CoreTypes Theory: From Jung’s Functions to Speech Analysis

CoreTypes is a proprietary typing system that identifies your personality type not from answers about your behavior, but from the semantics of your natural speech. At its core sits the model of information metabolism: sixteen stable ways a psyche filters and processes information. The roots of the system go back to Jung’s typology and Socionics. The diagnostic method is our own: we analyze not what people report about their habits, but the actual words they use.

This page lays out the full theory. Why questionnaires like the MBTI keep misfiring, what a functional stack is, how function words expose the architecture of thinking, what a free test can and cannot do, and why a recorded interview beats any automated tool. It’s a long read. In exchange, the question “why did three tests give me three different types” will stop being a mystery.

1. Why Behavioral Tests No Longer Work

The Myers-Briggs Type Indicator (MBTI) is the most commercially successful product in the history of typology. Four letters like INTJ or ESFJ show up in resumes and dating profiles alike, and HR managers pronounce them in interviews like a medical diagnosis. Corporations buy assessments in bulk, consultants build team-building programs on top of them; the assessment industry is worth billions a year.

Academic psychometrics, meanwhile, lost its patience with behavioral questionnaires long ago. Cognitive science, statistics, and computational linguistics have each stress-tested these tools, and their conclusions converge: the instruments measure the wrong layer. Surface behavior instead of deep information processing.

Let’s walk through where exactly they break. There are two failure points, and both are fundamental.

1.1 The Behavioral Trap of Classic Questionnaires

The problem isn’t the wording or the length of the form. The problem is the data source. A classic test relies entirely on what a person reports about themselves, and self-report about behavior is dirty data.

Here’s a scene every HR manager knows. A candidate takes a personality test during hiring. The question reads: “Do you stay calm under pressure?” The candidate knows why the test is there, knows which answer the employer wants to see, and has a fair idea of what happens to people who answer it wrong. And answers accordingly, sometimes deliberately, sometimes without noticing. The candidate is rarely lying, mind you. The format does the lying for them.

Psychologist Delroy Paulhus described this mechanism as Social Desirability Bias and broke it into two distinct components:

1. Impression Management. The conscious tailoring of one’s image to the expectations of an audience. In hiring, this isn’t a deviation but the default: candidates “fake good” because a job depends on it.
2. Self-Deception. An unconscious and entirely sincere belief in an idealized version of oneself. A person honestly writes “I always stay calm.” Their colleagues, the ones who watched them an hour before deadline, remember it differently.

The second component is the nastier one. Impression management can at least in theory be caught by lie scales. Self-deception can’t: the person isn’t lying, they’re sincerely wrong about themselves. The questionnaire records that error as fact.

Empirical estimates suggest social desirability can account for 10% to 75% of the variance in personality questionnaire responses. That range is wide, but even the lower bound is a serious problem for a method that claims to measure innate traits. The distortion is systematic: socially approved qualities get inflated, unwanted ones get masked.

The test asks what you do. Your social mask answers.

1.2 The Failure of Test-Retest Reliability

A scientific measuring instrument has one boring, non-negotiable property: reliability. A thermometer that gives two wildly different readings on the same healthy person within a day goes in the trash. Personality testing somehow gets more forgiveness, though the grounds for strictness are the same. If a method claims to identify an innate, lifelong type, its results have to be stable over time.

Organizational psychologist David Pittenger pulled together the longitudinal data on the MBTI. The picture is blunt: when people retake the test after five weeks, up to 50% of them come out as a different type on one or more scales. Five weeks. Personality does not change on that timescale. Workload changes, stress levels change, so does the quality of sleep the night before the test.

The question “do you plan ahead or act on the fly?” reads differently to a person on vacation and to the same person buried in a crunch. A momentary state gets recorded as an inborn trait. Hence the verdict: the test-retest reliability of behavioral tests sits consistently below the threshold psychometrics considers acceptable for diagnostic decisions.

Now picture a company using data like this to decide who gets promoted and who gets cut from a project. A coin toss would be cheaper, and in some cases about as accurate.

To be fair, MBTI advocates object that the test “was never meant for personnel selection.” Perhaps. But personnel selection is exactly where it gets used, and the psychometric critique targets the properties of the instrument, not the intentions of its authors.

