# Bayesian inference and the teleological argument

**URL:** https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696
**Category:** Side Conversation
**Tags:** Philosophy, Science, Design
**Created:** [May 6, 2024, 10:32pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696 "2024-05-06T22:32:52Z")
**Posts on this page:** 20
**Page:** 1

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 6, 2024, 10:32pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/1 "2024-05-06T22:32:52Z")

</div>

> [@The proper distinction between science and theology](https://discourse.peacefulscience.org/t/the-proper-distinction-between-science-and-theology/16636/99):
>
> All teleological arguments are, at their core, claims that P(Life|God) \>\> P(Life|~God), or at least that P([some feature of the universe]|God) \>\> P([some feature of the universe]|~God). But the fact is we don’t know nearly enough about either nature or God to be able to reliably calculate P(X|God) or P(X|~God). For this reason, I don’t think any teleological arguments support the existence of God. Even if it could be shown that evolution is utterly false, it wouldn’t help support the hypothesis of creation (though I don’t believe evolution and creation are contradictory anyway).

Given your criteria for supporting a claim how would you make any argument that supports a proposed cause of any origin event?

Take the origin of the carbon atom on earth as a test case.

---

<div class="post-metadata">

### Author: ![misterme987](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/misterme987/32/15207_2.png) [@misterme987](https://discourse.peacefulscience.org/u/misterme987)
#### Post date: [May 6, 2024, 10:37pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/2 "2024-05-06T22:37:32Z")

</div>

> [@colewd](#):
>
> Take the origin of the carbon atom on earth as a test case.

I don’t even know what you’re trying to say here. WTF is “the origin of the carbon atom on earth”???

This Bayesian approach can absolutely be applied to other problems, though. For example we could ask what is the relationship between P(ancestral convergence | common ancestry) and P(ancestral convergence | separate ancestry). As it turns out, P(AC|CA) is about 132 orders of magnitude larger than P(AC|SA), based on White et al. ([2013](https://doi.org/10.1371/journal.pone.0069924)).

I’m no statistician though; you’d be better off asking @Dan_Eastwood about Bayesian inference.

---

<div class="post-metadata">

### Author: ![Dan\_Eastwood](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/dan_eastwood/32/2368_2.png) [@Dan\_Eastwood](https://discourse.peacefulscience.org/u/Dan_Eastwood)
#### Post date: [May 6, 2024, 10:40pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/3 "2024-05-06T22:40:37Z")

</div>

@misterme987 I thought your stats were well stated. I don’t know what Bill means either.

@colewd Carbon atoms don’t originate on Earth; they originate from the fusion of helium atoms inside of stars. What this has to do this the topic is a mystery.

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 6, 2024, 10:52pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/4 "2024-05-06T22:52:39Z")

</div>

Hi Andrew  
You made a claim teleological arguments are P(Life/God etc). Then followed it up by the claim that they don’t support arguments for the existence of God.

I am trying to discover if we can make any argument about causes of origin events given your criteria that teleological arguments must be reduced to probability calculations.

Your probability argument for CA is based on the alternative SA happening solely from the laws of physics and chemistry and not the product of design or special creation.

Instead of a probability calculation could you use inference to the best explanation?

---

<div class="post-metadata">

### Author: ![misterme987](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/misterme987/32/15207_2.png) [@misterme987](https://discourse.peacefulscience.org/u/misterme987)
#### Post date: [May 6, 2024, 10:57pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/5 "2024-05-06T22:57:39Z")

</div>

> [@colewd](#):
>
> Your probability argument for CA is based on the alternative SA happening solely from the laws of physics and chemistry and not the product of design or special creation.

Okay, let’s try that. What’s the relationship between P(AC|CA) and P(AC|SC [special creation])? Well, we can’t quite determine that because we don’t know if the SC hypothesis involves a deceptive god or a non-deceptive god. Let’s split it into SC-NDG (non-deceptive) and SC-DG. P(AC|SC-NDG) will be ~0 because there’s no functional reason for ancestral convergence, it’s all about events in the past and not functions in the present. P(AC|SC-DG) is incalculable because we have no way of knowing the mind of a deceptive god. Therefore P(AC|CA) \>\>\> P(AC|SC-NDG) and P(AC|CA) ? P(AC|SC-DG).

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 6, 2024, 10:59pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/6 "2024-05-06T22:59:44Z")

</div>

> [@Dan\_Eastwood](#):
>
> @colewd Carbon atoms don’t originate on Earth; they originate from the fusion of helium atoms inside of stars. What this has to do this the topic is a mystery.

What is your probability calculation that shows this is true?

