Catch up on stories from the past week (and beyond) at the Slashdot story archive

typodupeerror
DEAL: For \$25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 Internet speed test! ×

## JournalJournal: /. users can't handle fuzzy logic2

There are a lot of dumb posts on /. Many readers seem willing to argue extreme positions, in violation of common sense. Obviously, this site is composed mainly of university-educated geeks -- the kind of people that you would normally think of as intelligent. So why are so many of the comments on /. so dumb? Why do they so frequently fly in the face of common sense? How can so many readers shirk reality in favour of wishful thinking?

After much pondering, I think I have discovered the answer. It is based on a profile that identifies the typical /. reader as:

1. University educated
2. Intelligent
3. Computer programmer

Points 1 and 2 are fairly obvious. University students tend to be idealists, and intelligent people are often good at rationalization, which is the misapplication of logic. But the "computer programmer" factor adds an additional factor that I hadn't noticed until recently: Boolean logic.

Computer programmers are used to dealing almost exclusively with Boolean logic. That is, after all, the way that computers work... digital, binary, precise. So I guess it should come as no surprise that many programmers have trouble debating real world issues, which tend to involve fuzzy logic, statistics, and shades of gray. (Actually, it comes as a little bit of surprise to me, since *I* am a computer programmer and *I* prefer statistical reasoning to Boolean logic.) This is not to say that members of the general public are likely to know anything about statistics, but I have higher expectations for someone who is educated in math and the sciences.

Perhaps the disctinction between the techniques is not clear to you, but it makes a big difference. Consider the following logical argument:

1. A || B
2. ~A
=> B

Translated into real world terms, "We have to choose either A or B. A is bad, so we had better choose B." This exhibits several characteristics of black and white thinking. Ignoring the possibility of a 3rd option C, this immediately assumes that you have to choose either A or B and not a compromise between the two. The second error is in assuming that just because A is bad, it can't be preferrable to B. Sometimes you have to choose between the lesser of two evils.

Try applying the above logic to any of the common arguments /. readers make about needing to have absolute freedoms, absolute rights to fair use, absolute anything.

Here's another one:

1. There exists some x:A such that A=>B
=> ~(~B)
=> B

Translated into plain English, "B is possible, therefore B is true." Here is another example where Boolean logic differs from statistical reasoning. To a statistician, one piece of anecdotal evidence doesn't prove anything, but in Boolean logic it proves that B is possible. For example, if one open source company out of 1,000,000 makes a profit, a Boolean logician will say that it is possible to make money selling open source software and a statistician will say that it isn't.

So what am I saying here... That now that we have recognized our differences we should celebrate our diversity? No, of course not. Boolean logic may apply to the limited field of computer programming, but it should not be used to decide social issues... So stop it already!

This has been another entry in my ongoing series of journal entries that critique the many false arguments you read every day on /.

-a

## JournalJournal: Where do they get their money?

Is /. now entirely Microsoft funded? I haven't seen a non-MSDN ad for over a week now.

-a

## JournalJournal: Why I don't argue by analogy on /.27

Over the last few years, I have noticed a disturbing trend: /. readers really like to argue by analogy. State an argument using facts or statistics or common sense and you're likely to get an analogy in return.

"Open source software is like a car." "Copying music is like reading a book." "Source code is speech."

Well, I'm sick of it. The problem with arguments by analogy is quite simply that they aren't logically valid. The fact that two things are alike in one aspect does not necessarily imply that they are alike in another. In many cases they are, but you have to understand that when you are debating a moot point, the "my analogy is better than your analogy" defence isn't going to cut it. A "preponderance of the analogies" is not going to produce a verdict.

This is not to say that analogies aren't good for anything; they make an effective educational tool. I had a high school physics teacher who was very fond of describing an electric circuit by analogy to a system of water pipes. The width of the pipes corresponded to the resistivity of the wire, pumps were like batteries, and the height of the water above ground was its voltage. But it is essential to understand the limits of the analogy. [Resistive] electric circuits and water circuits match up nicely because they are both first order systems; therefore, the analogy only holds so long as you are comparing first order effects. There is no guarantee that I'm going to get a sane result if I try to add inductors and capacitors to the water circuit or turbulence to the electric circuit.

Analogies work best in situations where the audience is receptive to your ideas, such as the above teacher-student relationship, but they are practically useless as a debating tool. Debates about social policy are the worst because they are so subjective. If you use an analogy then your opponent will nitpick it, likely claiming that it is a "false analogy" because it doesn't correspond in some minor aspect (that you consider irrelevant but he claims is paramount). If your opponent uses an analogy, chances are that you can reduce it to a straw man, a slippery slope, or plain old over-simplification. In an atmosphere such as this one, I feel it is best to avoid analogy-based arguments and stick to the specific issue at hand.

Let's take an example: "Sharing music is not stealing. Stealing involves depriving someone of property." On the face of it, this looks like a semantic argument. (Common usage of the word "steal" has changed over the years to the point where we can talk about "stealing an idea", but on /. it is a loaded word.) But if you look deeper, you will see that this is also a refutation of an analogy in disguise. What your opponent is really saying is that pirating music is not analagous to stealing toasters because of all-important reason X (where X is his semantic argument).

On one hand, you could attempt to prove to your opponent that his definition of stealing is arbitrary by making the perfectly reasonable observation that if you hired a guy to fix your roof and then refused to pay him, that wouldn't be depriving him of property either. (Most /.ers would consider this case to be morally wrong.) Unfortunately, that argument would be the wrong approach, not because your observation lacks merit, but simply because using it would invite an analogy war.

Of course, once you have avoided the analogy trap, you still have to follow through with a coherent argument. Instead of analogies, I would suggest all of the following approaches: facts ("Whether or not you call it 'stealing', the fact is that music piracy is illegal."), statistical analysis ("The observation that music sales increased during Napster's operation is misleading because it doesn't take into account the effect of statistical lag."), common sense ("Kazaa searches may generate a lot of false positives at the moment, but there are some obvious ways to improve their algorithm."), evaluation of human nature ("Why do you think Americans will donate money to support a band when they don't even donate 1% of their income to charity?"), and appeals to political orientation ("Times change. Intellectual property falls within the spirit of capitalism, regardless of whether Adam Smith ever wrote about it in the 1700s.").

Therefore, I hereby vow not to use analogies on /. Should I feel the temptation to analogize, I will rephrase the example to explicitly state my logic. If you catch me in analogy, flame on. And of course I encourage you to join the boycott.

Note: Misuse of analogies is only one of a wide variety of false arguments that are common on /.. The next worst offender is probably slippery slopes. I will discuss these in a future journal entry.

-a