James Reynierse has published (or co-published) a whole series of articles which attack MBTI type dynamics. These articles, published in the Journal of Psychological Type, available from CAPT behind a pay wall, include:
- Oct 2008Preference Multidimensionality and the Fallacy of Type Dynamics: Part 1 (Studies 1–3). James H. Reynierse and John B. Harker
- Nov 2008 Preference Multidimensionality and the Fallacy of Type Dynamics: Part 2 (Studies 4–6) James H. Reynierse, John B. Harker
- Jan 2009 The Case Against Type Dynamics, James H. Reynierse
- Mar 2012 Toward an Empirically Sound and Radically Revised Type Theory, James H. Reynierse
Since he's primarily attacking things from a empirical/statistical perspective, this may prove unconvincing to some (for example, see INTP's reaction to a mention of Reynierse in this thread).
Still, I think Reynierse's approach to be interesting and not unconvincing,in its way. However, I think it effectively adds more value to a Big Five (aka Five Factory Model or FFM) approach. With his latest "Toward an Empirically Sound and Radically Revised Type Theory" article, Reyneirse finally revealed what I thought his previous articles had been working for: his own view of what a sound MBTI-like type theory would be.
He lists the following five principles for his theory (not all of them are central to understanding his approach):
- Principle 1: All individual human beings differ with their own personal
identity and individuality formed by their own unique genetic, ontogenetic,
epigenetic, and experiential background.
- Principle 2: The individual MBTI preferences are the fundamental unit
of analysis for type theory.
- Principle 3: The individual preferences are arranged as sets of
- Principle 4: The individual preferences are free to combine with each other and in any order.
- Principle 5: The combination of individual preferences is additive
rather than interactive.
- Principle 6: The expression of psychological type is fundamentally
contextual and situational.
- Principle 7: MBTI preference scores matter and indicate strength of
- Principle 8: Type dominance is a function of strength of preference and the dominant preference is simply the independently high-value preference.
Independent preference scales
So, his fundamental argument is that each MBTI "dichotomy" is not a dichotomy and stands on its own. Reynierse argues that each of the four letter pairs of the MBTI represents a continuous, non-dichotomous spectrum of preference. That is, one isn't either "I" or "E" exclusively, but one's preference falls somewhere on the E/I scale (this is very similar to Big Five trait theory).
He also argues that each preference stands on its own equally. Therefore J/P is not a "pointer" to the orientation of functions, but instead is a stand-alone dimension that functions as independently as its peers do. In no sense is J/P "special" as a preference or scale (nor are the other scales uniquely affected by it).
So, when using the MBTI (the instrument, not the theory) he would argue that the numeric scores are significant as a rough approximate of strength of preference, and the strength of preference is important.
In Reynierse's view, every person does not have a "dominant function", but instead has a "dominant preference" (or preferences, if the two strongest preferences are roughly equal). The dominant preference is just to the strongest preference on any of the four preference scales. So one's strongest preference might even be "J" or "I".
Effects of preferences
Reynierse also argues that not only does each preference have an effect, but the combination of two or three preferences together can act additively. However, in his analysis he found that three preferences together only sometimes had an effect, and the addition of a fourth preference seemed to be generally statistically insignificant.
In addition, the effect of any single preference, preference dyad, or triad was often situational; ascribing the effect of a dyad or triad to a "whole type" just muddies the waters.
Reynierse gives examples of lexical descriptions and their correlations with ordered preferences—so order is important here, and association may be strongest with the additive qualities of two or three preferences. So, for example, EFP is associated with "fun loving" (compare with PE's association with "impulsive"), SJ with "conservative" (compare with JS's "organized") and JT with "thorough" (compare with TJ's "decisive").
Reynierse suggests a type code of preferences ordered strongest to weakest. He expands this into an eight-letter code (with the second half reversing the order and letter of the first so that one can tell at a glance where a preference falls). Under his type code, I would most likely be an "NPIFtejs", since my intuition and perceiving preferences are the most extreme ("INFP" becomes "NPIF" when arranged by strength of preference, and "tejs" being reversed and opposed form of "NPIF"... just "estj" rearranged to mirror).
This form of type code makes it easy to see which preferences have less influence (the ones in the middle), and which ones are more extreme (ones at the ends).
What's lost and what's gained
There are interesting pros and cons to Reynierse's approach (ignoring questions of absolute truth/truthiness). By making preference strength central, he effectively creates more sub-types of the 16 standard types. It also makes room for people who have an "X" in their code (even in the J/P slot) or who otherwise don't seem to follow type dynamics in practice.
One can even salvage some of the nature of the 8 functions by merely associating them with preference pairs. So, for example, Te just becomes the result of the additive qualities of T and J. Si becomes the additive qualifies of S and J. Ne of N + P, etc. However, such an approach would not seem to naturally yield Fe/Ti and Te/Fi pairings.
This also suggests it should be possible to write equally insightful descriptions of other preference pairs (and triads). And some of those would be more descriptive of an individual than the traditional eight functions (which ought to work best for peoples whose strongest two preferences are J/P and one of S/N/F/T).
However, in this model type dynamics goes by the wayside. No longer would there be Fe-doms and Ni-doms and the like. No longer would the there be any expected model of type development (at least not in any particular order for any given type).
As an extension of Big Five/FFM
In a way, Reynierse's model works better as an extension of the Five Factor Model/Big Five rather than as a reworking of MBTI itself. First of all, Reynierse tweaks the preference definitions to better match those of the Big Five, so that Feeling/Thinking becomes more like the Big Five's "Agreeable". So clearly Reyneirse is trying to shift things towards the Big Five definitionally.
Secondly, this model does add an element I think the Big Five traditionally lacks: a sense of the combinatorial influence of traits. Most Big Five based "type" descriptions just provide a bucket of description of individual traits, without providing any sense of how they interact. Reyneirse's model addresses that (at least for dyads and triads). It also makes sense that particular preferences, when strong, have bearings on particular situations. This is something that the MBTI does not capture particularly well (except when it defines individual preferences, the 8 functions, or the temperaments).
At any rate, it is kind of fun as a more statistically rigorous model that also provides a way think about preference strength as making a real difference (sort of like the light-hearted "The 24 Types of INTP" thread on intp central).
For the tl;dr crowd: an empiricist thinks MBTI type dynamics are bunk, that pairs and triplets of preference are where it's at, and proposes a less bunk-filled approach involving type codes ordered by strength of preference.
So, what are your ordered preferences?