Why MBTI doesn't work

[Ezra]

This thread aims to show those of you who don't and wish to know why MBTT doesn't work why it doesn't, and why it can't be correlated with socionics.

Essentially, the reason it doesn't work is because the methods of typing are confused. And when there is more than one way to determine type, that is when the trouble starts.

In MBTT, there are two ways of determining type:

1) Dichotomically (the most common method)
2) Functionally (closer to socionics)

Now, say you take a test that determines that your type is ESFJ. How did the test determine you type? It tested you dichotomically, based on each of the E/I, S/N, T/F, J/P dichotomies. Now, convert that into functions, and your functions should be Fe leading, Si secondary. Does this have to be the case? Well, for the theory to work properly, yes. Simply put though, it doesn't. Why? Because Fe leading, Si secondary necessarily has to be the case for one to be ESFJ, and it isn't necessarily the case. You could, for example, be on the S/N border, and what does that mean? Your Auxiliary function could just as easily be Ni as it could be Si? Furthermore, if you try correlating this with socionics, you're essentially purporting that what the ESFJ's socionical equivalent - the ESE - could easily be an EIE, which is bullshit. How could someone's creative function possibly be their potential PoLR? It couldn't.

[Jarno]

In MBTT, there are two ways of determining type:

1) Dichotomically (the most common method)
2) Functionally (closer to socionics)

I can't remember to have seen a nr 2 Functionally test. Where have you spotted that one?

Dichotomically test are pretty good though, although MBTI uses somewhat less correct definitions if compared to Socionics.
So these kind of tests are near as good as Socionics dichotomies.

Another thing to keep in mind is that altough MBTI may be less accurate because of their lack of information elements, their simplicity of dichotomy's excels in convenience.

[Jarno]

You could, for example, be on the S/N border,

This also applies to socionics I would think?

You could for example be on an Fe / Fi border.

[Expat]

In MBTT, there are two ways of determining type:

1) Dichotomically (the most common method)
2) Functionally (closer to socionics)

I can't remember to have seen a nr 2 Functionally test. Where have you spotted that one?

Ages ago such a test was posted here, I think (which is not very helpful of course).

Another thing to keep in mind is that altough MBTI may be less accurate because of their lack of information elements, their simplicity of dichotomy's excels in convenience.

But what's the point of having a "convenient" method that is less accurate?

You could, for example, be on the S/N border,

This also applies to socionics I would think?

You could for example be on an Fe / Fi border.

Well not really, although for some types (and some individuals of some types) it may be difficult to spot, and of course even more so if you are talking about tests.

[Ezra]

I can't remember to have seen a nr 2 Functionally test. Where have you spotted that one?

You could say this was one. But I was just point out that you can possibly use the functions to determine someone's type. Just not very effectively.

Another thing to keep in mind is that altough MBTI may be less accurate because of their lack of information elements, their simplicity of dichotomy's excels in convenience.

Yes, but then the correlations will never be ABCD = ABCd if this is the case.

This also applies to socionics I would think?

You could for example be on an Fe / Fi border.

But not the Se/Ni or Ne/Si border.

[Jarno]

But what's the point of having a "convenient" method that is less accurate?

Well many people on this forum seem to have great difficulty discovering their type.

I'm not sure, but the cause seems to be the difficulty of Information Elements.

Then one could ask himself isn't it more convenient to trade some accuratesse to get a result.

[Ezra]

But what's the point of having a "convenient" method that is less accurate?

Well many people on this forum seem to have great difficulty discovering their type.

I'm not sure, but the cause seems to be the difficulty of Information Elements.

Then one could ask himself isn't it more convenient to trade some accuratesse to get a result.

That's stupid though. That's like the Jack Daniel's company deciding suddenly to discontinue use of their unique slow mature method for their whisky and convert into a larger factory, just to produce more at a quicker rate, hence deteriorating the quality and complexity of it.

[Mediator Kam]

But what's the point of having a "convenient" method that is less accurate?

Well many people on this forum seem to have great difficulty discovering their type.

I'm not sure, but the cause seems to be the difficulty of Information Elements.

Then one could ask himself isn't it more convenient to trade some accuratesse to get a result.

