Simulacra Levels

There's an old saying: "There's a lion across the river." At level 1, this means there's a lion across the river. At level 2, it means "I don't want you to cross the river." At level 3, it means "I'm with the popular kids who are too cool to cross the river." At level 4, it means "A firm stance against trans-river expansionism focus grouped well with undecided voters in my constituency."

This taxonomy — adapted from Baudrillard's concept of simulacra by Ben Hoffman and developed extensively by Zvi Mowshowitz — is one of those frameworks that, once you internalize it, makes previously incomprehensible behavior snap into focus.¹

The Four Levels

Level 1 is straightforward: words describe reality. You say what you see. This is the naive baseline, the thing language was presumably built to do — maintain shared Maps And Territories so your tribe can navigate the environment together. "The food source is over there" means the food source is over there.

Level 2 introduces strategic lying. You still understand that words are about the world, but you're willing to manipulate the shared map for personal advantage. Say there's food across the river, but you tell everyone there's a lion there instead, so you can hoard the food. The statement is a means to an end. Crucially, though, you still care about what's actually true — you just don't care about communicating it. The map is now an instrument, not a mirror.

Level 3 is where things get genuinely strange. At this level, people aren't even interpreting statements as claims about the world. "There's a lion across the river" is parsed as a bid to associate with the group that says there's a lion across the river. Arguments for or against the proposition are conflated with arguments for or against the implied social move. An attempt to simply investigate the literal truth — "wait, has anyone actually seen a lion?" — is considered at best naive and at worst politically irresponsible.

Level 4 is pure game. The pseudostructure of levels 2 and 3 breaks down. Statements are moves in a positional competition where everyone is trying to think a bit ahead of everyone else. Nobody is even pretending to describe reality, signal tribal loyalty, or pursue a specific consequence. They're just surfing the meta-game.

The Covid Case Study

Zvi's extended application of this framework to the early pandemic is the most concrete demonstration of why it matters.¹ When people said "there's a pandemic headed our way from China" in January 2020, they were operating on all four levels simultaneously, and the interpretation depended on who was listening.

Level 1 speakers were just trying to communicate an epidemiological fact. Level 2 speakers might have been trying to sell you masks, generate clicks, or scare you into compliance — not because the pandemic wasn't real, but because the truth or falsity of the statement wasn't why they were saying it. Level 3 speakers were signaling group membership: "I'm the kind of person who takes COVID seriously" or "I'm the kind of person who doesn't panic." Level 4 speakers were positioning themselves for whatever the next meta-level move would be.

The problem isn't that any particular level is "wrong." The problem is that the levels interfere with each other. If your society has spent decades operating mostly at levels 3 and 4, then when a genuine lion shows up — when someone needs to communicate a real, urgent, object-level fact — nobody remembers how to parse level 1 statements. Every attempt to say "no, literally, there is a virus and it will kill people" gets heard as a social move. The category "sincere factual assertion" has atrophied.

The Five Communicators

Zvi classifies speakers into archetypes based on which levels they attend to.² The Oracle only looks at level 1 — they say what's true and don't care about consequences. The Trickster only looks at level 2 — they say whatever produces the consequences they want, truth be damned. The Sage looks at both: they'll tell you the truth, but only if telling you the truth also produces good consequences. If telling you the truth would be harmful, they stay silent.

The Pragmatist is the realistic version of the Sage: someone who mostly tells the truth but has a breaking point. Given enough bad consequences, they'll lie. And the Nihilist doesn't care about either truth or consequences — they say whatever and then eat at Arby's.

The interesting thing about these types is that all of them are trustable in some sense, once you know which type you're dealing with. An Oracle is trustable in the obvious way. But a Trickster is also trustable — you just interpret their statements as manipulations rather than observations. You reverse-engineer their likely goals and update accordingly. Even a Nihilist gives you Bayesian evidence about what someone with no skin in the game would say off the cuff.

The person who is genuinely untrustable is the one whose type you can't identify. And in a society where level 3 and 4 behavior dominates, type-identification becomes extremely hard, because everyone is wearing the same costume of whatever level of discourse happens to be fashionable.

Why It's Hard to Reverse

The degradation from level 1 to level 4 is not easily reversible. Once "there's a lion across the river" means "I'm with the cool kids," any new phrasing you invent to communicate the literal presence of a lion will eventually get captured by level 3. The cool kids will adopt your new phrasing. Soon it'll mean the same thing the old phrasing meant. There's no combination of sounds that reliably communicates "no, really, an actual lion."

This is the same dynamic that makes Common Knowledge so hard to establish. Everyone can individually know there's a virus. But for the knowledge to be actionable, it has to be common knowledge — everyone knows that everyone knows — and common knowledge requires a communication channel that people actually trust at level 1. If the channel is degraded to level 3, common knowledge can't form, because every broadcast is interpreted as a social move rather than a factual claim.

There's a deep connection here to the problems explored in Legibility And State Power. James Scott's argument about why states simplify is partly a story about level 1 collapsing into level 2: the state needs standardized surnames not because it cares about truth but because it needs to collect taxes. And the simulacra framework helps explain why those simplifications are so sticky — once the state's map becomes a tool of control, correcting the map becomes an act of resistance rather than a correction of fact.

The Saving Grace

If there's hope, it lies in the observation that level 1 accuracy still has survival value — more survival value than social positioning, when the stakes are high enough. If the lion actually eats you, being seen as cool for denying the lion doesn't help much. High-stakes environments select for level 1 behavior. Low-stakes environments breed level 3.

The pandemic was instructive because it started as a high-stakes environment where the lion was visible, briefly pulled discourse toward level 1, and then — as the crisis became politicized and the direct threat became abstract — drifted back toward levels 3 and 4 with breathtaking speed. The lesson isn't that we can't communicate about lions. It's that the window in which we can communicate about lions is narrow, and it closes faster than you'd think.