Cognitive functions tests: what the eight functions measure

What are the eight cognitive functions?

The tradition describes two kinds of mental activity. Perceiving functions take information in: Sensing attends to what is concrete and present, Intuition to patterns and possibilities. Judging functions decide what to do with it: Thinking weighs logic, Feeling weighs values and the effect on people. Each of the four can point outward, at the world of events, or inward, at a private world of impressions. Four functions in two directions gives eight.

• Extraverted Sensing. What is happening now: the room, the pace, the chance in front of you.
• Introverted Sensing. Stored experience: what worked last time, how a thing is usually done.
• Extraverted Intuition. Branching outward into possibilities and what a thing could become.
• Introverted Intuition. Converging inward on one pattern, or on where a situation is heading.
• Extraverted Thinking. Organising the outside world: measures, sequences, making a system work.
• Introverted Thinking. Building an internal model that has to hold together before you use it.
• Extraverted Feeling. Reading the mood of a group and tending it.
• Introverted Feeling. Checking a choice against a private sense of what matters.

The names come from Carl Jung's book on psychological types, published in the 1920s. Ranking the eight belongs to the later Myers-Briggs tradition. That is intellectual history rather than evidence, and the two are worth keeping apart.

What is a function stack, and what does a test ask?

A stack is an ordering. Each four-letter type is assigned a dominant function it leads with, an auxiliary that supports it, a tertiary that develops later, and an inferior function said to surface under pressure. Some versions extend the order to all eight and call the lower four shadow functions. Only sixteen orders exist, one per type, because the order comes from the letters rather than from your answers.

The tests ask you to rate statements written in the language of one function at a time. An Introverted Thinking item might say you need an idea to be internally consistent before you act on it. An Extraverted Feeling item might say you notice when the mood of a room drops. The items for each function are averaged, the averages ranked, and a type inferred from the top two.

Two features of that design matter. The items are written in the theory's own vocabulary, so agreeing with them partly measures how familiar you find the theory. And because the result is a ranking, functions separated by a fraction of a point are reported as first and second.

What does the evidence say about the functions and the stack?

The functions themselves have attracted little published testing. What has been tested repeatedly is the four-scale instrument the stacks are derived from, and those results set the ceiling for anything built on top. Scored alongside a five-factor inventory in a sample of 468 adults, the four Myers-Briggs scales lined up with four of the five broad traits and behaved as continuous dimensions rather than as truly dichotomous preferences or distinct types (McCrae & Costa, 1989). The authors proposed the five-factor model as a better framework for interpreting the four indices.

Furnham (1996) gave both instruments to one sample and found the same alignment, with nothing in the letters corresponding to Neuroticism, the dimension most closely tied to how steady a person feels under pressure. A review for practitioners drew the practical conclusion: the four-letter type formula does not support the inferences commonly drawn from it, and considerable caution is warranted before acting on one (Pittenger, 2005). A score near a cut point can fall either side of it, which makes the letters a less solid thing than they look.

There is a structural point underneath the psychometric one. Decades of lexical research, sifting the words people use to describe themselves and each other, produced a structure of five broad dimensions that survived repeated attempts to replace it (Goldberg, 1993), and a six-factor version that adds Honesty-Humility (Ashton & Lee, 2007). Eight ranked functions are a scheme applied to the data, not a structure that comes out of it. That does not make them useless as language, but a stack should not carry a decision about work or a relationship.

Why do the eight functions feel so accurate?

Because they are eight true things about human minds, and you do all eight. Everybody attends to the present sometimes and to patterns sometimes, everybody weighs logic sometimes and values sometimes. A description assembled from those parts is nearly certain to contain sentences you recognise. Forer showed the effect in a classroom: students handed an identical, generic sketch rated it as a highly accurate description of themselves (Forer, 1949). Recognition is not evidence of fit, because a description can suit almost everyone and still not tell you apart.

The stack adds a second reason. When part of the description misses, the theory supplies an explanation: that function sits lower in your order, or your inferior function had the upper hand that week. A system that absorbs any mismatch is comfortable and hard to check.

How do the eight functions map onto trait language?

If you know your stack, here is the closest reading in trait terms. Treat it as a reading, not a validated mapping: no study has measured the eight functions against the six dimensions and reported these correspondences. They are inferences from what each function is described as doing, and they point at which dimensions to look at, not where on them you sit.