That's stupid though. That's like the Jack Daniel's company deciding suddenly to discontinue use of their unique slow mature method for their whisky and convert into a larger factory, just to produce more at a quicker rate, hence deteriorating the quality and complexity of it.

True, but I believe we are taking Socionics/MBTI too seriously. For my friends who took the personality test I gave them, they were happy enough to see that the profile fit them well enough to talk about it. They would not care if the test itself was whatever percent inaccurate, as long as it gave an answer that made sense.

Now for us who take Socionics or MBTI seriously, we would like to find the most accurate/sensible dichotomy, but for the majority of people, MBTI is convenient and fast, if not a bit inaccurate.

[Ezra]

True, but I believe we are taking Socionics/MBTI too seriously. For my friends who took the personality test I gave them, they were happy enough to see that the profile fit them well enough to talk about it. They would not care if the test itself was whatever percent inaccurate, as long as it gave an answer that made sense.

Now for us who take Socionics or MBTI seriously, we would like to find the most accurate/sensible dichotomy, but for the majority of people, MBTI is convenient and fast, if not a bit inaccurate.

And that's the difference between you and me. I prefer to take time in order to achieve a more accurate result; you want quick results, whatever the cost.

[Expat]

But what's the point of having a "convenient" method that is less accurate?

Well many people on this forum seem to have great difficulty discovering their type.

I'm not sure, but the cause seems to be the difficulty of Information Elements.

Then one could ask himself isn't it more convenient to trade some accuratesse to get a result.

I agree with Ezra, I can't see the point of this reasoning -- what's the point of getting "a result" if it's not the correct one?

Sure, if someone just wants to be able to call themselves "some" type, then ok -- but in that case, just flip a coin, why bother with a test or whatever?

But if you really want to use socionics - as in understanding your type, other people's types, and why they (and you) interact the way you do -- the only way is the hard way, that is, understand everything, including the functions, quadras, etc. How are you going to use socionics if you don't understand it?

[Jarno]

You both seem to compare using dichotomy's as using a coin.

That's a bit exaggerated. I wouldn't say dichotomy's are as certain as a totally random event. They are just a bit less accurate, and a whole lot faster.

I agree in the end you should compare your type with descriptions, relations, quadra's etc. But to get started, dichotomy's can give you a decent insight.

[Phaedrus]

You both seem to compare using dichotomy's as using a coin.

That's a bit exaggerated. I wouldn't say dichotomy's are as certain as a totally random event. They are just a bit less accurate, and a whole lot faster.

I agree in the end you should compare your type with descriptions, relations, quadra's etc. But to get started, dichotomy's can give you a decent insight.

For most people, using other typing methods than dichotomies is much more like using a coin. The four dichotomies are absolutely central to both Socionics and MBTT, and everyone who disagrees on that is ignorant. The mystery here is how on earth people are able to not get that.

As Ganin has correctly pointed out, too many people on this forum are mistyped. And I am certain that one of the main reasons for that sad phenomenon is that they have disregarded the dichotomies -- on totally wrong grounds.

[Expat]

You both seem to compare using dichotomy's as using a coin.

That's a bit exaggerated. I wouldn't say dichotomy's are as certain as a totally random event. They are just a bit less accurate, and a whole lot faster.

No, you totally misunderstood what I said and meant. I never said nor implied that using dichotomies is like flipping a coin.

My point was that, if what it counts is to get a fast result even if less accurate, then you might as well flip a coin.

I agree in the end you should compare your type with descriptions, relations, quadra's etc. But to get started, dichotomy's can give you a decent insight.

Yes - to get started. Not to get the final understanding of your, or anyone else's, type.

[Jarno]

...you might as well...

That is the part where my comment was referring to, but let's not argue about these details. I guess we agree that dichotomy's are a decent starting point.

But I'm still figuring out if Information Elements are something a newcomer should be worrying about. As I see many cases where it creates more confusion then insights.

Maybe most IM descriptions are too abstract, which make them more difficult to apply then dichotomy's.

[Expat]

But I'm still figuring out if Information Elements are something a newcomer should be worrying about. As I see many cases where it creates more confusion then insights.

Maybe most IM descriptions are too abstract, which make them more difficult to apply then dichotomy's.

But the problem with this approach is the following: it does nothing to explain relationships.

