Beyond MBTI · Combinations

No combination is impossible

Somewhere on a forum, someone has already written that an INTJ cannot be an enneagram 2. Someone has already replied that INTJ 9 “makes no sense”, that INTJ 7 is a contradiction in terms, that a sanguine INTJ does not exist. The sentences carry the air of technical knowledge and they are not: none of them describes a mechanism, and none of them has anything behind it beyond who has seen what. This page dismantles the entire species of claim — and, along the way, shows the one place where incompatibility nearly happens, for a completely different reason.

The two systems being crossed here are not asking the same question. MBTI, in the dynamic reading this site uses, describes how you process: the order in which four mental processes come on stage — Ni, Te, Fi, Se. The enneagram describes why you act: a basic desire and a basic fear, in Riso and Hudson’s formulation. A processing order does not contain a motivation, and a motivation does not contain a processing order. That is why this page’s answer is always the same one, and always a structural one.

This page argues two things: that there is no impossible combination, and that no evidence base exists that could support an exclusion even in principle. It does not argue that every combination is equally frequent — some are visibly rare — nor that the crossings described here have been measured. They have not. The section on real tension is declared interpretation: one unvalidated map laid over another unvalidated map, written by someone who knows it. If you find a sentence here that reads like a diagnosis, that sentence is badly written.

The claim

What is actually being said when someone says “impossible”

The sentence always has the same shape: X cannot be Y because of Z, and the Z is almost always a function. “An INTJ cannot be a 2 because there is no Fe.” It sounds like a deduction from a mechanism, and it is not — the premise is a claim about processing and the conclusion is a claim about motivation. A step is missing in the middle, and that step is never supplied by anybody, anywhere.

This matters because the same sentence changes weight depending on who reads it. For someone settled in their own self-description it is forum noise. For someone who spent months assembling two languages that finally explained something, it is an eviction notice. And it is always the second person who takes the sentence seriously — the first has already scrolled past.

Worth noticing, too, is the asymmetry: a claim of impossibility never arrives with an alternative that costs the claimant anything. Nobody writes “impossible, therefore my model is wrong”. The one who gives ground is always the person who described herself.

The structural argument

One map says how. The other says why.

The cognitive stack describes the shape of your thinking, not its subject matter. Ni arrives whole, ahead of the justification; Te puts criteria into the external world; Fi keeps a value that is not up for negotiation; Se registers the physical present and arrives last. None of that says what moves you. Jung built the function-attitudes exactly that way — modes of apprehending and of judging, not the contents to be apprehended and judged.

The enneatype is the content. The 2 wants to be loved and fears being unwanted; the 5 wants to be capable and competent and fears being useless, helpless, incapable; the 8 wants not to be controlled. Riso and Hudson state every point as a desire/fear pair, and not one of the nine pairs mentions intuition, thinking, feeling or sensing. They are vocabularies that do not touch.

For “Ni–Te makes the 2 impossible” to be true, the order of the functions would have to determine the motivation. If it did, the enneagram would be redundant — you would read the four letters and already know what the person fears. That both systems coexist, and that people feel the need to use both, is already the practical admission that they answer different questions. Independent in the question, not in the numbers: what gets measured in real populations is correlated, and two sections from here you will see by how much. Different questions do not prevent correlated answers — they prevent one from dictating the other.

What the stack does do is give motivation its shape. A 2 running on Ni–Te helps differently from a 2 running on Ni–Fe: the same wanting, executed with other tools. That is the distance between “the 2 is impossible here” and “the 2 looks like something else here”. The whole page lives in that difference.

The evidence

There is no study that could forbid anything

Suppose someone wanted to sustain an exclusion empirically. There is nothing to do it with. The crossing of MBTI and the enneagram rests on essentially one clearly peer-reviewed comparison: Wagner and Walker (1983), with 390 people who already knew the enneagram well enough to type themselves. The study compared mean scale profiles across the nine enneatypes; it never built the table of MBTI type by enneatype. And that is what matters here, far more than the strength or weakness of any correlation: a design that compares means cannot, even in principle, deliver the empty cell an exclusion needs.

