Specify the kind of philosophical work
ChatGPT can be used for an explanation, a comparison of arguments or a conversation about a premise you are taking for granted. Those are different requests. “Teach me philosophy” leaves the tool to choose the task, level and tradition for you.
This guide is published by Sage, which offers a dedicated philosophical practice app. We have not established that our model replies outperform ChatGPT. The examples below are original suggested prompts, not measured outputs from ChatGPT.
Start with a question that can be examined
For an explanation, name the concept and the level: “Explain Aristotle's practical wisdom for a beginner. Separate a paraphrase of his account from a modern example.” Then ask what a serious objection or qualification would be.
For dialogue, describe the event and name your inference: “I think staying in my stable job proves I am a coward. Test that inference without deciding for me.” Ask for one question at a time if that helps you think. You can request a Stoic, Socratic or another perspective; a requested voice is still an interpretation, not the historical philosopher.
For a comparison, state a common problem and ask what the positions disagree about: “How might a Stoic and Aristotle reason about anger after a colleague takes credit? Distinguish their premises before suggesting an action.” A useful answer should preserve differences, rather than blending traditions into generic encouragement.
Challenge the answer in return
Look for a fact the answer assumes. Did you give evidence that the colleague intended harm? Did the tool infer your motivation or guarantee an outcome? Correct one assumption and see whether the answer changes.
Then ask for a counterexample: “When might staying be courageous rather than cowardly?” You can also ask which facts would make the advice wrong. These are ways to inspect reasoning; they do not guarantee the next answer will be accurate.
Open the source rather than trusting the citation
Ask for the work, book or section, translation and whether a line is a quotation or paraphrase. Open the primary source. Check the words and the surrounding argument. For Aristotle, the Nicomachean Ethics at MIT identifies W. D. Ross's translation. For Marcus, George Long's Meditations provides a complete public-domain text.
If a source cannot be located, remove the quotation marks and treat the sentence as unverified, or leave it out. A precise-looking reference is not itself proof. Sage's recorded Marcus source check shows why thematic similarity and verbatim attribution must be kept separate.
Carry a practice forward deliberately
Check OpenAI's current personalization documentation for available memory and instruction controls. It is inaccurate to assume every interaction always begins without continuity. Settings and eligibility matter, and stored context can be incomplete or wrong.
Whether you use a general assistant or Sage, keep one observation in your own words and return with what actually happened. A private note can do this too. The useful test is whether the next conversation becomes better informed, rather than whether the product uses the word “memory.”
When Sage may fit
Sage makes six named starting lenses easy to choose, keeps conversations and saved replies, and offers Book club and a seven-evening decision path. It does not have an exclusive ability to challenge you or discuss philosophy. Its own recorded replies make unsupported assumptions, which we show alongside useful reasoning.
Sage Basic includes 10 messages a day and 100 a month with all six mentors, no card. Sign-in is required for your own conversations. Plus is $9.99/month for unlimited text. Pro is $19.99/month and includes 60 voice minutes a month.
Compare the product workflows, read a real synthetic Socratic exchange, or choose a Sage mentor. For academic interpretation, a qualified teacher and the assigned scholarship remain important.