Re: [asa] Bayesian inference and design inference

From: Iain Strachan <>
Date: Sat Sep 19 2009 - 07:36:57 EDT


Thanks for responding. I'll try and address the issues you raised.

> A few questions/comments.
> First, a lot of what you say here seems to assume a person is coming at the
> question from an utterly agnostic point of view - where the existence or
> non-existence of any designers at the level in question (capable of
> designing species, life, planets, galaxies, universes, etc) is expressly
> unknown, and ID's purpose is primarily to help someone decide on that
> question. But what if you have independent reasons for suspecting, even
> strongly suspecting, the existence of some kind of designing entity
> beforehand? Now, I know the response here can be 'Well, if you already
> believe or strongly suspect that kind of designer exists, then ID is
> superfluous anyway.' But I'd disagree - in that case you'd be using ID as a
> kind of lens through which to view nature, identifying events throughout
> natural history that most prominently/obviously displayed the hallmarks of
> design, or guidance, etc. And this seems to be the case even if we can
> identify some, possibly all, of the nature causes for a given hallmark.

If I understand you correctly you are saying that if you already
suspect that the Designer exists then you might well make Design
inferences. (In terms of my argument you would say that there is a
finite prior probability of the existence of the Designer). I think
the distinction to be made is that even given you believe in God, as
the bible says "no-one has seen God" (John 1:18). We may believe (I
do believe) that God created the universe and everything in it - but
still accept that the working out of all of this came about by
evolution. But in the case of a watch, you actually know not only
that human beings capable of making watches exist, but you also have
positive evidence and witness that they actually do make watches. If
someone had actually seen an Intelligent Designer actually fabricate a
biological organism, then one would be justified in drawing the design
inference over other organisms.

Also it is interesting your choice of the word "hallmark". A hallmark
is usually an extra symbol that identifies the designer, like on a
piece of silver, or the artist's signature at the bottom of the
painting, or a serial number on the computer case. The hallmark itself
is not part of the mechanism of operation of the artefact. It would
be like some encoded extra message in the DNA of the organism (
perhaps like an Acrostic poem). It is not hard to see how this could
be contrived to happen, given the redundancy of DNA codons (64 codons,
21 amino acids). If one found something like that, rather like the
prime number sequence in Sagan's "Contact", then one might well have
found a "hallmark" of design.

> Second, let's put aside the question of independent reasons. Let's say we
> approach the design question without knowing about the existence, nature or
> capabilities of the designer in question. There still seems to be one
> problem: We can be certain, or at least supremely confident, in the
> existence of some designers - humans, of course. And while humans are
> nowhere near the capabilities of a designer on the level ID routinely
> investigates (Capable of designing life, possibly front-loading evolution,
> intervening in nature from outside of time, etc) we also have tremendous
> capabilities now, and everyone agrees those capabilities are growing. In
> principle, intelligent design of this nature can accomplish everything we
> see in nature. The problem is this: If we accept what was just said, then it
> seems we immediately privilege 'design' as the default explanation for all
> we see in nature, even if we have a fully mechanistic explanation on hand
> (designers can employ mechanisms as well, after all.) 'Unguided,
> unintelligent, wholly fortuitous nature' is something we have less evidence
> for - and it seems we can never have positive evidence for it in principle.

I'm not quite sure what you mean here. Is it that we have more
evidence for "Designers" (ie us) than for mechanistic processes
leading to design?

If so, I think the flaw is that we know we are not intelligent or
complex enough to design ourselves - the Dawkins argument that the
designer has to be more complex than the designed. So I don't see
that we have a default "design" explanation of Design.

(BTW I DON'T accept the Dawkins complexity of the designer argument in
full - it can only apply if the Designer is part of the material
universe, in which case the more complex the more improbable. However
if the designer is transcendent and part of a "higher reality" then
all bets are off about the relative probability of such a being

Just to respond briefly to David:

you wrote:
I followed this up until this sentence:

"In every case we
know about, information implies an information giver"

I reject that statement. Information requires an information source,
a form of low entropy. Not an information giver in the sense of an

Yes, I reject that statement too. Just to clarify, I was stating what
I understood to be the ID position, not my own view.


