Inductive Symmetry Audit

Choose it when an argument says an observed pattern supports a Christian or theistic conclusion. Common examples include "whatever begins has a cause," "complexity points to design," "religious testimony supports revelation," "miracle reports indicate intervention," or "moral obligation points to God."

The tool is less about whether the conclusion is emotionally plausible and more about whether the inference is being applied evenly. If similar patterns are ignored when they create trouble for the conclusion, the audit brings that into view.

The central question is: "Am I accepting this inductive move because it is genuinely strong, or because it points toward the conclusion I already prefer?" The app compares the accepted anchor rule with parallel inductions that use similar evidential habits but press in a less convenient direction.

If the anchor is granted high force while parallel inductions are weakened or rejected without a good differentiator, the tool treats that as unresolved asymmetry.

The anchor is the induction the apologetic argument wants to use. For example, "whatever begins to exist has a cause" may be the anchor in a cosmological argument. Parallel inductions are nearby observations that should also matter if ordinary empirical patterns are doing the work.

In that same case, parallels might include the dependence of known minds on brains, the embodiment of known creators, or the fact that known causes occur within space and time. The tool asks why one pattern gets promoted while the others get discounted.

A strong differentiator changes a real evidential variable. It might show that the cases differ in scope, evidence quality, reliability, defeaters, rival explanations, or specificity burden. It should be something a fair critic could understand without first accepting the desired conclusion.

A weak differentiator usually says, in effect, "this case is different because God is different." That may express a doctrine, but it does not yet explain why the empirical pattern should be treated differently.

Archetypes are quick stance profiles. They let users see how the same argument behaves under different habits of confidence, openness, and parallel treatment. They are not insults or personality diagnoses.

The comparison is pedagogical: it shows how much the result depends on force, defense mode, and willingness to let uncomfortable parallels count.

A high score means the favored induction is doing more work than similar rejected or weakened inductions are allowed to do. It identifies a problem in the argument's symmetry, not a mathematical proof that the conclusion is false.

A useful response is to repair the argument: lower the force of the anchor, accept more of the parallel pressure, add an independently defensible bridge premise, or state a more modest conclusion.

Treat it as a prompt bank. The list contains other empirical regularities that may be relevant when theistic arguments appeal to causation, mind, agency, communication, morality, miracles, or design. Not every item applies to every argument.

The best use is selective: choose the stances that are genuinely near the current argument, then ask whether they are being ignored for principled reasons or because they are inconvenient.

Use it as a second-stage tutor. The prompt carries the current numbered stance map, score drivers, differentiators, repairs, and follow-up questions into an AI assistant so the user can continue the analysis without losing context.

The most productive workflow is iterative: copy the prompt, ask for the strongest repair, return to the app, revise the inputs, and see whether the tension actually drops.

No. The app audits a narrower issue: whether a particular apologetic argument is applying inductive standards symmetrically. A high score does not prove that Christianity is false, and a low score does not prove that Christianity is true. The score indicates how much unresolved asymmetry remains between the favored induction and nearby parallel inductions.

Think of it as a consistency check. If an argument says one observed pattern strongly supports theism, the app asks whether similar observed patterns are allowed to count when they create pressure against the same conclusion.

Inductive symmetry means that similar patterns of reasoning should receive similar evidential permission unless there is a relevant difference. If an apologist says, "All observed beginnings have causes, so the universe needs a cause," then observations about minds, creators, agency, communication, causation, and embodiment also deserve attention when they bear on the same conclusion.

The point is not that every induction has equal force. The point is that unequal treatment needs to be paid for with a differentiator that actually changes the evidential situation.

A differentiator is the reason one induction should count more, less, or differently than another. Good differentiators identify a real evidential variable: scope, evidence quality, source independence, defeaters, rival hypotheses, projectibility, or specificity burden.

Weak differentiators merely restate the preferred conclusion. For example, "God is different" may be true within a theology, but it is not yet an evidential differentiator unless it explains which observed variable changes and why that change should affect the inference.

The score rises most when a parallel is highly similar to the anchor, receives much weaker treatment, and lacks a strong differentiator. The tool calculates this as permission gap × similarity × undefended remainder. A rejected parallel with high similarity creates more tension than a weakened parallel with low similarity because more of the evidential gap remains unsupported.