Let us take ENTj as example:

According to dichotomies, ok. E, N, T, j.

The ENTj's dual: ISFj: I, S, F, j.

The ENTj's conflictor: ISFp: I, S, F, p.

A newcomer should then ask, "what? So a change in ONE dichotomy is enough to turn the best relationship into the worst? And why only rationality/irrationality HAS to be the same, while all the others have to be the opposite?!"

Moving then on to two other good relationships, activity and mirror - ESFp and INTp.

The newcomer should then ask, "wtf? So these two are positive relationships, even though they are irrational like the ISFp (which so bad)?"

I hope you see what I'm getting at. If you focus on the 4 dichotomies, the relationships - especially if you start to think of all the others - can't be explained. The newcomer would think - with reason - that the relationships are arbitrary, or that they are based on observations but nobody can explain them.

The only way to properly understand how, and why, say, the ENFp is the ENTj's benefactor and the ESTp is the beneficiary and not vice-versa (or whatever) is by understanding the functions.

Unless the purpose is not to understand relationships - but in this case there is no point to going for socionics.

[Phaedrus]

Maybe most IM descriptions are too abstract, which make them more difficult to apply then dichotomy's.

Yes, certainly so for a newcomer.

But the problem with this approach is the following: it does nothing to explain relationships.

That's no problem at all, because that comes later. A newcomer doesn't need to understand relationships from the start. First he or she must determine his or her type by other means. It's a very big mistake to start with relationships theory before you have figured out your correct type.

A newcomer should then ask, "what? So a change in ONE dichotomy is enough to turn the best relationship into the worst?

Yes. Why not? Every one of the 16 types is a unique entity. That truth must be grasped first. And if that is understood, the question will never be asked.

And why only rationality/irrationality HAS to be the same, while all the others have to be the opposite?!"

Why not? Why should a newcomer ask those questions about something he or she has not understood yet?

Moving then on to two other good relationships, activity and mirror - ESFp and INTp.

The newcomer should then ask, "wtf? So these two are positive relationships, even though they are irrational like the ISFp (which so bad)?"

Why not? Again such a question is totally misguided, and there is no real problem.

I hope you see what I'm getting at. If you focus on the 4 dichotomies, the relationships - especially if you start to think of all the others - can't be explained.

They should not be explained immediately. First the newcomer needs to understand the types.

The newcomer would think - with reason - that the relationships are arbitrary, or that they are based on observations but nobody can explain them.

No. The newcomer would have no legitimate reason to think that. The types must be understood first, later comes the relationships theory. Every reasonable newcomer will understand that.

The only way to properly understand how, and why, say, the ENFp is the ENTj's benefactor and the ESTp is the beneficiary and not vice-versa (or whatever) is by understanding the functions.

Which also comes later. First the types, then the functions.

Unless the purpose is not to understand relationships - but in this case there is no point to going for socionics.

The ultimate purpose is of course to understand relationships. But a newcomer must try to understand the basics first. And the basics of Socionics are the types -- and the types are best explained by the four dichotomies.

[Jarno]

Yes, the bottom line is, dichotomy's should only be used to find someone's type.

When people are ready to discover things about relationships, people should find out what the dichotomy's actually consist of. Namely IM elements.

This was the path that I followed, so it shouldn't be a problem for anyone to approach socionics from this direction.

[Expat]

Yes, the bottom line is, dichotomy's should only be used to find someone's type.

When people are ready to discover things about relationships, people should find out what the dichotomy's actually consist of. Namely IM elements.

That is fine, but the greater point I was making is that if a mistake is made when typing by dichotomies - and mistakes will always happen - people should be ready to revise that typing if the understanding gotten from IM elements points in another direction.

This was the path that I followed, so it shouldn't be a problem for anyone to approach socionics from this direction.

Pardon my bluntness, but that's a rather rosy-eyed, not to say self-centred, conclusion. People are different and relate to definitions and descriptions differently.

[Ezra]

Jarno, you're a retard.

[munenori2]

But I'm still figuring out if Information Elements are something a newcomer should be worrying about. As I see many cases where it creates more confusion then insights.

Maybe most IM descriptions are too abstract, which make them more difficult to apply then dichotomy's.

But the problem with this approach is the following: it does nothing to explain relationships.