On the enneagram side, the systematic review by Hook and colleagues (2021) gathered 104 samples and produced the most uncomfortable problem, which is structural: factor analyses of enneagram questionnaires do not cleanly recover nine factors. If the nine points do not emerge sharply from the instrument that measures them, then a claim about the intersection of one of those points with an MBTI type is being made about a quantity whose measured outlines are still loose.

Put the two together and look at what would be needed to say “INTJ 2 does not exist”: a sample of INTJs with enneagram typings, a defensible typing procedure on both sides, and an empty cell that survived the sample size. Not one of the three exists. The exclusion is not false the way a wrong number is false — it is unsupported the way a question nobody asked is unsupported.

The data that does exist

None of the nine reaches zero

The only large public survey crossing the two systems is a self-selected, self-typed online compilation of roughly 120,000 responses (Enneagram Personality, 2026). Within the INTJ slice, the self-declared enneatypes fall out like this:

  • Type 1, the Reformer — 26.97%
  • Type 5, the Investigator — 17.12%
  • Type 6, the Loyalist — 15.14%
  • Type 8, the Challenger — 13.28%
  • Type 4, the Individualist — 8.38%
  • Type 9, the Peacemaker — 8.33%
  • Type 7, the Enthusiast — 5.23%
  • Type 2, the Helper — 2.96%
  • Type 3, the Achiever — 2.60%

How to read those numbers

Rare is a number larger than zero

Two readings come out of it, and the first is the one that matters: nothing is zero. The two smallest are precisely the two the internet likes to declare impossible — the 3, at 2.60%, and the 2 just above it, at 2.96%. Rare, and rare is a number larger than zero. The exclusion argument needs an empty cell to work, and there is not one here.

The second reading is a warning about the data itself. The ordering is strange enough to distrust: the 1 comes first, at almost 27%, above the 5 — the opposite of nearly everything the community literature has been repeating about INTJs for twenty years. That neither corrects the community literature nor is corrected by it. It is simply a good reminder that the table describes people who answered a quiz and chose both of their own labels.

So: a self-selected sample with the MBTI label brought from home, with nothing resembling probability behind it and no verification of either typing. It is not a population base rate and must not be read as one. It is good for exactly one thing — showing that all nine cells have people inside them — and for that one thing it is good enough.

Where the tension is real

Four combinations that take work, and none that is forbidden

Before a single word of this section: nothing here has been measured. There is no sample of INTJ 2s, no study of INTJ 9s, nothing at all. What follows is a reading — an attempt to describe what happens when a motivation and a processing order pull in different directions inside the same person. Tension is not an obstruction; it is where the work sits. And that is exactly why this section must not become an exclusion list by the back door.

INTJ 2. The 2 wants to be needed: the strategy is to become indispensable by reading what the other person needs before they ask. That strategy runs natively on Fe, which the INTJ does not have on the conscious floor. What the INTJ has is Te and Fi — so the care comes out as a solution and as silent loyalty. In practice, this is the INTJ who reorganises someone’s finances without being invited to, builds the spreadsheet, genuinely solves the problem — and is sincerely wounded on discovering that the person wanted a hug. The underlying friction is larger than the linguistic one: the machine was built for autonomy and the 2 fears being unwanted. Holding both is difficult. Difficult is not impossible, and the matured version of it is rare and very good to know: someone whose help actually works.

INTJ 9. The 9 avoids conflict and inner disturbance, and does it by settling into other people’s agendas until they have forgotten their own. Te exists for the precise opposite: to put a criterion on the table, out loud, and hold it. INTJ 9 is an engine that produces sharp conclusions coupled to a motivation that does not want to defend them. In practice: you saw clearly where the plan breaks, said nothing in the meeting, and ended the day with a headache that has no apparent cause. Or: a remarkably agreeable INTJ, whose “fine by me” is not agreement but withdrawal. The work here has a short address — the conclusion already exists; what is missing is saying it.

INTJ 7. The 7 keeps the doors open; the fear is limitation, deprivation, being trapped in pain. Ni does the opposite of keeping doors open: it converges, discards, arrives at a single outcome — and the J wants to close. INTJ 7 is someone who plans a project up to the point where it has already given all the pleasure it could give, and then abandons it: the classic Ni loss of interest once the idea is clear, multiplied by a motivation that treats commitment as loss. It is the variant most often mistaken for ENTP from the outside, and the one that most often hears “you are wasting your talent”, a sentence that helps nobody and describes nothing. The real tension is between a head that wants to finish and a wanting that refuses to choose.

