The Second Emergency All working notes

Working Note 01

The second emergency has no owner

Published 19 September 2026

Every emergency has two layers. The planning cycle is strongest at managing the hazard and restoring the systems it disrupts.

The first layer happens to systems. The grid, the network, the water plant, the dispatch centre. Emergency management addresses it across five components, prevention, mitigation, preparedness, response and recovery, and it addresses it well.

The second happens to people. It is not the failure itself. It is what the failure does to the person who needed something that day.

The five components do not ignore people. The problem is that responsibility is divided by service, sector and mandate, and nobody necessarily owns the consequences as a whole. Recovery reaches furthest into the future, but it is still organized through services, programs and responsible bodies. Consequences that cross all three can remain visible to everyone and owned by no one.

The distinction that matters is not timing. That is the thing I had wrong when I began writing this.

The second emergency can be immediate

On 17 September 2026, 911 service was unavailable overnight across Ontario and in three other provinces. As of this note, no cause has been announced. I do not know the cause and assign none. I draw no connection to any other event, and I use this here to describe the shape of a problem rather than to say anything about its origin.

What it shows is that for this class of failure there is no lag. Somebody needed help that night and the route to help was gone. That is not an aftermath. It began in the same minute as the first emergency.

The same is true of payments. When the card does not work and the deposit does not arrive, the consequence is not pending. It is happening at the till.

The same is true of medication. A person may still have insulin in the fridge. But renewing a prescription, confirming coverage, obtaining pump supplies or reaching remote support may depend on several connected systems working at once. When those systems stop together, an administrative failure can become a medical one.

Many familiar hazards provide observable cues, mapped areas, forecasts or established response triggers. A connected-systems failure can remove service across several sectors and jurisdictions at once. Here there is no ramp. It is on, and then it is off.

The second emergency can also take years

At the other end of the range is the pandemic, the largest recent test of this problem.

Pandemic planning, and I worked on some of it, was necessarily organized around the disease and the continuity of essential services. It did not adequately anticipate the scale of what followed: prolonged isolation, persistent learning losses, and rising violence in some homes. By 2021, police-reported intimate partner violence was 4 per cent higher than in 2019, and family violence against children and youth was 9 per cent higher.

And then the part that should unsettle anyone who plans for a living. The most confidently forecast indirect harm did not arrive. In the spring of 2020, serious researchers predicted that pandemic unemployment would produce a surge in suicide. The surge did not appear in the Canadian mortality data. A peer-reviewed study found that suicide mortality fell during the first pandemic year. The Public Health Agency of Canada's current series describes suicide mortality as relatively stable from 2020 through 2023, with the exception of a peak in 2022. CMAJ later asked why the predicted tsunami of mental health crises had not materialized as expected.

That is not reassuring. It means the forecast increase did not arrive in the period when it was expected, while harms receiving far less attention accumulated in schools and in homes.

A plan built as a list of anticipated consequences will therefore be confidently and specifically wrong.

Why the layer goes unowned

There is a structural reason, and it is not negligence.

During an event, a municipality may activate its Emergency Control Group to coordinate priorities, resources, information and service continuity. Each member may own part of the consequence, but the table may still have no single person responsible for seeing how those consequences interact.

Some consequences begin immediately. Others continue arriving after stand-down, into a room that has been emptied by design.

And the effects do not sit in one place. They land in social services, public health, schools, housing and income support, which in Ontario are spread across different departments, different boards and different orders of government. No single office can see the whole of it, and none of them is in breach of anything.

What follows

If the consequences cannot be forecast, the plan cannot be a catalogue of them. It has to be a structure. Who convenes, on what trigger, watching which indicators, how often, and with what authority to act.

Which produces two questions that any municipality can ask itself today, at no cost.

In the first hour, who at the table is responsible for the people rather than the system?

Thirty days after it is declared over, who calls the meeting, and under what authority?

If nobody in the building can answer either one, the gap is already located.

What comes next

A later note will deal with a second problem, and I think it is larger than the one above.

A credible public warning is itself an event. When people are told a system may fail, some act on the warning, and some of those actions can produce the failure that was forecast. Robert K. Merton named this the self-fulfilling prophecy in 1948. His opening illustration was a sound bank destroyed by depositors acting on a false belief that it was failing.

Merton's loop still moved through conversations, newspapers and people physically arriving at the bank. Today a warning can reach millions of people, and produce coordinated action, before an institution has issued its first response.

Three questions follow, and this note answers none of them.

What a population actually does when given a dated warning of technological failure. Y2K offers a rare large-scale test of this and should be examined from that perspective.

Whether the findings of disaster research extend to this case. Since the 1950s, work associated with Enrico Quarantelli and the Disaster Research Center has repeatedly found that mass panic is rare, social order usually persists, and cooperative behaviour commonly predominates. That evidence comes largely from bounded events with a recognizable hazard, an affected area and an eventual end. It does not answer the question posed here: how people behave after a credible warning that the connected systems supporting ordinary life may fail indefinitely.

And whether the first question changes the threat model. If a response to a warning can produce a failure, then the capability required to cause harm is not the capability to break a system. It is the capability to make the break credible. Those are very different numbers.

None of this concerns whether any particular forecast about artificial intelligence is correct. I take no position on that and this project never has. It concerns what people do, which happens either way. Through 2026 a United Kingdom campaign called Pull the Plug has organized protests against data centres, including a London march in February covered by MIT Technology Review. Whatever one thinks of their position, public behaviour is already moving ahead of any infrastructure failure.

What I do not know

This note is dated and published because the argument should be on the record, not because it is finished.

I do not know whether any Ontario municipality has a standing body that survives the stand-down, or indicators it watches through a recovery period, or a named owner for the consequence layer during an event. I do not know whether this already exists somewhere under a name I am not using. That has happened to me before on this file, which is why I ask rather than assert.

If you work in a municipality, in public health, in a school board or in social services and you know the answer, I would like to hear it, particularly if the answer is that this is handled and I have missed it.

Write to angela@lindow.ca.

A second volume will follow. What goes into it depends on what comes back.

Angela Lindow

Sources

If you are in Canada and need support, call or text 988.

All working notes Print or save as PDF Back to the documents

Free to use, adapt and reproduce. No permission required, no fee.

Working notes are dated when published. If one is revised, the revision is noted on its page. Notes are never silently edited and never backdated.
TheSecondEmergency.com  ·  angela@lindow.ca