Gains:
- Ability to put an argument into numbered premise-conclusion standard form and distinguish implicit premises
- Being able to consciously use the distinction between validity (structure) and soundness (accuracy of premises)
- Ability to make the judgment of validity and soundness without leaving it to artificial intelligence and to build the counter-argument against the strongest version
The heart of philosophy is argument. Understanding a philosophical text means being able to see how it reasons, beyond knowing what it says. The concepts of argument (a reasoning structure consisting of premises put forward to support a conclusion - a claim), premise (an intermediate claim accepted or defended to support the conclusion) and conclusion (the final claim derived from the premises) are the building blocks of this analysis. Argument mapping (making the skeleton of reasoning visible by visually/structurally extracting an argument as premises, intermediate conclusions, and final conclusion) is the most powerful method of revealing the logical backbone of a dense text. The AI can quickly produce a first draft of this skeleton; but it is up to man to decide whether the argument is valid or sound—that is, to make a logical judgment.
In this unit, we'll see how to map an argument step by step, how to use AI at it, and why you should never delegate judgment of validity and soundness.
Anatomy of the argument
Let's first separate two basic concepts, because that's what all argument evaluation depends on.
Validity: An argument is valid if its conclusion is necessarily true if its premises are true. Validity is only about structure, not whether the premises are actually true. The argument "All cats fly; Tabby is a cat; therefore Tabby flies" is valid, but its premise is false.
Soundness: An argument is sound if it is both valid and all its premises are actually true. Soundness requires both the structure and the truth of the premises.
AI is good at extracting the structure of an argument (listing the premises and conclusion). But “is this argument valid?” Even the question "Is it solid?" is sometimes mistaken. question, he cannot evaluate the actual truth of the premises — because this requires knowledge of the world, conceptual sensitivity, and judgment. This is exactly why you will have the AI draw the map and evaluate it yourself.
Hint: Ask AI for the argument in “standard form”: numbered premises (S1, S2, ...), intermediate conclusions if any, and final conclusion (S). This format makes it easier to both verify the argument and find the weak link.
Mapping steps
1. Find the result. What is the author trying to prove most? Expressions such as "in that case", "therefore", "therefore" usually indicate the result.
2. Collect the premises. What are the arguments that support the conclusion? The expressions "because", "because", "because" refer to premises.
3. Add implicit premises. What are the unspoken premises (hidden assumptions) that are necessary for the argument to work? Making these visible often opens up the strongest point of criticism.
4. Separate intermediate results. In long arguments, a premise itself may be deduced from lower premises. Separate these layers.
5. Test validity. If the premises are accepted as true, does the conclusion seem necessary? If not, what is the missing premise or is there a leap in inference?
6. Test durability. Are the premises really true? Which is controversial? This is exactly where your expert judgment comes into play.
Caution: AI may confirm an argument as "valid"; Never consider this approval as the final word. AI can produce convincing but inaccurate explanations of logical validity. Check validity by assuming the premises are true and considering for yourself whether the conclusion is inevitable.
three mini cases
Case 1 — Hidden premise revealed. A PhD student mapped the argument in a moral philosophy paper. AI derived 3 premises and 1 conclusion in standard form. The student noticed a 4th premise (a hidden value assumption) that was needed for the argument to work but was never mentioned in the text. This hidden premise was the most controversial point of the paper and became the center of the student's critique. AI gave the skeleton; The real invention belonged to the student.
Case 2 — AI incorrectly confirmed validity. An instructor had the AI evaluate an argument in a student paper. AI said "the argument is valid". When the lecturer checked, he saw that there was a "middle term" error in the transition to the conclusion (counting the two concepts in the premises as the same even though they are actually different). The argument was not valid. The AI's fluid confirmation concealed a flaw in logic; human control has caught up.
Case 3 — Robustness debate. A seminar group mapped a political philosophy argument. The structure was valid. But the debate boils down to the truth of a central premise: "Every rational individual maximizes his own interest." The group demonstrated in a 20-minute discussion that this premise is empirically and normatively controversial. AI couldn't make this argument; he had only provided the map. The judgment of soundness belonged to the group.