1.3 The Evolution of Testing: From Surface Actions to Cognitive Triangulation

None of this means online tests belong in the trash. It means the questions have to change.

The CoreTypes test is built around semantic choice. Instead of “do you enjoy loud parties,” it offers tasks where you pick phrasings, words, and ways of describing situations. Which words you reach for to describe a task, a conflict, a day off: that’s an indirect linguistic probe. It targets the way you process information, not behavioral habits you’d have to judge for yourself anyway.

Verification runs on two tiers:

• The CoreTypes Cognitive Test: a psycholinguistically optimized questionnaire probing how you structure ideas and filter incoming information.
• The Functional Confirmation Test: a separate follow-up block that measures the activation of specific elements of your functional stack, for example the balance between your Leading function and your Vulnerable function.

Hence the word “triangulation” in the title: the type gets confirmed from several independent angles. The first test produces a hypothesis, the confirming test checks it function by function, and the final verification is done by an expert working with live speech. An error at one level gets caught at the next.

Now the honest part, which we repeat across the site. An automated test, ours included, is a screening tool. It builds a hypothesis and narrows the options, and it does that considerably better than behavioral questionnaires. But the final precision comes from a different format: an in-depth interview with a certified analyst, recorded or face to face. Only in free, spontaneous speech do semantic markers accumulate enough statistics for a conclusion with accuracy close to 100%. That’s why the golden standard of diagnosis in CoreTypes is the recorded interview, and the tests are the road to it.

1.4 The Fallacy of Behavioral Determinism

Old typologies, and classical Socionics in particular, carry one more systemic sin: totality. Many schools tried to explain everything through type: what career you’d pick, how you dress, why your marriage failed, in extreme cases even the shape of your face. There were manuals in the spirit of “how to spot a type by its gait.” I wish that were an exaggeration. It isn’t; such texts exist.

Modern cognitive science rejects this without reservations. Human behavior is shaped by genetics, neurobiology, cultural environment, upbringing, trauma, chance, and conscious decisions. Type is one factor on that list, not the list’s owner. The moment a theory promises to predict a life path from a type, it stops being a model and becomes a horoscope.

CoreTypes draws the line hard. Type is a way of processing information. Not a fate, not a ceiling, not a life script.

1.5 The CoreTypes Paradigm: Information Metabolism and Semantics

Now the main part: what we inherited from our predecessors, and what we built ourselves.

From Jung and Socionics, CoreTypes inherits the structural model: 16 types of information metabolism, eight functions in each stack, the regularities of intertype relations. Decades of practice haven’t broken this architecture, so we saw no reason to.

What changed is the diagnostic method. Classical Socionics identified types however it could: questionnaires, observation, dichotomy walkthroughs, freehand interviews, in some schools even physical appearance. Hence the famous school wars, where three socionists give one person three different types. Speech semantics surfaced in Socionics now and then, but never became the field’s standard. In CoreTypes it isn’t “one of the cues.” It’s the foundation of the method: the type is determined by which semantic categories a person uses and how they connect them in context. That, not the renamed types, is the real difference between CoreTypes and Socionics.

The type itself, meanwhile, isn’t a set of behavioral traits. Your habits, career, and hobbies form a plastic socio-behavioral phenotype: statistically correlated with your type, but never 100% determined by it. The type is a cognitive genotype, a stable way of filtering and structuring incoming signals. The functions of the stack steer your focus of attention. And the focus of attention leaves measurable traces in speech, which is what the system runs on.

Instead of caricatures (“all INTJs are cold logicians”), CoreTypes describes the deep cognitive interfaces of the types: the Grandmaster (INTJ), the Beacon (ESFJ), the Catalyst (ENTP), the Harmonist (ISFP), and the other twelve. The chapters below take the mechanism apart, from Pennebaker’s function words to context-aware neural models.

2. The Psyche as a Cognitive Processor: The Core of the System

Here we need the central CoreTypes metaphor, which also happens to be the most practical one. The brain is a computer. Habits, interests, skills, and career are software you install and uninstall throughout your life. Your psychological type is the processor: the thing all that software runs on. Software changes. Hardware doesn’t.

The metaphor is crude. Metaphors are. But it captures the system’s core distinction honestly, so it stays.

