The Bandwagon of Information Theory

The Talk

This is perhaps my most famous talk, in which I warned against the excessive application of information theory to fields where it didn’t belong.

The Warning

“Recently there has been a great deal of talk about ‘information theory’… The field has been overworked… We should like to issue a warning and at the same time a reassurance.”

I expressed concern that information theory was being applied to:

• Psychology
• Linguistics
• Biology
• Economics
• Even religion!

The Reassurance

I clarified that information theory was a mathematical tool with specific applications, not a universal explanation for everything.

What I Said

1. Information is not meaning: The mathematical concept of information is about uncertainty reduction, not semantic content
2. Don’t force fit: Not every problem is an information theory problem
3. Stay rigorous: Maintain mathematical standards

Reception

The talk was controversial. Some appreciated the caution, others felt I was being too restrictive.

In retrospect, I think I was mostly right - many early applications of information theory were overblown, though some (like bioinformatics) turned out to be genuinely useful.

---

Not every problem needs information theory. Some problems are just problems.