This topic is vague and open-ended.  I'm leaving it that way deliberately.  Perhaps some interesting, better defined topics will grow out of it.  Or perhaps it's too far afield from the concept of less wrong cognition to be of interest here.  So I view this topic as exploratory rather than as an attempt to solve a specific problem.

What useful purposes does religion serve?  Are any of these purposes non-supernaturalistic in nature?  What is success for a religion and what elements of a religion tend to cause it to become successful?   How would you design a "rational religion", if such an entity is possible?  How and why would a religion with that design become successful and serve a useful purpose?  What are the relationships between aspects of a religion, and outcomes involving that religion?  For example, Catholicism discourages birth control.  Lack of birth control encourages higher birthrates among Catholics.  This encourages there to be a larger number of Catholics in the next generation than would otherwise be the case,  Surely there are other relationships like this?  How do aspects of religion cause them to evolve differently over time?

New to LessWrong?

New Comment
32 comments, sorted by Click to highlight new comments since: Today at 9:35 AM

If you use "engineering" in the title, you're going to have to be much more precise. What goals and measurements are you interested in? Why are you considering "religion" as a unitary thing, distinct from other belief and social conformity mechanisms?

What is success for a religion?

I'd say it seems pretty fundamental that Religion is a kind of meme. One measure of success for memes is their ability to spread (virulence). If that is your only measure, you're likely to have some ethically terrible things going on. It seems (to me) like an obvious constraint that the meme must spread without causing any obvious harm in its wake (except to related memes that may be in competition).

What useful purposes does religion serve?

I consider [create an in-group] to be a pretty central (usually unstated) optimization point (not exactly a purpose) in most religions. This has some well-studied psychological effects on said group that can benefit all its members socially, psychologically, and (further down the causal line) medically. Although it tends to lead to some unfortunate side-effects for the out-group.

How would you design a "rational religion", if such an entity is possible?

Problem is that I've seen more than one source define "religion" as something like "systematic belief in the supernatural". I'm not convinced that such belief can be "rational" (optimal for making more effective decisions). Perhaps as a stepping stone -- use a religious-type belief system as an infection vector, then strip the supernatural elements away bit by bit; but that seems awfully deceitful to me.

This seems like a promising direction. So let's say that by religion I mean a useful meme. The meme consists of a doctrine. That is, a collection of statements regarding human belief or action. A person infected with the meme lives in accordance with the doctrine. The doctrine is designed to cause a useful effect, apart from its own flourishing. The effect is caused by changes in behavior of the people infected by the meme. The effect need not be explicitly stated within the doctrine. This a broad definition of religion, as it doesn't require that the doctrine contain any statements about the supernatural. But at this point I want to keep things open. As some posts have suggested, there's a lot of argument about what does or doesn't constitute a religion.

In order to be successful, a useful meme must be sufficiently:

  1. Transmissible, meaning that it must be pragmatically possible to expose many people to the meme.
  2. Infectious, meaning that a significant number of people exposed to the meme must become infected.
  3. Contagious, meaning that it must spread of its own accord, until it has infected a threshold number of carriers. The threshold may differ from meme to meme, and is whatever threshold is needed to enable the successful creation of the desired useful effect.
  4. Viable, meaning that it must be possible for a person to survive and prosper when infected with the meme.
  5. Robust, meaning that once a given person is infected, it must resist eradication within that individual.
  6. Hardy, meaning that it must be able survive, and possibly flourish, in a variety of intellectual environments, and there must be no commonly found environment in which it cannot survive.
  7. Resistant. Meaning that beyond a certain degree of spread, there cannot be an obvious method by which a (presumably non-infected) person or group can eradicate it.
  8. Stable, meaning that it must not change its nature significantly over time. Changes that continue to allow it to flourish and to cause the desired effect are permissible.

This list is not meant to be definitive or exhaustive. And I don't claim to be using the best terminology.

For example, some religious doctrines contain the idea that if you cease to believe in any part of the doctrine, you will suffer in Hell upon death. This idea might enhance the robustness of the religion by discouraging the development of disbelief. Others contain the idea that it's your duty, or that you're rewarded in some fashion, for converting non-believers. This idea might enhance the contagiousness of the religion by encouraging those who are already infected by the meme to work to infect others.