---

<div class="post-metadata">

### Author: ![Faizal\_Ali](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/faizal_ali/32/12416_2.png) [@Faizal\_Ali](https://discourse.peacefulscience.org/u/Faizal_Ali)
#### Post date: [May 6, 2024, 11:00pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/7 "2024-05-06T23:00:42Z")

</div>

Bill never disappoints.

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 6, 2024, 11:10pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/8 "2024-05-06T23:10:53Z")

</div>

> [@misterme987](#):
>
> Okay, let’s try that. What’s the relationship between P(AC|CA) and P(AC|SC [special creation])? Well, we can’t quite determine that because we don’t know if the SC hypothesis involves a deceptive god or a non-deceptive god. Let’s split it into SC-NDG (non-deceptive) and SC-DG. P(AC|SC-NDG) will be ~0 because there’s no functional reason for ancestral convergence, it’s all about events in the past and not functions in the present. P(AC|SC-DG) is incalculable because we have no way of knowing the mind of a deceptive god. Therefore P(AC|CA) \>\>\> P(AC|SC-NDG) and P(AC|CA) ? P(AC|SC-DG).

Let’s get rid of special creation. What is the probability that two species with different chromosome counts are reproductively related? Probably not doable with realistic assumptions.

My point is we are often using inference to the best explanation in our analysis vs reducing all inferences to probability calculations.

---

<div class="post-metadata">

### Author: ![misterme987](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/misterme987/32/15207_2.png) [@misterme987](https://discourse.peacefulscience.org/u/misterme987)
#### Post date: [May 6, 2024, 11:17pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/9 "2024-05-06T23:17:01Z")

</div>

> [@colewd](#):
>
> What is the probability that two species with different chromosome counts are reproductively related? Probably not doable with realistic assumptions.

Well, you’ve advanced a hypothesis. Let’s call this hypothesis 2SR (two species related). The null hypothesis is ~2SR. We can’t assess the probability of the hypothesis without any data.

---

<div class="post-metadata">

### Author: ![Dan\_Eastwood](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/dan_eastwood/32/2368_2.png) [@Dan\_Eastwood](https://discourse.peacefulscience.org/u/Dan_Eastwood)
#### Post date: [May 6, 2024, 11:46pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/10 "2024-05-06T23:46:13Z")

</div>

> [@colewd](#):
>
> > [@Dan\_Eastwood](#):
> >
> > @colewd Carbon atoms don’t originate on Earth; they originate from the fusion of helium atoms inside of stars. What this has to do this the topic is a mystery.
> 
> What is your probability calculation that shows this is true?

Don’t be silly.

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 7, 2024, 1:44am UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/12 "2024-05-07T01:44:53Z")

</div>

> [@Dan\_Eastwood](#):
>
> Don’t be silly.

> [@colewd](#):
>
> My point is we are often using inference to the best explanation in our analysis vs reducing all inferences to probability calculations.

I think we agree.

---

<div class="post-metadata">

### Author: ![Gisteron](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/gisteron/32/15845_2.png) [@Gisteron](https://discourse.peacefulscience.org/u/Gisteron)
#### Post date: [May 7, 2024, 4:51am UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/13 "2024-05-07T04:51:36Z")

</div>

> [@colewd](#):
>
> My point is we are often using inference to the best explanation in our analysis vs reducing all inferences to probability calculations.

What exactly is an “explanation” to you? What is it supposed to accomplish and how? When is one explanation “better” than another, enough to where we can start looking for the “best” of them all?

---

<div class="post-metadata">

### Author: ![Mercer](https://avatars.discourse-cdn.com/v4/letter/m/e274bd/32.png) [@Mercer](https://discourse.peacefulscience.org/u/Mercer)
#### Post date: [May 7, 2024, 6:23am UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/14 "2024-05-07T06:23:06Z")

</div>

> [@colewd](#):
>
> What is the probability that two species with different chromosome counts are reproductively related?

What is the probability that two individuals of the same species with different chromosome counts are reproductively related?

Show your calculations.

---

<div class="post-metadata">

### Author: ![Gisteron](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/gisteron/32/15845_2.png) [@Gisteron](https://discourse.peacefulscience.org/u/Gisteron)
#### Post date: [May 7, 2024, 7:40am UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/15 "2024-05-07T07:40:02Z")

</div>

Heck, for that matter, @colewd, what is the probability that a mother and her mutant offspring with a different chromosome count are reproductively related?

