While we often think of type as just a single model-a single way of looking at human interactions-it is actually quite rich because the whole of "type" is a cluster of related compatible models: 16 whole types, functions in their attitudes, type development, temperament, interaction styles and so on. They are part of the "type family."

The multiple models within type theory are not reducible to a single model. They are compatible but distinct and complementary without contradicting each other.

• 4 temperament with 4 interaction styles = 16 types
• Pairs of 8 functions in their attitudes = 16 types
• 4 MBTI scales * polar opposites = 16 types

Additional models, like type development and the "things in common" across the temperaments, are also used in a compatible way. All these models "fit together" or "interlock."

Why not use just one model? Isn't there a perfect model out there? We must, and automatically do, use multiple models. Each model is a different kind of model or looks at a different level of personality or from a different perspective.

• temperament = core needs and values of a type (whole patterns model)
• functions = cognitive processes of a type (parts and processes model)
• whole type = whole type themes (unique themes model)

Furthermore, because a model is by definition a generalization, additional models are needed to capture mechanisms that lead to apparent exceptions and contradictions.

We should strive for completeness and consistency in how we describe our models (portraits, self-portraits, descriptions, etc.) in order for our models to be useful to everyone.

None of the individual models by themselves is the whole picture. The notion of type by itself doesn't describe everything.

Do not try to compress different types of information into one model. No single model can be a universal model, explaining everything. And while diversity of models is important, be careful about mixing and matching.

Good models are not arbitrary; neither are the relationships between different models that fit together. The very nature of things in the universe mandates certain relationships. This is seen in all of nature. And because of natural laws, some things are possible, others are not possible, and some things actually exist, while others do not. Examples:

• The 8 mental processes do not pair arbitrarily, but pair to meet developmental demands - at least one for the inner world and one for the outer world, and at least one for perception and one for judgment.
• The relationships between the theories of temperament and type: each temperament tends to draw on certain functions to fulfill its needs, and there are 4 types that come out of (or match up with) each temperament.

Thus, models interlock in a certain way, naturally. (This is a consequence of the systems notion of "structure.") If the theory of how two models relate says one thing, and the apparent result is something different, a misalignment, then there is error or aberration of some kind.

When we go out and apply a model, we gather and use information. This introduces conflict, ambiguity, and other kinds of "measurement error." Error is inevitable. "Triangulation" is a way around error.

• A surveyor, hiker or navigator can pinpoint a location by using at least three landmarks or positions and "triangulating" the true position from those.
• In family systems theory, triangulation refers to three people in a relationship. As an analogy to using multiple models, a third person introduced to a system may be needed to resolve conflict between two people.

Sometimes there appears to be a conflict between two otherwise compatible models. One model suggests A, another B. The conflict is actually lack of adequate information. Introducing a third model can resolve and clarify. Thus, a body of theories must always be made up of at least 3 models, to escape this bind of conflict and error.

Exceptions also seem to occur because a model has been applied too generally or in an incorrect way, or has been described inadequately, or possibly is incomplete. Remember, however, that at least 5% error is absolutely inevitable!

Models should be objective - usable by different people under different circumstances with similar results. At the same time, remember that each model is a mental measuring device - it is a filter/lens etc. for interpreting the real world that we cannot perceive directly. It is a map and not the world itself. And the act of measuring causes interference.

Good models in science both explain (retrospective) and predict (predictive), and when a model does both of these then it is called a "control device."

• Many people use type as an explanatory tool.
• A trained facilitator can use type as a tool to actively intervene and guide a situation (the model becomes a so-called control device.)

Practice using type to predict - this is how we learn if our explanations are correct, because human beings are a little too good at explaining practically anything in retrospect.

The manner of measurement also varies with the model. A model about processes, for example, requires a different manner of measurement than a pattern-based model.

Try to explain "how" something happened or predict "how" it will happen, not "why." How means step-by-step, or in more detail, or some kind of causal, behavioral or relationship picture. Why's often end up as blanket statements. Compare these:

• He did it because he's an ENFP and ENFPs are like that.
• He did it because he has a preference for Intuiting.
• He did it because his Idealist temperament drew on the Intuiting function to meet his Idealist needs in this situation, and the ENFP whole-type theme of "life is like a story" explained it for him in his situation.
• He did it because he wanted to marry his girlfriend and she didn't want to marry yet.
• He did it because he's a typical impulsive teenager.
• He did it because that's the way the world is.

Example six is a universal model, and is useless. The fourth and fifth reasons appear to be common reasons people give that are mostly unrelated to type (but why does he want to marry her? And why is he impulsive? There may be type-development explanations.) The third example is the most detailed and is a good example of using multiple models. The first two are common uses of type that rely only on a single model.

Models cannot change themselves. However, Jung once talked about the idea that someday humans could evolve new mental functions as-yet undreamed of. But type contains no way for us to predict this. However, just as introducing a third model can resolve an apparent conflict between two other models, so too can one model suggest opportunities for clarification in another model. For example, temperament has helped us flesh out the difference between sensing in its introverted and extraverted attitudes. And there are aspects of personality, such as body type and personality, that may one day be brought "into the fold" to complement type. Oh, and a model that can change itself is a living system!

Living next-door to the type family is several other models in various stages of development.

• The Enneagram does have themes, parts and processes in a greater or lesser extent, but people who espouse the Enneagram cannot seem to decide what it is modeling. (Virtues and vices? Personalities? Emotions? etc.) If it turns out to be about virtues and emotions, then it may add to type. The Enneagram's triangular nature may also spark ideas about personal development.
• The 5-Factor model is an attempt to capture in 1 model the qualities of numerous other models: temperament, type, functions, development, emotional state all in one. However, the 5-Factor model cannot be used in all the ways the multiple models of type are used, because by trying to be a more universal model, it loses the quality of having different perspectives and having different kinds of models (themes, processes, parts, wholes, etc.) Usefulness is lost.
• We do not go beyond 16 types to subtypes, because the rough models we have beyond the 16 types level (gender, culture, etc.) are not consistent or complete, and there are too many exceptions, overlaps, etc. between them at this time.