Bloom's taxonomy

A committee of educators chaired by Benjamin Bloom published the original taxonomy in 1956, as the first handbook of Taxonomy of Educational Objectives, covering the cognitive domain. The purpose was unglamorous and practical: examiners at different institutions wanted a shared language for what a question was actually asking a student to do. In 2001 a team led by Lorin Anderson, a former student of Bloom's, and David Krathwohl, one of the original co-authors, published a revision. It turned the noun labels into verbs, reordered the top two levels, and added a second dimension for types of knowledge.

Most people who use it in workplace training only ever meet the verb lists. That is a shame, because the verb list is the least interesting part.

The six levels of the revised Bloom's taxonomy

Read down that middle column rather than the verb column. The verbs are a prompt, not a diagnosis. What separates the levels is the amount of the answer the learner has to supply themselves.

Writing a learning objective that survives contact with a room

A usable objective has three parts: a verb from the level you actually want, an object it acts on, and the conditions or standard the performance has to meet. The pattern that works:

By the end of this session, participants will be able to [verb] [what] [under what conditions, to what standard].

Two tests before you keep an objective. Could an observer sitting at the back agree on whether it happened? And could you build the assessment now, before you design the content? If either answer is no, the verb is doing no work.

One objective, one verb. Objectives with 'and' in the middle are usually two objectives wearing a coat, and they hide the fact that the second half never gets taught.

The second dimension most people skip

The 2001 revision paired the six cognitive levels with four kinds of knowledge, producing a grid rather than a ladder. This is the part that actually changes a training design:

• Factual. The terminology and specifics. Policy names, product facts, the five steps of your escalation path.
• Conceptual. Models, categories, principles, and how they relate. Why the escalation path is built that way.
• Procedural. Methods, techniques and the criteria for choosing between them. How to run the conversation, and when to use this method rather than that one.
• Metacognitive. Knowledge about one's own thinking. Knowing that you go quiet under pressure and having a plan for it.

Most corporate training sits in the top-left corner of that grid: remembering factual knowledge. Most of the behaviour people are asked to change lives in the bottom rows, in procedural and metacognitive knowledge, which cannot be delivered by slides. If you plot your programme on the grid and everything clusters in one corner, you have found your problem without needing a survey.

What changed in the revision

1. Nouns became verbs. Knowledge became remember, comprehension became understand. Objectives describe actions, not possessions.
2. The top two swapped. In the original, evaluation sat at the top, above synthesis. In the revision, create sits at the top and evaluate below it.
3. Knowledge became its own dimension. The old 'knowledge' level was carved out into the four types above, so a level and a knowledge type are specified separately.
4. Subcategories were tightened. Each level gained clearer sub-processes, which is what makes assessment design possible rather than decorative.

Both versions are still in circulation and both are defensible. Pick one and use it consistently across a curriculum. Mixing them mid-programme produces objective sets nobody can compare.

How to audit an existing programme with it

This takes about an hour for a one-day workshop and it is the fastest value the model offers.

1. List every stated objective, then underline the verb in each. If there is no verb, the objective is a topic heading.
2. Tag each verb with a level and a knowledge type. Two columns, nothing fancy.
3. Count where the weight sits. Compare the distribution to what the business actually asked for.
4. For each objective, write down the evidence you would accept. Anything with no possible evidence gets cut or rewritten.
5. Check the timetable against the tags. Apply, analyse and create objectives need practice time and feedback time. If ninety per cent of the agenda is presentation, those objectives are decoration.

Where the model breaks down

Bloom's taxonomy is a classification scheme, not a theory of how people learn. Four honest limits:

1. The pyramid is not Bloom's. The familiar triangle graphic came later, from other hands. It smuggles in the idea that higher is better and that each level must be conquered before the next. A team that cannot recall the safety rule under pressure does not need a creating exercise, it needs recall that holds when stressed.
2. The hierarchy is not cleanly cumulative. Real work mixes levels constantly. People routinely create competent things while still shaky on the underlying facts, and learn facts through creating. Treating the order as a strict prerequisite chain produces programmes that never get past the foundations.
3. Verbs are not levels. 'Analyse the customer complaint' is an analyse-level task the first time and a remember-level task the fifth time, once the learner is reciting an analysis someone else built. The level lives in the relationship between the task and the learner's prior knowledge, never in the word alone. Verb lists invite exactly this mistake.
4. It says nothing about motivation, behaviour or transfer. The taxonomy classifies cognitive intent. It does not tell you whether anyone will behave differently on Monday. For soft skills in particular, most failures are not cognitive at all. People know what good feedback looks like and still avoid the conversation.

There are companion taxonomies for the other two domains. The affective domain, covering attitudes and values, was published by Krathwohl, Bloom and Bertram Masia in 1964. A psychomotor domain was named by Bloom's group but never completed by them, and the versions in circulation were written by others later. If your objective is really about willingness rather than capability, the cognitive taxonomy is the wrong tool and no amount of verb-swapping will fix it.

Using it without becoming a bureaucrat

• Write objectives at the level the job needs, not the highest level available. Aiming everything at create is a fashion, not a design.
• Let the level dictate the method. Remember and understand can survive reading and explanation. Apply and above need attempts, consequences and feedback.
• Assess at the level you claimed. A multiple-choice test after an apply-level objective quietly downgrades the whole session to remember.
• Keep the objective list short. Four defensible objectives beat twelve aspirational ones, and the twelve-item list is usually a sign nobody chose.

The gap the taxonomy exposes is easiest to see in a room rather than on a spreadsheet. Put a group through a short, real challenge and you routinely find people who can explain a model fluently and then fail to use it the moment the pressure is on. That is not a knowledge problem, it is the distance between understand and apply, and it only closes with attempts and feedback. This is the whole argument for experiential learning: the higher levels of the taxonomy are not deliverable by explanation, because the thing being learned is a performance. Design the objective honestly, then give people somewhere to practise it badly first.

Tour De Force runs this as live, experiential training for teams worldwide — online and in person. Talk to us, or play Gamified learning appsThe Weekly Challenge to see the method in ten minutes.