“Modeling, it should be clear, is an art form. It depends on the experience and taste of the modeler. In this it is much like cartooning, especially political cartooning. The modeler (cartoonist) must decide which features to make salient (exaggerate), and which features to eliminate (avoid), in order to answer the questions (make the political point)” —John Holland, Hidden Order.
Are models in science better for prediction or understanding, and why?
A model approach in science is simply a necessity. A natural phenomenon in all its diversity is most often not amenable to analysis. Therefore, the object is replaced by a model that neglects some characteristics of the object. In principle, this is not a problem. The problem arises when the solution for the model is unreasonably applied to other objects, and crazy conclusions are made. The art of creating of model most often is not to reflect all the essential features of the object, but to obtain a solvable model ...
That experience is unlike any other I could describe.This is the best that can happen to a scientist (s): the moment of realizing that something that has happened in his (her) mind exactly matches what is happening in nature.When this happens to you, you are amazed. One is surprised that one construct of his mind is truly being realized out there, in the most serious, God's universe.Great is the shock, and great, great joy. Leo Kadanoff (James Gleick, Chaos)
Dragoljub Šarović
I call such situations "the victory of science over reason".
Dear Dragoljub Šarović
Interesting question. Models can arise using different approaches. They can be based on: mental contemplation (Western philosophical approach), on our physical senses or extensions of them, including the observation of behavior (scientific approach), or they arise from our inner experiences giving us a direct insight into our nature and the nature of reality. The last leads to understanding. Science is always statistical hence it is predictive (regardless of the model), and Western philosophy is always speculative. Maybe it is time we combined approaches to get a deeper understanding of reality.
Warm regards Tina
Dear Tina Lindhard ,
Thanks for the inspirational and helpful answer.
Kind regards
Dear Dragoljub Šarović ,
We human always use internal picture - model of Universe to plan actions and predict future. The value of model is measured by its ability to incorporate all known facts. The models are not static, but evolve with knowledge.
Best regards,
Konstantin
To Mr.Konstantin M. Golubev
Of course, I agree.There is no "best" model because there are different appropriate modeled for different contexts and different purposes. Asked question relates to the usefulness of the model in forecasting and understanding.
In my opinion, models are generally better for understanding than for prediction.
Most (if not all) models used in engineering practice incorporates assumptions in order to simplify complex theoretical material behaviour trends. The value of models developed for general application is based on the understanding of these assumptions and hence, the intrinsic limitations associated with these assumptions. A true understanding of the assumptions built into a model allows the experienced user to make adjustments in order to accurately simulate behaviour patterns. The danger is when such models are perceived and used without this knowledge. The model should always be evaluated according to a simplistic quote: "If your theory does not explain what the universe looks like, you better change your theory because the universe is not going to change".
Models and what they simulate are part of the universe.
To Gerrit Jordaan ,
Models are always an approximation of reality. Just this deviation from reality, significantly reduces the ability of models to predict.
- Badges
- Science topic
- Similar topics
- Engineering
- Modeling