Jan 4, 2017

4 Clues that Socionics Theory Is at Least Partly False

Major Theoretical Fallacy #1: According to Model A, information is conveyed from person to person via matching functions. In other words, an ESI conveys information about introverted ethics to an SLI via the 1st function of first and the 6th function of the second. It is this information that determines the nature of the relationship between them.

No, that's just not how things work. If the SLI and the ESI are sitting together, and neither of them talks, one or both might begin to feel tense. How do you explain that socionically? If one starts talking about something of little interest to the other and goes on and on without stopping, the other will become annoyed. Which information element is being affected? If the two types are in a relationship and the ESI has a prestigious and successful career while the SLI is struggling in their much more low-paying job, this will lead to tension and a feeling that the ESI could "do better." Which function is being affected? If the SLI is more physically attractive than the ESI, the ESI will struggle more with feelings of jealousy. Which function is "acting up?"

My point is, the objective and measurable reasons of relationship tension that have been, and are being discovered, have nothing at all to do with so-called information metabolism. 

Major Red Flag #2: Socionics has been around for 40 years, and the issues being debated among professionals in the field have not changed significantly in at least 15-25 years. What's the cause of this? Are socionists not smart enough to make a big breakthrough? Is lack of funding the problem? 

I invite you to try to make a case that the cause of the stagnation of the field is anything other than the fact that the entire field is based upon a set of unprovable axioms. 

My conviction is that real progress is only possible if researchers ignore the unprovable axioms themselves, but then it's simply psychology research, not socionics research. Socionics can only exist if it has its axioms. And having those axioms makes scientific research and thus real breakthroughs impossible. 

Obvious Red Flag #3: It is virtually impossible to get even a small group of socionists to come to a unanimous agreement on someone's type. Existing tests give conflicting results. Each test was created by a socionist who thought long and hard within the context of their own understanding of socionics, which differs from that of other socionists. 

At the very least, this tells us that socionic type is often far from obvious. How obvious does type need for socionics to be useful? What is the threshold of "nonobviousness" beyond which the application of socionics loses all practicality?

Hidden Red Flag #4: Among socionics enthusiasts, a certain percentage — perhaps 30-40% — are more or less widely recognized as belonging to a particular type. These are the people you think of when reading a type description. The rest are in a gray zone where they are either 1) never fully certain of their own type due to a plethora of opinions on the matter, or 2) unable to convince everyone else that their type is indeed X. What does it say about socionics if only 30-40% of people easily identify with a certain type and are recognized as such by their associates in the community? 

I'll answer that rhetorical question for you. It means one of two things. 1) (optimistic scenario) Socionics theory itself is correct or nearly correct, but type descriptions, tests, and socionists themselves are "not good enough." 2) (pessimistic scenario) Socionics theory itself is incorrect in some significant way, and the type identification problem is the natural result of fundamental errors in the theory.

I'm a pessimist. Even if we accept scenario #1 above, what does it say about socionics if the combined efforts of hundreds of high-IQ individuals is not enough to overcome the stated problem — at least locally, within a single socionics school, — assuming the theory itself is correct? 

In other words, any way you look at it, there is something wrong with the theory (see point #1). 

12 comments:

aestrivex said...

You could reject the specificity of the assumptions inherent in point #1, as I do and have done for years, without dismissing some of the other concepts in socionics such as quadra values.

Your view that the assumptions of socionics are inherently untestable, as we have discussed before, seems needlessly limited to me. Very few people have criticized the proposed methodology of my kickstarter proposal other than to argue about the scales used for the testing method (which seems insurmountable, other than to do video interviews with all the participants) or to caution that it may not work.

Rick said...

What do quadra values derive from then, in your opinion?

If I were to concede that it's actually perfectly okay for socionics to be untestable, how would that affect points #2-4? These characteristics don't budge at all in response to debates and changes of opinion regarding the foundations of socionics. It doesn't matter what any of us think about the issue.

aestrivex said...

I don't know what you mean by "what do the quadra values derive from." They are certainly more fundamental than the specific intertype relations which form a superset of the assumptions needed for basic assumptions about compatibility based on shared quadra values.

I don't know if the assumptions about compatibility based on quadra values would truly work when tested rigorously in practice, but the hypothesis of same quadra = higher compatibility and opposite quadra = lower compatibility is straightforward enough to operationalize in principle.

Issues #2-4 which you address have to do with human error, they have nothing to do with science. I agree that the only defensible view of socionics is that the huge majority of socionists are incompetent. Unfortunately I find is considerable evidence that this may indeed be the case.

Rick said...