Using this framework, perhaps the original post might be improved a bit. Putting some of the questions asked in the original post into the new framework, we get:

  1. What useful purposes does religion serve? That is, what kinds of "useful purposes" can be designed into such memes?
  2. Are any of these purposes non-supernaturalistic in nature? That is, can a completely non-supernatural religion flourish and create a useful effect?
  3. What is success for a religion and what elements of a religion tend to cause it to become successful? That is, what elements of currently existing religious doctrines have helped them to become successful?
  4. How would you design a "rational religion", if such an entity is possible? That is, is it possible to design a religion that encourages "less wrong" cognition?
  5. What are the relationships between aspects of a religion, and outcomes involving that religion? For example, Catholic doctrine includes elements that discourage birth control. Lack of birth control encourages higher birthrates among Catholics. This encourages there to be a larger number of Catholics in the next generation than would otherwise be the case. Thus the toolkit now contains one element: the prohibition of birth control as a means of increasing the contagiousness of a religion.
  6. How do aspects of religion cause them to evolve differently over time? For example, Catholicism contains a permanent, authoritarian hierarchy of individuals who are dependent on the church for their survival and satisfaction. Without the Church, a Cardinal with no other job skills might starve or suffer disrespect from the community. Thus he is incentivized to increase the authority of the Catholic church in order to help secure his own survival. Over time, Catholic doctrine adopted the idea that the Pope is infallible. It might be possible to draw a line between the two phenomena and say that in a doctrine which supports a social structure that includes a permanent, authoritarian hierarchy of individuals who are dependent on believers for their survival and satisfaction, the doctrine will tend to evolve towards the accumulation of power for the hierarchy so as to ensure and increase their survival and satisfaction. Possibly another element for the toolkit.

What other such heuristics exist? Would a large enough collection of such heuristics aid in the analysis and design of religious movements?

I think Effective Altruism fits the eight criteria you gave.

I don't think rationality!CFAR currently has all eight at the moment but I think there a good chance that it will get them in the future.

In both cases calling them religions is likely not helpful.

I'm doing a little reading on both of them now. Big question: how to make them successful as social movements. I wonder if their elements can be modeled in a fashion similar to that which I did above. And if so, if there's anything that such an application can tell us about how to improve their chances for success.

I don't think it's even helpful to model them in that way. Your list for example doesn't contain a point for "community". It doesn't consider the need for charismatic leaders.

There not much reason to focus on issues like contraception because a good new ideology can spread much faster if the community can recruit well and give new members a lot of value.

This sort of problem is susceptible to the red queen's race -- if you solve it, that just means that lots of people will copy you, and you will find yourself in competition with a new crop of viral religions. You don't make much long-term progress by being innovative, so perhaps your best bet is to copy some moderately successful cults and hope that power won't corrupt you. The Hare Krishnas are a good template -- just avoid the racketeering, mail fraud, murder, and child abuse, and you should do fine.

I'm new here and not sure exactly what you expect when someone links, but it seems like you guys are generally intelligent so:

http://www.enotes.com/research-starters/sociological-theories-religion-structural

It sounds like what you're asking (with regards to the function of religion) is something that has been covered a great deal by the structural-functionalist sociological approach. If you're willing to read up on it there's a lot of information out there on the topic. Hope that helps! If you'd prefer I answer your question on here more directly feel free to ask, I'm in the middle of finals and haven't read up too much on the topic myself so I'd have to do some research before getting back to you.

but it seems like you guys are generally intelligent

Nah, we just fake it :-P

I feel like intelligence is similar to logic or grammar and faces the dunning kurger effects full force essentially. As they state in the abstract of their work: Their lack of skill deprives them not only of the ability to produce correct responses, but also of the expertise necessary to surmise that they are not producing them.

If you're able to "fake" being intelligent, you require the ability to produce the "intelligent" response, and the ability to recognize when you're not being intelligent. So if you don't have it, you can't really fake it... I mean, unless you're moderately skilled and meticulously research and craft your responses specifically for effect, but even then that means you're able to do so effectively...

Drats, foiled again!

:-)

Religion serves numerous purposes, some of which have been mentioned already by other commenters. I want to add two others:

1) a mechanism for preserving Intersubjective Truths, that is, truths that it is not possible to re-derive from first principles in a reasonable amount of time.

2) a connection to the spiritual side of life and spiritual experiences.

Also, with so many different purposes it is tempting to design different religions to fill all these roles. I suspect that is harder than it seems. Since whatever fills at least some of these roles will attempt to expand to fill all of them.

Since whatever fills at least some of these roles will attempt to expand to fill all of them.

In the real world I don't see that any religion tries to fill the same roles. Christianity doesn't try to expand to fill law-making in the same way that Islam does.

Christianity doesn't try to expand to fill law-making in the same way that Islam does.

The existence of the Vatican state today, and that of the Holy Roman Empire in Medieval times, I think proves you wrong.
Christianity has always attempted to impose itself in law-making and state politics, and was very good at doing so for almost two millennia. It just has lost (arguably) in Europe and in America in the last century, although it's still fighting for power whenever and however it can.

The existence of the Vatican state today, and that of the Holy Roman Empire in Medieval times, I think proves you wrong.

