The Problem of Individuation
By Malcolm Chisholm on January 22, 2011
If you deal with master data management (MDM) and data modeling for long enough you start to ask yourself questions, and, if you search diligently enough, you can sometimes find that non-data modelers have asked themselves the same questions. Maybe these other groups have created some nuggets of wisdom that we data modelers can use? I think it's worth a look. In what comes below I am relying heavily on a paper by Peter King [Theoria 66 (2000), 159-189].
MDM is about managing data for individuals. Often we say MDM is about "things", but trying to define what "thing" means is difficult, and we will leave that for now. Sometimes we say "instance" instead of "individual", but I will use "individual" here.
The Scholastic philosophers of the Middle Ages thought their way through highly abstract problems. They were largely dismissed after the triumph of the natural sciences following the Renaissance produced enormous amounts of new knowledge. However, it is interesting to look at some of the Scholastics' works, and one of the problems they set themselves was to explain the "individuality" that is found in the individual - generally called the Problem of Individuation. They broke it down into five main questions:
- What makes something the thing it is?
- What makes something the kind of thing it is?
- What makes something the same as others of the same kind?
- What makes something different from others of the same kind?
- What makes something different from others of different kinds?
I don't know about you, but I find these questions are very clear statements of part of what may be going on in my rather fuzzy mental processes when I am doing data modeling. In fact, I can't wait to try them out on some unsuspecting consultant in my next data model review (just kidding...maybe).
There is an allied question:
- What makes this thing the same as a thing I have encountered before or will in the future
However, this is more to do with the Problem of Identity than with the Problem of Individuation, so we will leave it for now.
So the above set of questions is a nice tool to have. Did the Scholastics answer them? They thought that the answers to (2), (3), and (5) lay with form. Form is the type of something. In a data model it corresponds to entities - which should more properly be called "entity types". So for the data modeler to answer questions (2), (3), and (5), he or she must have a sufficient definition of entity types.
For questions (1) and (4) the Scholastics could answer at the level of concepts. Logical division - what is loosely termed subtyping today - allowed general concepts to be subdivided into specific ones. An individual concept could be defined by using the formula "Species = Genus + Differentiae". There is no room to go into that more here, so let's just say the Scholastics generally agreed on an answer. But they could not figure out how to do this for individuals. What made Socrates Socrates? No final answer. Interestingly, the Scholastics did recognize three sub-questions for question 1:
1a. What makes a given individual an individual?
1b. What makes a given individual this very individual?
1c. What makes an individual to be one?
For the Scholastics, 1a was the core question in the Problem of Individuation. They could not solve it. Over the centuries they came up with various explanations, only to have them shot down. But I think that what they wrote can still be mined for approaches to data modeling in MDM.
One issue I often encounter in MDM that may be connected with question 1a involves trying to understand when is something an individual. In a Customer table for a retail store, I know that each record will correspond to an actual individual. But in the Product entity, I understand that each record will correspond to a Product Type. E.g. Product ID "X123" may be for a particular kind of teapot, but it is not for an actual individual teapot in the same way that any record in the Customer table is for a Customer. In other words, the Customer table stores data about individuals, whereas the Product table stores data about concepts. Does it matter? I am not sure, but I think it might. And at least I can pose a number of questions a bit more clearly and distinctly thanks to knowing a little of the work of the medieval philosophers of Europe and the Middle East.
Malcolm Chisholm, Ph.D. has over 25 years of experience in enterprise information management and data management and has worked in a wide range of sectors. He specializes in setting up and developing enterprise information management units, master data management, and business rules.
There have been no comments yet.