Weak prompt / Strong prompt
Weak prompt:
Is the argument of this text good?
This claim is both vague and dangerous: the AI makes a judgment like "good/bad" for you, confuses validity with soundness, and does not show how it arrived at it.
Powerful prompt:
Your role: a logic-oriented argument analyst. Extract the argument of the following text into standard form: - Numbered premises (S1, S2, ...), intermediate conclusions (if any), final conclusion (S). - Show separately the implicit premises that are not in the text but are necessary, with the label [implicit]. - Mark the sentence on which each premise is based. Do not judge: do not say "valid/sound". Instead, present to me as a QUESTION the step of inference that needs to be tested for validity and the premise that may be controversial for soundness. I will decide.Text: [paste text here]
This prompt omits the structure but leaves the judgment to you; It unpacks implicit premises and makes them verifiable.
Four copyable templates
1) Standard form mapping:
Put the following argument in standard form: numbered premises, intermediate conclusions, final conclusion. Link each element to the sentence in the text. Add comments, just extract the structure. Text: [text]
2) Hunting for implicit premises:
List implicit premises that are necessary for this argument to reach its conclusion but are not stated in the text. For each: briefly state why it is necessary and how questionable its accuracy is. Just rely on the text. Text: [text]
3) Weak link diagnosis (not yours to judge):
In this argument, show (a) the step where the inference is weakest, (b) the most controversial premise in QUESTION form. Do not say "valid/invalid" or "true/false"; Just point out where I should look. I will decide.Argument: [paste standard form]
4) Creating counter-arguments:
Generate the 2 strongest objections to the following argument. Specify which premise or inference each objection targets. Don't make a straw man out of objections; target the strongest version of the argument.Argument: [paste standard form]
Validity/robustness table
Question
What does he ask?
Role of AI
whose decision
What is structure?
Antecedents and conclusion
takes out the skeleton
common
Is there an implicit premise?
hidden assumptions
candidate lists
human beings correct
Is it valid?
Are premises→conclusion necessary?
Shows the test point
human judgments
Is it solid?
Are the premises really true?
Marks controversial
human judgments
Common mistakes
- Confusing validity with robustness. The premises of a valid argument may be false; Just because the AI says "valid" does not mean "correct".
- Accepting the AI's judgment. Leaving the "this argument is good/bad" type evaluation to the AI; The logical decision is yours.
- Omitting the implicit premise. The strongest criticism is often in the unspoken assumption; Don't settle for the skeleton given by the AI, look for the hidden premise.
- Setting up a straw-man. When producing a counter-argument, aim for the strongest version of the argument, not a weak caricature of it (principle of charity).
- Not connecting the map to the text. Accepting the map as true without showing the basis of each premise in the text.
In summary
Argument mapping makes the logical backbone of a philosophical text visible. AI can quickly extract the antecedent-consequence skeleton and implicit antecedent candidates; This is a huge time saver. But the two critical judgments—validity (structure) and soundness (truth of premises)—remain human. Never confuse validity with robustness, do not accept the AI's "good/bad" judgment, be sure to look for implicit premises and build counter-arguments against the strongest. AI is a cartographer; You are the passenger who evaluates the road.
Application task
- Choose a single, clear argument from an article or text.
- Extract the premise-consequence skeleton with the "standard form mapping" template.
- Find hidden assumptions and flag the most controversial ones with the “implicit premise hunt” template.
- Test the validity for yourself: assuming the premises are true, does the conclusion seem necessary?
- Write your own one-sentence objection to the weakest premise.
checklist
- [ ] I put the argument into standard numbered form.
- [ ] I marked the implicit premises separately.
- [ ] I used the distinction between validity and soundness consciously.
- [ ] I left the validity and robustness decision to myself, not to the AI.
- [ ] I linked each premise to the support in the text.
- [ ] I built the counter-argument against the strongest state, not a straw man.