by David Leech

Worldview affects cause selection and EA should be open to worldview stress testing. Based on a presentation given by David Leech at an EA for Christians Academic Workshop in April 2023.

Summary

• EA is a worldview or has a worldview

• This affects cause selection

• EA should be open to worldview stress testing and worldview exploration

• Philosophers of religion are a relevant expert class for worldview questions, so there is a prima facie case for EAs to take the philosophy of religion seriously

In this series of three posts I consider some articles on the Effective Altruism Forum which touch indirectly on the philosophy of religion, and I use these as a springboard for suggesting that the secular EA community can, and perhaps should, be open in principle to a range of non-naturalistic worldview possibilities, including but not limited to theism. These posts will be somewhat schematic, as is the limited engagement with philosophy of religion on the Forum so far.1

In this first post, I propose that EA is or has a worldview, that worldviews affect cause selection, and that philosophers of religion are an expert class for worldview questions, so that we might expect EAs to take the philosophy of religion seriously, all things being equal.

The inescapability of having a worldview which favours some or other kinds of causes

In James Fodor’s post Effective Altruism is an ideology, not (just) a Question, the philosophy of religion is used as an example of a potential EA cause area implicitly ignored (he argues) by EA worldview2 presuppositions. Fodor’s argument, in short, is: you can make a case for the philosophy of religion, just like, say, AI safety, being a cause which is important, neglected, and tractable, since there are a significant minority of philosophers who hold religious beliefs;3 many religions teach that salvation or (some kind of) liberation is the highest good; the philosophy of religion is relatively neglected; and the expected impact would be very high (salvation; liberation etc.). However, it is an example of a cause area which is not identified by the EA community as a deserving one.

Fodor’s point is not that the philosophy of religion is a good cause area for EAs to consider, but rather that these kinds of cause area candidates are ‘not generally discussed by EAs’. His explanation for this is that EAs’ naturalistic worldview presuppositions implicitly bias them towards a focus on cause areas like AI safety, new technologies etc. rather than (e.g.) the philosophy of religion:

‘it is not as if effective altruists have carefully considered these possible cause areas and come to the reasoned conclusion that they are not the highest priorities. Rather, they have simply not been considered. They have not even been on the radar, or at best barely on the radar … I have not found any evidence that th[ese] choice[s are] the result of early investigations demonstrating that emerging technologies are far superior to the cause areas I mention. Instead, it appears to be mostly the result of disinterest in the sorts of topics I identify, and a much greater ex ante interest in emerging technologies over other causes…This, in turn, is the result of the underlying ideological commitments of many effective altruists.’

Holden Karnofsky defines ‘worldview’ as ‘a set of highly debatable (and perhaps impossible to evaluate) beliefs that favour a certain kind of giving’. To take an obvious example: EAs with a Christian worldview might favour in-church giving or giving to mission projects as a way of doing the most good, whereas EAs with a secular worldview wouldn’t (since God or salvation are not part of the latter’s worldview).  

Fodor lists five ‘core tenets’ of what he considers to be a fair sketch of the standard EA worldview, allowing for individual variation etc. It will suffice just to mention tenets one and four here:

1. The natural world is all that exists, or at least all that should be of concern to us when deciding how to act. In particular, most EAs are highly dismissive of religious or other non-naturalistic worldviews, and tend to just assume without further discussion that views like dualism, reincarnation, or theism cannot be true…

1. The best way to approach a problem is to think very abstractly about that problem, construct computational or mathematical models of the relevant problem area, and ultimately (if possible) test these models using experiments. The model appears to be of how research is approached in physics with some influence from analytic philosophy. The methodologies of other disciplines are largely ignored.

Fodor is not suggesting that EAs shouldn’t have a worldview, or that the current one (as he presents it) is false, but rather that having a worldview is inescapable (‘we can’t really get anywhere in rational investigation without certain starting assumptions’); also, we should be clear about what our worldview is, and about how it ‘focus[es] on certain viewpoints and answers while largely ignoring others’ – which determine cause area identification and prioritisation. He adds that it would be consistent with the standard EA worldview – evidentialism, commitment to scout mindset etc. – to stress test their own worldview with the kind of thinking they consider the best way to approach any issue which matters.4

A prima facie case for why EAs should take the philosophy of religion seriously as an expert class

With Fodor, I find it convincing that secular EA is a worldview (‘ideology’) and not just a question; I take it that having a worldview is inescapable and there is no neutral view from nowhere, and therefore that EAs do have a worldview. How to evaluate different worldviews is beyond the scope of Fodor’s article, but in the matter of worldview evaluation, a candidate for a or the relevant expert class here – alongside some nearby expert classes like analytic metaphysics and philosophy of mind – is presumably analytic philosophy of religion. (I stress analytic philosophy of religion, because by Fodor’s tenet four – which I do think characterises the standard EA worldview, broadly speaking – we can expect that EAs would take analytic philosophy of religion more seriously than non-analytic forms.5)

If we grant with Fodor that EA should, in principle, stress test their own worldview with the kind of critical thinking they apply to anything else which matters; and that worldview choice does indeed matter a great deal, since it implicitly determines cause area identification and prioritisation, we seem to have prima facie reason to think that EAs should take analytic philosophy of religion seriously, for the following reasons. Firstly, EAs have a high esteem for analytic philosophy – indeed, many leading EAs in the field/community (Macaskill, Ord etc.) are analytic philosophers – and analytic philosophers of religion are a subset of analytic philosophers. Secondly, on the face of it, philosophers of religion constitute an expert class whose proper subject matter is the critical evaluation of worldview beliefs (theism, atheism, but also pantheism, religious naturalism etc.).

In a second post I discuss why some EAs do not take the philosophy of religion seriously as an expert class, and I offer some further reasons why they can, and perhaps should, take philosophy of religion seriously.

1To date, references to philosophy of religion on the Forum are quite limited. A search on ‘philosophy of religion’, ‘philosopher of religion’, ‘philosophers of religion’ (22 June 2023) yielded 10 results.

2Fodor’s own preferred term is ‘ideological’.

3In Bourget and Chalmers’ 2020 PhilPapers Survey of philosophers (results, paper), 19% (target group) endorsed theism. See also Rob Bensinger’s EA forum post 2020 PhilPapers Survey Results for discussion.

4‘[EAs should] critically analyse this ideology, understand its strengths and weaknesses, and then to the extent to which we think this set of ideological beliefs is correct, defend it against rebuttals and competing ideological perspectives. This is essentially what all other ideologies do – it is how the exchange of ideas works. Effective altruists should engage critically in this ideological discussion, and not pretend they are aloof from it by resorting to the refrain that ‘EA is a question, not an ideology’’.

5

Consequently, when I refer to philosophy of religion in this post, I understand analytic philosophy of religion.