I can out-think ya, I can out-fight ya, I can out-philosophize ya. -- Max Cady (Robert de Niro), "Cape Fear" (1991)
A term used to describe designers obsessed with progressively higher level Meta Data. One level of Meta Data is expected. Two levels is excessive. Three levels and Analysis Paralysis can be declared. Also used as a term of ridicule for philosophers who have lost any foundation by overly-aggressively abstracting.
Cognitive metaphor (Meta Thinking):
Level 1: Thinking about concrete items
Level 2: Thinking about thinking
Level 3: Thinking about thinking about thinking
This way of thought also leads to Too Much Abstraction.
Level 1 is science. Level 2 is philosophy. Level 3 is BS, or Mental Masturbation.
Level 2 is BS as well. Disagreed; philosophy is a useful intellectual endeavor. Though admittedly it can be overdone.
Philosophy is BS only if you are stuck on Level 1. You confuse the symbol with the academic department.
In college I had a professor who explained that level 3, 4, 5, 6, and so on are all just level 2, because they're all thinking about thinking. Arguing along these same lines, if you have a generic way of managing metadata, you should be able to use the same mechanism for handling data about data about data as you use for handling data about data, since after all, it doesn't really matter whether the data you have data about is metadata or not.
In the same manner, Smalltalk has class Class and class Meta Class. It does not, however, have class Meta Meta Class, class Meta Meta Meta Class, and so on....
Abstraction is whether one can perceive a view which is "about" another. It is essential to learning and improvement and includes the concept of Yes and No.
From the socio-cognitive point of view (Adam M. Gadomski, 1994), in the TOGA (Top-down Object-based Goal-oriented Approach) model of abstract intelligent agent (AIA), 3 basic meta-levels of agent's reasoning/thinking are distinguished:
Level 1: Domain-dependent reasoning, according to scheme: Information --> Preferences (choice of goal) ---> Knowledge (acting)
Level 2: Reasoning about and changing its own domain-dependent preferences/goals or knowledge (learning)
Level 3: Reasoning about how domain-dependent preferences can be changed (using moral criteria), and about agent's learning methods and their modifications (teaching).
According to this cognitive structure, usually, in concrete situation, AIA increases its level of thinking if its reasoning on the lower meta-level is not successful. page 
See original on c2.com