Conversation and Signaling

What is conversation?

“The primary way of sharing information via language use” you might say.

Great answer! A+. We are done here. Maybe just some totally true aspects of conversation that completely adhere to the “information sharing” explanation of conversation.

1. We love to listen

If conversation is information sharing, surely the optimal strategy is to listen. Assuming that information acquisition is costly, and I am getting it for free in a conversation, listening is the closest we get to a “free lunch”

2. We hate to talk

Talking in a conversation is the worst idea. It costs you on two fronts: 1. You are giving away information that you might be better off keeping to yourself and using when beneficial. 2. The opportunity cost of talking is immense: the time you used to give away valuable information for free took away time for you to acquire free information.

3. We have a strict conversation logbook

Since we all love to listen and hate to talk, we need to find some contractual agreement for our conversation where we make sure not to get cheated. If I give the other person valuable information, I expect valuable information back! Thus before I meet anyone for coffee to tell them about very important stuff I rate the importance on a scale of 1-10. I send a little whatsapp saying “hi Juashimir, I have important information (scale 8) which might interest you. If you want to hear it I expect a 9-information back. Last week when I gave you the 4-information, you only came back with a 3 in my opinion. Pls compensate for my previous invaluable information”

4. Conversation is about information, not topic

Last time meeting Juaschiemizu we had a great conversation, Information flow was excellent.

Me: Hallo

Joshi: Hi

Me: Economists think at the margin

Juaschimir: Like the Möbius strip, the Klein bottle is a two-dimensional manifold which is not orientable.

Me: Although pre-twentieth-century naturalists such as Charles Darwin made game-theoretic kinds of statements, the use of game-theoretic analysis in biology began with Ronald Fisher's studies of animal behavior during the 1930s.

Jaeiousch: Beginning in late 1895, Herzl wrote Der Judenstaat, which was published February 1896 to immediate acclaim and controversy.

Me: Thanks, Bye.

J: Bye.

Getting Real

Imagine a world in which people don’t listen, are constantly eager to speak, disequilibria of conversation time and information value accrues, and worst of all people focus on topics rather than acquisition of information. It’s a world in which conversation is about signalling, not information.

Convos go way more like this:

Me: HI Juaschimir, whats up? I was super smashed yesterday! Were you there?

J: LOOL, of course I was there, but guess what I didn’t see you either. I was talking to this hot chic about god knows what.

Me: oh now I remember, the blond one. Lisa?

J: Yes Lisa, the sister of Emil.

Me: Ow how is Emil? I heard he works for the Trump campaign!

J: Who in the right mind works for Trump?

M: Are the election results out yet?

J: I have no clue, I don’t follow this stuff anymore.

M: Me neither. I’d rather talk to chicks like Lisa…

Rather than sharing information, conversation is an arbitrary moving from one topic, to a tangentially related topic, back to the original topic, and sideways to an associated subject. If you pay close attention. It happens every sentence. A person says a sentence. The counterpart latches on to one word or a phrase and moves to a slightly different topic with another sentence. This is then repeated until the conversation ends. In the above convo the pivoting keywords were: smashed, meeting, chic, Lisa, Brother, Trump, election, interest, Lisa. Notice that these are different but related topics, all containing very little and useless information.

So why do we talk like that? And why are we so wrong to believe that conversation is about information?

Maybe a little topical tangent here: Chimpanzees pick lice off other chimps. If you ask why Chimps pick lice off each other, you might say because they want to get rid of lice. That’s actually not true: There is no correlation between those who need lice picking and those who get lice picking, and the amount of lice picking does not rely on reciprocity. Rather mating patterns and hierarchy among chimps explains who pics lice and gets lice picked.

Chimps signal their interest or loyalty by picking lice. It’s a costly action that sends a credible signal. Humans do the same.by wanting to talk, in the way we talk (tangential topics), we signal to our friends, family, potential spouse, that we know stuff, that we can assert ourselves and that we spent time and effort doing so. Dancing around a random topic in the fashion described above is costly, it takes mental capacity and organization.

Just like the chimp doesn’t pick lice, we don’t share information in conversation. We both signal loyalty and interest.

(This blog post is more or less a rewritten summary of the conversation chapter of "Elephant in the Brain", by Economist Robin Hanson)