03

Cicero and the Question at Issue

Primary classical sources: [Cicero, *De Inventione*](https://atlas.perseus.tufts.edu/library/urn%3Acts%3AlatinLit%3Aphi0474.phi036/) and [*Partitiones Oratoriae*](https://catalog.perseus.tufts.edu/catalog/urn%3Acts%3AlatinLit%3Aphi0474.phi038.opp-lat1) • Modern selection: [*How to Win an Argument*, Cicero passages selected and translated by James M. May](https://assets.press.princeton.edu/chapters/p10830.pdf)

Begin with invention, not wording

Invention is the rhetorical work of discovering the available arguments before deciding how to phrase them. Cicero treats it as a disciplined search for material, not the invention of facts.

When people draft too soon, the first attractive sentence fixes their attention on one theory of the case. Invention delays polish while you identify the question, possible proofs, audience beliefs, definitions, causes, comparisons, and objections. The result is a wider field from which to choose.

Use a blank page with columns for for, against, uncertain, and needs definition. Generate before judging. Then choose only material that bears on the decisive question.

Five-minute inventory: What happened? What is it called? How should it be judged? What action follows? Who has authority? What would the other side say? This modern checklist echoes the classical habit of searching a case systematically.

Understand stasis as a stopping point

Stasis is the point where opposing accounts first come to rest on a definite contested question. Finding it matters because proof aimed before or after that point will not resolve the actual disagreement.

Imagine a customer refund dispute. Both parties agree a payment occurred. They disagree whether the delivered service counts as completion under the contract. The issue is not payment as fact, nor yet compensation as remedy; it is definition: what qualifies as completed service?

Cicero’s works use classifications shaped by forensic and civic cases, and later rhetorical teaching names related status questions in different ways. Use the framework as a diagnostic tool, not as a claim that every modern conflict fits one rigid list. The practical sequence is: conjecture about fact, definition of the act, quality or justification, and procedure or authority.

This map complements the modern fact, definition, value, and policy map used for general disagreements. Conjecture roughly corresponds to fact; definition remains definition; and the qualitative question roughly overlaps value by asking how an admitted act should be judged. Policy asks what should be done and may contain several stases at once, while procedure asks who may decide and how. The modern map sorts the kinds of issue present; stasis analysis locates the decisive stopping point in a contested case. They serve different purposes rather than competing for one universal classification.

Stasis sentence: “The parties agree ___; they disagree whether ___; answering that question determines ___.” If the final clause is weak, the question may not be decisive.

Test the conjectural question

A conjectural question asks whether something happened, exists, caused an outcome, or is likely to occur. “Conjectural” here means inferential: the answer must be reconstructed from signs and testimony rather than merely stipulated.

Ask what observation would distinguish rival accounts. Separate direct evidence from inference and identify missing records. If two explanations predict the same evidence, that evidence cannot decide between them.

For a delayed project, one account blames unclear requirements; another blames insufficient staffing. Ask what each would predict: revision history, idle work, workload, defect type, or decision latency. Gather evidence that could count against the favored account.

Conjecture worksheet: alleged event or cause; agreed observations; competing accounts; evidence each predicts; source reliability; uncertainty remaining. Do not jump to blame while causal structure is still unresolved.

Test the definitional question

A definitional question asks which category properly contains an admitted act or situation. Definitions matter because a label often carries consequences: error, negligence, harassment, investment, or experiment.

Build a definition from criteria rather than emotional force. State necessary features, useful indicators, and boundary cases. Then compare the present facts with the criteria. Avoid defining the category so broadly that every disliked case enters or so narrowly that no real case can qualify.

StepPromptOutput
Name categoryWhat label is disputed?“Material breach”
State criteriaWhat must be true?Obligation, failure, substantial effect
Compare factsWhich criteria are met?Evidence for each
Test boundariesWhich nearby case should not count?Minor, rapidly cured defect
State consequenceWhat follows from the label?Remedy becomes available

Definitions may be stipulated by law or contract, conventional within a profession, or proposed for the conversation. Say which kind you are using. A dictionary can describe usage, but it does not automatically settle a specialized or normative boundary.

Test the qualitative question

A qualitative question asks how an admitted and defined act should be judged: justified, excusable, harmful, beneficial, proportionate, or blameworthy. “Quality” in this rhetorical sense concerns evaluation, not manufacturing quality.