This is why the mini-bars and "Why did my score change?" section matter. They show which numbered stances are carrying the most tension, so the user can focus on the few variables that actually move the diagnosis.

Choose Accept when the parallel deserves roughly the same evidential permission as the anchor. Choose Weaken when the parallel matters, but some real difference reduces its force. Choose Reject only when the parallel should not count for a clear evidential reason.

If rejecting a parallel feels necessary mainly because accepting it would threaten the conclusion, that is a warning sign. The app is designed to make that kind of motivated asymmetry visible.

The force slider measures how much work the accepted rule is being asked to do in the argument. It is not a measure of personal confidence in God, atheism, Christianity, or skepticism. It is local to the selected argument.

A high force setting means the anchor rule is being treated as strong enough to support a substantial conclusion. If similar parallels are then rejected, the tension rises. A lower force setting makes the argument more modest and usually easier to defend.

Defense mode marks the type of move being made from evidence to conclusion. Ordinary induction carries one burden. Best-explanation reasoning carries a rival-comparison burden. Modal, metaphysical, analytic, and rhetorical modes add still more burden because they ask the argument to do more than report a pattern in experience.

The safest practice is to choose the weakest honest mode. If the argument is merely saying "this pattern makes theism somewhat plausible," Inductive may be enough. If it says "God must exist," Modal or Metaphysical burdens become unavoidable.

Modal smuggling happens when an argument quietly moves from observed regularity to necessity. For example, "things in our experience have causes" is an empirical pattern. "Whatever begins to exist must have a cause" is a stronger modal claim. That stronger claim needs a bridge premise.

The app flags this because many apologetic arguments start with ordinary induction but then protect the preferred conclusion by treating the rule as necessary while treating contrary inductions as merely empirical.

Scope drift occurs when an argument begins inside one domain and then uses that evidence outside the domain where it was observed. Causes inside space-time, creators inside the universe, and minds linked to brains are familiar observations. Moving from those observations to a spaceless cause, universe-external creator, or disembodied mind changes scope.

Scope drift is not always fatal, but it creates a bridge burden. The argument must explain why a pattern observed in one domain projects into the new domain.

Specificity inflation occurs when modest evidence is used to support a much more specific conclusion than it licenses. A feature of the universe might raise a question about explanation. That is weaker than concluding a personal, disembodied, timeless, omnipotent, morally perfect, communicative Christian deity.

The repair panel often lowers specificity. This can feel less dramatic, but it is usually a sign that the revised conclusion is more proportionate to the evidence.

The additional list is a prompt library, not a claim that every item defeats every argument. It gives users more inductive patterns to consider when an apologetic argument relies heavily on selected empirical regularities.

The list is intentionally broad because cherry-picking often appears only after multiple neglected patterns are placed side by side. Users can decide which additional stances are relevant to the selected argument and ignore the rest.

Use it as a sharpening tool. The app is most useful when the goal is not to win a point but to find out which premise is doing too much work. A strong argument should survive clearer scope, cleaner bridge premises, and fair treatment of parallels.

If the app raises the score, the constructive response is not defensiveness. It is to ask: should the conclusion be more modest, should the anchor force be lowered, or is there a better differentiator that has independent support?

A skeptic should avoid treating the score as a disproof. The audit identifies asymmetry in a route to a conclusion, not the impossibility of the conclusion. A high score says, "This inference needs repair," not "This worldview is impossible."

The strongest skeptical use is diagnostic and comparative: show which standards are being applied unevenly, invite a better differentiator, and separate what the evidence supports from what the argument wants to conclude.

Paste the prompt into an AI assistant and ask it to analyze the current numbered stance map. The prompt includes the claim, evidence, accepted rule, treatments, differentiators, risk score, repairs, and follow-up prompts.

The best next step is to answer the assistant's questions, then return to the app and revise the settings. If a better differentiator emerges, enter it and watch whether the score meaningfully changes.

A successful repair lowers the mismatch between the conclusion and the evidence. It might accept a parallel, reduce the force assigned to the anchor, add a bridge premise, restrict scope, or replace a doctrinally loaded claim with a more modest one.

The repaired argument does not have to be boring. It just needs to avoid making one induction carry more theological weight than comparable inductions are allowed to carry.