OK, then I'll rephrase the question: what are quadra values and why do they exist in the first place? Why are there four quadras of four types with a particular uniform pattern of intertype relations and not, say, 4 quadras of different sizes?

>> Issues #2-4 which you address have to do with human error, they have nothing to do with science. I agree that the only defensible view of socionics is that the huge majority of socionists are incompetent.

How do you tell a competent socionist from an incompetent one? Or, if no foolproof way to tell them apart currently exists, can you propose a way that would work given advances in the field or in science in general?

aestrivex said...

OK, then I'll rephrase the question: what are quadra values and why do they exist in the first place? Why are there four quadras of four types with a particular uniform pattern of intertype relations and not, say, 4 quadras of different sizes?

I don't know why there are four quadras; I am content with the explanation that Augusta suggested this model. Why does it matter? My scientific hypotheses are not focused on the individual types within quadras; maybe those individual types are guideposts to help us diagnose the individuals and make the critical predictions about compatibility, but the compatibility tests do not depend on the diagnoses.

If you were to suggest that there are many varieties of beta introvert, more than we currently categorize, but that the delta introverts are really unnecessarily separated and are basically only deserving of one category of "delta introvertness", I would have no problem with that, it does not affect any of the operational assumptions I am making.

As I have said several times and in many other comments to your similar objections on this blog, I totally dismiss, and have dismissed for years, the "particular uniform pattern of intertype relations" you mentioned, as totally inconsistent with my observations. Although of course many other socionists who I respect, and also many who I don't respect, disagree with that.

How do you tell a competent socionist from an incompetent one? Or, if no foolproof way to tell them apart currently exists, can you propose a way that would work given advances in the field or in science in general?

Your question is extremely broad, but to be brief a competent socionist could be described as a socionist that describes something that works. That is, more formally, anyone who puts forward an tractable operationalization that shows some relationship to interpersonal compatibility. By default, then, a competent socionist virtually needs to have an understanding that socionics is not science, which excludes probably 90% of Russian and Ukrainian socionists including people like Ekaterina Filatova and Aleksandr Bukalov, who have unambiguously demonstrated that they do *not* understand what is science.

Other socionists like Olga Tangemann, who do clearly understand what is science and that socionics isn't that, can be seen as incompetent in the sense that she doesn't talk or think about science, she freely acknowledges that her ideas only have to do with her own biased emotions.

But Ashton Boone, who has a splinter socionics model far outside of the mainstream, would be perfectly able to demonstrate his competency by running the same tests.

So I agree with you in a formal way that there is not currently a "foolproof" way without better operationalizations that actually test for interpersonal compatibility.

But in practice that doesn't prevent you from telling which socionists have no grasp of scientific reasoning, make no effort to present their work objectively to an independent observer at all, and are just generally pretty obviously full of shit. And those issues are obviously very common.

Jason Menzies said...

I'll give one big flaw that I found: according to Model A, it is not even possible in principle to find five people who can't get along. Why? Assume there are four quadras, and that everyone from the same quadra gets along. Now find five people who don't get along. The first four must go into a different quadra, right? - because everyone from the same quadra must get along, so they all must end up in different quadras. Where then, does the fifth person go? They can't be with the Alpha, Beta, Delta, or Gamma, because, as we said, they don't get along with any of them and to be in the same quadra, you must get along. Therefore, this is not even possible in theory in socionics! However, how likely is it in real life to find five, six, seven, eight (or even fifty) people who mutually don't get along? Therefore, this whole notion of exactly four quadras of people must be wrong, and so there is something seriously flawed here! Just my two cents...

Joshua Hilderberg said...

Hello,

First, I want to thank you for bringing up arguments against socionics. I think you have raised some good points overall since you left socionics.

But I don't understand why in your #1 point, you assume that informations can only be conveyed through verbal means, and not through attitudes or physical behaviors. The simple fact that someone speaks or doesn't speak, can convey a lot of meanings. Not counting what he is looking at, on what he directs his attention, wether his behavior is calm or troubled, etc. Also, even if you don't know someone, you can still read "his" attitudes, by projecting behaviors from past experiences or fantasized ones.

In all those non verbal situations, you can still process informations from what you perceive. If your husband thinks you are not working enough, he doesn't need to remind you constantly. You know what he thinks, how he thinks and those aspects of his personality can still affect you, even though he says nothing. By the way, I think that the jealousy is often more evident through behaviors (wether it is expressed through Fi behavior, Se behavior, etc. or any combination of information elements) than through directly telling you he is jealous. So jealousy would rather be a counter-argument to your point.

