Search for your constitution online and you will find questionnaires — some reasonable, some not. They are worth something, and they cannot do the one thing people most want from them. This article is about what an eight constitution self-test can and cannot establish, and why confirmation works differently.
In Summary
- Questionnaires sort by the axes you would expect: build, digestion, sweating and temperature, bowel and bladder, temperament, disease tendency, food preference, recovery speed.
- Those axes are real, and a good questionnaire is genuinely worth consulting.
- But confirmation comes from pulse diagnosis and from how the body responds to constitutional acupuncture — not from a form.
- Generative AI tools currently produce constitution questionnaires that look plausible while getting the constitution names and the underlying concepts wrong. Treat AI-generated versions with caution.
- Body shape misleads more than people expect. Each constitution has a standard build, but posture and environment shape the body heavily.
- Pancreotonia, for instance, comes in lean builds, builds developed only in the upper body, builds developed only at the abdomen — a range.
- Weight change can shift what a questionnaire returns, which is one reason readers report different answers at different clinics.
- A clinic that confirms in one or two visits usually had a clear case. Longer observation is more accurate, and checking elsewhere is reasonable.
What the Questionnaires Actually Ask
Most self-tests, whatever their source, sort along a similar set of axes. Typically:
- Build and impression — large frame and strong bone structure with a loud voice, versus lean or wiry with a softer voice
- Digestion — strong digestion and a good appetite, versus weak digestion and fullness after very little
- Sweating and temperature — feels the heat and sweats readily, versus feels the cold with cold hands and feet
- Bowel and bladder — passing comfortably and completely, versus loose stools, incomplete emptying, frequent or weak urination
- Temperament — direct and quick with strong leadership, versus careful, inward and deliberate
- Disease tendency — vulnerability in blood pressure and the cardiovascular or hepatobiliary systems, versus in the digestive, respiratory or renal systems
- Food preference — meat, rich food and spice, versus vegetables, fruit and plainer cooking
- Recovery — bounces back quickly from illness, versus frequent minor complaints and slow recovery
These axes are not arbitrary. They map onto real differences in organ hierarchy, and a well-constructed questionnaire has genuine value as a starting point.
A Note on AI-Generated Tests
People increasingly ask a chatbot to produce a constitution questionnaire, so this is worth saying plainly.
When I have tested current generative AI tools on this, the questionnaires they produce look plausible on the surface. But continue the conversation and it becomes clear that the model has the constitution names and the underlying concepts substantially wrong. Longer, more elaborate versions were worse rather than better.
One tool declined to give an answer where the responses were ambiguous and suggested visiting a clinic instead, which is a more honest failure mode. Another was most useful when it simply pointed to an existing, well-made questionnaire rather than generating its own.
That last approach is currently the better one: an established questionnaire beats a freshly generated one. I make this observation about the tools as they stand; models change, and this may not hold for long.
Why Body Shape Misleads
This is where self-testing most often goes wrong, and it comes up constantly in the questions readers send me.
Each constitution does have a standard build. But body shape is heavily influenced by posture and environment, so within a single constitution there is real variety.
Take Pancreotonia. It appears as a lean build, as a build developed mainly in the upper body, as a build developed mainly at the abdomen — and more besides. Someone whose temperament reads clearly as Pancreotonia while their frame looks like Vesicotonia is not a contradiction requiring resolution. It is ordinary.
So a questionnaire that leans on build will misroute a substantial number of people.
Why the Answer Sometimes Changes
Readers describe being assessed as one constitution at one clinic and another elsewhere, or getting different results before and after a change in weight. One reader was read as Colonotonia when lighter and as Cholecystonia or even Hepatonia after gaining weight, while a questionnaire returned Renotonia.
That is frustrating, and it is also informative. Body changes shift the surface signs that questionnaires and brief examinations rely on. The organ hierarchy underneath does not change — but what is visible on top of it does.
Which is precisely why confirmation cannot rest on surface signs alone.
How Confirmation Actually Works
Constitution is confirmed through pulse diagnosis and through observing how the body responds to constitutional acupuncture.
That is the substance of it, and the companion article to this one describes the process in detail. The short version: an initial reading is taken from interview and questionnaire, then treatment is applied and the body’s response is watched. Confirmation typically emerges over three to ten sessions.
Two practical points from that process:
A clinic that confirms in one or two visits usually had an unambiguous case. That happens, and it is not a sign of carelessness.
But longer observation is more accurate, and having the reading checked somewhere else is a reasonable thing to do rather than a sign of distrust. I say this to readers regularly.
So Is the Questionnaire Worth Doing?
Yes, with the right expectation attached.
A self-test tells you which constitutions are worth considering and gives you vocabulary for what you have been noticing about yourself. That is genuinely useful, and it makes the first consultation more productive.
What it cannot do is settle the matter. Please confirm through pulse diagnosis and constitutional acupuncture — and hold the questionnaire result loosely until then.
Summary
The axes these tests use are real; the conclusions they draw are provisional. Build misleads, weight shifts the picture, and current AI tools get the concepts wrong often enough to be worth avoiding for this purpose.
Use the questionnaire as a question, not an answer.
Related: How to Follow a Constitutional Diet Without Failing