INTJ 3. The 3 is the most externally referenced motivation there is: self-worth depends on what has been achieved and on what has been seen. Fi is the most internally referenced feeling function there is: a value that asks nobody for approval. Putting both in the same person produces a genuinely productive tension — the INTJ 3 builds an image out of real competence and despises the image while building it, and that internal suspicion usually keeps them clear of the hollow version of the 3. The risk has an appointed hour: the day the achieving stops. Leave, illness, redundancy, the project delivered, retirement. If Fi has gone twenty years without being consulted about what it wanted, there is nobody in the room when Te switches off. Worth knowing beforehand rather than afterwards.

Saying it again, because the pull to read this as a ranking is strong: none of these four is worse than the other five, and none of them predicts anything about anyone’s life. They are places where the map shows two arrows pointing in different directions. Everybody has some.

Translation

Three things “impossible” usually means

When someone says a combination does not exist, they are almost always saying one of three things. Only the third is useful — and the third is a completely different sentence from the one that was said.

  • “I have never seen it.” An argument from absence, made inside a tiny and biased sample: the people who write about typology on the internet. If the 2 turns up in around 3% of a self-typing compilation, you would need to know a great many INTJs — and know them well enough to know what moves each one — before “never seen it” meant anything.
  • “The stereotypes clash.” And they do: the caricature INTJ is cold and distant, the caricature 2 is warm and self-effacing. Except that the collision is between two cartoons, not between two people. Neither the order of the functions nor the desire/fear pair says anything about warmth of manner: that is a third variable neither system measures.
  • “One of your two typings is probably wrong.” This is the useful version, and it is almost always what the person should have said. It is not a claim about what can exist, it is a claim about the probability of measurement error — and it deserves to be taken seriously, because typing error is common. Pittenger (2005) collects evidence that MBTI itself is unstable over a few weeks: a large share of people receive a different type when they retake the test. If one instrument wobbles on its own, two self-typings combined wobble more.

The worked example

You arrived at “INTJ 2”. What to check, in order.

Mistyping is far more likely than impossibility — and “more likely” is a tool, not a verdict. From the most probable explanation to the least, the list runs like this:

  • You are an INFJ. Ni–Fe is the architecture on which the 2’s project runs effortlessly: Fe is literally the function that reads what the environment needs and supplies it. And the INTJ/INFJ distinction is the most fragile in the code — both stacks open with Ni and close with Se, and what changes is the middle pair, Te–Fi on one side and Fe–Ti on the other, crossed by the T/F letter, which is the only dichotomy on which men and women answer markedly differently in the technical manual’s samples (Myers et al., 1998). It is the exact point at which a typing slips.
  • You are a 1 with a 2 wing. The 1’s moral demand plus the impulse to correct by helping produces behaviour that, from the outside, is almost indistinguishable from a 2 — and from the inside is something else: what moves you is that things should be right, not that you should be needed.
  • You are a 6. The 6’s anxious loyalty anticipates other people’s needs with a precision that looks like a 2’s care, and comes from somewhere else: security. One technical note settles a lot of confusion here — the 6 has no 2 wing. A 6’s wings are 5 and 7. If what recognised you in the description of the 2 was the caring, it is not coming from the wing.
  • You have a 2 in your tritype and are reading the tritype as the core. In the arrangement proposed by Katherine Fauvre — her own extension, not part of the classical enneagram — each person would have one type in each of the three centres. An <strong>INTJ</strong> 5-2-9 has the 2 in the heart centre and is still a 5: the caring shows up in the form, not in the engine. Recognising yourself in a type’s description is not the same as having that type as your core, and the tritype is the commonest source of that slip.
  • You are, in fact, an INTJ 2. This possibility does not leave the list — it is merely less likely than the four above, and “less likely” is an entirely different sentence from “impossible”.

