Short Answer

Historically, dying in office was extremely common. In the modern era, it is much less so — but still happens.
Roughly 49 justices have died while serving, but the rate has dropped sharply since the 20th century.

1. Historical Pattern: Death in Office Was the Norm

Before 1900, about two‑thirds of all justices died in office.
Reasons included:

shorter life expectancy
lack of retirement benefits
limited medical care
financial need to continue working

This made death in office the default exit path for early justices.

2. Modern Shift: Retirement Became the Norm

The pattern changed in the 20th century when Congress created a pension system for federal judges.
This allowed justices to retire without financial penalty, dramatically reducing deaths in office.

Additional structural factors:

increased life expectancy
availability of “senior status” for federal judges (semi‑retirement)
cultural shift toward planned retirement

3. Modern Era: Death in Office Still Happens, but Rarely

Recent examples include:

William Rehnquist (2005)
Antonin Scalia (2016)
Ruth Bader Ginsburg (2020)

These cases are notable precisely because they are exceptions in the modern pattern.

4. Quantitative Overview

Based on the historical record:

49 justices have died in office since 1789.
Most pre‑1900 justices died in office.
Most post‑1950 justices retire voluntarily.

This marks a clear structural transition from death‑based exit to retirement‑based exit.

5. Structural Interpretation (Audience‑Facing)

From a relational‑anthropology perspective, the shift reflects:

the move from structural necessity (no pensions → must serve until death)
to structural discretion (pensions + senior status → controlled exit)

In the modern era, death in office is no longer a structural inevitability — it is a choice, often tied to:

ideological timing
political alignment with the appointing administration
personal health and capacity

This aligns with research showing justices sometimes time their retirement for political reasons.

6. Bottom Line

Death in office used to be the dominant pattern. Today, it is uncommon but still structurally possible and occasionally chosen.

If you want, I can produce:

a timeline visualization of deaths vs. retirements
a structural analysis of how life tenure interacts with functional prohibition/consent
a projection model for likely future exits based on age + historical patterns