2.1 What Is Information Metabolism?

The term was coined by the Polish psychiatrist Antoni Kepinski, and not for typology. He noticed that the psyche “digests” information much the way the body digests food. Signals, images, and words arrive, get sorted, get absorbed or discarded. When the stream is chaotic and exceeds capacity, mental digestion breaks down. Kepinski saw it clinically in his patients. Three hours of doomscrolling will get you a decent home demonstration.

The closest concept in modern psychology is cognitive style: a built-in way of allocating attention, sorting, and remembering information. One important detail: cognitive style says nothing about intelligence. It says which filters the psyche runs by default. Two equally smart people can process the same situation in entirely different ways, and each will be right in their own plane.

2.2 Genotype and Phenotype: Why You Are Not a Stereotype

Take a Grandmaster: a processor tuned for Structural Logic plus Intuition of Possibilities. By the old schools’ logic, such a person belongs in chess, code, or a lab coat. In real life, a Grandmaster can turn out to be an artist, a business consultant, or an attentive family therapist, and regularly does. What stays systematic and analytical is the way they think. What they do with it is decided by biography.

Hence the working rule we repeat across all CoreTypes materials: the type explains part of a person, not the whole person. The genotype defines how you process information. The phenotype shows what you do with it. Confusing the two is a short road back to the horoscopes of chapter 1.4.

The rule has a practical consequence for predictions. A type won’t tell you which profession a person will choose. It will tell you how they’ll master whichever one they choose: from which angle they’ll enter a problem, what will come easily, where they will start to grind.

2.3 The Functional Stack: The Operating System of the Mind

Attention is a hard-limited resource: the psyche can’t process every signal at once. Try listening to a podcast, reading a contract, and watching your kid on a playground at once; the psyche will drop one of the three without asking. It needs a dispatcher to decide what gets priority, what runs in the background, and what gets ignored. That dispatcher is the functional stack. Each of the 16 types has its own, made of eight functions. The four main ones:

• The Leading function: your main window on the world. It runs almost constantly and at maximum energy, setting the very optics of perception. For the Grandmaster it’s Structural Logic: systems, rules, patterns. For the Beacon it’s Ethics of Emotions: emotional climate and atmosphere.
• The Creative function: your main working tool. If the Leading function defines what you want to see, the Creative one defines how you act to get it.
• The Suggestive function: your zone of trust. The psyche happily absorbs information on this channel from other people, with almost no critique.
• The Vulnerable function: the point of least resistance. Signals here overload the system, so the brain pushes them out of focus. Not out of spite, but to save resources.

The full stack includes four more functions, but these are enough to grasp the mechanics. There are strong channels where information flows freely, and weak ones where it passes with effort and cost.

2.4 Intertype Relations: The Physics of Communication Friction

In old Socionics, relations between types were often served up as destiny: happiness with this type, war with that one. CoreTypes strips out the mysticism and keeps the mechanics. Compatibility is the level of cognitive friction in the information exchange between two stacks.

When a Grandmaster talks to a Duke (ESFP), the Leading function of one lands squarely on the Vulnerable function of the other. Both spend their energy on internal translation and defense; fatigue arrives before understanding does. That’s high friction, in our classification Conflict relations.

When the same Grandmaster talks to a Beacon, the strong functions of one feed the receptive channels of the other. Information passes without translation. That’s low friction, Dual relations.

Low friction isn’t a poetic exaggeration; it has a measurable correlate. In comfortable conversation the speech of two people starts to synchronize: function words drift toward each other, rhythm and style follow. The phenomenon is called Language Style Matching (LSM), it replicates in experiments, and it tracks the level of trust between speakers. Even conversational “chemistry,” it turns out, shows up in word frequencies.

Each of the 14 kinds of intertype relations has its own mechanics and its own typical mistakes; we cover them one by one in separate blog articles, from Dual to Conflict. Here the principle is enough: relations between types are the physics of information exchange, not a verdict and not a guarantee.

3. Speech Semantics: How Words Reveal the Mind

You judge a processor by its output. For the psyche, the output is spontaneous speech.

Questionnaire answers are under a person’s control and, as chapter 1 showed, get distorted. But speech style is nearly impossible to control: it’s built from thousands of micro-choices the psyche makes automatically, dozens of times a minute. That makes speech the most honest data source about cognitive style we have access to. Let’s unpack what that claim stands on.