No, I don't think they do. Vatican is a state only in the name, and as the famous quip about the Holy Roman Empire goes, it was neither Holy, nor Roman, nor an Empire :-)

The relationship between Christianity and state power is a complicated one and not reducible to a yes/no question.

I also think that if you replace the wiggly "impose itself in law-making and state politics" with much more direct "seek to rule, that is, solely control the state power", you'll find that Islam and Christianity are quite different in that respect. The notion that Islam and Christianity have essentially same attitude towards state is nonsense.

The existence of the Vatican state today, and that of the Holy Roman Empire in Medieval times, I think proves you wrong.

The Holy Roman Empire had a deal whereby you have worldly rulers on the one hand and the pope being the spiritual leader on the other hand. Popes supported the concept of the devine right of kings that was supposed to be given out by God to King David and passed down from David via inherentance to the present kings.

Islam on the other hand considers only God a valid source of laws and doesn't think that God has given human kings the right to make laws the way Christianity thinks with the devine right of kings.

If you think that the devine right of kings wasn't important in Medieval times I think you are massively misreading history.

Disclosure: I haven't read the full string of comments

I'm pretty sure you're a bit off on the Islamic side of things though.

a valid source of laws and doesn't think that God has given human kings the right to make laws the way Christianity thinks with the devine right of kings.

Kings' claim to rule seems to be fairly similar to that of an Islamic caliph, who are supposed to be prophets selected by God himself, and are able to create laws/etc. as he would want basically.

Kings' claim to rule seems to be fairly similar to that of an Islamic caliph, who are supposed to be prophets selected by God himself, and are able to create laws/etc. as he would want basically.

No. The core idea of Islam is that a caliph is not a person who can create laws and if he tries to do so his followers have an obligation to fight him over it. God gave humanity the sharia and the caliph can't simply deviate from it because he thinks that it would be better if man and woman have the same right as far as inheritance goes.

The caliph can make rulings that interpret the koran (fatwa's) but he can't make laws.

Worldly issues such as alimony rules and inheritance rules are hardlocked in the sharia. In our cities with sizable Muslim populations you have Muslims running parallel legal system. In Berlin we have problems that arise from German law considering alimony to be a regular payment while sharia law considers it to be made in a large one time payment.

That conflict of alimony payments is something that leftish Muslims in Berlin worry about because their religion dictates them to live differently than German law.

Christianity doesn't work that way. If you take the Western idea of what religion happens to be, that not the cluster filled by Islam. Islam adds the sharia with doesn't have an equivalent. While not every Muslim believes in cutting of hands, the civil rules for alimony and inheritance are a core part of Islam as practiced by liberal Western Muslims.

Apart from the issue of lawmaking the caliph is the religious leader of his nation while a king isn't.

I think this is correct, but there is a difference even so. Islam was something like a state religion from a very early point in its history, and there is much in the Qur'an and hadith that reflects this. Christianity got started among mostly-poor mostly-powerless subjects of the Roman Empire, and the New Testament reflects this. It's only hundreds of years later that it became a state religion.

I suspect (but don't know) that this makes it easier to be a reasonably conventional Christian without feeling that your religion should be in control of the state, than to be a reasonably conventional Muslim without feeling that your religion should be in control of the state.

Islam was something like a state religion from a very early point in its history

Not "something like", but "a classical full-blown dialed-to-eleven" state religion. Note, for example, that early Islam knows no distinction between religious laws and state laws. The idea that they could be different would be treated as an idiocy.

As kithpendragon has already wrote, religions from a memetic point of view must contain tools for its propagation and defence, such as ways to infect other agents and fight aggressions.
I also want to propose another question: the purposes served by religions are better served by what other social constructs?

the purposes served by religions are better served by what other social constructs?

The secular parallel to religion is called an ideology.

What does "better" mean?

I guess it depends on the purpose. If I was designing a new religion, I could ask: a religion spreads through such and such mechanism, how we could design a faster spreader?
I'm not pointing at any particular direction, I was just saying that it's a better question than simply analyzing the state of the art, so to speak.

it's a better question than simply analyzing the state of the art

Yes, I agree. But still, the answer very much depends on the specific criterion you're interested in. Social constructs that are extra virulent (spread fast) will be different from social constructs that are, say, extra robust (survive long).

One way of defining things is when you inherently let signifcant parts of your life be dictated by trust. Normally there are requirements to earning trust such as being good in a field of knowledge to be able to function as a expert regarding such subject allowing others to skip details.

You can obvoiusly enter into dynamics that other dynamics are not allowed to move into such easily. Science or courts need some level of proof. Military power is dependent on weapon production economies. If you just trust it will be a good thing if you do it you don't need any of those complications.

For example a trust based approach to prisoners dilemma can just have the players cooperate while agents wanting to be game theorethically competent must make up all kinds of excuses to cooperate or suffer contraoperation.