It is true that most relationship tensions are not only caused by information metabolism, but that fact cannot discredit information metabolism altogether. I am glad that relationships are not only determined by socionics, or there would be a number of people I could never be friends with :-)

Joshua Hilderberg

Jonathan Rabson said...

Hi Rick, ...Was just curious what you've been up to and thought I'd look at your site again. You've become a good spokesperson for the shortcomings of Socionics, which I've always felt has been full of shortcomings, ambiguities, and flaws, though I continue to find the basic structure more interesting and relevant than perhaps you do. But I think some of your examples in this post are possibly unfair. As to point #1, it's common knowledge among any kind of typologists that there could be problems in an interaction not specifically related to type; but of course a Socionist would see some of those things you mention in Socionic terms. For example, you say "one starts talking about something of little interest to the other." Of course the Socionist would look for what information aspect that interest had to do with, and might conclude that the disconnect was because this interest wasn't in the other person's quadra. But you don't provide information on what the topic in the conversation was, so of course you're able to conclude that obviously the lack of connection was due to a non-Socionics reason. In that case, it seems like a foregone conclusion; the Socionics side has no chance, because you're in control of the example, so as evidence it's inconclusive.

The key thing, though, is what you say in point #2, "unprovable." That's really the crux. It seems to me that Socionics is exceedingly testable, because it purports to make predictions about intertype relations. Why then is there perhaps a lack of good statistical psychological research on it? Well, it wouldn't be particularly surprising. Great psychology research is kind of hard to come by...good research design is difficult, and there are a lot of studies on various things that are not really well designed. Why would all these smart Socionics people not realize that what they need to do is come up with very large random samples of people ignorant of Socionics that they can type in various competing ways to see if any of those accurately predict interactions with others? Perhaps because of their background, they're disinclined to do statistical research? I have little doubt though that if someone were inclined to do the right kind of study, that person could either debunk or prove Socionics once and for all. But I also wouldn't be surprised if no one ever attempts such a thing. :)

Rick said...

Joshua Hilderberg, sorry for taking so long to reply.

My point in #1 can be summarized as follows. Intertype Relations describe the interaction pattern that take place between people when all other factors are equal (same level of skill, intellect, rights, power, and even same gender, etc.). The problem is, these other factors are never equal. And even when they seem to be, there are usually "additional factors" that prevent the classical intertype relation from playing out in full.

Jonathan Rabson, likewise sorry for the delayed reply. Jonathan, it's been a long time! :)

>>Of course the Socionist would look for what information aspect that interest had to do with, and might conclude that the disconnect was because this interest wasn't in the other person's quadra.

My point here is that there are almost always far more concrete, proximate causes to look to that explain people's responses to each other than information exchange. 9 times out of 10, socionists' explanations to me seem totally contrived.

On provability, people have been saying what you've said here for years. Where is that conclusive study? Any socionics study must by necessity begin by typing the participants of the study, and that's where everything falls apart. If socionist X types the participants and demonstrates positive correlation Y between two parameters (say, quadra and subjectively reported interaction comfort), that only shows that X is able to predict with reliability Y how people will feel around each other. To be robust X would have to show similar results with a number of different participant sets. There's no direct proof of socionics anywhere in there.

Next, we bring in more socionists to type the participants independently of each other. Each produces a different correlation. We test the socionists with several batches of participants and find out whose correlation is higher and whose is lower. We still haven't proven anything about socionics, just that the level of corroboration is [presumably] quite low between socionists.

Now imagine specialist Z comes into the picture who knows nothing about socionics and uses some other typological system and is able to predict even better than the best socionists who will get along with whom. Again, this says nothing about socionics, because there is no way of demonstrating that any of the socionists types the participants in accordance with socionics theory.

Jonathan Rabson said...

Hi Rick,
Regarding "no way of demonstrating that any of the Socionists types the participants in accordance with socionics theory," I think what you're referring to is the general problem of construct validity. It affects not just Socionics but any attempt at psychological measurement...e.g., if you have a questionnaire or other testing mechanism that has good face validity, that is still not 100% proof that it corresponds with the theoretical concept attached to it. This doesn't mean that the results are useless, just that at some level, what is being measured is the testing instrument itself. The correspondence with theory can only be indirectly corroborated by concrete results that indicate some predictive power by the theory.
Socionists probably make this harder if they insist upon typing by experts instead of using some more mechanistic, less subjective testing process. If what they're doing is having an expert type people and then report correlations regarding how the subjects reported interaction comfort, that raises some research design red flags. Unless the expert viewed or met the subjects individually and did not see any of them interact with each other, the typing could be influenced by the expert noticing who got along with each other. Even if the expert did not witness any interactions between subjects, the typing could still be tainted by noticing a variety of extraneous similarities between the subjects.
To do the research right is very hard. I don't think one needs a high correlation between typing methods or experts though. One only needs to find if there's a single typing method that consistently finds the expected intertype relation correlations (or other correlations predicted by the theory). One could test a variety of methods to see if any meets that standard and discard the ones that don't. I know, easier said than done. :)

Ibrahim Tencer said...

