The Dependence Contest Instrument (DCI) — Version 0.1

 

The Dependence Contest Instrument (DCI) — Version 0.1

Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.

Instrument architecture: Dave Kelly.

---

I. The Beginning

K1. Everyone wants happiness. Incomplete happiness is not worth accepting if complete happiness is available. Complete happiness must therefore lie wholly within our control. Only assent lies wholly within our control. Happiness is therefore correct assent, correct when it matches the truth about value. The truth about value is that only virtue is good and only vice is evil. Everything else is indifferent.

K2. Denying Th 20 or Th 21 would undermine much of the Stoic view of positive happiness, but would not obviously damage the views on virtue or on avoiding unhappiness. (Sterling, Core Stoicism, closing comment, as stated in The Traditionalist Case and the CPA Record v0.2, B2)

K3. “I don’t follow those arguments, because I’ve never seen them. Even using the term ‘arguments’ fairly broadly, I haven’t seen a single one. I have never seen anything at all to show why I need to accept anything the ancient Stoics said about Physics or Logic (again, beyond broad theories like ‘good and bad exist’) in order to consistently hold to Stoic Ethics.” (Sterling, “Physics, Logic, and Ethics,” International Stoic Forum, 2 March 2022)

K4. “Now if all you meant was that some Physics or other was necessary, and some Logic or other, then that’s obviously true. And, I agree, not just any Physics or Logic will do. … The problem is that I see no reason why the Physics and the Logic have to look anything even remotely like ancient Stoic Physics and Logic.” (Same post)

---

II. Purpose

The DCI takes one non-corpus argument that Stoic ethics depends on Stoic physics or theology. It tests whether the argument succeeds. It does not audit the arguer, survey the scholarship, or settle what the ancient school held.

---

III. Definitions

Df1. Dependence in argument: a physical or theological premise used as a step in the author’s argument for an ethical claim.

Df2. Logical dependence: removing the physical or theological premise removes a premise the ethical claim needs. Here, “needs” means that no other route to the claim exists in the corpus.

Df3. Connection: physics or theology that explains, motivates or grounds an ethical claim without being a step toward it.

Df4. The three domains (K2): virtue, freedom from unhappiness, and positive happiness.

---

IV. Input Requirements

An argument for dependence, quoted or set out from a source text and cited, with its author named or marked anonymous. HALT if no ethical conclusion can be identified, or if the input is a survey of several arguments. A survey is split, with one run per argument.

---

V. Steps

Step 0, Activation. Name the source, quote the argument, and list the corpus documents opened.

Self-audit: is the argument quoted or set out from its source, not reconstructed from memory?

Step 1, Reconstruction. Set the argument out as numbered premises and a conclusion. Apply charity: choose the strongest reading the text allows. Mark each premise as physical, theological or ethical.

Self-audit: would the argument’s author accept the reconstruction?

HALT if no physical or theological premise appears. The argument is then not an argument for dependence. If the only such premise is of the broad kind K4 grants, such as that good and bad exist, the verdict is Broad physics only, granted (K4), and the run proceeds to Step 6.

Step 2, Target. Name the ethical claim the argument concludes to: the link of the chain or the theorem. Name its domain (Df4).

Self-audit: is the target the claim the argument actually needs, not a weaker one?

Step 3, Step test (Df1). Is the physical or theological premise a step toward the target, or does it stand beside it? If it stands beside, the verdict is Connection, and the run proceeds to Step 6.

Step 4, Removal test (Df2). Remove the premise. Is there a corpus route to the target without it? Cite the route from opened corpus documents: Sterling’s text, a theorem, or a ratified reading. If a route exists, the verdict is Dependence in argument only. If none exists, the verdict is Dependence shown, and it is reported to the operator as a finding against K3.

Self-audit: is the route cited from an opened document, not recalled?

Step 5, Domain check. If the target lies in positive happiness, test it against K2. Where the corpus grants dependence there, the verdict is Granted, positive happiness only.

Step 6, Report. Give the verdict, the route or its absence, the opened sources, and what the run does not claim.

---

VI. Verdicts

• Not an argument for dependence
• Broad physics only, granted (K4)
• Connection
• Dependence in argument only
• Granted, positive happiness only
• Dependence shown

---

VII. Failure Modes

• Straw man. Reconstructing the argument weaker than the text allows. Guarded by Step 1 charity.
• Recalled route. Citing a corpus route from memory. Guarded by the Step 4 self-audit.
• Domain slide. Treating a positive-happiness dependence as a dependence of virtue. Guarded by Step 5.
• D2 pattern-match. A route that sounds like the corpus but does not actually reach the target. Undetectable inside the run, so every verdict is held for the operator’s ratification.

---

VIII. Limits

The DCI cannot show that no argument for dependence succeeds. It tests one argument at a time. A Dependence-shown verdict is reported, not suppressed.

---

Version Note

v0.1.

---

Theoretical foundations: Grant C. Sterling (Eastern Illinois University). Analysis and synthesis: Dave Kelly. Prose rendering: Claude (Anthropic). 2026.