Keirsey's "Mirror Temperaments" and the MBTI:
A New Bridge between Type and Temperament

An important way of comparing and contrasting the 16 MBTI types is to arrange them into four groups of four types each. The basic way that Isabel Myers and Mary McCaulley constructed four-type groups was to combine the preferences from any two MBTI dimensions. For instance, from the dimensions of extraversion-introversion and sensing-intuiting, they created the extraverted sensors, extraverted intuitives, introverted sensors, and introverted intuitives -- or, to abbreviate, the ES, EN, IS, and IN grouping. With four dimensions of MBTI preferences, six such groupings are possible, and these are listed on pages 32-38 of the Manual. These type groupings are all symmetrical or "regular," since two sets of preferences are used to generate them.

Some critics have claimed that all these groupings are just so much "alphabet soup," with little or no connection or application to the real world. Over the past several decades, however, numerous surveys and other studies have revealed clear-cut patterns among the types, and some four-type groupings have become especially prominent as a result. One of the most noteworthy, given special prominence in the Manual, is the function-combinations, generated by combining the dimensions of sensing-intuiting and thinking-feeling. This produces the sensing thinkers, sensing feelers, intuitive thinkers, and intuitive feelers -- the ST, SF, NT, and NF grouping.

Of the other five symmetrical groupings, one in particular has received little attention, and it is the focus of my other talk here at this conference. In "Achilles Tendencies," I explore the combination of thinking-feeling with judging-perceiving, or the thinking judgers, thinking perceivers, feeling judgers, and feeling perceivers -- the TJ, TP, FJ, and FP grouping, which is type group #4 in the Manual. Since each of these four type groups is the combination of a judging function with the attitude of judging or perceiving, I call them the "Judging Attitudes," and my mission to this conference is to show how they are anything but "alphabet soup."

The Judging Attitudes have a lot to teach us about human behavior and experience, especially defensive communication and codependent personal relationships. My "Achilles Tendencies" talk focused on this area, and showed how techniques and ideas from family therapy and the 12-Step movement can help with problems that arise out of unhealthy Judging Attitudes.

The Judging Attitudes also shed new light on how personality relates to age-old ethical concerns such as freedom vs. control and the moral vs. the practical. One of the main points of this talk is to show how David Keirsey addresses these concerns with his temperament model in a way that points directly to the Judging Attitudes. Specifically, it is the "Mirror Temperaments" implicit in his discussion of role-directiveness and role-informativeness that reveal the Judging Attitudes, MBTI type grouping #4, to be a very helpful bridge between type and temperament.

In contrast with the Judging Attitudes, Keirsey's four temperaments -- the SPs, SJs, NFs, and NTs -- are a non-symmetrical grouping of types, and they have attracted a great deal of attention. Critics charge that Keirsey's temperament model is arbitrarily cobbled together, making a hash out of type groupings #2 and 3 in the Manual. Further, it could be pointed out, while Keirsey says some interesting things about cooperative vs. utilitarian behavioral styles in relation to his four temperaments, he seems to abandon or de-emphasize the temperaments, splitting them down the middle, in discussing another very interesting set of behavioral styles, the role-directives and role-informatives.

Here, again, there is a lot to learn that has not been realized previously. Keirsey has four key ethical concepts that he proposes to illuminate and explain with his temperament model, and a close look at what he does with them will provide some unexpected new connections between temperament and type theory.

Running throughout Keirsey's work is his basic system of four temperaments and an acknowledged strong connection of those temperaments to the 16 types of the Myers-Briggs Type Indicator. The Rationals are equated with the four NT types, the Idealists with the four NF types, the Guardians with the four SJ types and the Artisans with the four SP types. For ease of reference, I will simply refer to the NT, NF, SJ, and SP temperaments, on the understanding that each of them encompasses the four MBTI types that contain those letters.

Keirsey says that both the NTs and SPs are relatively more pragmatic than the other two temperaments -- more concerned with what works, what is practical or utilitarian. Both the NFs and SJs, on the other hand, are relatively more cooperative -- more concerned with what is right, ethical, or moral. Keirsey emphasizes that these are not all-or-nothing categories, just relative tendencies or inclinations owing to inborn temperamental differences.

This is an extremely important contribution to personality theory, though it needs to be sharpened and focused even more. Before doing that, however, let's turn first to the other set of important behavioral ideas Keirsey presents in Portraits of Temperament.

In developing his concepts of "role-directive" and "role-informative," Keirsey points out that each temperament is divided down the middle, with two of the types being role-directive and two role-informative. One might wonder what these leaves of his temperaments. What is left, I suggest, is the best-kept secret of temperament theory, Keirsey's hidden "mirror temperaments."

(Just to clarify: when I say "Keirsey's," I don't mean that he personally approves of or perhaps even knows about the "mirror temperaments" -- merely that they are a real, important part of his temperament model. And when I say "best-kept secret" or "hidden," I don't mean to imply any duplicity or sneakiness on Keirsey's part, merely to dramatize the fact that vast riches yet to be uncovered in his system apparently have not even been suspected as being there!)

Let's develop this "mirror temperament" idea. Keirsey begins with the temperaments -- NT, NF, SJ, and SP -- and he sees that each of them contain some role-directives and some role-informatives. He looks at the four NTs and sees that the two NTJs (ENTJ & INTJ) are role-directives, while the two NTPs (ENTP & INTP) are role-informatives. Of the four NFs, the two NFJs (ENFJ & INFJ) are role-directives, while the two NFPs (ENFP & INFP) are role-informatives. Of the four SJs, the two STJs (ESTJ & ISTJ) are role-directives, while the two SFJs (ESFJ & ISFJ) are role-informatives. And of the four SPs, the two STPs (ESTP & ISTP) are role-directives, while the two SFPs (ESFP & ISFP) are role-informatives.

So, Keirsey has two new categories. One is the role-directives, which include the two NTJs (ENTJ & INTJ), the two NFJs (ENFJ & INFJ), the two STJs (ESTJ & ISTJ), and the two STPs (ESTP & ISTP). The other is the role-informatives, which include the two NTPs (ENTP & INTP), the two NFPs (ENFP & INFP), the two SFJs (ESFJ & ISFJ), and the two SFPs (ESFP & ISFP). That is where his analysis stops -- and where mine begins. (It's highly significant that his analysis stops here, for in his most recent book, he has greatly de-emphasized the issue of role-directive vs. role-informative, compared to that of cooperative vs. utilitarian. This is most unfortunate, considering that control vs. freedom is an ethical issue of comparable importance to the issue of moral vs. practical.)

Now, note that among the eight role-directive types are four NJs and four STs, and that among the eight role-informatives are four NPs and four SFs. In terms of shared preferences, the role-directive NJs and STs are polar-opposites of Keirsey's SP and NF temperaments, and the role-informative NPs and SFs are polar-opposites of Keirsey's SJ and NT temperaments.

One might well wonder why Keirsey didn't just use the SPs and NFs for his role-informatives. The answer to that is simple: there is overlap between the SPs and the role-directive STs, and between the NFs and the role-directive NJs. Two of the SPs are STs, and two of the NFs are NJs. So that wouldn't work as a way of sorting the 16 types into two separate groups of role-directive and role-informative. But it is still highly important that the SPs and NFs are polar-opposites to the role-directive NJs and STs, as will become evident. Similarly, in answer to the question: why didn't Keirsey just use the SJs and NTs for his role-directives, we note that the overlap of those temperaments with the SFs and NPs would make the role-directives overlap the role-informatives. But again, it is very significant that the SJs and NTs are polar-opposites to the role-informative NPs and SFs, as we will see shortly.

Now, in presenting his concepts of role-directive and role-informative, Keirsey has unexpectedly created four new type groups, the NJs, STs, NPs, and SFs, which are polar-opposites of his four temperaments, the SPs, NFs, SJs, and NTs. We can learn a lot from these four new groups just because of their being polar-opposites to the original four temperaments. First, however, I will explain how I came up with the name "mirror temperaments."

Another useful way to look at Keirsey's four new groups is to think of S and N as being like left and right, or as "mirror reflections" of one another. In this respect, each of the four new groups can be thought of as a mirror-reflection of one of the original four temperaments: NJ is the mirror of SJ, NP is the mirror of SP, ST is the mirror of NT, and SF is the mirror of NF. That is why I call these four new groups the "mirror temperaments." (Thus, each of the four new groups is the "mirror" of one of the original four temperaments and the "polar opposite" of another one of the original four. E.g., the NJ mirror temperament is a mirror to the original SJ temperament and a polar-opposite to the original SP temperament. Both relationships are important, as we will see.)

One might now ask: if the NJ and SF mirror-temperaments are polar-opposites to the pragmatic SP and NT temperaments, why not regard the NJ and SF mirror-temperaments as cooperatives? Similarly, if the NP and ST mirror-temperaments are polar-opposites to the cooperative SJ and NF temperaments, why not regard the NP and ST mirror-temperaments as pragmatists? Now there is no overlap: the NJs and SFs vs. the NPs and STs as cooperatives vs. pragmatists are a complete and non-overlapping breakdown of the 16 types. Logically, at least, Keirsey could have used the mirror-temperaments to express all four ideas: cooperative, pragmatic, role-directive, and role-informative.

But with equal logic, Keirsey could have used the original temperaments to express all four ideas. If the original SP and NF temperaments are polar-opposites to the role-directive NJ and ST mirror-temperaments, why not regard the SP and NF temperaments as role-informatives? Similarly, if the original SJ and NT temperaments are polar-opposites to the role-informative NP and SF mirror-temperaments, why not regard the SJ and NT temperaments as role-directives? Again, there is no overlap: the SPs and NFs vs. the SJs and NTs as role-informatives vs. role-directives is a complete and non-overlapping breakdown of the 16 types. Thus, Keirsey could just as well have used his original temperaments to express all four of his basic behavioral ideas.

There is a problem in using either the original temperaments or the mirror-temperaments exclusively to illustrate the Keirsey's behavioral concepts. There is no logical way to set up a two-by-two matrix to show the type groups as opposites to one another. If you use the original temperaments and make the rows cooperative and pragmatic and the columns role-directive and informative, you find that the SJs are opposite the SPs and the NTs opposite the NFs. This makes no sense because the supposed opposites share one of their two explicit preferences. The same result occurs with the mirror temperaments: the STs are opposite the SFs and the NJs opposite the NPs -- the same logical defect that resulted in the first matrix.

These logical conflicts may be the reason that Keirsey -- knowingly or not -- kept two of his behavioral ideas attached to the original four temperaments and the other two associated with the (unrecognized) mirror-temperaments and did not try to combine them in a single diagram. Instead, the four-part diagram he used for the temperaments made the rows cooperative and pragmatic and the columns sensing and intuitive, which puts the NFs opposite the SPs and the NTs opposite the SJs.

But while Keirsey's temperament matrix does not contain an explicit conflict, since apples and oranges are being compared (F vs. P and T vs. J), there is still an implicit conflict. Some of the supposedly opposite types have some shared preferences. For instance, ENFP and ESFP share three preferences, so one wonders how they could be said to be opposite in any significant way, yet Keirsey regards the former as cooperative and the latter as pragmatic. Or, consider INTJ and ISTJ, which again share three preferences, yet the former is supposedly pragmatic and the latter cooperative.

With all of these anomalies and logical glitches piling up, it's time to reveal the remaining secret, the x-factor that vastly simplifies the discussion of Keirsey's behavioral concepts and shows how elegantly type theory captures what he was getting at. There are several ways to uncover this missing link between temperament and type, but I first noticed it when I read Portraits of Temperament and looked carefully at Keirsey's mirror-temperaments.

Consider Keirsey's eight role-directive types -- the four NJs (ENTJ, INTJ, ENFJ, and INFJ) and the four STs (ESTJ, ISTJ, ESTP, and ISTP). Note that fully 50% of these types are TJs (and that no other combination amounts to that large of a percentage). Does that produce an "aha" reaction? Numerous management studies reported in the APT Journal have revealed that the vast majority of managers -- and who are role-directive, if not managers? -- are TJ types.

We can go even further. Keirsey's eight role-informative types -- the four NPs (ENTP, INTP, ENFP, and INFP) and the four SFs (ESFJ, ISFJ, ESFP, and ISFP) -- include four FP types (again, 50% of the total). One of the management studies referred to above shows that it is specifically the command-and-control function of line management where TJs are the largest type group (80.1%) and FPs the smallest (3.8%). Here is a true opposition in types, with no overlap, between role-directive and role-informative: TJ vs. FP. It is obvious that the biggest single factor influencing the character of the role-directive NJs and STs is the presence of the four TJs -- and the biggest single influence on the role-informative NPs and SFs is the four FPs.

(Recall, now, my earlier claim that Keirsey could have derived role-directive and role-informative from his original temperaments? I suggested that since the NTs and SJs are polar-opposites to the role-informative SFs and NPs, the NTs and SJs could be considered role-directive. Now we can see just how appropriate that substitution would have been, because among the four NTs and the four SJs are four TJs (ENTJ, INTJ, ESTJ, and ISTJ). Similarly, among the four NFs and the four SPs are four FPs (ENFP, INFP, ESFP, and ISFP), which makes it just as appropriate to think of the NFs and SPs as role-informative. What is now clear is why those substitutions would work. The biggest common factor in both Keirsey's role-directive NJs and STs and my substitute role-directive SJs and NTs is the presence of four TJ types -- and the biggest common factor in both Keirsey's role-informative NPs and SFs and my substitute role-informative SPs and NFs is the four FP types.)

At this point, our bridge from temperament theory to type theory is now half built, and already we can clearly see where it is going, how it connects to type theory. We have distilled the essence of role-directiveness from Keirsey's NJ and ST groups and found it to be the four TJ types, and we have distilled the essence of role-informativeness from his NP and SF groups and found it to be the four FP types. This points directly to type grouping #4 in Myers' and McCaulley's MBTI manual: the TJ, TP, FJ, and FP grouping, which is based on combining the Thinking and Feeling preferences with the Judging and Perceiving preferences. Building the other half of the bridge will require us to see how the TP and FJ groupings are also found, in hidden form, in Keirsey's temperament model.

Let's return to Keirsey's sorting of the original four temperaments as to how they "implement goals." Consider now what the utilitarian or pragmatist NTs and SPs have in common -- i.e., what is the single largest contingent of types among them. There are four TPs, again fully 50% of the total, with no other grouping accounting for such a large proportion. Similarly, the cooperative or ethical NFs and SJs include four FJs, the largest grouping among their eight types.

(Note in passing that because the ST and NP mirror temperaments, being polar-opposites of the ethical or cooperative NF and SJ temperaments, could well be regarded as a stand-in for the pragmatist or utilitarian NT and SP temperaments. A second, very important reason why this is so: like the NTs and SPs, the STs and NPs include four TPs among them. Similarly, the SF and NJ mirror temperaments, being polar-opposites of the pragmatist or utilitarian NT and SP temperaments, are stand-ins for the ethical or cooperative NF and SJ temperaments. Again, the NFs and SJs, the SFs and NJs include four FJs among them. So, just as Keirsey could have derived role-directive and role-informative from his original temperaments, he could have derived cooperative and utilitarian from his unacknowledged mirror temperaments.)

All roads lead to Myers-McCaulley type grouping #4: the TJ, FJ, TP, FP types. Whether we start from Keirsey's original temperaments and observe their commonalities, or instead focus on the core of his mirror-temperaments, we reach the same insight. The quintessential role-directives are, in Myers and McCaulley's words, the "executive, tough-minded" TJs, the role-informatives are the "gentle, affiliative" FPs, the cooperatives are the "benevolent, need-oriented" FJs, and the pragmatists are the "impersonal, adaptable" TPs (Manual, pp. 36-7).

Further, the essence of issue of the moral vs. the practical is contained in the opposition between the cooperative or ethicist FJ types and the utilitarian or pragmatist TP types. And the essence of the issue of freedom vs. control is contained in the opposition between the role-informative FP types and the role-directive TJ types.

Thus, Myers-McCaulley type grouping #4, the Judging Attitudes, not only serves as a direct bridge between MBTI type theory and Keirsey's temperament system, but also contains in distilled, concentrated form the essence of the two fundamental ethical-behavioral issues of human social existence. In this respect, Keirsey's temperaments and mirror-temperaments are an astonishingly accurate "first approximation" of the Judging Attitudes contained in MBTI type grouping #4. Furthermore, the growing rapprochement between type and temperament theory receives a significant increase in credibility, in light of the fact that Keirsey's four behavioral-ethical concepts of "role-informative," "role-directive," "cooperative," and "utilitarian" are a near-perfect fit for the characteristics of the Judging Attitudes in type group #4.

2. Bissell, Roger E. (1999), "Keirsey's Mirror Temperaments and the MBTI, a New Bridge Between Type and Temperament," APT XIII International Conference, full text available online at: http://www.rogerbissell.com/achillestendencies/atmirror.html

3. Choiniere Ray and David Keirsey (1992), Presidential Temperament: the Unfolding of Character in the Forty Presidents of the United States, Del Mar, CA: Prometheus Nemesis Book Co.

4. Keirsey, David (1987), Portraits of Temperament, Gnosology Books, Ltd. (distributed by Prometheus Nemesis Book Co., Del Mar, CA).

5. _________(1998), Please Understand Me II: Temperament, Character, Intelligence, Del Mar, CA: Prometheus Nemesis Book Co.

6. Keirsey, David and Marilyn Bates (1978), Please Understand Me: Character & Temperament Types, Gnosology Books, Ltd. (distributed by Prometheus Nemesis Book Co., Del Mar, CA).

7. Montgomery, Stephen (1989), The Artisan, volume one of The Pygmalion Project: Love and Coercion Among the Types, Del Mar, CA: Prometheus Nemesis Book Co.

8. ________ (1990), The Guardian, volume two of The Pygmalion Project.

9. ________ (1993), The Idealist, volume three of The Pygmalion Project.

10. Myers, Isabel Briggs and Mary H. McCaulley (1985), Manual: a Guide to the Development and Use of the Myers-Briggs Type Indicator, Palo Alto, CA: Consulting Psychologists Press. 

11. Reynierse, J. H. and J. B. Harker (1995), The Psychological Type of Line and Staff Management: Implications for the J-P Preference. Journal of Psychological Type, 34, pp. 8-16.