TOK essay guide: how to unpack a prescribed title and build an argument that holds

A good TOK essay answers the prescribed title it was set, and only that title. You find the knowledge question inside the title, take a clear position on it, test that position against real, specific examples from the areas of knowledge the title asks for, give the strongest counterclaim a fair hearing, and end with a judgement about how far your position survives.

What is the TOK essay, and how much does it count?

The essay is the externally assessed half of TOK. Each exam session the IB issues six prescribed titles; you choose one and write an essay of up to 1,600 words. It is marked by IB examiners and counts for two-thirds of your TOK grade. The other third is the TOK exhibition, marked by your teacher and moderated by the IB.

• Word limit: 1,600. Examiners stop reading at the limit, so anything after it earns nothing. Quotations count; references, bibliography and visuals do not.
• The title: use it exactly as worded. You choose how to interpret it, not what it says.
• Planning: your interactions with your teacher are recorded on a TOK planning and progress form (widely reported as three interactions). It isn't marked, but it is required.
• What's at stake: your TOK grade (A–E) and your EE grade combine into up to 3 bonus points. An E in TOK means no Diploma. The full rules are in our guide to how IB points work.

How is the TOK essay marked?

Not with separate criteria. The essay is widely reported to be marked out of 10 by global impression: the examiner reads the whole essay and places it in one of five levels (from rudimentary up to excellent), then picks a mark within that level. Check the current descriptors in the TOK guide with your teacher.

In practice the examiner is asking one overall question: how well does this essay explore this title? Three things move an essay up the levels:

• It stays on the title. Every paragraph does work for the question that was set, not a nearby question the student found easier.
• The argument holds together. Claims build on each other, examples are analysed rather than described, and the conclusion follows from what came before.
• It weighs things. Counterclaims are real, implications are drawn out, and the student says how much their examples actually prove.

A global-impression mark also means there's no box-ticking. An essay that "includes a counterclaim" because a checklist said so, without that counterclaim changing anything, gains little.

How do I unpack a prescribed title?

Spend the first week on unpacking, not writing. Do it for two or three titles before you commit to one. Four steps:

1. Mark the key terms and decide what they mean here

Circle every word that carries weight: the concepts (knowledge, evidence, certainty, expert, bias, value) and the qualifiers (always, only, more, should, to what extent). Then write one or two sentences on what each concept means in the context of this title. Don't paste a dictionary definition; show the examiner which sense you're using and why.

2. Find the knowledge question inside the title

Every prescribed title is a knowledge question, or contains one. Rephrase it in your own words as an open question about knowledge: how it is produced, justified, shared or limited. If your rephrasing could be answered with a fact ("Yes, scientists use peer review"), it isn't a knowledge question yet.

3. Spot the assumptions

Ask what the title takes for granted. A title asking whether X "should" be trusted more than Y assumes that trust can be ranked. A title that opposes two things assumes they are separate. Questioning an assumption is often where the best paragraph of the essay comes from.

4. Choose areas of knowledge that pull in different directions

If the title asks for two areas of knowledge, choose two where the answer is likely to differ. The five areas are history, the human sciences, the natural sciences, the arts and mathematics. An essay in which both areas give the same answer usually has a thin counterclaim.

A worked unpacking of one title

"Is knowledge that anyone could check more trustworthy than knowledge that depends on expert judgement? Discuss with reference to two areas of knowledge."

Key terms

Knowledge question

Does the possibility of independent checking give us better grounds for confidence than the judgement of people with specialist experience?

Assumptions worth questioning

• That the two kinds of knowledge are separate. In practice, "checking" usually means checking by experts, which collapses the contrast.
• That checkable knowledge actually gets checked. Being open to checking and being checked are different things.

Areas of knowledge

The natural sciences (where replication is the ideal) against mathematics (where proof is, in principle, checkable by anyone, but some proofs can only be followed by a handful of people). History or the arts would also work, as the examples below show.

A thesis you could defend

Checkability makes knowledge more trustworthy only when checking actually happens; where checks depend on a tiny group, the line between "checkable" and "expert" knowledge disappears.

How do I build claims and counterclaims with real examples?

Each body section needs a claim, a specific real-world example, analysis of what that example shows (and doesn't), and a counterclaim with its own example. "Real-world" means a named case you could look up: who, what, when. "A scientist does an experiment" is not an example; it's a placeholder.

Natural sciences: a claim

Claim: because scientific results can be repeated by others, errors get caught. Example: in 2011 the OPERA experiment, timing neutrinos sent from CERN to a detector under Gran Sasso in Italy, reported them arriving about 60 nanoseconds faster than light allows. The team published and asked others to check. By 2012 the anomaly was traced to equipment faults, including a loose fibre-optic connection, and a separate experiment at the same laboratory measured speeds consistent with light. Analysis: the result was trustworthy-in-the-end because it was open to checking. But notice who checked: specialists with access to a particle detector. "Anyone" is doing a lot of work in the title.

Human sciences: a counterclaim to the claim above

Counterclaim: checkable is not the same as checked. In 2015 a large collaboration repeated 100 published psychology studies; only about a third of the repeats produced a statistically significant result again, compared with almost all of the originals. These studies had always been open to replication; for years, few people had done it. Analysis: checkability gives potential trustworthiness, and the potential only pays off if a community has reasons and resources to check.

Mathematics: where the contrast breaks down

Andrew Wiles announced a proof of Fermat's Last Theorem in 1993; referees found a gap, which Wiles and Richard Taylor repaired, and the proof was published in 1995. Any trained mathematician could in principle follow it; in practice its correctness was established by a small group of specialists. A sharper case: Shinichi Mochizuki's claimed proof of the abc conjecture (released in 2012) was challenged in 2018 by Peter Scholze and Jakob Stix, who argued that a key step fails. The proof was published in 2021 in a journal Mochizuki edits, and most number theorists remain unconvinced. A proof is the most "checkable" form of knowledge there is, yet here the community is split along expert lines.

Which structure should I use?

Both of these work. Choose the one that fits your argument; don't force your argument into a template.

Whichever you choose, end each body section with a sentence that ties it back to the title. If the examiner can't see what a paragraph proved, it proved nothing.

Where do perspectives and implications come in?

Perspectives are the different positions from which knowers approach the question. In the worked title, a patient deciding whether to trust a surgeon cannot check the surgeon's judgement; a peer reviewer can. A museum visitor reading an attribution label and the specialist who wrote it have very different access to the evidence. Bring in a perspective when it changes the answer, not as a separate box to tick.

Implications answer "so what?": what follows if your claim is right. If checkable knowledge is only trustworthy when it is actually checked, then a discipline's trustworthiness depends partly on whether checking is rewarded (published, funded, valued). That is a sharper sentence than "this shows that knowledge is complex".

How do I write a conclusion that actually evaluates?

The conclusion is where many essays lose their level. A summary ("In this essay I have discussed…") repeats what the examiner has just read. An evaluative conclusion does three things:

1. Gives a judgement on the title: yes, no, or under which conditions.
2. Weighs the sections against each other: which example carried more weight, and why.
3. Admits a limit, then says what it means for the answer.

For more ways essays slip at the last step, see TOK essay mistakes.

What does a sensible TOK essay timeline look like?

For May candidates, titles are reported to reach schools around September of the final year (around March for November candidates). Your school sets its own internal deadline before the IB's submission date. A realistic plan from release to final:

The essay overlaps with the busiest stretch of the Diploma: EE write-up, IAs and mock exams. Block the TOK weeks into the same plan as your revision plan so they don't get squeezed out.