Let us take ENTj as example:

According to dichotomies, ok. E, N, T, j.

The ENTj's dual: ISFj: I, S, F, j.

The ENTj's conflictor: ISFp: I, S, F, p.

A newcomer should then ask, "what? So a change in ONE dichotomy is enough to turn the best relationship into the worst? And why only rationality/irrationality HAS to be the same, while all the others have to be the opposite?!"

Moving then on to two other good relationships, activity and mirror - ESFp and INTp.

The newcomer should then ask, "wtf? So these two are positive relationships, even though they are irrational like the ISFp (which so bad)?"

I hope you see what I'm getting at. If you focus on the 4 dichotomies, the relationships - especially if you start to think of all the others - can't be explained. The newcomer would think - with reason - that the relationships are arbitrary, or that they are based on observations but nobody can explain them.

The only way to properly understand how, and why, say, the ENFp is the ENTj's benefactor and the ESTp is the beneficiary and not vice-versa (or whatever) is by understanding the functions.

Unless the purpose is not to understand relationships - but in this case there is no point to going for socionics.

Hammering away at all the pertinent points, I see.

[Expat]

Hammering away at all the pertinent points, I see.

Just making the point that in socionics, the types are inseparable from how they interrelate. But it's a point lost on some people.

[Phaedrus]

Jarno, you're a retard.

That's one of the most stupid things you have said on this forum. I can somewhat understand why people call me a retard or similar things, but if a person like Jarno also gets the same kind of reaction, then one has to start to seriously question the reasons for people's behaviour on this forum. Why? Simply because most of the things Jarno (to my knowledge) has said are things that you simply have to agree with if you know what you are talking about. It is true that he and I say similar things, but I am more blunt than he is, am I not? If you can't express the simplest of truths in the simplest of manners without risk of being called a "retard", then the objective truth is that you are being treated very unfairly.

Jarno is simply right about most of the things he has said in this thread (and many other threads as well). And if you don't understand that, you have things to learn. The most embarrassing thing (for you and others) is that the things you have to learn belong to the basics of Socionics and the 16 types in general.

[Ezra]

Jarno, you're a retard.

That's one of the most stupid things you have said on this forum. I can somewhat understand why people call me a retard or similar things, but if a person like Jarno also gets the same kind of reaction, then one has to start to seriously question the reasons for people's behaviour on this forum. Why? Simply because most of the things Jarno (to my knowledge) has said are things that you simply have to agree with if you know what you are talking about. It is true that he and I say similar things, but I am more blunt than he is, am I not? If you can't express the simplest of truths in the simplest of manners without risk of being called a "retard", then the objective truth is that you are being treated very unfairly.

Jarno is simply right about most of the things he has said in this thread (and many other threads as well). And if you don't understand that, you have things to learn. The most embarrassing thing (for you and others) is that the things you have to learn belong to the basics of Socionics and the 16 types in general.

He's self-aggrandising to the point of tedium (as are you) and is not open-minded or willing to accept the fucking facts in a factual situation. You're so wrapped up in your little systems and explanations that you can't see the fucking simple truth right in front of your eyes.

[Phaedrus]

He's self-aggrandising to the point of tedium (as are you) and is not open-minded or willing to accept the fucking facts in a factual situation. You're so wrapped up in your little systems and explanations that you can't see the fucking simple truth right in front of your eyes.

The fact is that both Jarno and I are saying uncontroversial things. It's actually quite incomprehensible that people are questioning them, because that must mean that they don't read basic articles on Socionics. Nothing in what we say is in clear disagreement with Socionics as that theory is described by leading Socionists.

When you are disagreeing with what we say, you are also disagreeing with for example, Dmitri Lytov and many others of the most well-known socionists. And that's very irritating, because neither you nor anyone else seem to have the courage to criticize him. The only fairly well-known socionist that I have seen beeing openly criticized and implicitly called incompetent is Sergei Ganin. And also in his case the critique seems to be almost totally unfounded.

[Jarno]

He's self-aggrandising to the point of tedium (as are you) and is not open-minded or willing to accept the fucking facts in a factual situation. You're so wrapped up in your little systems and explanations that you can't see the fucking simple truth right in front of your eyes.