The other systems

Temperaments, instincts, socionics: the same answer

The four temperaments, in the modern rereading that translates them into reactivity and pace — how much one reacts, how fast, for how long — do not describe the order in which your head processes. That rereading belongs to Eysenck, Kagan and Rothbart, not to Galen, whose doctrine of mixtures was another thing entirely. The twelve dominant-and-secondary combinations that circulate in the popular reading are LaHaye’s (1984), and all of them turn up in INTJs. The rarest is sanguine dominance, and the rarity has a stated reason: in the Eysenckian mapping, sanguine is the extraverted and stable quadrant, defined by seeking social stimulus, and being INTJ means having an introverted function in command. The two pull in opposite directions — but pulling in opposite directions is tension, and tension is not prohibition. The sanguine INTJ exists, is genuinely funny in front of a group, and goes home empty. He is usually the person in the room most confused by his own description.

The instincts — self-preservation, social, sexual — are more clearly orthogonal still, and the misunderstanding has a name: “social” does not mean sociable. It means your survival attention goes to the group, to your position inside it, to who is in and who is out — something an INTJ tends to do from the outside, silently, with a precision that is occasionally embarrassing. A social-dominant INTJ is no contradiction at all: it is someone who reads organisational politics better than average and still does not want to go to the pub afterwards.

And socionics. The correspondence between INTJ and ILI is a conversion convention between two maps that define their elements differently — and there are two rival conventions: converting by the order of the functions gives ILI, converting letter by letter gives LII. Typing, moreover, is done by incomparable procedures: a self-administered questionnaire on one side, interview and observation on the other. Coming out INTJ in one system and something else in the other is not a contradiction to be explained — it is the expected result of two different measurements of similar things.

The one exception

Big Five: and even here, the problem is a different one

There is exactly one case in which the sentence “that combination does not add up” has content, and it is not about the enneagram. It is this one: an INTJ with very high Extraversion on the Big Five.

The reason is not that it is forbidden. It is that McCrae and Costa (1989, N = 468) found a correlation of around −0.7 between the MBTI’s I–E scale and Big Five Extraversion — the two scales are, to a large extent, measuring the same variable with the sign reversed. In the same study, S–N and openness run almost as closely coupled (0.72); only T–F with agreeableness (0.44) and J–P with conscientiousness (−0.49) genuinely loosen. Saying “I am an INTJ and I scored very high on Extraversion” is rather like saying that two thermometers on the same arm disagree by twenty degrees. It is not an impossible person: it is a reading to check.

Keep the distinction, because it is the most useful thing on this page. Incompatibility of type would be: your motivation and your processing cannot coexist. Redundancy of measurement is: your two results are the same datum, and they do not match. The first is a claim about you, and nobody has the evidence to make it. The second is a claim about two instruments, and it is checkable — take them again, see which scale landed near the cut, and remember that the cut is arbitrary: the MBTI dichotomies are lines drawn inside continuous distributions, and whoever falls near a line changes letters on very little (Pittenger, 2005).

Neither case licenses telling anybody that they do not exist. One asks for silence for want of evidence; the other asks for a second measurement. Neither asks for a verdict.

What an exclusion costs

Every claim of impossibility charges somebody their own self-description — and never the person making it. The one who pays is almost always the person who already suspected they did not fit: the one who spent months assembling two languages that finally explained something, and who will now spend a few more months deciding which of the two to hand back. In exchange, the sentence delivers at most a tidier taxonomy. The foundation created to look after the ethical use of the instrument — which is not the body that publishes it, but is the body that writes the rules of use — is explicit about it: all types are equally valuable, results are never to be used to limit anyone, and the instrument was not designed to select people, because it does not measure ability, competence or skill. An exclusion is a selection made with a tool that has no permission to select. If you want to be genuinely useful to someone whose self-description sounds odd to you, there is a better question than any verdict, and it is a short one: tell me how you got there.

Sources

Not every claim about type carries the same weight. The tag beside each work says where it comes from — and what counts as evidence varies a great deal between them.

