Why Milwaukee dispatchers are learning sign language and text protocols

For most of the history of the emergency number, if you could not speak on a telephone you could not reach help directly. That is a remarkable sentence and it was true until very recently.

Text to 911 has changed it, and the change is still working its way through, unevenly, centre by centre.

What text to 911 actually is

You send a text message to the digits 911 and it arrives at a dispatch centre, where a call taker handles it as a conversation.

It is not available everywhere. Availability depends on whether the local centre has implemented it, and the Federal Communications Commission maintains a list of which jurisdictions have. Coverage has expanded substantially and there are still gaps.

If it is not available where you are, you get an automatic reply telling you to make a voice call. That bounce back message is required, which is a small piece of good regulation.

Who it is for

The obvious group is people who are deaf, hard of hearing, or have a speech disability, and for them this is a fundamental change in access rather than a convenience.

The group nobody designed it for, and who use it heavily, is people who cannot safely speak. Someone hiding from an intruder. Someone in a car with the person they are reporting. Domestic violence callers, who are a substantial share of text traffic in centres that have it.

It also works when a voice call will not connect but a text will squeeze through a congested network, which happens in mass events.

Why it is harder than it sounds

Everything that makes a voice call work is missing.

Location. Voice calls carry location information with varying precision. Text location data has historically been less precise, which is why every piece of guidance says to give your address first, in the first message.

Tone. A call taker gets a great deal from a voice: panic, intoxication, injury, whether someone else is in the room. Text gives you none of it.

Speed. A conversation that takes ninety seconds by voice takes several minutes by text, and the call taker is typing while dispatch is waiting.

Ambiguity. Short messages, abbreviations, autocorrect. Call takers have had to learn to ask clarifying questions in a way that does not waste exchanges.

So centres have written specific protocols: short questions, one at a time, confirm the address early, never assume, and a defined way to handle a conversation that stops.

The sign language part

Video relay and video remote interpreting let a deaf caller sign to an interpreter who speaks to the dispatcher. That infrastructure exists and it works.

Dispatchers learning some sign language is a different thing and it is mostly about the in person side of the job and about responder training, and about not being entirely dependent on a relay that takes time to connect.

It is not a replacement for professional interpretation and nobody in the field claims it is.

What people should know

Call if you can, text if you cannot. Voice remains faster and carries better location data.

If you text, put the address in the first message. Not your name, not what happened. The address.

Then a short description of what is needed. Police, fire, ambulance.

Do not use abbreviations or emoji. Do not send pictures or video, which most centres cannot receive.

Do not text and drive, which sounds ridiculous in this context and is a real problem.

And do not send a group text. It will not work.

What I think about it

I was on the desk when this was coming in and there was resistance, some of it reasonable, about speed and about location.

The resistance missed the point. For a category of people, the alternative to a slow imperfect text conversation was no direct access to emergency services at all, in a country where that service is treated as universal.

Slower and available beats faster and unavailable. That is the whole argument and it took a surprisingly long time to win.