Weighted Definition

For complex systems, concepts, or tools, it is often very difficult to form a discrete definition along the lines of "must have A and B and one or more of C, D, or E". Rather than try to force a "Boolean" definition, some kind of weighted definition may be more appropriate and better reflect the "organic" real world. Typically there would be a minimum requirements (prerequisites) and then a weighted list of common features. The score has to be above a certain amount to qualify, in addition to meeting the prerequisites. There's an example in DefinitionOfLife that I provided near the middle of the topic. Also, the cut-off point is optional, for it may be a tool to measure degree of X-ness where X is the definition being sought. --top

What purpose has this?

Like it says above, it may better reflect the "organic" real world. Humans tend not to classify things the same way a computer program or "simple" AI system might. Definitions are closely tied to human psychology for good or bad, and this weighted approach better reflects how humans classify things in my observation. We tend to use something that resembles FuzzyLogic.

What purpose is there going to the effort of creating a WeightedDefinition? Good academic and scientific work will clearly establish working definitions rather than rely on assumed definitions, let alone debate over them. (Except in some fields like biology, which are known for taxonomic debates, but they're obviously a special case.) Otherwise, this seems like a rather tedious effort merely to support what are (here, at least) the textual equivalent of water-cooler debates.

First, definitions are generally defined by usage, not by "expert panels". Second, academia has had a poor track record of producing precise and clear definitions. Often they are form of, "if you reformat or cram the thing into this notation/model, then we can define it as...." without much justification or testing of the reformatting process. Plus, it then carries the extra baggage of the container model, which may require 500 pages of dry reading to make sense of. The reader then has 2 models to figure out instead of just the one for the definition itself. And finally, many definitions depend on the observers' interpretation of things, and WetWare is an area that "pure" science wants to avoid. -t

Some examples of WeightedDefinitions:

CategoryDefinition, CategoryAbstraction

View edit of September 29, 2010 or FindPage with title or text search