3.1 Unconscious Language: The “Invisible Words”

Social psychologist James Pennebaker split language into two layers. Content words, the nouns, verbs, and adjectives (“table,” “run,” “blue”): we choose them consciously, to fit the topic. And function words, the pronouns, prepositions, conjunctions, and particles (“I,” “under,” “because”). Function words make up less than 0.1% of the vocabulary, yet they account for more than half of everything we say and write.

The brain runs the function-word layer on autopilot. Which is exactly why it’s worth listening to: autopilot reflects where attention actually goes, not where the speaker wants you to think. A classic example from Pennebaker’s work: people under strong stress, and people sliding into depression, start using “I,” “me,” and “my” noticeably more often. Attention turns inward. The pronouns give it away long before the person is ready to admit anything. In other experiments of his, pronoun distribution even exposed the status dynamics of a correspondence: who leads in a pair of writers and who adapts.

The individual speech pattern is also stable. Research on linguistic styles shows that an adult keeps a recognizable manner of distributing function words for decades. Faking it over any real distance is practically impossible. No amount of self-control is enough to police every preposition in live conversation.

3.2 Tested on Big Data: From Blogs to Facebook

The link between words and psyche wasn’t validated on a dozen student volunteers.

Tal Yarkoni analyzed the writing of hundreds of bloggers and matched their natural language against the authors’ personality profiles. The result was a wealth of stable mathematical correlations between vocabulary preferences and cognitive traits. Some styles consistently avoid harsh lexicon and lean toward communal words, others toward analytical and achievement vocabulary.

Andrew Schwartz’s team scaled the approach up: 700 million words from 75,000 volunteers on social media. An open-vocabulary approach, no predefined categories: the data itself showed that different psychological profiles use language in stably different ways.

And it goes further. In 2015, a University of Pennsylvania group showed that personality assessment from social media language is comparable in validity to traditional questionnaires and converges with ratings from people who know the subject. And a study in PNAS recorded a fact uncomfortable for human intuition: an algorithm trained on digital footprints judges personality more accurately than the person’s colleagues and friends do.

For us, these studies are the foundation. CoreTypes didn’t invent the link between speech and psyche. We apply it to one specific problem: determining the type of information metabolism.

3.3 Semantic Fields of the Functional Stack

Time to tie the two threads together. Every function drags its own vocabulary behind it; that cloud of words and constructions is what we call a semantic field. Speech analysis then becomes an inventory job: which fields are packed, which sit half-empty, and what kind of stack would produce exactly that distribution.

A few examples:

• Structural Logic. A speaker with Leading Structural Logic (the Grandmaster) saturates their speech with words of organization: “rule,” “classification,” “consequently,” “structure,” “systematize.” Not a smart-guy pose; the processor genuinely runs reality through systems.
• Ethics of Emotions. The Beacon’s Leading function. Expressive speech with constant monitoring of the atmosphere: “mood,” “excitement,” “drive,” “the vibe in the team,” “everyone’s on edge.”
• Experiential Sensing. Creative for the Beacon, Leading for the Harmonist. Speech full of texture: “cozy,” “soft,” “texture,” “taste,” “aroma.”

Empty fields are as informative as crowded ones. If an hour of spontaneous speech contains almost no vocabulary of physical sensation, that locates sensing in the stack as surely as an excess would.

A single marker proves nothing: everyone says “consequently” once in a while. The diagnosis emerges from the frequency picture across all fields at once, on a sufficient volume of speech. The more spontaneous text, the more reliable the conclusion. Keep that dependency in mind; it returns in chapter 7.

3.4 Context Decides: Escaping the “Bag of Words” Trap

The naive way to analyze speech is to count words against a dictionary. That’s how first-generation tools like the classic LIWC system worked, and the ceiling showed up fast. “I am not happy” landed in the happiness category. Sarcasm wrecked the statistics. A lawyer’s professional jargon got booked as personality traits.

CoreTypes uses context-aware neural models of the BERT family. Models of this class read not isolated words but the connections between them: the syntactic role of a word in its sentence, the effect of negations (“not,” “never,” “except”), the stylistic register. Jargon, slang, and direct quotes get filtered out so they don’t contaminate the profile.