To be honest, I do believe in everything you guys do. I like information elements, they are the core of a type. I like relationships and model A compatibility even more.

All I said is that you shouldn't get rid of dichotomy's to quickly.

My posts are sincere and humble. When I say, I can use dichotomys so everyone else can use them, is a way of saying: if even I can do it, everybody can.

I guess you are seriously misreading my posts.

And who's selfaggrandising? is it me because I have an opinion, or is it you because you have an opinion?

You better start thinking before you post. Have I called you names? Never...

Jarno, you're a retard.

What nice of you to say. Do you have many friends?

Calling me retard is such a lousy comment, only to evade the ongoing discussion or try to win it in a wimp sort of way.

[UDP]

Hammering away at all the pertinent points, I see.

Just making the point that in socionics, the types are inseparable from how they interrelate. But it's a point lost on some people.

Isn't that really what socionics was founded on anyways?
Inter-type relations?

That is how I always saw it.


And, though perhaps slower than you, I remember you saying how what ultimately made you realize you were not MBTI INTJ --> Socionics LII was your intertype relations, which has been my gradual course of understanding.

[Expat]

Isn't that really what socionics was founded on anyways?
Inter-type relations?

That's what the very conception of the socionics types is founded on, if you follow how Augusta even came to think of them. She didn't first notice types -- she first noticed the relationships, and from those she deduced the types.

[UDP]

.... which makes a great deal of sense.

[Jarno]

And, though perhaps slower than you, I remember you saying how what ultimately made you realize you were not MBTI INTJ --> Socionics LII was your intertype relations, which has been my gradual course of understanding.

Yes relationships finally confirmed my type too. The descriptions were too confusing. And "intuition in time" was something I also wouldn't have recognized as being me.

[Phaedrus]

She didn't first notice types -- she first noticed the relationships, and from those she deduced the types.

That is not entirely correct. Augusta was interested in human relations, and in order to understand them better she studied Freud, Kretschmer, Sheldon, and others, but she found Jung's theory the most promising. It is true that she deduced the 16 types from their relationships, but before she even became aware of the intertype relations she determined the 8 Jungian types by other means. The original 8 types of Jung was her starting point, they were the primary objects between which she started to observe certain patterns that are now described as intertype relations.

[Ezra]

When you are disagreeing with what we say, you are also disagreeing with for example, Dmitri Lytov and many others of the most well-known socionists. And that's very irritating, because neither you nor anyone else seem to have the courage to criticize him. The only fairly well-known socionist that I have seen beeing openly criticized and implicitly called incompetent is Sergei Ganin. And also in his case the critique seems to be almost totally unfounded.

I'm not interested in Lytov's take on socionics. I'm interested in what Augusta herself said.

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Do you have many friends?

What does that have to do with anything?

Calling me retard is such a lousy comment, only to evade the ongoing discussion or try to win it in a wimp sort of way.

There's nothing wimpy about it, it's just not intellectually simulating.

[Jarno]

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Dichotomies are also well known in the Socionics theory, they aren't limited only as MBTI tools.

IMO it's helpful to type someone using dichotomy's when for example, the used information elements aren't clearly visible.

Choosing an information element demands of you to make a direct choice from 8 options. A dichotomy uses smaller steps, which are easyer to identify.

Ofcourse it's up to you to use whatever you are most familiar with.

[Expat]

IMO it's helpful to type someone using dichotomy's when for example, the used information elements aren't clearly visible.

The problem with this is that it goes both ways.

Let's take an ESTp, especially those that would be described as Se subtype. Rather than often appearing coldly analytical and calculating, they are impulsive, playful, and like all ESTps they welcome Fe - that is, they like positive emotional feedback from an individual and perhaps even more so from a crowd. Especially immature ones - or not very clever ones - may appear obnoxious, especially when they are trying to appear "cool" and be admired - perhaps laughing too loud, telling rude jokes, etc etc.

Now, if you go via dichotomies, especially if you don't know the person well very much (and sometimes even if you do), it would be easy to mistype this ESTp as an "ethical type" since what you see is extroversion, sensing, irrationality and a greater focus on "feelings" (even if not always competently) than "logic". So you might end up typing him as ESFP instead.

