Re: [Phys-L] how research is done : exploring a maze using only local information

[John Denker <jsd@av8n.com>]

On 09/17/2015 07:34 PM, Joseph Bellina wrote:

> I do a lot of work with elementary teachers to help them understand
> what science is really about 

A commendable endeavor.

> because nothing in the education provides that. 

That's for sure.

Forsooth, it's even worse than that.  The usual education
plants some horrific misconceptions about what science is
and how science is done.

> The foundational idea is the being wrong is an
> opportunity to learn. It that's them a while to accept that but once
> they do risk taking becomes much easier

That's entirely true as stated.

Without meaning to disagree with that, let me suggest
that it should be the /second/ step.  There's a simpler,
less-demanding step that comes first, namely this:

   In many cases, you can learn just as much without
   being wrong.  You can and should learn from mistakes,
   but it's even better to learn without making very
   many mistakes.

I say to the class:  I'm going to do an experiment by
tossing this coin.  Clicker question:  How many of you ...
  a) ... hypothesize that it will come up heads?
  b) ... hypothesize that it will come up tails?
  c) ... think that (a) and (b) are stupid questions?

My answer is (c).  In real life, including science as well
as business, military, farming, et cetera, you should consider
*all* the plausible scenarios.  When the Boy Scouts plan a
campout, they consider the possibility that it will rain,
even if they hope that it doesn't.  Scientists are smart
enough to do the same.  If you consider *all* the possible
outcomes, you aren't guessing, so you can't be wrong.  Some 
of the possibilities on your list will turn out to be 
consistent with the outcome of the experiment, and some of
them won't, but since you didn't make a guess, you didn't
make a mistake.

Science is not some demented guessing game where the objective
is to predict some unpredictable event.

A hypothesis is something to be considered, nothing more,
nothing less.  You may have been told that a hypothesis is
a quote «scientific guess» ... but that's nonsense.  It's
not a guess.  Any given hypothesis might be likely, unlikely,
known true, known false, or whatever.  For example, in a 
proof by contradiction, the hypothesis is expected to be 
proved false.

    Since the definition of "hypothesis" is often disputed,
    I tend to avoid the word, and instead speak in terms of 
    /scenarios/ and /possible outcomes/.

When exploring a maze, when you come to an N-way branch point,
you should consider all N possibilities.  If you decide to
explore one of the branches and it turns out to be a dead 
end, that may be disappointing, but it was not a mistake;
it was not wrong; it was not a bad decision.  It did not 
/directly/ achieve the most valuable result (i.e. finding 
the cheese) but still it provided some value (i.e. some
information).  Information is valuable, because it increases
your odds of finding the cheese on a later step.  The cost
of obtaining information is part of the cost of doing business.

There is physics in this!  Information is a real, quantifiable
thing.  It is intimately related to entropy.

  Of course there is such a thing as a mistake.  You should
  definitely learn from mistakes ... but it's even better to
  learn without making very many mistakes.

  Typically, a mistake means you paid too much for too 
  little information.  It makes you sadder but wiser. (In 
  contrast, a non-mistake makes you wiser but /not/ sadder.)

Whenever you see somebody talking about "the" hypothesis
associated with a research experiment, you can be sure that
they have no idea what they're talking about.  In any 
experiment worthy of the name, there are multiple possible
outcomes, not just one.  There are hypotheses, plural.
Otherwise it might be a demonstration or an engineering 
exercise, but it's not a scientific experiment, and certainly
not research.  People do experiments to obtain information,
but if the outcome is known in advance, it provides zero 
information.

Last but not least, it must be emphasized that research
does not require explicit hypotheses at all.  For example,
when Lewis and Clark set out, they were not considering 
the hypothesis that there would be bighorn sheep in the
Rocky Mountains ... obviously not, because they had never
heard of either of those things.  Despite the lack of a
hypothesis, they were able to discover both of those things.

To be sure, they had a plan ... but it was not a detailed
plan.  Sometimes it is better to plan for flexibility than
to obsess over details.