You can teach yourself philosophy, and the reading half of the problem is already solved — the texts are free, the reading orders are good, and half a dozen blogs will give you a sequence tonight. Here is the one they broadly agree on:
- Read one survey history, to learn who the players are and what they were arguing about.
- Read one short modern introduction by a living philosopher, to see what doing philosophy actually feels like.
- Pick one primary text and read it slowly, with something knowledgeable beside you.
- Learn to take an argument apart — premises, conclusion, the move between them — as you read, not as a prerequisite.
- Build a review habit from week one, or the first book will be gone by the time you reach the third.
That is the easy half. The hard half is what happens on page 40, when a paragraph stops making sense and there is nobody in the room to ask, and six months later when you cannot honestly say whether you understood Plato or just enjoyed him. Almost nothing written for self-learners addresses that, so most of this article does.
How do you start learning philosophy on your own?
Start with the map, not the territory. Nearly every guide written for self-learners opens the same way, and they are right: read a history of philosophy before you read a philosopher. Knowledge Lust’s seven-step roadmap puts a history book at step one, then intro courses, then learning how to read philosophical texts, then basic logic, and only then the great works. The Autodidact Society guide recommends “refresher books” first and names Will Durant and Bertrand Russell as accessible entry points.
Then choose a route through the primary texts. There are two, and they pull against each other. Chronological order — Greeks forward — gives you the arguments in the sequence they were actually made, which matters more in philosophy than in most fields, since Aristotle is largely a reply to Plato and Kant credited Hume with waking him up. The other route is to start wherever your curiosity already is. One reading guide argues hard for the second, on the grounds that “chronological order is not the same as pedagogical order” and that after a survey you should go to “whichever pulled at you while reading the narrative” — its verdict is that “curiosity is a better guide than chronology.” That is good advice for a plan with no deadline behind it. Pick the route you will still be following in March.
Whichever you pick, the same free scaffolding is available to everyone: the pre-twentieth-century canon is out of copyright and legally free online, the Stanford Encyclopedia of Philosophy is a genuine academic reference that costs nothing, and there are university lecture series on YouTube. Take all of it.
Where every self-study roadmap breaks down
Follow any of those roadmaps to step three and you will notice they stop being able to help you at exactly two points.
The first is discussion. Knowledge Lust names it honestly and then hands it back: “Consider finding a friend or partner and meeting up every few weeks to talk about what you’ve read,” because studying philosophy “is best when you’re doing it with other people, having lively debates.” True — and a dependency on people the solo learner does not, by definition, have. The Autodidact Society’s “complete guide” does not raise discussion partners at all. Neither does Brain Fodder’s six tips.
The second is knowing whether you got it right. Knowledge Lust’s answer is a commonplace book — record “questions, favorite quotes, reflections, and summaries.” That is a fine habit and it is storage, not a check. The Autodidact Society’s advice on verifying comprehension amounts to reviewing your notes the same day. Neither tells you how to catch yourself being wrong.
The most honest page on this subject is the one written by an actual philosopher. On 13 May 2017, answering a reader who wanted to study alone, Gideon Smith-Jones at Ask a Philosopher asked the question the roadmaps avoid: “how can you tell whether you’re making headway, when you only have yourself to judge your progress?” He warned that a solo reader can settle into an idiosyncratic interpretation without ever noticing, and named the real obstacle: “Who is going to read what you write?” — “the biggest stumbling block,” because the central advantage of formal study is “to have the opportunity to have your work assessed by persons qualified to judge.”
That is the shape of the whole problem. A philosophy education is two things: a reading list, and a room where somebody who knows the text checks your reading and reads what you write. The internet has cloned the reading list a hundred times over. It has barely touched the room.
Look at what the guides recommend you equip yourself with — a philosophy dictionary, Coursera courses, the Stanford Encyclopedia, Philosophy Tube and Wireless Philosophy, Philosophy Bites, five-minute audio lessons, audiobooks. Every one of those talks. Not one of them listens. There is nothing on the standard list that responds to your reading of your paragraph.
How to read a primary text when there is nobody to ask
Everyone agrees on the technique and it is worth stating plainly, because it is unusual advice in a culture of speed-reading. Brain Fodder asks for “a slow, close read of the text digesting one paragraph and concept completely.” The Autodidact Society says read “closely and slowly…until you understand it completely,” and warns that reading quickly means you “most likely won’t learn much.” Philosophical prose is compressed argument; a paragraph of Hume can carry more inferential load than a chapter of most books. Two pages in an evening is a normal night’s work.
In practice, close reading is three questions asked of every stretch of text:
- What is the claim? State it in your own words, without the author’s vocabulary.
- What is the argument for it? Which sentences are doing the supporting, and is the move valid?
- What would have to be true for this to be false? If you cannot say, you have not located the claim yet.
Then comes the wall — a passage where the technical vocabulary has quietly shifted, or a step is doing work you cannot see. The standard advice for the wall is re-read it, look the term up in a dictionary, or find the relevant Stanford Encyclopedia entry. All three help, and all three are about the literature in general rather than the sentence in front of you. The encyclopedia entry explains what “transcendental” means in Kant’s system; it does not explain what this paragraph is doing with it.
