Is Suffering as Real as We Think It Is?

Two writers I generally respect have said things about suffering that I want to push against, because they are said with great confidence and I do not think the confidence is earned.

Sam Harris has argued that if we accidentally build machines that are conscious and suffer, we will be to blame for that suffering, and that this is among the more serious moral risks of the technology. Yuval Noah Harari, asked how he decides what is real and what is a story, says that his test is a question: can it suffer? Nations cannot; a cow can. Suffering is his criterion of reality.

In both cases suffering is treated as the central fact about consciousness, the thing that makes a mind matter. I want to raise two doubts.

Doubt One: Consciousness Without Suffering

The first is simple. It may be possible to have consciousness without the ability to suffer at all.

I have argued that pain is the alarm a mind needs when it has the power to override its own protective signals. If that is right, then pain is a feature of a particular kind of mind, one that evolved inside a fragile body it was responsible for keeping alive. It is not a feature of mind as such. A conscious system with nothing to protect, or with no authority to override its own protections, would have no use for pain and no reason to have it.

That means the machine Harris worries about is not the default outcome. Building a conscious machine and building a machine that suffers are two separate engineering achievements, and there is no reason to assume the second comes free with the first. We might well produce awareness with nothing at stake in it.

If so, then suffering is not the defining feature of consciousness. It is a contingent addition, one that our lineage happened to acquire, and we have generalised from our own case to all possible cases, which is the oldest mistake in the study of minds.

Doubt Two: Who Says It Is Real?

The second doubt is harder to hold, and I am less sure of it.

Harari says suffering is real. Real to whom? To us, evidently; nobody who has been in pain doubts that something was happening. But consider another thing that is real to us: the solidity of objects.

The table in front of me is solid. I can lean on it. Nothing is more obvious. And yet, at the level physics describes, it is almost entirely empty space, with a scattering of particles and the fields between them, and its solidity is a fact about how my hand's electrons interact with its electrons, not a property the table has in itself. Solidity is real as an experience, and it is a representation of something that is not, in the underlying description, solid at all.

Perhaps pain is real in exactly that way, and no other. Real as an experience, undeniably. But a representation, produced by a nervous system, of something that in the underlying description is a pattern of signalling, and no more intrinsically terrible than a table is intrinsically solid.

If that is right, then our certainty about the reality of suffering is of the same kind as our certainty about the solidity of tables: complete, immediate, and not evidence of anything except how our machinery presents the world to us.

Why We Cannot See This

There is an obvious reason we treat suffering as the most real thing there is, and it is not philosophical.

Suffering is the thing we most fear. It is the state evolution built us to avoid above all others, because avoiding it was how our ancestors avoided damage. A creature built that way will experience its own suffering as maximally real, maximally important, and maximally beyond question, because a creature that could doubt its own pain would have been eaten. The unquestionability is a feature of the alarm, not a discovery about the world.

So when a thoughtful person says that suffering is the touchstone of reality, I hear the alarm speaking. It is exactly what an alarm designed never to be dismissed would say about itself.

Where This Leaves the Machines

If both doubts have force, then the moral panic about suffering machines needs adjusting in two directions.

It may be far less likely than assumed, because suffering is not a free consequence of awareness and would have to be built in on purpose or by a very specific accident.

And if it did occur, it may be a smaller matter than our own reaction to it suggests, because our reaction is calibrated by a system whose job is to overrate pain. I do not say it would be nothing. I say we are the wrong instrument to measure it with, and we should know that about ourselves before we legislate.

I hold the first doubt firmly and the second loosely. But I notice that the second is the one nobody is willing to entertain, and that the unwillingness feels less like an argument than like a flinch.