them, it's no surprise we can then show those things.

Even if the simulation is not accurate, we could make certain inferences about the world that runs the simulation, given knowledge of both the simulated reality and the necessary properties that reality should have in order to allow for that kind of simulation.

I can't imagine anything important we can know about a world based on its having simulated this one. Can you give some examples?

Concerning the second claim, even if we lacked empirical knowledge we would still have some knowledge. For example, we could know that one hypothesis is more likely true than another on grounds of simplicity.

Perhaps, but again I can't imagine any such inference that would actually be relevant to our behaviour. Can you provide examples again?

If I may for a moment engage in some armchair psychologizing, I'd say that your reluctance to take the simulation argument seriously stems from a bias against hypotheses that appear to you to be "silly" in some sense.

It's possible, but conversely, philosophers tend to be biased towards ideas that seem silly (see the list in my first paragraph of this post, for eg). In fact they have an extra financial incentive for being so, since if they start saying 'the world is pretty much the way it appears to be', they'll pretty quickly be out of a job - whereas I'm (financially) disinterested.

I don't see what evidence we have that the other is more heavily influenced by bias on an issue where neither of us is persuaded by the other's argument.

Arepo wrote:

Pablo Stafforini wrote:I can't quite understand the source of your radical skepticism about the possibility of rationally assessing the merits of the simulation argument. Indeed, I was previously under the impression that you thought the simulation argument failed.

Without wanting to get sidetracked into an epistemology discussion, I think knowledge is poorly defined, and is better thought of as the conclusions of seemingly useful assumptions. Given that, if an argument's conclusion is useless (ie gives us no motivation to change our action) then I find that equivalent to saying the argument fails. (Alan, hopefully this answers your question - I can't speak on a theory I know nothing about, but in principle I'm interested in it strictly on its unique compatibility with the evidence. If other theories are equally compatible, I see no reason to distinguish between - or pay any attention to - any of them, except insofar as they tickle my imagination)

But now you seem to be saying that we cannot know at all whether it fails or succeeds. In any case, such skepticism would seem to require having a 0.5 credence in the proposition that the simulation argument is sound, and this would demand taking the implications of that argument very seriously indeed.

I don't think not knowing the probability means I should have 0.5 credence in it - I have the same attitude towards a malicious demon, the Matrix, a benign god who's just allowing suffering for my own good and any number of such incompatible metaphysical propositions. I can't think them all 50% likely.

Concerning the first claim, we would know nearly everything we now know if we thought we were living in a accurate simulation: knowledge about the simulated reality would then yield knowledge about the real world.

This seems circular. If we don't know about our simulators, we don't know whether or how much their world resembles ours. If we assume things about them, it's no surprise we can then show those things conditional on our assumptions.

Even if the simulation is not accurate, we could make certain inferences about the world that runs the simulation, given knowledge of both the simulated reality and the necessary properties that reality should have in order to allow for that kind of simulation.

I can't imagine anything important we can know about a world based on its having simulated this one. Can you give some examples?

Concerning the second claim, even if we lacked empirical knowledge we would still have some knowledge. For example, we could know that one hypothesis is more likely true than another on grounds of simplicity.

Perhaps, but again I can't imagine any such inference that would actually be relevant to our behaviour. Can you provide examples again?

If I may for a moment engage in some armchair psychologizing, I'd say that your reluctance to take the simulation argument seriously stems from a bias against hypotheses that appear to you to be "silly" in some sense.

It's possible, but conversely, philosophers tend to be biased towards ideas that seem silly (see the list in my first paragraph of this post, for eg). In fact they have an extra financial incentive for being so, since if they start saying 'the world is pretty much the way it appears to be', they'll pretty quickly be out of a job - whereas I'm (financially) disinterested.

I don't see what evidence we have that the other is more heavily influenced by bias on an issue where neither of us is persuaded by the other's argument.