Shadow work is a method for noticing the parts of yourself you disowned, using your disproportionate reactions to other people as the instrument. The practice is straightforward: log your reactions for a fortnight, find the repeated trait, work out what suppressing it bought you, then run one small experiment. Everything else is elaboration, and most of the elaboration is where it goes wrong.
What you are actually doing
Jung’s shadow is everything about yourself that did not fit the person you needed to be. Some of it is unattractive. A great deal of it is not — ambition in someone raised to be modest, anger in someone raised to be pleasant, need in someone who had to be self-sufficient early.
If you are not sure what yours holds, there is a structured way in: seven signs and a self-assessment works from observable evidence rather than from introspection, which is the part that tends to produce polished answers.
The reason the practice works on other people rather than on yourself directly is that you cannot inspect what you have hidden from yourself. What you can observe is your reactions, and disowned traits announce themselves as disproportionate irritation when someone else expresses them freely. That is the whole mechanism, and it does not require Jung’s metaphysics: if you paid a real cost to suppress something, watching someone get away with it is genuinely provoking.
Worth stating plainly before you start. This is a useful framework, not a validated theory. There is no measurement of a shadow and no experiment that would disprove it. It earns its place by pointing at an observable pattern, not by being demonstrated.
How to do shadow work, step by step
- Log reactions for two weeks. Do not interpret. One line per disproportionate response — irritation, envy, contempt, sudden dislike. Facts only: what happened, how strongly you reacted, what it actually warranted. Interpreting while logging turns the exercise into an argument with yourself, and the two-week delay is what makes the next step possible.
- Find the repeated trait. After a fortnight you will have three or four recurring themes rather than twenty incidents. The repetition is the finding. One entry is noise; the same trait appearing in five unrelated people is data.
- Ask what it would have cost. For each theme: where did you learn this was not allowed, who reacted, and what did going without it get you? There is nearly always a real benefit — safety, approval, being the reliable one — which is why it held. A trait with no payoff for hiding it does not stay hidden.
- Find the five per cent. Not the extreme version. If the disowned trait is self-assertion, the reclaimed version is one direct sentence in one conversation, not a confrontation. People reach for the dramatic version, fail at it, and conclude the material was too deep.
- Write the prediction, then run the experiment. Before you do it, write down what you expect to happen. Then do it, within the week, and write what actually happened. The gap between those two is what changes behaviour — not the insight.
- Stop at twenty minutes, every time. Bounded sessions are the single thing that keeps this from becoming brooding. Set a timer and stop mid-sentence if you have to.
The printable versions of the log, the projection sheet and the experiment tracker are in the worksheets, set out in full rather than behind an email form.
Shadow work exercises worth doing
Five that produce material reliably. Use one at a time.
The disproportion audit. The core exercise, described above. Everything else is a supplement to it.
The envy inventory. List everyone whose situation you envy, then name the specific component — the money, the attention, the ease, the fact that they asked. Envy is a more accurate map of what you want than your stated goals are, because it does not go through the filter of what you have decided is acceptable to want.
The eulogy of the disowned trait. Take a trait you dislike in yourself and write two hundred words on what it is good for. Not to celebrate it — to find the function. Most disowned traits are adaptive responses that outlived the situation that required them.
The 5% experiment. Named above and worth repeating because it is the one that changes anything. Insight without a behavioural test has a poor record.
The reread. Four weeks of entries, read in one sitting, looking only for repetition. This is where the actual finding lives, and it is the step people skip because it produces nothing new to write.
For a wider bank of questions organised by what they are for, see the prompts.
A worked example
Concrete beats abstract here, so: mine.
The log entry. Wednesday. A colleague described their project at length in a standup. My reaction: seven. What it warranted: one.
The repeated trait. Across a fortnight, four separate people, all doing versions of the same thing — saying their work was good, out loud, unprompted.
What it cost. I had a well-established principle that good work announces itself and that self-promotion is vulgar. Written down, the convenient part became obvious: the principle meant I never had to claim my work was good and risk someone disagreeing. It was not a value. It was insurance.
The five per cent. One email describing a finished project, without hedging.
The prediction. That it would read as boastful and that someone would think less of me.
What happened. Two people replied saying it was useful. Nobody mentioned the tone. The prediction was not slightly off, it was entirely wrong, and that gap did more than the insight had.