  1. Academic psychology Eysenck, H. J., & Eysenck, M. W. (1985). Personality and Individual Differences: A Natural Science ApproachPlenum Press.
  2. Academic psychology Kagan, J. (1994). Galen’s Prophecy: Temperament in Human NatureBasic Books.
  3. Academic psychology McCrae, R. R., & Costa, P. T. (1989). Reinterpreting the Myers-Briggs Type Indicator from the perspective of the five-factor model of personalityJournal of Personality, 57(1), 17–40. N = 468.
  4. Academic psychology Rothbart, M. K. (2011). Becoming Who We Are: Temperament and Personality in DevelopmentGuilford Press.
  5. Methodological critique Hook, J. N., Hall, T. W., Davis, D. E., Van Tongeren, D. R., & Conner, M. (2021). The Enneagram: A systematic review of the literature and directions for future researchJournal of Clinical Psychology, 77(4), 865–883.
  6. Methodological critique Pittenger, D. J. (2005). Cautionary comments regarding the Myers-Briggs Type IndicatorConsulting Psychology Journal: Practice and Research, 57(3), 210–221.
  7. Methodological critique Wagner, J. P., & Walker, R. E. (1983). Reliability and validity study of a Sufi personality typology: The EnneagramJournal of Clinical Psychology, 39(5), 712–717. N = 390.
  8. Jung’s own work Jung, C. G. (1921). Psychologische TypenRascher. Ed. inglesa: Psychological Types, Collected Works 6, Princeton UP, 1971.
  9. Official MBTI Myers, I. B., McCaulley, M. H., Quenk, N. L., & Hammer, A. L. (1998). MBTI Manual: A Guide to the Development and Use of the Myers-Briggs Type Indicator3ª ed., Consulting Psychologists Press. Amostra nacional representativa, N = 3.009.
  10. Official MBTI The Myers & Briggs Foundation (s.d.). Ethical Use of the MBTI® Assessment · All Types are Equally Valuablemyersbriggs.org, acesso em 2026. Duas páginas distintas.
  11. Official MBTI The Myers & Briggs Foundation (s.d.). The Processes of Type Dynamicsmyersbriggs.org, acesso em 2026.
  12. Official MBTI The Myers-Briggs Company (s.d.). The MBTI assessment (themyersbriggs.com)Empresa que publica e licencia o instrumento, distinta da Myers & Briggs Foundation, que zela pelo uso ético.
  13. Specific model Augustinavičiūtė, A. (1998). Socionika: VvedenieAST / Terra Fantastica. Obra escrita em Vilnius, anos 1970–80.
  14. Specific model Fauvre, K. C. (1994–). What is Tritype®?katherinefauvre.com/tritype. Marca registrada da autora: USPTO 6.474.587 (2021) e 6.558.145 (2021).
  15. Specific model LaHaye, T. (1984). Why You Act the Way You DoTyndale House. Publicado em 1984 como Your Temperament: Discover Its Potential e retitulado na edição Living Books (1988). As doze combinações dominante/secundário.
  16. Specific model Naranjo, C. (1994). Character and Neurosis: An Integrative ViewGateways/IDHHB.
  17. Specific model Riso, D. R., & Hudson, R. (1999). The Wisdom of the EnneagramBantam.
  18. Specific model Wikipedia (2026). Socionicsen.wikipedia.org, acesso em ago. 2026. O próprio artigo carrega aviso de que depende excessivamente de fontes ligadas ao tema.
  19. Specific model Wikisocion (2021). Intuitive Logical Introtim (ILI) · Logical Intuitive Introtim (LII)wikisocion.github.io. Arquivo estático de wikisocion.net, congelado num backup de março de 2021; o site original saiu do ar em agosto de 2021.
  20. Antiquity Galeno (c. 170 d.C.). Mixtures (De temperamentis)Em Galen: Works on Human Nature, vol. 1, ed. e trad. P. N. Singer & P. J. van der Eijk, com P. Tassinari, Cambridge Galen Translations, CUP, 2018. Nove misturas.
  21. Online survey Enneagram Personality (2026). MBTI & Enneagram Correlation: Data From 120,000 Peopleenneagram-personality.com. Amostra autosselecionada: 60.000 homens e 60.000 mulheres que responderam ao teste de eneagrama do próprio site; o tipo do eneagrama é a saída desse teste e a origem do tipo MBTI não é informada. Respostas ambíguas foram descartadas (tipo principal >6% acima do segundo), o que infla as correlações.