I am really not a sociologist, so someone correct me if I what I'll say is totally wrong, but it seems to me that there are at least two quite distinct types of religion (and a continuum of possibilities in-between), the first one consisting of those religions where "religion" religion (gods, clergy, etc.) is almost one and the same thing as something like civil religion of a community (for example, if you found out that a tribe adds various religious chants to their local "judical process" which otherwise is very similar to a Western judical process you would not hesitate to call it a religious ritual, even though chanting part may be inessential), and another where those are two different things. In my mind the first type roughly corresponds to paganism, and the second one to religions similar to Christianity. I think that religions of the first type may be useful to "grease the wheels" of society (especially not a very sophisticated one), and its leaders may not even be that interested in spreading it, except for personal gains. However, note that this is a vague guess, I would need to read much more about pagan societies to understand if it is at least partially the case. It is unclear to me what does religions of the second type do, because greasing the wheels of society is covered by a civic religion.

Another possible benefit of some types of religion is providing some incentives to get some (although not all) things correct. If you believe that god judges you whether or not you your thinking about the world is correct, you might feel and behave as if you have "skin in the game" and thus you might end up with higher motivation to avoid deceiving yourself (and others) for personal gain [1]. In many contemporary societies personal belief in god seems isolated from most beliefs that have practical consequences. If your society practices trial by combat, and your belief in god makes you willing to fight against a stronger person believing that you will win just because you are in the right, your belief will have practical personal consequences. However, in a contemporary society belief in god is usually harmful only indirectly. Thus, a question is, does feeling that you have "skin in the game of believing the truth about god's creation" lead to enough correct beliefs so as to outweigh having incorrect beliefs about god? I think that it is likely that at least in some cases might do. In addition to that perhaps in some cases religious beliefs might be personally helpful if they are harmless and they displace potentially dangerous beliefs who are new and have not yet shed their most extreme parts.

I think that in both cases religion/belief in god is not strictly a necessity, but in some cases it may (or may not) turn out to be somewhat useful if there are no better alternatives available at that time.

[1] I don't like it. Also, I guess that it is probably not very good for society in the long run.

A successful religion must serve Moloch, and your designer choice is how much and what exactly are you going to sacrifice first.

Yeah, I realize this is a fully general argument against anything. Maybe the reasonable part is that I wouldn't want to compete at a religion market. I'd rather stay at a level of communities. Yeah, seems like a debate over terminology, but those two words have widely different connotations for me. In communities people stay freely; different communities are not necessarily enemies; etc.

A successful religion must serve Moloch, and your designer choice is how much and what exactly are you going to sacrifice first.

Not a bad first approximation, however, let's examine Gnon, or Moloch if you insist on that terminology, a little bit first. Notice that of the four of Gnon's sub-processes all but Cthulhu naturally, if in some cases rather brutally, converge towards making people believe true things or at least having an accurate working model of local reality.

Cthulhu is different, it causes people to engage in signalling completions that may very well result in them competing to believe ever more false things. In a way this signals "I'm so high status I don't need to have an accurate model of what the peasants are doing". One thing religion can do, when it's working right, is keeping a throttle on this kind of status competition, by accusing anyone who starts saying anything too outlandish of heresy. Of course, there are several ways this can go wrong. For example:

1) some outlandish thing may turn out to be true.

2) once it ceases to be conservative (as is happening with the quasi-religion of "social justice") it starts accusing anyone saying insufficiently outlandish gets accused of heresy.

Could you define what you mean with the word religion?

Unfortunately, I can't give a good definition for this except by example. Christianity, Judaism, Islam, Buddhism, Daoism, Hinduism, Atheism. I suppose that it might even be possible to ask the questions assuming that science is a religion. The focus of the questions is intended to be on the engineering and social aspects, rather than on a question like "Should Atheism be considered a religion?" I understand that the vagueness makes this a less than perfect delineation of a topic.

The focus of the questions is intended to be on the engineering and social aspects, rather than on a question like "Should Atheism be considered a religion?" I understand that the vagueness makes this a less than perfect delineation of a topic.

Atheism shouldn't be thought of as a (single) religion for the same reason non-apples aren't a (single) type of fruit.

Do you consider New Atheism the same religion as Communist Atheism in the former Soviet Union?

Other questions: Is Effective Altruism a religion? Is Intentional Insights? Is Yoga a religion? Is the Landmark forum?

The focus of the questions is intended to be on the engineering and social aspects

The social aspects of Islam that regulates how inheritance and child support payments are supposed to work is different than Buddhism which as a religion doesn't care for that and might as David Chapman suggests not include any ethical imperatives.

Before you engineer something you have to decide on what kind of thing you actually want to engineer.

Given that you said you are working on a conlang, it's useful for you to take a mental step back and ask whether the word religion is a useful word for thinking about the domain you want to talk about.