Scientists say the secret of a good conversation is 'pink noise'
Researchers at Dartmouth College in the US found that more enjoyable, more connected conversations show more 'pink noise', a balance between stability and change. The study is published in Science Advances.

Researchers at Dartmouth College in New Hampshire, US, have concluded that the quality of a conversation can be described using 'pink noise', a concept normally used by physicists. The findings, published in the journal Science Advances, help explain what makes a chat enjoyable.
What pink noise is
First author Kathryn O'Nell explains that in physics pink noise describes a state on the edge of chaos: a system can stay at a stable point without being stuck there for ever. In conversation, that means lingering on a topic long enough to get to know each other and understand it, but moving on before it turns stale.
The team contrasts three kinds of conversation. White noise arises when people fail to follow what was just said: the talk jumps around, or one person dominates and seems oblivious to the other's replies. Red noise arises when people mirror each other's language and fall into a rut. Pink noise sits between them, as exploring a topic alternates with bringing in fresh ideas.
How the study worked
The scientists used an AI program called MPNet to score every turn of each conversation on 768 criteria, including meaning, tone and context. They analysed more than 1,000 conversations between strangers, who answered questions about the chats afterwards. Conversations with more pink noise were enjoyed more, and participants wanted to keep talking and felt a stronger connection.
A similar effect appeared in conversations between friends about their formative experiences. In speed-dating conversations, more pink noise predicted which pairs met for a second date.
The team also analysed podcasts with two hosts and found that the pinkness of their conversations increased over the first 50 episodes. O'Nell suspects the skill can be learned.
A note of caution
Prof Elizabeth Stokoe, a conversation analyst at the London School of Economics, praised the move away from the idea that more similarity is better. She added, however, that asking whether a conversation is enjoyable risks missing the point, since many of the most important conversations in life are not enjoyable.