Make the standard explicit. Was the action necessary? Were safer alternatives available? Was the response proportionate to the risk? Which duties conflicted? What could a reasonable person know at the time, rather than with hindsight?

Two people can agree on facts and definitions yet choose different values. One prioritizes uniform treatment; another prioritizes need. The productive move is to compare standards and consequences, not repeat factual proof.

Quality matrix: standard / why legitimate / evidence it applies / competing standard / tradeoff / proportionate response. Include the interests of people absent from the room.

Test procedure and authority

A procedural question asks whether this forum, person, time, or process is appropriate for deciding the case. It prevents substantive debate from hiding a threshold problem such as missing authority or an unsafe process.

Examples include: Does this manager have authority to change compensation? Should a complaint be investigated before a public conclusion? Is the deadline valid? Does confidentiality limit what can be discussed here? Procedure is not automatically evasion; fair process can protect evidence, participation, and review.

But procedure can also be abused to postpone indefinitely. Ask whether the rule serves a legitimate purpose, applies consistently, and offers a real path to decision. “Not here” must be paired with “where, by whom, and when.”

Procedure checklist: decision-maker; jurisdiction or role; notice; access to evidence; opportunity to respond; conflict of interest; timeline; appeal or review; temporary protection while unresolved.

When a process is unsafe or coercive, pause argument and seek appropriate support. Rhetorical skill does not replace legal rights, safeguarding, or professional investigation.

Move to the earliest decisive issue

The earliest decisive issue is the first disputed question whose answer changes what must be considered next. Starting there reduces wasted proof and prevents an argument from smuggling in unsettled premises.

If parties dispute whether an email was sent, arguing whether its tone was proportionate is premature. If sending is admitted but “approval” is disputed, the definition comes first. If both are settled, evaluate justification. If the wrong body is deciding, procedure may precede everything.

Use an issue tree:

  1. Is the relevant event or causal relation disputed?
  2. If not, is its classification disputed?
  3. If not, is its justification or seriousness disputed?
  4. Does authority or procedure block a valid decision?
  5. Which answer controls the remedy?

The sequence is diagnostic rather than mechanical. Procedure can come first, and a policy argument may contain parallel factual and value questions. Choose the question that most efficiently separates the live alternatives.

Meeting line: “Before we compare remedies, can we confirm whether the live issue is what occurred, what it counts as, or how it should be judged?”

Run a stasis conference

A stasis conference is a short structured conversation for agreeing on the question before presenting full cases. It is useful whenever people keep repeating evidence without convergence.

Give each side two minutes to state: agreed facts, disputed proposition, preferred question, and consequence of its answer. A neutral recorder writes the candidate questions. Together, classify each as conjectural, definitional, qualitative, or procedural, then rank by decisiveness.

Use this template:

  • Shared ground:
  • Candidate fact or cause question:
  • Candidate definition question:
  • Candidate quality or justification question:
  • Candidate procedure or authority question:
  • Earliest decisive issue:
  • Evidence or standard needed:
  • What becomes irrelevant after each possible answer:

Finish by drafting both sides of the issue in symmetrical language: “Whether X occurred” rather than “Why X obviously occurred.” Assign evidence and a time to revisit the map.

The exercise succeeds when each participant can explain why this question comes first and what answer would move the case. Cicero’s enduring lesson is practical: eloquence begins with discovering what one is actually speaking about.

After the conference, test the chosen issue against two rival formulations. Ask what evidence each wording would admit, which party it burdens, and whether it quietly assumes the conclusion. For example, “Did the team ignore the warning?” presupposes that the message qualified as a warning and that the team received it; “What notice did the team receive, and what response did it require?” keeps those matters open. Prefer the formulation that exposes rather than settles contested premises.

Keep an issue log when the case develops. Record the wording, classification, agreed facts, evidence needed, and reason for any change. Changing the question is legitimate when new information reveals a prior mistake; changing it silently after losing ground is not. The log makes that distinction visible and gives a newcomer a reliable route into the dispute.

Close by naming what the framework cannot decide. Stasis identifies the question and suitable proof, but it does not guarantee honest evidence, a fair forum, or a morally sound standard. Those remain responsibilities of the people and institutions using it.