The difference is fundamental. What gets counted isn’t “how many times a word was said” but “what that word is doing in that sentence.” The accuracy of semantic typing rests on this move from dictionary to context.

Let’s be direct about the limits too. Semantic analysis determines the type of information metabolism, and only that. It doesn’t measure intelligence, doesn’t read minds, doesn’t evaluate values, and won’t tell you whether someone is a good person. The model answers one question, how this psyche processes information, and that’s precisely why it answers it reliably.

4. Age Limits: Why We Don’t Type Children

This is the question parents ask most: can you determine my child’s type, to help with school, activities, a future profession? The wish is understandable. The answer is no: CoreTypes doesn’t type people under 16. That’s a limitation of the method, not platform greed, and here are its reasons.

4.1 A Brain Under Assembly

The prefrontal cortex, the zone of planning, self-control, and complex decisions, matures at around 20 to 25 years of age. Until then the brain is physically rebuilding itself, and a child’s cognitive style stays fluid. Pinning down a type before it stabilizes is like benchmarking a processor that is still being soldered.

Why does the boundary sit at 16 rather than 20? A compromise between neurobiology and linguistics: the cortex formally matures later, but a person’s own speech style settles earlier, and by 16 the semantic analysis already has something of the person’s own to work with. For borderline cases we recommend an interview with an analyst rather than an automated test.

4.2 The Chameleon Effect: Teenage Mimicry

The second reason is linguistic. Teenagers absorb the speech of their environment faster than any other age group. Linguistics calls it communication convergence: the unconscious copying of the rhythm, slang, and vocabulary of whoever’s acceptance one wants. If the group’s norm is the language of loud emotion, a teenager will speak it, whatever their own type happens to be. Speech analysis at that age captures the fashion of the peer group, not the cognitive style.

What should parents do instead of typing? Observe without labels: what charges the child, what drains them, which problems they return to on their own. Those observations will serve both of you, and the type can wait until the speech becomes the person’s own.

4.3 Stability in Adulthood

By age 16 to 18, the core cognitive style generally settles. From then on, a person’s manner of writing and speaking holds for decades: views change, professions change, vocabulary grows, but the structural pattern of speech stays recognizable. Linguists have relied on this stability for a long time, up to and including the attribution of anonymous texts by authorial style.

From that point, semantic diagnostics works at full power, and its result stays meaningful for the rest of adult life.

5. The 16 CoreTypes: A Gallery of Archetypes

A four-letter formula says little to someone far from typology. INTJ needs decoding, while the word “Grandmaster” paints a picture instantly. So each of the 16 types carries an archetype name and a physical symbol. Not decoration but mnemonics: an image that makes it easier to remember how a given processor allocates attention in daily life.

Next to our names, the table lists the MBTI codes and the Socionics pseudonyms. That’s a bridge for people arriving from other systems: if you have spent years thinking of yourself as an INTP or a Balzac, you’ll find yourself in the Seer’s row. From here on, CoreTypes materials use our names.

The full gallery:

5.1 How Different Processors Run: Four Examples

To feel the difference between stacks, look at what four different types focus on and how it sounds in their speech.

• The Grandmaster (INTJ), Rook. Driven by Structural Logic: the mind automatically hunts for rules, hierarchies, and system connections. In speech, everything gets sorted into place: “this falls into three categories,” “consequently,” “going by the rules.”
• The Catalyst (ENTP), Spark. Runs on Intuition of Possibilities: the processor keeps cycling through alternatives and hidden potential. Fast speech, full of thought experiments: “concept,” “maybe,” “alternative,” “what if.”
• The Beacon (ESFJ), Lantern. Governed by Ethics of Emotions: attention tuned to the emotional field of the people around. Warm, expressive language, always about the atmosphere: “the vibe was incredible,” “everyone was thrilled.”
• The Harmonist (ISFP), Flower. Lives through Experiential Sensing: the present moment, aesthetics, the quality of sensation. Speech rich in physical texture: “soft texture,” “a cozy place,” “a balanced flavor.”

Four people will describe the same evening in four different vocabularies. Semantic diagnostics stands on that effect.

5.2 Different, Not Better

In the old behavioral systems, an unspoken hierarchy of types grew on its own: the “promising” ones and everyone else. Logical types collected compliments about intellect; ethical types got a condescending “but they’re warm.” We don’t support that approach, and not out of politeness.

