Through the Keyhole

For many years I have been interested in the general problem of how easy it is to delude yourself with  logic. As a logician and computer programmer I am often involved in situations where there are literally millions or billions of things involved in a particular problem, and, of those million things, maybe one or two are slightly wrong and then only sometimes.

Solving these types of problems have taught me one thing of particular interest.

I call it the "keyhole dilemma".

Let's suppose I am in one room looking through an old fashioned keyhole in a locked door into another room.  On the wall next to me is a lever.  When I pull this lever a bell above (on my side of the door) is supposed to ring.  The keyhole only affords me a partial view of what's in the room.  Inside the other room is some kind of machinery about which I know only what I can see through the keyhole that causes the bell to ring in response to pulling the lever.

The bell does not always ring and my job is to fix it.

Ideally I would like to fix the problem by only using the keyhole - perhaps by inserting a wire and poking some switch or control that resolves the bell ringing reliability problem - this would use the least of my time.  However, I can spend more money and time cutting a larger hole (as well as fixing the mess) so that I can see more of what's going on if I think that will more quickly and cheaply solve the overall problem.   On the other hand I could make such a larger opening only to find that it was unnecessary and I only needed to use the keyhole.

Then again, part of the bell ringing machinery may be attached to the door, or worse require that the door be present and closed, in order to work.  So I might have to go through the wall instead - which is an even bigger mess and takes more time.  Then again if a monkey is ringing the bell tearing the door or wall apart might scare him off completely leaving not only the bell not working but me having to figure out what was on the platform and what did it do in order to ring the bell.

The real issue is then is that, through the keyhole, do I see enough of the problem to formulate a reliable solution.

Now inside the room there can be all manner of bizarre nonsense to ring the bell: a monkey on shelf who is poked by the lever that is trained to pull a string on the bell,  a string directly from the lever to the bell, some complex Rube Goldberg machine involving mice, bowling balls, a set of fun-house mirrors that invert my keyhole perspective, etc. - literally anything.

So now, standing before the door, what do I do to fix the problem?

On solution is to use science.  Science uses a process called the "scientific method".  Basically you collect observations about the problem until you can form a hypothesis about what you see.  You then reason about your hypothesis and conduct experiments to prove your hypothesis.

So, for example, I might see part of a rope through the keyhole and observe that, when I pull the lever, the rope becomes taut and then the bell rings.  I might also see that sometimes the rope does not become taut and the bell does not ring.

From this I would form the hypothesis that the rope is not solidly attached to the lever and that sometimes when I pull the lever the rope slips and does not ring the bell.

I may then sit down and try pulling the lever twenty times to see if that's indeed what's happening after which I take whatever steps necessary to attach the rope more solidly to the lever (perhaps by cutting a hole in the wall, etc.)

This seems reasonable, doesn't it?

Well, only if twenty tries is actually a useful amount of testing...

The problem, of course, is that through the keyhole I do no see that the lever is actually poking a monkey who pulls the rope to ring the bell and the monkey, being lazy, does not always do his job when poked by the lever.  Cutting a hole in the wall frightens the monkey who runs out the window leaving me with nothing to fix...

So while science might provide answers it can only do so if the perspective through the keyhole is wide enough to expose what is actually going on.  Otherwise science can only provide a perspective that is bound by the limits of what I actually observe - which in fact might be very little of the true problem.

So as I wrote yesterday in "Failing the Future" we have things like contraceptive hormones.

In this case the "keyhole" is the action of becoming pregnant.  The observation is that without ovulation pregnancy cannot occur so the "fix" is to block ovulation.

What is not "seen" through this keyhole is what else is involved in ovulation - the bigger picture as it were - involving selecting mates which would appear to fundamentally affect the most basic genetic aspects of humanity.  And then there is the impact on society as a whole - what is that impact and is it an improvement, and is it an improvement as compared to what else...?

Similarly for "anti-depressants" - through the keyhole we see "sad people" who we think should be happy.  We create drugs to make them happy but do not realize that perhaps there is a sound reason for their unhappiness and that it in fact serves a useful purpose.

Or for pain - we create oxycodone - but why do we have pain in the first place?   Through the keyhole we see only someone sad - so we make them feel better.  Did we foresee their subsequent addiction problem, its impact on their spouse, family or children?  On their life?  Did we trade two days of a sore back for twenty years of heartbreak? 

The "delusion" here is that we humans always see enough of the problem to create a safe solution.

In fact we don't - we can't even predict the weather reliably.

Human arrogance always tells the inventor or scientist that the product or solution he is creating will be the "be all, end all" of whatever situation is being addressed.  In the case of something like hormone contraceptives most of the scientists from the 1960's that created the technology are probably long dead. 

Yet as generations pass affected by consequences of this the true consequences are only revealed.

The reason I am writing this today is that modern society is more and more focused only on the view through the keyhole.  Solve what we see of the problem right now.  "Look, see, isn't the solution wonderful!" while in fact the solution is slowing and insideously destroying those who embrace it.

The next iteration of "science" will not be to refine the existing scientific method but instead to step back and study the "keyhole dilemma" - how do I make sure  that what I am doing is not harmful and in fact useful, how do I make sure it reveals "enough" of the real problems, what is "enough" in order to go forward with one idea over another.

Just because you can does not mean that you should...