Patterns (was Re: Birthdays)

Roger Burton West roger at firedrake.org
Thu Dec 16 17:42:42 EST 2004


On Thu, Dec 16, 2004 at 10:30:56PM +0000, minnow at belfry.org.uk wrote:

>So isn't it that the patterns are *there* in the randomness that is the
>beginning of science?  Not turning randomness into patterns, but suddenly
>noticing a pattern of some sort in what had previously appeared random, and
>saying "gosh, that's interesting.  I wonder if it *always* follows that
>rule, or if there are exceptions?"  I would have thought that *making*
>patterns was not science, because that would be causing the data to fit the
>thesis, and shouldn't it be the thesis that is developed from the data?

The problem is that the human brain is _extremely_ well-suited to taking
data and finding patterns in them, whether the patterns are there or
not. (I assume that this is because a false positive on "is there a
sabre-tooth hiding in that bush" is less of a problem than a false
negative.)

When one thinks one observes a pattern it is vitally important to step
back from it and say "OK, I've observed this, but can it be backed up
by any objective method?".

So really it depends on what you mean by "making patterns" - people who
reckon they have found patterns in data that really are purely random
don't think they're making them up.

R
--
To unsubscribe, email dwj-request at suberic.net with the body "unsubscribe".
Visit the archives at http://suberic.net/dwj/list/



More information about the Dwj mailing list