Yet, if you go by functions, you'd clearly see that what he's doing is to generate a positive Fe input in others and towards himself, even as he does not realize that it may not be working -- so at the very least you'd see that he values Fe, and so ESFp is less likely.

So, likewise, I can say that it's helpful to to type using functions when the dichotomies aren't clear - or when they are misleading.

[Phaedrus]

I'm not interested in Lytov's take on socionics. I'm interested in what Augusta herself said.

In that case you are a ... you know the word. It starts with an "r".

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Which proves that you have managed to get the basics all wrong, despite my warnings not to let them fool you. When shall you start to think critically? When will you grasp the fundamentals? You are still young, I know, but there are some limits to how stupid one can be. Socionis is based on the four dichotomies. First you type by dichotomies, then you understand by functions. Get it?

[Jarno]

So, likewise, I can say that it's helpful to to type using functions when the dichotomies aren't clear - or when they are misleading.

Yes, I certainly agree with that too.

[Expat]

Yes, I certainly agree with that too.

Ah cool.

Let's now take Kristiina's case. She thought she was INTj, then ENTj for quite some time. In MBTI-like tests, she still gets strong results as "T" - very strong. At first she thought she was clearly a logical type. Because she's clueless? Because she is, indeed, a logical type and she's wrong now about being ENFj? But both Rick and I typed her as ENFj in person - Olga got ISTj in her test, but she uses quadra values besides her own theories.

One of the reasons why Kristiina thought she was "T" is because most of the descriptions of "F" refer either to Alpha Fe or to Fi; even the socionics descriptions of Fe have a heavy bias towards Alpha Fe over Beta Fe. So it's not suprising that an ENFj might see herself/himself as a "T" type. This may be Ashton's case too.

These are the kinds of problems that the "dichotomy" crowd has to accept are real, and not just with easy answers like "she had no clue when typing herself as T then" or "she can't be ENFj then". That's a cop-out.

@Kristiina: I hope you don't mind my using your case as example.

[Slacker]

I'm not interested in Lytov's take on socionics. I'm interested in what Augusta herself said.

In that case you are a ... you know the word. It starts with an "r".

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Which proves that you have managed to get the basics all wrong, despite my warnings not to let them fool you. When shall you start to think critically? When will you grasp the fundamentals? You are still young, I know, but there are some limits to how stupid one can be. Socionis is based on the four dichotomies. First you type by dichotomies, then you understand by functions. Get it?

No. Socionics is based on functions, not dichotomies. You really need to start your own typing system rather than continually re-writing this one.

[Jarno]

I'm not interested in Lytov's take on socionics. I'm interested in what Augusta herself said.

In that case you are a ... you know the word. It starts with an "r".

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Which proves that you have managed to get the basics all wrong, despite my warnings not to let them fool you. When shall you start to think critically? When will you grasp the fundamentals? You are still young, I know, but there are some limits to how stupid one can be. Socionis is based on the four dichotomies. First you type by dichotomies, then you understand by functions. Get it?

No. Socionics is based on functions, not dichotomies. You really need to start your own typing system rather than continually re-writing this one.

Not only functions. Dichotomies are a necessary attribute of any typology

source: http://wikisocion.org/en/index.php?title=Dichotomies

[Slacker]

I'm not interested in Lytov's take on socionics. I'm interested in what Augusta herself said.

In that case you are a ... you know the word. It starts with an "r".

All I said is that you shouldn't get rid of dichotomy's to quickly.

Why not? Otherwise you end up typing by both functions and dichotomies, and then it's just another MBTT.

Which proves that you have managed to get the basics all wrong, despite my warnings not to let them fool you. When shall you start to think critically? When will you grasp the fundamentals? You are still young, I know, but there are some limits to how stupid one can be. Socionis is based on the four dichotomies. First you type by dichotomies, then you understand by functions. Get it?

No. Socionics is based on functions, not dichotomies. You really need to start your own typing system rather than continually re-writing this one.

Not only functions. Dichotomies are a necessary attribute of any typology

source: http://wikisocion.org/en/index.php?title=Dichotomies

Again, Socionics is BASED ON function use. Dichotomies are secondary. As opposed to what Phaedrus said, which is roughly the opposite.

[Jarno]

Dichotomies are secondary.

Yes this sounds fair enough.