Voice, exit, and loyalty (Hirschman)

Albert Hirschman (1970) described how members respond to declining organisations: voice (attempt reform), exit (leave), or loyalty (stay and endure). FPE reframes this as \(\Gamma\) and \(\mathcal{D}_{KL}\) management under changing \(\Psi\).

\(\Psi\) as organisational shelter

Workplaces, families, states are level-\(L+1\) nodes that shield level-\(L\) agents from raw \(\mathcal{E}_\Sigma\):

predictable rules
resource routing
physical security

When \(\mathcal{R}^{(L+1)}\) falls, \(\Psi \to 0\): the individual’s denominator rises even if they did nothing wrong (Tainter, Diamond).

Three strategies

Strategy	Behaviour	FPE
Voice	Speak up, propose fix, negotiate	Attempt to lower \(\mathcal{D}_{KL}\) and \(\Gamma\) inside coupling
Exit	Quit, divorce, emigrate	Decouple to remove \(\Gamma\) when voice fails
Loyalty	Stay, hope, silence	Defer exit; often hide \(\mathcal{D}_{KL}\) to avoid \(\Gamma\) spike

None is morally supreme. The Persistence Ratio asks: does this coupling still allow \(\mathcal{R} \ge 1\) for you?

When voice is rational

Voice pays when:

\(\mathcal{D}_{KL}\) is shared error (fixable process, not predation)
counter-parties engage in operation 1 + 2 (truth + friction cut)
your \(\Phi\) can sustain the conflict cost

Voice fails when Persecutor dynamics punish honesty → loyalty becomes people-pleasing at scale.

When exit is rational

Exit is the \(\Gamma\)-optimal move when:

unresolved set \(C\) is large and partners refuse repair
\(\mathcal{R}^{(L+1)}\) below your viability threshold
voice raises your \(\mathcal{D}_{KL}\) without changing \(P\)

Attachment shame (“quitters fail”) ignores Lemma 3: no node is self-caused. Leaving a collapsing \(\Psi\) is substrate preservation.

Finnish irtipäästäminen (“letting go”) is exit framed as energy conservation, not moral failure.

When loyalty is rational

Loyalty pays when:

\(\mathcal{R}^{(L+1)}\) is high or recoverable
short-term \(\Gamma\) spike buys long-term \(\Psi\)
voice is already changing the super-node

Pathological loyalty = loyalty when \(\mathcal{D}_{KL}\) about the system is high and rising — staying to avoid exit cost while \(\Phi\) bleeds.

Drama triangle link

Victim often uses loyalty without voice.
Rescuer uses voice for others but not exit from toxic \(\Psi\).
Persecutor punishes voice → forces loyalty or exit drama.

Practical decision tree

Estimate \(\mathcal{D}_{KL}\): is the system’s story accurate?
Estimate \(d\Gamma/dt\): getting better or worse?
Check \(\Phi\): can you afford another voice cycle?
If voice fails twice with no repair → exit dominates expected \(\mathcal{R}\).
If \(\mathcal{R}^{(L+1)}\) recovers → loyalty reinforces \(\Psi\) for all.

Civic and ecological scale

Agents invested in own persistence are structurally invested in \(\Psi\) (biosphere, institutions). Undermining \(\mathcal{R}^{(L+1)}\) for private numerator gain is medium-term self-harm — not external ethics bolted on, but fractal accounting.