---

<div class="post-metadata">

### Author: ![Dan\_Eastwood](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/dan_eastwood/32/2368_2.png) [@Dan\_Eastwood](https://discourse.peacefulscience.org/u/Dan_Eastwood)
#### Post date: [May 7, 2024, 3:52pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/16 "2024-05-07T15:52:29Z")

</div>

A technical note: Statistical inference requires two probabilities, one under a null hypothesis and another undersome alternative. The ratio of these probabilities leads to a Likelihood Ratio test, or with a little more work a Bayes Factor. You wouldn’t learn about this ratio in an intro stats class; it’s a more general method that is at the heart of statistical theory.

Typically in IDCreationist arguments involving probabilities, they make an unstated assumption that one of those probabilities is 1.0. You CAN do this with a Bayes Factor as a Prior assumption, and it results in concluding your prior assumption is true, no matter what the data says.

---

<div class="post-metadata">

### Author: ![Dan\_Eastwood](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/dan_eastwood/32/2368_2.png) [@Dan\_Eastwood](https://discourse.peacefulscience.org/u/Dan_Eastwood)
#### Post date: [May 7, 2024, 3:59pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/17 "2024-05-07T15:59:56Z")

</div>

> [@Gisteron](#):
>
> > [@colewd](#):
> >
> > My point is we are often using inference to the best explanation in our analysis vs reducing all inferences to probability calculations.
> 
> What exactly is an “explanation” to you? What is it supposed to accomplish and how? When is one explanation “better” than another, enough to where we can start looking for the “best” of them all?

@Bill: Specifically, _what value of which parameter_ (and what distribution) would support Design as a better explanation?

Hint: That isn’t answerable without testable hypotheses about Design or the Designer.

IMO, questions that amount to “What is the probability of God?” make a mockery of faith. Bad arguments do nothing to further your cause.

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 7, 2024, 4:05pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/18 "2024-05-07T16:05:45Z")

</div>

> [@Dan\_Eastwood](#):
>
> IMO, questions that amount to “What is the probability of God?” make a mockery of faith. Bad arguments do nothing to further your cause.

Hi Dan  
I agree with you here.

Most historical arguments are based on inference to the best explanation and not a probability calculation.

---

<div class="post-metadata">

### Author: ![Dan\_Eastwood](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/dan_eastwood/32/2368_2.png) [@Dan\_Eastwood](https://discourse.peacefulscience.org/u/Dan_Eastwood)
#### Post date: [May 7, 2024, 4:39pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/19 "2024-05-07T16:39:09Z")

</div>

> [@colewd](#):
>
> Most historical arguments are based on inference to the best explanation and not a probability calculation.

@Bill: Specifically, _what value of which parameter_ (and what distribution) would support Design as a better explanation?

---

<div class="post-metadata">

### Author: ![Gisteron](https://sea2.discourse-cdn.com/flex016/user_avatar/discourse.peacefulscience.org/gisteron/32/15845_2.png) [@Gisteron](https://discourse.peacefulscience.org/u/Gisteron)
#### Post date: [May 7, 2024, 4:45pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/20 "2024-05-07T16:45:33Z")

</div>

> [@Dan\_Eastwood](#):
>
> > [@colewd](#):
> >
> > Most historical arguments are based on inference to the best explanation and not a probability calculation.
> 
> @Bill: Specifically, _what value of which parameter_ (and what distribution) would support Design as a better explanation?

@colewd or even more generally, what does it mean for something to be an “explanation” at all, and how do we decide which is “best” between two or more of them? Give us _something_ to work with here, Bill, _any_ way at all to understand how your use of “inference to the best explanation” means anything other than “arbitrary insistence on what ever I personally want to be true”.

---

<div class="post-metadata">

### Author: ![colewd](https://avatars.discourse-cdn.com/v4/letter/c/ecd19e/32.png) [@colewd](https://discourse.peacefulscience.org/u/colewd)
#### Post date: [May 7, 2024, 9:07pm UTC](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696/21 "2024-05-07T21:07:45Z")

</div>

> [@Dan\_Eastwood](#):
>
> @Bill: Specifically, _what value of which parameter_ (and what distribution) would support Design as a better explanation?

Hi Dan  
Can you be more specific? What observation is generating the potential inference? The design inference becomes stronger when it is hard to assign the laws of physics, chemistry and in addition chance as a cause of what we are observing.  
– Can you infer planning from the observation?  
–Does the cause of what you are observing likely require intelligence?

As an example the origin of a living genome is difficult to explain with the laws of physics and chemistry. The origin of carbon is easier to explain as you cited a star as the potential origin of carbon.

[Next page](https://discourse.peacefulscience.org/t/bayesian-inference-and-the-teleological-argument/16696.md?page=2)
