In security

The more locks I add to a door, the harder it becomes to open. But that difficulty is blind; should I lose my keys, then I too will be locked out. Therefore, the more locks I have, the more necessary it is to have spare keys somewhere else. However, if there is another set not with me, this always opens up the possibility that someone else will find them and use them to unlock the door. So I hide them. The more secure I wish to be, the more difficult I make it to find them. But again, the impartiality of my acts rebound on me: if I lose my keys, and have hidden the second set well, I too may never retrieve them.   I may then be tempted to hide a third set in a more accessible place - and weaken the security at a stroke. This is the dilemma of locks. 

There is an equivalent dilemma in the use of computer passwords, which are a kind of verbal key. To prevent others gaining unauthorised access to my system, I may require that a password be given. This will be a word or phrase chosen by me, and known only by me.  Without this password access cannot be gained. So were I to forget this word, I would myself be locked out. I must therefore choose a password that is immediately obvious to me; but if it is too obvious - my name, date of birth, telephone number - then it will be nearly as obvious to someone else who can then gain access as if they were me. If it is not so obvious to me, I run the risk of losing it amidst the flood of thousands or millions of other words that are not quite central to my life. 

If I have a piece of equipment, I may be worried that some delicate and crucial component is faulty, and is therefore vitiating the equipment's performance. But to test that part, it often happens that I need first to disassemble the equipment. In doing so, I may know that there is a small but real probability that I am damaging the part that prompted the disassembly in the first place. At the end of the test I am therefore no more secure in my knowledge about the state of the crucial component: after disassembly I know it was undamaged, but the very process of testing and re-assembly may well have induced precisely the fault I first feared and sought. Just as I can never be sure that my locks or my passwords are totally secure and safe, so I must accept that I can never be certain about the state of this piece of equipment, 

Even the brain is subject to these types of dilemmas.  When I press the switch on the wall, the lightbulb is illuminated. Every time I have pressed the switch, the lightbulb - if functioning properly - has been illuminated; it has never been a hit or miss affair.  And so I might reasonably assume that pressing the light switch is linked causally to the illumination of the bulb.  But to do so is to make a huge leap that what has always happened in the past will always happen in the future.  In fact, we have no grounds for such a leap. No matter how often we return to the switch, press it and observe the illumination of the lightbulb, it still tells us nothing about the next occasion we do so. And each time we press the switch, and observe the illumination, we have not proved that induction was right, because this test of the future immediately becomes part of the past, just another example from which no conclusion can be drawn.  This is the dilemma of induction, the ultimate insecurity. 

(17.12.89)

Index