The unexpected part: my dislike of the colleague dropped by about half within a month. I had not been trying to change my feelings about them at all.
How to start without doing damage
Four rules, in order of how much they matter.
Write about events, not about your character. “I was short with my sister on Tuesday when she mentioned the house” leads somewhere. “I am not a patient person” is a verdict, and verdicts have no next step.
Bound it. Twenty minutes, two or three times a week. Daily shadow journalling is where this most often turns into brooding with better stationery.
Watch the tone. The single most reliable test of whether this is working is whether the internal voice is interested or prosecuting. Curious means you are examining something. Prosecuting means you are building a case against yourself, and more of it does not eventually turn into insight.
Do not do it harshly. This is the part most shadow work material omits and it determines whether the process helps at all. Meeting disowned material with contempt does not integrate it — it re-suppresses it with shame on top. Self-compassion is strongly associated with lower psychopathology across a large meta-analysis, while suppression carries a measurable physiological cost. The severity that feels like rigour is the thing stopping it from working.
The failure mode: when it becomes rumination
Shadow work and brooding are indistinguishable from outside. Both involve sustained attention to uncomfortable material about yourself. One helps; the other reliably makes things worse.
Rumination is associated with prolonging depressive episodes rather than resolving them, and the work separating brooding from reflection found the two behave quite differently. Quick tests: shadow work asks what happened, rumination asks why you are like this. Shadow work produces a specific observation, rumination a general verdict. Shadow work ends in a changed action, rumination ends when you are tired.
The full set of tests and a two-week protocol are in how to tell shadow work apart from rumination, which also covers the specific circumstances in which this practice is genuinely risky.
One mechanical fix helps more than any amount of vigilance: write in the past tense about a specific occasion, or in the third person. Self-distanced reflection produces less distress and more insight than immersed reflection.
What shadow work is not
- It is not therapy. Where the material is trauma, this is work to do with a professional, not a notebook — the boundary is set out in shadow work and therapy.
- It is not a licence. Naming an impulse as shadow and then acting on it is the inverse of integration. Integration means knowing the impulse is yours and still choosing what you do.
- It is not universal projection. Some people really are behaving badly. Treating every judgement as your own material removes your ability to assess anything, including who is safe.
- It is not catharsis. Expressing anger to discharge it has a poor record — Bushman found in a 2002 experiment that venting increased subsequent aggression rather than reducing it.
- It does not finish. New material surfaces whenever your circumstances or roles change. Anyone describing completed integration is describing a self-image, which is the thing that generates shadows.
Where to go from here
The natural sequence: start with the worksheets for structure, use the prompts when the page stalls, and read the rumination distinction before you have been at it a month.
If the pattern turns out to be inherited rather than personal — the same trait suppressed in a parent and a grandparent — breaking family patterns covers what actually passes down. If it centres on anger specifically, what actually builds confidence is the more useful neighbour. And the wider psychological frame, including where Jung holds up and where he does not, is in spiritual psychology.
Frequently asked questions
How do I start shadow work as a beginner?
Keep a reaction log for two weeks and do nothing else. One line per disproportionate reaction, facts only, no interpretation. It is the least dramatic entry point and the only one that reliably produces material specific to you.
How long does shadow work take?
A single recurring pattern can shift within weeks once you act on it. The practice as a whole does not end, because circumstances keep producing new material.
Can I do shadow work on my own?
Yes, for ordinary material, with bounded sessions and an eye on the tone. Not for trauma, and not while in a depressive episode, where solo excavation is most likely to become rumination.
Do I need to know anything about Jung?
No. The method rests on an observation — that disproportionate reactions point at suppressed traits — that you can test directly. The theory is optional and, strictly, unfalsifiable.
What if nothing comes up?
Then you are probably starting from introspection rather than from observation. The log works because it uses your reactions to other people, which are available even when self-knowledge is not.
Is it normal to feel worse?
Briefly, yes — it involves looking at things you avoided for reasons. Worse for a few days is expected. Worse for weeks, heavy rather than uncomfortable, means it has become rumination and the answer is to stop rather than dig harder.
General information, not psychological advice, and not a substitute for therapy. If this surfaces trauma or persistent low mood, please speak to a qualified professional. Sourcing standards are in the editorial policy.