This is the one place a machine can actually stand in for the room, if it is pointed at the right thing. An AI tutor that answers from general internet knowledge is worse than useless here — it will confidently give you the received summary of a philosopher instead of an account of the text you are holding, which is precisely the idiosyncratic-interpretation risk running in reverse. An AI grounded strictly in the source you uploaded is a different tool: you can ask it what a paragraph is claiming and get an answer traceable to that translation, that edition, that argument. Study Junkie builds the course from your upload and keeps the tutor grounded in that source rather than the internet’s opinion of it — and it never replaces the text with a summary, which matters more in philosophy than anywhere else, because in philosophy the argument is the wording.
How do you know you actually understood it?
By trying to produce the argument, not by recognising it.
Rereading a passage until it feels familiar is the classic trap: fluency feels like understanding and is not. The test is retrieval — close the book and reconstruct the argument, then check. If you can state Descartes’s cogito in your own words, say what it is supposed to establish, and name the objection Hobbes raised against it, you know something. If you can only nod along while reading it, you do not yet.
Two things make this work alone. The first is writing something that gets read back — an argument reconstruction, an objection, an answer to a question you did not set yourself. That is Smith-Jones’s stumbling block, and it is the gap practice assignments with feedback are for: not grading, since you have no exam and no grade, but somebody telling you that your reconstruction dropped a premise. The second is spacing the attempts out, because a check you run immediately after reading is measuring your short-term memory. We wrote about running spaced repetition on a book without building flashcards — the same mechanism applies to a primary text, and philosophy needs it badly, since you meet Plato in month one and need him intact when Aristotle starts arguing with him in month four.
How do you keep going past month three?
Most self-study philosophy plans die not at a hard passage but at an ordinary week — travel, work, a bad cold — after which the thread is gone and picking it back up means rereading 80 pages you no longer remember.
Three things make a plan survivable, and none of them are willpower:
- Small units. One argument per sitting, not one chapter. A finished unit is resumable; an abandoned chapter is not.
- A review schedule you did not have to design. The reason commonplace books fail is not that notes are useless — it is that nothing brings the note back to you on the day it would have helped.
- A second mode. Some evenings you cannot do close reading. Listening to a chapter you have already read, at 1.5×, on a walk, keeps the thread alive on a night that would otherwise be a zero. Study Junkie generates audio narration at 0.75×–2× from the same upload for exactly this.
Where Study Junkie fits — and where it doesn’t
Concretely: you upload the PDF, DOCX or text file of the primary text your reading order names, and it becomes a chaptered course — a few minutes of processing even for a 500-page book. You get the tutor grounded in that text, assignments with feedback on what you write, scheduled recall so month one survives to month six, and narration for the evenings when reading is not going to happen. It works in eight languages, and you can upload in one and study in another, which is useful when the best available edition of a text is not in your first language. Signing up gives you 1,500 free Study Credits with no card; after that it is one-time credit packs from $10 that never expire, not a subscription — a plan measured in years should not carry a monthly bill.
Now the honest part.
An AI grounded in a text is not a philosopher and not a peer. It closes the “nobody to ask about this paragraph” half of Smith-Jones’s problem much better than a dictionary does. It does not fully close the other half: having your thinking assessed by someone qualified to judge, who has read the secondary literature and can tell you that your objection is a well-known one, and that it was answered decades ago. If you can get into a real reading group, a seminar, or a distance programme — Smith-Jones points readers to Pathways to Philosophy — take it. That is better than what any software will give you.
The reading is free and this is not. The texts, the Stanford Encyclopedia, the lecture series and the podcasts cost nothing. If your budget is zero, the roadmaps above are a complete and honest answer, and you should follow them.
Nothing removes the slow reading. No tool makes the Nicomachean Ethics a quick evening. The guides are right about this and we are not going to pretend otherwise. What you can remove is the part where you stall on a paragraph for a week with nobody to ask, and the part where six months of reading quietly evaporates.
If that is the part you keep losing, start a course from the first book on your list and see whether the second half of the education is what was missing.
FAQ
Should I read primary texts or histories and introductions first? Both, in that order — a survey history to learn the landscape, then the primary text itself. Smith-Jones’s own recommendation is to read the primary text first and then check your interpretation against secondary sources, which is a good discipline: it stops you arriving at Descartes already knowing what you are supposed to think about him.
Chronological or by interest? Chronological gives you the arguments in the order they were made, and philosophy is unusually conversational — most major works are replies to earlier ones. Interest-first gives you momentum, and it is what the reading guides tend to recommend for beginners on the grounds that historical order is not teaching order. If you have failed at a reading plan before, take interest-first; finishing beats completeness.
Do I need to learn formal logic first? No. Learn to identify a claim, the premises supporting it, and whether the step between them holds. Formal logic sharpens that and is worth a short book at some point, but treating it as a prerequisite is a common way to spend three months not reading philosophy.
Where should I start if I have never read any philosophy? Not Kant. Start with a survey history and one accessible primary work — Plato’s dialogues are the traditional first stop because they are readable and they are the conversation everything after is having. Save the Critique of Pure Reason for when you have argument-reading under your belt, or it will end the project in its first ten pages.
Can you really teach yourself philosophy without a degree? Yes, for understanding — which is what most people want. The reading is public, the encyclopaedias are free, and nothing about the subject requires supervision to comprehend. What a degree gives you that is genuinely hard to replicate is the assessment: people who will read your writing and tell you where you went wrong. Close that gap however you can — a group, a forum, a programme, or a tutor that at least reads what you produce — and the rest of it is available to anyone willing to read slowly.