June 25th, 2020
Many years ago I looked carefully at two formal personality models from the world of psychology (OCEAN and MBTI), and rejected them as overly concerned with the more mundane aspects of personality. Occasionally over the years, in response to the prodding of others, I have reconsidered them and rejected them. However, my rejections were always based on contemplation and intuition. In this essay, I propose to attempt a more detailed analysis of three systems:
1. The 3-dimension model I have long preferred, consisting of three dimensions Bad_Good, False_Honest, Timid_Dominant, as well as their perceived values, their circumferential values, and their accordance values.
2. The Five Factor model, also known as the OCEAN model, consisting of five dimensions: Openness to Experience, Conscientiousness, Extraversion, Agreeableness, and Neuroticism. These could be augmented with perceived values and accordance values.
3. The Myers-Briggs Type Indicators, consisting of four dimensions: Extraversion_Introversion, Sensing_Intuition, Thinking_Feeling, and Judging_Perceiving. Again, these could be augmented with perceived values and accordance values.
Let’s assume that all three systems possess adequate orthogonality. The true test of a system is, in my opinion, its ease of application in creating algorithms. So I set up a few test cases and see how the systems would be used.
Test Case 1
The trigger event is: "Mary tells Jane that she (Mary) likes Tom by positive amount X.” Jane’s options are:
1. “I also like Tom by positive amount X.” (Agreement)
2. “I do not like Tom by negative amount Y.” (Disagreement)
3. Evade the subject of Tom.
Test Case 2:
The trigger event is “Joe insults Fred.” Fred’s options are:
1. Ignore the insult.
2. Retaliate verbally.
3. Retaliate physically.
As soon as I started working on inclination equations for these test cases, I ran into the killer problem, a problem so obvious that I kick myself for not noticing it earlier.
Neither the OCEAN model nor the MBTI model permit emotional relationships. In my model, for example, the internal attribute Bad_Good has a corresponding perceived value pBad_Good that specifies how “Bad_Good” one person perceives the other to be—which is a good substitute for how much the perceiver likes the perceived. In the same manner, pFalse_Honest corresponds to trust.
But what would pNeuroticism denote? Or pSensing_Intuition? None of the personality attributes used in either of these models translate well into any kind of relationship, and they certainly don’t permit any way to express two of the most important human relationships: affection and trust.
So we’re stuck with the Crawford 3-dimensional model. I’ve never been particularly enamored of it—there are a number of flaws that I just can’t correct. But it remains the best we’ve got.
