A Theory of Pattern

We often think of "pattern" as something drawn out of observations.  

• We experience something, and then we abstract from it some regularity or description, and call our expression of this abstraction a pattern.
• These ad hoc expressions of "pattern" may be formalized with logic or mathematics, or they may remain intuitions or natural language, but regardless, there is no framework to understand coherently and uniformly what the pattern itself is.

In pattern language, we take the opposite of this observational approach to "pattern".  

• We begin with a pattern whose structure and operation is entirely established.
• Then we project that pattern onto our experience to see if there is a match between our experience and the pattern's expression.
• Because we started with the pattern framework, we retain a full understanding of what the pattern means.  
• All that remains to demonstrate is whether the pattern is expressive of our experience.

 This alternative meaning of "pattern" and approach to modeling and understanding experience, is important because a uniform and coherent expression in pattern is always computable.

Examples of automating the observational approach to pattern include neural networks, machine learning and large language models.  While these technologies are clearly powerful and of proven utility, they are not explainable.

The pattern language approach has explainable meaning in every aspect of its structure and operation.  It is an entirely different route to computable and explainable intelligence.