matt (hbergeronx) wrote,

  • Mood:

on modeling

The only useful thing that my thesis adviser ever said to me was as follows- "if you make a model with enough adjustible parameters, you can get it to fit any situation. The point of modeling is not to convey inner truths or to make grand theories. Simple models which don't have any basis in reality and are obviously wrong from an ab-initio standpoint are still very useful if they have predictive power."

I've noticed a lot of people I've been spending time online around using tests and assesments to determine what type they are or sorting their temperament. I wanted to write down my feelings on these things. First of all, I don't buy into assessment because it has little in the way of predictive power. Knowing that I'm a type 5 does not tell you something about how I make or assess things internally, and it doesn't tell you how I'm going to handle situations. Some might argue the very fact that I disagree analytically with being categorized as a type 5 proves how type 5 I really am. That's fine, but in order to relate to me on that level, you need to explain to me why this model is wrong from an ab-initio standpoint and how you intend to use the model to have predictive power. One of the key personal principles I've betrayed for the last eight years or so is a failure to begin from first principles.
If there's anything that quantum mechanics has taught me well, it's that models are always inherently wrong, and first principles are a simplicity that belies complexity. The uncertainty principle, in physics, is not about things are uncertain. It says that all knowledge is a tradeoff- for particles, it is the tradeoff between knowledge of position and velocity. It is not that you can't know things- you just have to understand and accept what you trade off as the price of knowledge. We use tools to observe systems, and tools have impact on the systems we study.
I will also mercilessly attack anyone who uses quantum theory, or indeed, any system of study such as language analysis or numerology or whatever if I perceive the slightest lack of rigor or care. I need to do something about carefully destroying the Pirsigs and the Robert Anton Wilsons of the universe, especially my own. Destroy all systems! Destroy all nations! Destroy! Kill! Kill! Kill!

  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic

    Your reply will be screened

    Your IP address will be recorded