Who lived what they taught?

Why this method fits this school

Qualitative comparative analysis, developed by the sociologist Charles Ragin, treats causes as combinations. Instead of asking how much one variable moves an outcome on average, it asks which configurations of conditions are necessary or sufficient for it, and it allows more than one road to the same place. It works best with a few dozen well-known cases, which is roughly how many Stoics we know well enough to say anything about.

The Stoics already reasoned in this shape. The claim that virtue is sufficient for happiness is a sufficiency claim in the strict sense. The unity of the virtues says the conditions only work together. And the school's famous division of humanity into sages and fools is a crisp set: a man drowning an inch below the surface drowns as surely as one a fathom down. Yet the doctrine of moral progress, prokopē, admits degrees. Fuzzy-set QCA, where membership runs from 0 to 1, is in a sense the progressor's view of the world.

The question

Which combinations of life circumstances go with a Stoic being remembered as living consistently with what he taught? The outcome is judged from the ancient testimony, not from anyone's modern reputation. Seneca and Dionysius the Renegade matter as much as Epictetus here, because a method that only studies successes can't tell you what separates them from failures.

How each set is scored

CONS Lived consistently with the doctrine (outcome)

1: the sources treat his life as matching his teaching under a real test. 0.67: broadly consistent, minor criticism. 0.33: serious contemporary charges of hypocrisy. 0: abandoned or betrayed the doctrine.

POW Political power

1: supreme or consular authority. 0.67: senior office or chief adviser to a ruler. 0.33: civic or ambassadorial role. 0: none.

WLTH Wealth

1: among the richest of his society. 0.67: propertied elite. 0.33: modest means. 0: poverty or slavery.

ADV Severe adversity

1: slavery, exile, or condemnation. 0.67: major loss, illness, or catastrophe. 0.33: episodic hardship. 0: none recorded.

TEACH Stoic lineage

1: trained directly under a recognized Stoic teacher. 0.67: a Stoic mentor in the household or sustained contact. 0.33: indirect or mixed influence. 0: none recorded.

PROF Philosopher by vocation

1: taught philosophy as his career. 0: career in public life.

COURT Courtier

1: served at the pleasure of a monarch as adviser or officer. 0.67: partly so. 0: held office in his own right, opposed the throne, or was the monarch himself.

COURT was not in the first design. It was added after the first truth table put Persaeus and Seneca in the same row as Cato and Rutilius Rufus, with opposite outcomes. A contradictory row means a condition is missing, and what the two failures share is that both served at the pleasure of a king.

What the published run found

These results come from the R QCA package on the published scores. Edits you make above change the tables but not this summary.

Adding COURT resolves the contradiction. Rutilius, Cato, and Marcus land in a row with consistency 1.00. Persaeus and Seneca fall to 0.49.

Adversity outside a court is the strongest single path. ADV × ~COURT reaches 0.93 consistency and covers 0.68 of the outcome. The school's strongest witnesses, from Cleanthes drawing water at night to Epictetus enslaved and lame to Rutilius in exile, faced hardship without a throne to answer to. The narrower statesman's path, POW × ADV × ~COURT × ~PROF, is perfectly consistent but covers only 0.32.

The recipe for failure is not the mirror image of the recipe for success. Analyzed on its own, inconsistency comes through ADV × COURT (0.80), adversity while serving an autocrat, and through a missing Stoic lineage without adversity, which is Egnatius Celer's row alone.

Being a trained professional philosopher is weaker evidence than it first looked. In the crisp version, TEACH × PROF seemed nearly sufficient. In fuzzy sets it reaches only 0.71, because most teachers of the school score 0.67 rather than 1 on the outcome. Their lives were rarely put to a public test, so the sources give us less to judge.

Nothing is necessary on its own. Not being a courtier scores 0.92 as a necessary condition, but its relevance is low (0.42), since almost no one in the sample was a courtier. It is a trivial necessity, not a finding.

What this can't show

Observability. Consistency is hardest to see when nothing tests it, so the outcome partly depends on one of the conditions. A check restricted to cases whose conduct under a specific pressure is recorded gives nearly the same fit for ADV (0.84 against 0.85), which is reassuring but not decisive.

Survivorship. The ancient sources remember remarkable lives. Obscure Stoics who lived well or badly in private are invisible to this design.

Too few cases for the space. Six conditions define 64 possible combinations and the sample fills 10 of them. The parsimonious solution comes out as three equally good models, which is a sign to treat these as hypotheses rather than conclusions.

Boundary cases and judgment calls. Thrasea Paetus and Helvidius Priscus are Stoics partly by later attribution. Dropping them leaves ADV × ~COURT in the solution set. Several scores, including Egnatius Celer's wealth and Athenodorus's court service, are judgment calls, and each case lists its sources so the scoring can be argued with.

Data and code

Everything needed to reproduce or dispute the analysis is here. The R script reads the Scores sheet of the workbook directly.

• Codebook and scores Excel
• Analysis script R
• Full output of the published run Text

Method reference: Charles Ragin, Redesigning Social Inquiry: Fuzzy Sets and Beyond (2008); Adrian Duşa, QCA with R (2019).