#1 is pretty easy to answer using "classical" socionics theory.

First of all, like Jonathan mentions, socionics does not require any assumption that all interpersonal comfort stems from the intertype relation (in this example the benefit relation). If that's what you're saying, it seems like a strawman-type argument, whether intentional or not. Maybe that's how people see it in the Russian community but I have never assumed anything like this.

Interpersonal comfort in general is a theme that involves Si, Fi, and Fe. Socionics says that people find certain types of IM input to be more psychologically pleasing (the valued functions), and other ones to be displeasing. So we might expect in an ESI-SLI relationship that the SLI will find the ESI's Se ouput to be annoying and his Fi to be valuable on some level. In my experience these kinds of interactions do generally happen as socionics predicts.

"If one starts talking about something of little interest to the other and goes on and on without stopping, the other will become annoyed. Which information element is being affected?"

This would be a combination of Fe and Ne.

"If the two types are in a relationship and the ESI has a prestigious and successful career while the SLI is struggling in their much more low-paying job, this will lead to tension and a feeling that the ESI could "do better." Which function is being affected?"

Se and Fi.

"If the SLI is more physically attractive than the ESI, the ESI will struggle more with feelings of jealousy. Which function is "acting up?"

Again Se and Fi -- jealousy is an SeFi theme. Moreover I would expect Se ego types (or types with "high Se") to be more strongly affected by feelings of jealousy of material wealth. (Personally I could care less if someone has a higher- or lower-paying job than me, the idea seems ridiculous.)

But none of this is necessarily due to the specific intertype relationship (benefit). Obviously, these kinds of generic issues will affect many relationship pairings regardless of the intertype relation involved. Not everything is about intertype relationships, but everything IS about information metabolism. This is a subtle but important point. As I recall you yourself were tending more towards viewing interaction as fundamental rather than intertype relations. My own view of the theory takes this very seriously and I don't even attempt to make a ton of detailed predictions about individual relationships, they're just outlines.

As for the other points, I would agree that most people are not very competent at socionics. Typing people is hard, and no one has figured it out completely.

Rick said...

Jonathan,

I fully agree. Nothing much to add there.

Ibrahim,

I have some disjointed comments:

1. "Socionics says that people find certain types of IM input to be more psychologically pleasing (the valued functions), and other ones to be displeasing. "

Yes, and I no longer buy this. I think this is an overly complicated way of viewing things. What is relevant is "pleasing," and relevance changes on an hour-to-hour basis. What used to be pleasing becomes mundane and tiresome. Sleep 2 hours more or less than usual, and different kinds of information will become more or less relevant and thus pleasing. A neurophysiological approach is a lot simpler IMO.

2. "If one starts talking about something of little interest to the other and goes on and on without stopping, the other will become annoyed. Which information element is being affected?"

This would be a combination of Fe and Ne.


What if the person is going on and on about bricklaying methods? Why would that affect Fe and Ne?

3. (Personally I could care less if someone has a higher- or lower-paying job than me, the idea seems ridiculous.)

Could this be because you haven't experienced it yet and simply haven't "gotten in touch" with your jealous side? Or because at this stage in life your career is not your top priority? I have found over the years that I've experienced a lot of things that I had thought weren't of any importance to me.

Jealousy in particular has been quite well studied and can be expressed in formulaic terms. Stick yourself in the formula and it's easy to model a situation that would arouse jealousy.

About information metabolism. I view our exercises in information elements as one of the most fascinating parts of my socionics exploration. The way I view things now is that there are many different states of mind, most of which are accessible to most people to some degree. Each state of mind comes with a corresponding set of neurophysiological parameters, including things like neurotransmitters, hormones, mental processing speed, and others.

Our exercises demonstrated that each person has a fairly limited set of "favorite" states of mind that they return to habitually. Getting into other states of mind may require unusual circumstances, but it's possible. Over time, changing life circumstances affect the probabilities for falling into different states of mind.

I no longer think states of mind can be equated with socionics functions. The flavors of states of mind are too varied, and I've found the differences of state of mind to be very great between people who are supposedly the same type.

I think it's a worthy pursuit to train yourself to experience a wider away of pleasurable, positive states of mind, as well as the flexibility to be able to respond adequately and confidently to lots of different situations. I just don't see how this can be expressed in socionics terms.