Conversation Design: The Cooperative Principle

For most people, it is natural to talk. We do not even have to think about it. Speech is what separates us from most animals.

We have these amazing communication techniques that we apply all day long that we are not even aware of. We use words, facial expressions, gestures and we consider context to share the message.

Cooperation is a big part of that. When we communicate we always try to help each other out and be cooperative. We provide information that’s truthful, we provide the right amount of information and we do it with the right tone of voice. Paul Grice is the researcher that identified this. He called it the Cooperative Principle.

The Cooperative Principle

Paul Grice identified 4 maxims that make up the cooperative principle. You could call these 4 maxims collectively the engine of communication. They are the natural force that powers our conversations. They are so crucial that we should list them:

Maxim of Quality

When people communicate, we assume that we are being truthful. To move a conversation in the right direction, we have to provide truthful information of high quality. Otherwise, the conversation is not going anywhere.

Maxim of Quantity

When people talk, we always provide as much information as needed. This comes natural to humans, but it is actually difficult to do for bots. For example, if you would ask the question:

User: do you know the time?
Bot: it’s 4 pm

This would be the right amount of information. But if we were dealing with a bot that is not being cooperative, he could answer the question by saying ‘yes’ or maybe say ‘in Amsterdam it’s 6 pm, in London it’s 5 pm, in NYC it’s 12 pm’.

In those two cases, the bot is actually answering the question, but he is not offering the right amount of information. A human would never do this.

Maxim of Relevance

Everything that is being said in a conversation is assumed to be relevant. Sometimes we might say something that seems a little off, but then our brain allows us to process the information, skip a few steps, and respond accordingly. It allows us to imply certain information.

John: Do you have everything packed?
Jane: Aren’t we leaving tomorrow?
John: Uber will be here in 20 minutes… get it done now!

In this example, John assumes that Jane has not packed yet and therefore tries to rush her since the Uber will be there soon. A bot would never understand Jane’s respond in this dialogue, because it has trouble with inferring meaning.

Maxim of Manner

When humans speak, we adjust our tone. We use the right tone of voice for the situation and we adjust accordingly. Think about it, you probably speak differently to children than you do to colleagues or adults in general.

Simple for humans, difficult for bots…

These components make up the engine of communication. The Cooperative Principle is what powers human conversations and as conversation designers, we have to be aware of this.

But now we have to wonder what speeds up this engine. We have to identify the gearbox of this engine.

Turn-taking as the gearbox of conversations

Turn-taking is what speeds up the conversation and it is an important component of conversation design for chatbots and voice assistants. By letting each other know when it is our turn to speak, we can speed up the conversation and make it more efficient.

The best way to let the other person know it is their turn to speak is by using a prompt. A prompt is a question that leads to action. Asking a question lets the other person know it is his or her turn to speak.

That is a prompt that lets the user know it’s his turn to speak. This makes the conversation nice and efficient. There is little room for error.

A mistake we often see companies make, especially when it comes to chatbots… Is that they add extra information after asking a question. This creates noise in the interaction.

Here is the thing, if you ask a question then people want to answer it. So you ask a question and your user gets ready to answer, because it feels most natural, but now you are adding additional information which confuses the user.

Short sentences with simple questions are more efficient. It allows for better turn-taking, which speeds up the conversation… Like a gearbox!

Let’s look at a full example.

Final thoughts on the Cooperative Principle and Turn-Taking

As conversation designers, we have to understand how humans communicate. Only when we understand how the Cooperative Principle powers human communication, we can start deconstructing. When we comprehend its parts, we can start designing for it.

Since we want to keep it simple for our users, we can use Turn-Taking. This is a solid principle that keeps the engagement simple and efficient.

Like we said in the intro: the cooperative principle is the engine that powers human conversations, turn-taking is what speeds up the conversation.

Originally published at https://blog.robocopy.io on August 29, 2019.