Baxnet Ideas · Founder note
The Typing Indicator Is a Tiny Psychological Thriller
TL;DR: The typing indicator contains almost no information. That leaves just enough space for the recipient to invent the rest.
A friend writes, “Can I tell you something weird?”
You say yes.
The typing indicator appears. It disappears. It returns for long enough to suggest that the first attempt was not quite right. Then it disappears again.
Within forty seconds, you have written six possible endings. Somebody is moving abroad. Your name came up at dinner. There has been an incident. You are about to be asked for money, forgiveness, or a lift to the airport.
The eventual message is: “Never mind. Sorted.”
The typing indicator contains almost no information. That is what makes it so effective.
It reveals that something is happening while withholding what it is. The interface supplies three animated dots. The recipient supplies the plot.
The dots announce work, not meaning
Typing indicators are easy to treat as a small piece of decoration around the real conversation. Technically, they are a separate event.
Signal describes them as optional indicators that let people see and share when messages are being typed. They can be switched off, and both people need the feature enabled for the status to appear.
The distinction becomes even clearer in the XMPP standard for chat-state notifications. It defines composing and paused as separate states. Composing means that someone is interacting with the message input. Paused means they had been composing and have stopped.
This is usefully mundane. Neither state says what the person is writing, why they stopped, whether they deleted anything, or whether a message will arrive.
The sender may be producing a careful answer. They may also be correcting autocorrect, changing a restaurant booking, being interrupted by the kettle, or deciding that the joke was not as good as it seemed eleven seconds ago.
A typing indicator is not a promise of a message. It is evidence that, briefly, there was activity near the input box.
We write the missing message ourselves
Researchers who study computer-mediated communication treat interface details like these as social cues. A 2020 paper on how cues work in emerging media notes that even minimal information, including an indication that someone is typing, can be processed by the receiver as part of the message.
That does not mean the cue has one stable meaning.
After “What time shall we meet?”, the dots usually promise logistics.
After “I think we should talk”, the same dots have secured funding for a full psychological thriller.
If they appear beneath a manager’s name after you ask about your performance, every pause looks deliberate. If three people begin typing after a questionable message in a family group, the interface seems to announce an incoming tribunal. Nobody has said anything yet, but the atmosphere has already changed.
The recent turn, the relationship, and the stakes of the subject do most of the work. The indicator merely gives our anticipation somewhere to attach itself.
There is a good reason for that anticipation. In face-to-face conversation, we can see someone preparing to speak. We have posture, eye contact, breathing, hesitation, and sometimes the mercy of being able to leave the room. Messaging removes most of those signals. The typing indicator returns one narrow form of presence: the other person is still here and may be forming a response.
Research on early text-based online communities observed that response delay could create ambiguity and alarm. A missing reply could look like agreement, disagreement, indifference, absence, or technical trouble. The dots reduce one uncertainty—some activity is taking place—while creating another: what activity?
The most dramatic draft may be the one nobody sends
It is tempting to interpret the duration of an indicator.
A long spell of typing feels like evidence of care, anger, difficulty, or a very substantial paragraph. Then the person sends “yeah”.
Perhaps they wrote a substantial paragraph and deleted it. Perhaps they received a phone call. Perhaps they typed “yeah” with the concentration of someone defusing a bomb. The visible thread cannot tell us.
This matters if we later analyse a conversation. Once the dots disappear, they are normally absent from the chat history we can inspect. The final message remains. The waiting does not.
That missing experience can be important. The person receiving “yeah” after ninety seconds may have felt a very different exchange from the person reading the same word in an export months later. But missing context is not permission to invent the abandoned draft. It is a reason to be modest about what the record can support.
For personal chat intelligence, the boundary is simple. A sent message is evidence. A typing indicator, if it is available at all, is evidence of a temporary interaction state. Neither is a reliable transcript of everything the sender considered saying.
The typing indicator tells us one small thing: somebody began composing.
The psychological thriller starts when we mistake the rest of our imagination for evidence.
Further reading: Typing Indicators from Signal Support, XEP-0085: Chat State Notifications, and Explicating Cues: A Typology for Understanding Emerging Media Technologies.