Each of the 16 processors is a specialized instrument, honed for its own class of problems. A society needs the Grandmaster’s systemic depth as much as the Beacon’s warmth; it needs the Catalyst throwing ideas at the wall, and it needs the Harmonist to remember the world is physical, not conceptual. There’s no best type. There’s a type better suited to a given problem, and there are problems your type is better suited to.

6. Practical Application: From Career to Relationships

Knowing your type changes nothing by itself. What changes things is how you use it. Three areas where the payoff is most visible: career, team, close relationships.

6.1 Career: Run on Your Own Hardware

Chronic burnout is often simpler than it looks: for years, a person keeps launching software their hardware was never designed for.

A Grandmaster in chaotic sales, where the job demands constant forceful pressure, meaning Volitional Sensing, their Vulnerable function, runs dry within a year or two, and no time-management course fixes it. The mirror case: a Beacon in a role with no live human contact for weeks, only spreadsheets and regulations, wilts just as reliably, only more quietly. In both cases the difference isn’t discipline or motivation. It’s which channels of the stack carry the main load.

The working rule: build the career around the strong functions. Carriers of Structural Logic get systems, data, and order-making. Carriers of Intuition of Possibilities get idea generation, alternative-hunting, early project stages. Carriers of Ethics of Emotions get communication, teams, and work with atmosphere. This isn’t a ban on other professions but a map of energy costs: work through your own functions recharges you, work through foreign ones drains the resource even when it looks successful from outside.

6.2 Team Chemistry: The Right Processors Together

Teams usually fall apart not because the people are incompetent but because their processors clash. The reverse also holds: a well-assembled chain works like a conveyor. The Catalyst pours out raw ideas. The Grandmaster picks the viable ones and builds them into a system. The Executor (ENTJ) turns the system into a profitable plan: that is Practical Logic, his element. The Beacon sells the vision to the team and keeps the working climate alive. Four different stacks, minimal duplication, minimal friction.

The practical takeaway for a manager: hire not “another one just like us” but the function the chain is missing. Five identical stacks will produce five identical blind spots.

6.3 Relationships: Understanding the Friction

The feeling of “we speak different languages” in a couple or a family is often not a figure of speech but an accurate description: cognitively, the languages really are different. Knowing each other’s types changes two things.

First, expectations. If your partner’s Leading function is Structural Logic and Relational Ethics sits in the weak part of their stack (as with the Grandmaster), their dryness isn’t cooling off. Their attention simply doesn’t catch subtle emotional signals by default. With such a partner a direct request works; resentment over what went unguessed doesn’t.

Second, the reaction to conflict. When someone’s remark hits your Vulnerable function, it helps to recognize what is happening as stack friction rather than an attack. “Our processors are wired differently” sounds prosaic. It also helps you exhale before something unnecessary gets said.

7. How to Accurately Determine Your Type

Everything above matters on one condition: the type is identified correctly. Recall the dependency from chapter 3: the more spontaneous speech the analysis gets, the more accurate the conclusion. The CoreTypes diagnostic ladder is built on it, from quick estimate to expert review.

1. Free Express Test. The entry point. The test is built on semantic choice, not questions about habits. It isn’t all-powerful, but it produces a working hypothesis about your type and starts the process.
2. Paid Confirming Test. The next step if you want intermediate proof: checking the hypothesis function by function, narrowing the options, sharpening the initial result. Especially useful when the first test lands in a fork between two neighboring types.
3. Recorded Interview. The flagship format. You record video or audio answers to a set of questions, and a live analyst works through the semantics of your spontaneous speech. Expert analysis removes the distortions of automation and delivers accuracy close to 100%.
4. Live Session. For those who’d rather talk it through: a real-time session with an expert, questions as they come up, and advice tailored to whatever brought you here: career, marriage, a team that keeps stalling.

What to Do With the Result?

Learning your code is half the job. The other half is applying it: at work, in business, in relationships. For that we keep a shelf of applied materials: books that take each type apart in detail, plus step-by-step guides to intertype communication, the kind that tell you where friction with a particular type will come from and what to do about it. Start with the free test. Then pick up the book on your type. Somewhere around its second chapter, the typology starts paying for itself.