> On Fri, Sep 18, 2009 at 4:12 PM, Iain Strachan <>
> wrote:
>> I wanted to outline my reason for not accepting the Intelligent Design
>> inference, which relates to Bayesian inference, which is at the heart
>> of modern probabilistic methods.  Although it's a technical term, I'll
>> try and explain it by way of example, without the maths.  I think even
>> if we don't directly use Bayes's Theorem to make inferences, we
>> implicitly do something like it, which can't carry over into making
>> the kind of design inference that the ID community want to make.
>> Here's my (somewhat light-hearted) example to explain the process:
>> You walk into a room where there is a computer with the monitor
>> switched off, a scientist and two "subjects".  One of the subjects is
>> a human and the other one is a monkey.  The scientist tells you that
>> what is now on the screen of the monitor was input using the keyboard
>> from one of the two subjects.  Which one was it?  Not being able to
>> see what's on the screen (is it a sentence or gibberish like asdfga0s
>> dua0s9df d0 ads09' ), you are unable to say - it's 50-50.
>> Then the scientist switches on the monitor and the screen displays the
>> So you're now thinking it's definitely the human.  How could a monkey
>> have typed an intelligible sentence?  But when you say that, what you
>> had assumed before you saw the screen was that it was equally likely
>> to be the human or the monkey.
>> Then the scientist tells you the rule for selection of the subject.
>> She took a fair coin and tossed it ten times.  The rule was if it came
>> up heads all ten times, the human would be selected, otherwise the
>> monkey.
>> You're still thinking it's the human.  A 1024:1 shot isn't that remote
>> compared with the possibility of a monkey typing that sentence.
>> Suppose she tells you she tossed the coin a thousand times and the
>> human would be selected if they all came up heads, otherwise the
>> monkey.
>> I guess the first thing you do is examine the coin to see if it really
>> is fair.  You toss it a few times  Heads, Heads, Tails, Tails, Heads
>> Tails.  Seems pretty fair.  You run a lie detector on the scientist.
>> She's telling the truth.
>> So now you're thinking, incredible as it seems, that it's the monkey.
>> Maybe monkeys can be trained to do such a feat.
>> What you're implicitly doing is estimating a _prior probability_ on
>> which subject was chosen.  Ten coin tosses all needing to be heads
>> makes a prior probability of 1/1024 of it being the human and
>> 1023/1024 of it being the monkey.  Then you get more data, and as a
>> result you recalculate your estimate of which of the subjects typed
>> the sentence.  From the observations, you modify your prior
>> probabilities to get _posterior_ probabilities.  The nature of the
>> evidence might swing your estimate right round.  If one wants to do
>> this calculation rigorously (although you are implicitly doing it
>> roughly anyway), you would use Bayes's theorem to compute the
>> posterior probabilities.  This is the way it's done in all sorts of
>> modern expert systems, for example for medical diagnosis.  I was told
>> by my PhD supervisor, a director at Microsoft Research Labs in
>> Cambridge, that a little Bayesian inference engine powers the "printer
>> troubleshooter" in MS windows.  (Though it's never helped me much!!)
>> (Or maybe it was the infamous "paper clip" cartoon character that gave
>> you hints all the time that you didn't want to know - can't remember
>> exactly).
>> Now take an example that has been cited by Dembski as "design
>> detection".  It relates to the case (if I recall the details
>> correctly) of Nicholas Caputo, who was in charge of arranging the
>> ballot tickets in elections in one particular ward (state? - I don't
>> know the correct term).  It is well-known that the name that appears
>> first on the list of a ballot paper has an unfair advantage because
>> lots of people are too stupid to check and just put an X on the first
>> name on the list, irrespective of the party.  Hence lots would be
>> drawn to determine which party was at the top of the list each time.
>> It was found that under Caputo's direction, a democrat had appeared 40
>> times out of 40.  He was convicted on the grounds that this was so
>> unlikely to happen by chance, that he must have rigged it.  This is
>> cited as a clear instance of design detection.
>> However, even here, implicitly one is using a Bayesian technique.
>> This is because you have concrete independent evidence that humans can
>> be corrupt and prone to rigging elections.  Most people are honest but
>> a small minority are corrupt.  You've therefore got a reasonable idea
>> of the prior probability of a corrupt person determining the first
>> name on the ballot paper.  The 40 out of 40 democrats is further
>> evidence that modifies your posterior probability that Caputo is
>> corrupt.
>> There are other pieces of evidence that could swing it back the other
>> way.  Suppose that Caputo, like Pinocchio has an affliction that
>> causes his nose to grow every time he tells a lie.  He stands up in
>> court and swears that he the way he conducted the drawing of lots was
>> above board, and the result was just a freak; he was as surprised as
>> anyone else to see the result.  His nose stays the same size.  That
>> would make you more likely to accept the freak result, because your
>> prior probability of his telling the truth just took a massive
>> increase.
>> Now take the case of the intelligent design inference.  In the
>> publicity for his new book, Stephen C. Meyer states that DNA is like a
>> computer code with immense amounts of information.  In every case we
>> know about, information implies an information giver, and a program
>> requires a programmer.  Hence design (the nature of the designer
>> remaining unknown and unspecified) is the best explanation.
>> But it seems to me that this is entirely different from the Caputo
>> case.  In the Caputo case you have independent verifiable evidence
>> that people exist who rig elections.  Many of them, hopefully, are in
>> jail!  So you can assign a prior probability.  But by definition, you
>> don't KNOW about the existence of an unspecified designer - the fact
>> that you don't say anything about the identity of the designer
>> undermines the whole argument.  There is therefore no meaningful way
>> to assign a prior probability.  Indeed what you are trying to do is to
>> infer the _existence_ of a designer responsible for the perceived
>> design.  In the Caputo case, the software programmer case etc, you
>> already know that corrupt people, computer programmers, etc exist -
>> you are trying to determine whether your evidence is explained by one
>> of these people whom you know exist, or by coincidence.
>> In a nutshell:   Inferences we make are all implicitly Bayesian,
>> because we have a prior idea of the probabilities of the different
>> inferences that could be made.  But with the Design inference, where
>> the nature of the Designer is unknown (as is tenaciously held by the
>> ID community), then you can't assign a prior probability and hence
>> can't begin to make a meaningful inference.
>> Discuss.
>> Iain
>> --
>> -----------
>> Non timeo sed caveo
>> (\__/)
>> (='.'=)
>> (")_(") This is a bunny copy him into your signature so he can gain
>> world domination
>> -----------
>> To unsubscribe, send a message to with
>> "unsubscribe asa" (no quotes) as the body of the message.

Non timeo sed caveo
(")_(") This is a bunny copy him into your signature so he can gain
world domination
To unsubscribe, send a message to with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Sat Sep 19 07:37:37 2009

This archive was generated by hypermail 2.1.8 : Sat Sep 19 2009 - 07:37:37 EDT