How do you get self-understanding from a trait profile?

The reason to take a functions test is rarely taxonomy. It is that you want a better account of why you work the way you do.

• Ask for degree, not direction. Not whether you lead with Thinking or Feeling, but how far from average your Agreeableness sits.
• Look inside a dimension. High orderliness with low follow-through, or high compassion with low politeness. The facets are where that lives.
• Check it against someone. People are the better judge of their own inner life, such as how anxious they feel, while friends judge evaluative traits such as intellect more accurately (Vazire, 2010).

Broad traits also carry a predictive record, forecasting outcomes such as occupational attainment, relationship stability and longevity about as well as socioeconomic status or measured cognitive ability (Roberts et al., 2007). Those are group-level associations, not forecasts about one person, but they are more than a stack has shown.

Why does Wellington use scores and bands instead of a stack?

The traits a function description gestures at are matters of degree. Measured across a large sample, each falls into a bell shape, most people near the average and a thinning tail on either side. Everybody has an amount of each trait.

Because the curve is smooth there is no natural place to cut it, so any either-or line is put there by whoever wrote the test. The cut costs the distance: two people either side of it are reported as opposites, and somebody near it can rate the same statements slightly differently another day and get a different result. Scales of this kind, measured against a trait inventory, proved to be continuous dimensions without distinct types underneath (McCrae & Costa, 1989), on the broad structure decades of descriptive research produced (Goldberg, 1993).

Wellington reports the distance instead of hiding it. Each dimension becomes a percentile against a reference sample, and three bands sit over it with published cut-offs: Quiet as far as the 30th percentile, Balanced as far as the 70th, High above it, on norms provisional until the norming sample is complete. The bands and the Portrait names exist to make the result easier to talk about; the percentile is the measurement, and it is always shown alongside.

The name follows the numbers rather than the other way round. A named Portrait summarises the pattern of bands in one phrase, for practicality, and the percentiles stay visible beneath it as the measurement.

How does Wellington measure this?

Wellington is a science-based personality assessment from Therabot Labs LLC. It measures personality as continuous traits, built on the Big Five and HEXACO models, and writes the results back as a personal report rather than a type.

There are no functions and no stack. The free Snapshot asks 60 questions, takes about 7 minutes, and scores 6 continuous dimensions: Extraversion, Agreeableness, Conscientiousness, Emotional Stability, Openness to Experience and Honesty-Humility. Every statement is rated on a 5-point accuracy scale, each dimension is the average of several statements, and that average is placed against a reference sample as a percentile and reported in one of three bands: Quiet below the 30th percentile, Balanced to the 70th, High above it. Those norms are provisional while the norming sample is completed.

In the written report, each dimension has three pages, one per band, prepared in advance by the writing and psychology team, and your answers decide which one you read. The Wellington Membership adds the facets beneath each dimension and the 24 VIA character strengths, 90 traits in all.

The first ten Snapshot questions are answered before anything is asked of you; after question ten you sign in with an email address so answers are saved and results can be shown. No card is needed. Wellington is a wellness tool for reflection, not a medical or psychological diagnosis, and it does not replace care from a qualified professional. It is not built for hiring, and you can export or delete your data.

Questions people ask

Are cognitive functions scientifically supported?

The eight functions themselves have very little published testing. The four-scale instrument they are derived from has a great deal, and it shows those scales relating to four of the five broad traits while behaving as continuous dimensions rather than types (McCrae & Costa, 1989). A review for practitioners concluded that the four-letter type formula does not support the inferences commonly drawn from it (Pittenger, 2005).

Why does my function stack keep changing between tests?

Because a stack is a ranking of eight averages, and adjacent functions are often separated by a fraction of a point. A small change in mood or wording can swap two positions and hand you a different dominant function, and so a different type. For the same reason, a review of the four-letter format concluded that a type formula cannot carry the inferences drawn from it (Pittenger, 2005).

Is a cognitive functions test the same as the MBTI?

No. The official Myers-Briggs instrument scores four either-or scales and reports a four-letter type. Cognitive function tests score eight functions and infer a type from the top two. They share a lineage and most of their vocabulary, and the published evidence sits with the four-scale version rather than the stacks built from it.