The Circularity of dFSCI

(updated for greater clarity – I hope)

This short essay is a response to some comments by my good friend Gpuccio on Uncommon Descent.   I am arguing that in a sense his use of the concept of dFSCI as an indicator of design is circular.

First a clarification. The argument:

"I cannot see how this can have arisen through necessity or chance therefore it must have been designed"

is not circular. It is a lousy argument, but it is not circular and it not the one I am picking on. The argument I find circular is:

"All objects except biological ones with dFSCI are known to be designed. Therefore, biological objects with dFSCI are almost certainly designed."

Now I will try and explain why it is circular (in a more detailed fashion than I have before).

First my summary of your argument. I hope this demonstrates I understand your argument. I am sure you will tell me if you disagree

[start summary]

You classify the world into outcomes which exhibit dFSCI and those which do not. Also the world includes outcomes which fall into three groups:

(A) we know them to be designed (because we observed humans design them or have reliable reports of humans designing them) e.g. Brooklyn Bridge

(B) we known them not to be designed (because we know how they were produced) e.g. the pattern of iron filings on a magnetic field

(C) we do not know for sure whether they are designed or not because we do not know how they were produced (which includes aspects of life such as proteins, but also such things as old marks on rocks).

You argue there is a correlation between dFSCI and group A.   So if dFSCI is present in outcomes in group C we can conclude that they also are designed.

It is similar to the logic of a medical test. We observe that a high PSA rating is highly correlated with prostate cancer so if we observe a high PSA rating elsewhere we conclude prostate cancer is likely (I know the PSA test is not actually very good – but let’s assume it is)

[end summary]

* Update – in this comment Gpuccio has indicated he does not agree with this summary.  Nevertheless  I think it is a reasonable account of how we should proceed if we wish to establish dFSCI as indicator of design

Now look at each of the defining characteristics of dFSCI that you supply in your comment and see to what extent it correlates with groups A and B. Is each feature rarely absent or rarely present in either group?  Obviously if a feature it is often present and often absent in both groups then it is not much use as an indicator.

1) String of digital values. This feature is often present and also often absent in both A and B. There are many non-digital things that have been designed and many non-designed things that are a string of digital values e.g. any complex molecule. (A note on "digital". In the case of man-made artefacts the digital nature of things is something we impose on the object. The script of Hamlet is a collection of analogue marks. We divide them into discrete categories i.e. letters. To get to something that is intrinsically digital you have to get to atomic levels whether the item be in group A, B or C)

2) Scarcely compressible. (Remember that in general we can only estimate if something is uncompressible – this should perhaps be phrased – appears to be scarcely compressible). Again this feature is often absent and often present in both groups A and B. We sometimes design things that are compressible, but also design things such as random number tables or zip files which are not compressible as far as we know. The orbit of a heavenly body is not designed but compressible. A jumble of rocks is not designed but not apparently not compressible.

3) Conveys information for a well-defined function. This threw me a bit. Do you mean "performs a well-defined function" as haemoglobin performs the function of carrying oxygen. Or is there some new subtlety here? There appear to be things that perform well-defined functions and others that do not in both groups A and B. We design art works that have no function and features such as the water cycle perform the well defined function of delivering water to land and thus sustaining life on land.

So we are left with:

4) Complexity is greater than 150 bits


5) No explanatory necessity mechanism is known

I should note that for (4) as well as (5) we should add the phrase – "as far as we know". If some non-design mechanism is discovered which makes the outcome probable than the complexity drops does it not? Perhaps I may combine those as – no known plausible non-design mechanism?

Here at last we have something which is only found in one of the two groups and might act as an indicator of design. Group A includes things for which this feature is present as well as things for which it is absent. We often make things that could have plausibly been produced without intervention. But group B – stuff we know is not designed – always has a known plausible non-design mechanism.  It never has this feature. 

I can summarise this as a table with examples in each cell.

Feature   Group A
(known design)
Group B
(known not design)
String of digital values Present Hamlet Any complex molecule
  Absent Mona Lisa Clouds
Scarcely compressible Present Zip file Jumble of fallen rock
  Absent Wallpaper pattern, “Caputo” ballot papers (see Dembksi NFL) Orbit of a planet or comet
Performs a well defined function Present Watch Coast of Norway
  Absent Jackson Pollock painting Water cycle
No plausible non-design mechanism Present Hamlet ????
  Absent Artificial hill Hurricane

But hang on – how did we know the stuff in group B was not designed? We allocated it to group B because we knew it was  produced by a non-design mechanism! There cannot be anything in that red cell by definition. Here lies the key circularity.

A couple of refinements

1) I have shown that none of the individual features of dFSCI can be used as an indicator of design.  But maybe a combination will work? Perhaps the combination of digital values, incompressibility and well-defined function will do the job?  Well there is little doubt that it is possible for humans to design things that have and have not these features.  So the combination is both present and absent in group A.  There are also plenty of things in group B that do not have all three features.  But are there things in group B that have all three features?  Yes – there are crystals whose molecular structure is compressible and who serve a function in preserving life – salt for example. Also note that the CSI (as opposed to dFSCI) only has the feature – no plausible non-design mechanism – and this generally taken as an indicator of design in the ID literature.

2) Maybe what you are saying is that often things in group C have the feature "no apparent non-design cause" and then it turns out later that they were designed.  This would be legitimate and non-circular.  Unfortunately it is not true. History is stuffed with things for which there was no apparent non-design cause which later turned out to have a natural cause – the pattern that iron filings make when they fall in a magnetic field is one example!

If nothing else I hope I have shown that the concept of dFSCI is far from straightforward. And I haven’t got into the complexities arising from defining the function specification.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: