It was the latter phenomenon that prompted us to design the FD33. Four years ago, we discovered that the J/P designation in the MBTI was faulty. It outlaws, in an a priori (ie, 'by definition') manner, the possibility of, for instance, an 'introverted intuitive' type (IN) who prefers 'spontaneity' and does not seek closure (P). This type, BY DEFINITION does not exist, according to the MBTI, and cannot even be designated using the MBTI nomenclature - since INFJs and INTJs, the introverted intuitives, are all 'J' types, and neither the INFPs nor the INTPs are introverted intuitives! For lack of a better way of speaking about these types, rendered virtually invisible by MBTI assumptions, we began to speak of iNfps and inFjs.

We asked, "Might not some of the INs that we expected to see in zone 4 be iNfps, but rendered as INFPs (ie, IFs) by the MBTI?" How could one tell? Well, we might look at the individual INFP's MBTI scores, and see if he/she actually scored higher on the N scale than the F scale, despite having also scored higher in 'P' than 'J'. This might suggest the presence of an iNfp. We knew of other tests based on Jungian personality theory, ones which define the scales independent of each other. Perhaps these could be used. What would be the critical thing to look for in a test? Given the opportunity, what would WE want to test for? The answer seemed to be 'functional preference order', and this led to the experimental instrument that we arrived at, the FD33.

Here was our reasoning. It was very clear to us that the design flaw in the MBTI J/P designation was that the 'J' and 'P' in any given type name are definitionally ambiguous. In other words, the 'J' in INTJ for example is made to play two distinct roles that are not necessarily consistent. On the one hand the 'J' in INTJ means that the individual 'seeks closure'. On the other hand, it means that the second letter in the type name (N) is the dominant function. But are the two synonymous? - which they would have to be, in order for the letter 'J' to remain unambiguous and consistent. Or are there instances in which the two meanings come into contradiction with each other?

The way the J/P designation is used in the MBTI results in a state of affairs that is not unlike the one that would result from using the word 'democrat' in two senses - 1) to identify an individual's official political party affiliation, and to 2) identify the individual as one who believes in 'democracy' as an ideal form of government. In some cases those people whom we call 'democrats' in the first sense are also democrats in the second sense of the word. But there are cases in which an individual is a democrat in the first sense but not the second, or vice versa. And this same state of affairs is basically also what we have in the case of the J/P designation as it is used in the MBTI!

One way to clarify and remedy the situation is to SEPARATE OUT the two senses of the ambiguous word, and provide different terms for each meaning. Essentially, this is what we have done by suggesting the nomenclature that distinguishes between the 'iNfp' and the 'inFp'. The 'p' in this nomenclature simply and consistently refers only to 'spontaneity-seeking' (P) or 'closure-seeking' behavior (J) and is reserved for that purpose, and it is left to the capital letter in the name to designate which function is dominant. This language permits us, finally, to make the distinction that cannot be made using the original MBTI nomenclature. And it thus enables us to identify types that were previously rendered invisible - iNfp, the inFj, the eStj, and so forth.

Using the MBTI strictly according to MBTI rules, it should be theoretically possible to determine an individuals MBTI type simply by identifying 1) whether he or she is an 'I' or an 'E', and 2) his/her 'functional preference order'. By definition, the INFP is the only introverted group with an F-N-S-T preference order. To identify as an INFP is to say that he/she is an I/F-N-S-T, and vice versa.

But if you were to actually go about trying to do this - using some method independent of the MBTI test to identify the individual's preference order - you'd discover that there are individuals who are introverts and have an F-N-S-T preference order, but do not score as INFPs using the MBTI instrument. In fact, you'd discover that in addition to the 8 preference order patterns that are DEFINED as acceptable ones, there are [(4*3*2 - 8)=] 16 additional functional permutations that are definitionally 'outlawed', for BOTH the E and I orientation! The I/T-F-S-N, for instance, is just one of these 32 outlawed options.

Given this state of affairs, an very interesting question arises. As everyone who takes the MBTI will score as one of the 16 permitted types, which of the 16 MBTI types will each of the 32 'outlaw' types that we can identify in the way described above be rendered as using the MBTI? The definitive answer to this question is not obvious. Might the I/T-F-S-N, for instance, score as an INTP (who by definition has dominant T)? But what if that individual happens to have strong 'J' qualities? Will he therefore score as an INTJ? What other good option is there for him? But if he does score as an INTJ, this categorization will no doubt make for a troublesome label, as the individual will be assumed to have a dominant N function, whereas in reality, N is least prefered.

How about some of the other 'non-traditiional' preference orders - the I/T-N-F-S, for example? Or, for that matter, the traditional I/N-F-T-S preference order pattern - whom one would expect to score as an INFJ? Will a strong 'p' preference in such an individual render him as an 'INFP', despite an N-F-T-S preference order?

If so, won't various MBTI types be distributing across the enneagram in some rather peculiar ways? Wouldn't there be a certain amount of 'noise' thus created in the statistical data?

Well, what if we were to put a simple test together, one that attempted to discern 'functional preference order' patterns? If nothing else, our attempt to do so would demonstrate the phenomenon that we are interested in bringing to the attention of others who are exploring the relationship between the two typologies. Such an instrument might give us data with which we could assess whether those individuals who have uncommon Enneagram-MBTI combinations might also have uncommon functional preference order patterns. And, at the very best, the data might be of a level-of-specificity that brings into relief previously unperceived distribution patterns. We were trying, in other words, to demonstrate the kind of 'data lens' that might be most useful in this inquiry into how, and why, MBTI types distribute across the Enneagram.

In addition, we also trying to do something else with the FD33. We had previously discerned 5-levels of development of each of the four Jungian 'functions'. We have come to believe that individuals who use a function as their dominant function are more likely to develop that function than if they use it as an inferior or tertiary or perhaps even an auxiliary functions. Furthermore, individuals who have developed a function to a higher stage will tend to conceive of (or experience) that function in a different way. We see our 5-level model as an extension of Von Franz's very useful, but rudimentary, distinction between inferior and developed functions, and we have informally used the 5-level model, in coordition with a method of analysis that 'nests' function-related explanatory frames in such a way as to successfully type not only individuals, but supra-personal entities - organizations, theories, and so forth - in a way that emphasis on mere 'preference order' does not permit. We hope to have the chance of explaining this feature of our model in a future paper.

Would it be possible to formally incorporate some of these features or our model into the format of a simple 33 question survey - one which would render data that might be particularly useful in understanding the distribution of MBTI type across the Enneagram? To be honest, we don't know the answer to this question, but we are trying to find out. One thing that is clear to us, though, is that the FD33 scores should be taken with a grain of salt at this point, and should definitely not be taken as a 'substitute' for the MBTI. Indeed, for reasons that one can deduce from what we have said above, even if the questions were to be perfectly accurate in identifying functional-preference order, its results would not always be consistent with the functional-preference order predicted by MBTI type. We would expect that in some cases the preference orders identified by the FD33 would agree with the functional-preference orders associated with MBTI type. But we would also expect that some individuals, with 'non-traditional' preference orders, will nonetheless fall into one or another of the traditional MBTI categories, which will, by definition, have a different preference order from the one identified by the FD33.

Section One: General Observations

skip to section two

As yet, we have performed no statistical analyses on the data that we have accumulated via survey returns. We have made some general observations, and we've taken a preliminary look at a random sampling of 33 questionnaires from among the 309 that we have received to date. We are not sharing our thoughts and observations with you because we have come up with definitive answers, or even conclusions, but because some of you are eager to hear about the status of the project. In the same collaborative spirit that those who decided to share their results with us demonstrated, we have chosen to include you in on our thinking and experience as we slowly move from one step to the next.

The returns that are coming in are split pretty much right down the middle, genderwise. Out of the first 141 returns, there were 67 men and 74 women who had responded. Only 51 individuals specified EITHER their Enneagram or MBTI type (24 men and 26 women), and only 18 individuals (14 men and 4 women) gave us SOME indication of both of their types. And even amongst this rather small group who seemed to know both types, there were a few who were ambivalent, mostly in regard to Enneagram type, it would seem (although a few also report confusion about their MBTI type).

In other words, approximately 2/3 of the individuals who responded filled in neither of the type boxes. As we were originally interested in studying the distribution of MBTI type across the Enneagram, this of course meant that 2 out of every 3 returns provided us with no usable data regarding that issue. Nevertheless, these returns do provide us with data that can be reviewed with respect to the distribution of preference-orders, and to evaluate responses to particular questions.

After having received the first 141 returns, we began to take a closer look at the returns -and decided to quickly make some changes in the questionnaire. We added a 'don't know' box and a 'won't tell' box, in addition to the two types boxes. In subsequent returns, this additional information has helped us to see that many individuals simply don't know their types, although a small handful seem reluctant to report them, for whatever reason.

In beginning to analyse the first batch of returns we also discovered some anomalies in the computer program which did the scoring, which resulted in some errors when individual chose specific answers. Although we were able to correct this problem and adjust scores accordingly, some individuals who used the on-line program may not have have received entirely accurate scores. For this we apologize.

In addition, on one question in particular, the computer was over-writing the answer with the answer to another question, corrupting the data in a way that made the original answer irretrievable. Although this will have effected only a small portion of the scores, it is a source of error.

It was pointed out to us that the phrasing of some questions was ambiguous or cumbersome, so when we made the changes to the computer program, we also substituted clearer questions, and some typos that had previously escaped our attention.

The second batch of questionnaires - approximately the same size at the first, have roughly similar patterns as those mentioned above.

A Sample of 33 Returns

skip to section three

When I began to write this paper, I wanted to take a closer preliminary look at the the data. I chose the last 33 returns that we had received. Such a small sample will not, of course, render statistically signficant results. But I was not looking for results, I was trying to answer the question, 'what kind of patterns am I looking for to begin with?', and this exercise proved helpful. I share it with you in the spirit of collaborative inquiry.

In this group there was again about a third (13) who knew either one or the other of their types. But twice as many individuals (approximately 20%) knew both of their types.

Out of the 13 who knew either one or the other of their types,

A total of 5 had preference orders that agreed with the preference orders associated with their MBTI type, and the two who knew both types fell into the 'most common' MBTI-Enneagram combinations -

mbti type	order	enneatype
infj	n-f-t-s	4
intp	t-n-s-f	5
intp	t-n-s-f	don't know
enfp	n-f-t-s	don't know
enfp	n-f-t-s	don't know

3 had preference orders that inverted the second and first MBTI function, but remained within 'traditionally' allowed MBTI patterns -

mbti type	order	enneatype
intp	n-t-f-s (associated traditionally with intj)	don't know
entp	t-n-s-f (associated traditionally with entj)	7
infj	f-n-s-t (associated traditionally with infp)	don't know

And 5 out of the 13 had more radically different non-traditional preference order patterns, which also were inconsistent with their mbti type. Interestingly, in this group, the uncommon preference orders seem to make sense in the context of the individual's enneatype -

mbti type	order	enneatype
infp	n-t-s-f	4 (for which IN is prototypical)
infj	f-s-n-t	2 (for which EF is prototypical)
intp	n-s-f-t	4 (for which IN is prototypical)
infp	n-t-s-f	don't know
esfp	f-n-s-t	don't know

Might a less common Enneatype and MBTI combination in an individual indicate that that individual is more likely to actually be a 'non-traditional' MBTI type, with a non-traditional preference order? This small sample proves nothing. But does it suggest a testable hypothesis?

Is it possible that some of the 'uncommon' combinations are neither reporting mistakes attributable to the individual, nor signs that the individual is involved in some kind of inner 'conflict' between different 'parts' of the individual (ala Wyman), but an artifact that results from trying to fit an individual with a non-typical preference order (or preference order/orientation combination) into one of the sixteen categories made for individuals with permitted preference orders? Might this account for why the MBTI seems to some individuals particularly awkward and unhelpful, just as the Enneagram is similarly experienced by others?

Recently I have come across a number of people who have been having difficulty establishing (or feeling comfortable with) either their Enneagram or their MBTI type, or in coming to grips with less common combinations of both. Are we doing a disservice to them if we don't take seriously the possibility that it may not be the foot that doesn't fit the shoe, but the shoe that doesn't fit the foot? It may seem like six of one or half dozen of the other, but when we take the former approach, people inevitably feel that it is they who lack something. Understanding what is happening at the margins, or in the interface between systems, perhaps we can cause these systems to expand to accomodate those who are otherwise effectively excluded.

Well, those are some of the questions I ask myself, anyhow. Onward now to a more precise analysis of the data, which may take some time to perform.