[FoRK] [fonc] People and evidence

Chris Warburton chriswarbo at googlemail.com
Mon Sep 9 08:54:29 PDT 2013

Alan Kay <alan.nemo at yahoo.com> writes:

> Many of the commenters on this list have missed that "evidence" and
> "data" requires a fruitful context -- even to consider them! -- and
> that better tools and data will only tend to help those who are
> already set up epistemologically to use them wisely. (And don't forget
> the scientists I mentioned who have been shown to be deeply influenced
> by the context of their own employers.)
> The fault is not in our stars ...

This is certainly true; giving large numbers of people (eg. via
schooling) the *inclination* to question their reasoning processes would
be a revolutionary change. Unfortunately I wouldn't know where to begin
to bring such a change about.

What I do know, from my computer science/software engineering background
and my own experience in trying to justify my everyday decisions, is
that there is still a lot of potential for incremental hill-climbing for
those few of us who already have this inclination (see lesswrong.com for
a growing community).

At the moment I usually conclude that it's not worth the investment to
sanity-check my day-to-day decisions in depth, but this balance can be
tipped by technological support, allowing me to do it out of habit.

Another example of an incremental improvement, again for those who
recognise the need, could be to have (electronic) lab books prompt the
user for justifications when they spot certain cues in their input. For

User: "Events were modelled as a markov chain..."
Tooltip: "Model" spotted. Why is "markov chain" a realistic model?
User: Established by http://my-journal.com/it-really-is-markov.pdf
System: Thanks.

The point of such a system wouldn't be to verify what we tell it, but
simply to make us think twice about what we're doing.

Even though these incremental improvements will hopefully be thrown away
by some revolutionary change, they make it easier for us to come up with
that change. This is my method of tackling unsolvable problems. For
example, if I were stuck in a sub-optimal language and couldn't envisage
what a better language might look like, I would spend my time improving
the infrastructure (compiler, IDE, etc.) so that when anyone does come
up with an idea, it's easy for them to build and try it.

fonc mailing list
fonc at vpri.org

More information about the FoRK mailing list