Saturday, July 21, 2012

Well, I had another day or so of thinking about the conceptual model I was developing - originally for stipulative and legislative definitions.  Actually, a few minutes rather than a whole day was what I had, but such is life when you have a job.  But even in that limited time I realized that I had not got the idea of communities in the model. 

So I went back to the conceptual model and put in communities as shown in Figure 1.

Figure 1: Relations of Concept and Term

Speech Community: a group of individuals who share a vocabulary that describes a concept system.  [See Wikipedia for more - - although it seems you have to belong to the Speech Community Speech Community to understand this entry.]

Semantic Community: a group of individuals who share a common understanding of a set of concepts and relationships (a concept system), irrespective of the terms used to describe them.
I understand that both of these definitions are preliminary and need more work.

The Interpretant must belong to a Speech Community in order to use the term in communication.  If the Interpretant is a member of such a community, they will be able (or will have a way to be able)to recognize the sign as signifying a concept.  Recognition is the important element here.  It is, of course, possible that the Interpretant may not recognize a particular term that is used within the Speech Community (e.g. if the interpretant is new to the community and still learning its vocabulary).  However, the Interpretant is in communication with at least some other members of the Speech Community, and has the opportunity to find out from them what the term means to them  - how they recognize it.

Recognition seems to be very little discussed in the semantics literature (which I am not well versed in, so I can easily be wrong).  However, my background is biology  - and recognition is perhaps the most fundamental concept in biology.  The standard analogy here is that of a lock and a key.  One biologically active molecule, e.g. an antigen, is recognized by another biologically active molecule, e.g. an antibody.  The antibody meets the antigen and neutralizes it.  This is recapitulated in various ways across the entire science. 

Back in our diagram.  The key is the term and the lock is the recognition process (part of signification) that identifies the concept to the interpretant.  What is important is that recognition can be learned among the members of a Speech Community.

New Types of Existing Term

I have also done some more work on Existing Term.  An Existing Term can be differentiated into the following subordinate genera:

Known Term: a term that is in use in a particular Speech Community

Unknown Term: a term that is not in use in a particular Speech Community

What is interesting here is that a New Term is not in use in any Speech Community.  But how would you know if a given term, T, is not in use in any Speech Community?  You could not unless you knew about all speech communities.  Obviously, this is impossible, or at least unimaginably difficult.  It would seem therefore that a given interpretant could not distinguish a New Term from an Unknown Term (an Unknown Term relative to the Speech Communities that the interpretant belongs to).  This has important practical implications.

New Types of Previously Existing Concept

I have also differentiated Previously Existing Concept into Known Concept and Unknown Concept.

Known Concept: a Previously Existing Concept that is known to a Semantic Community

Unknown Concept: a Previously Existing Concept that is unknown to a Semantic Community

An example here might be "quasar" (derived from "quasi-stellar object").  Quasars existed long before astronomers identified them.   So at one time they were an Unknown Concept, that then became a Known Concept within the Semantic Community of astronomers.

Contrast this with Mortgage-Backed Security (MBS).  No MBS existed at all prior to the 1960's [I think this date is accurate, but I am not a historian of financial theory, so maybe the 1930's is a safer decade to use].  Neither instances of MBS, not the theoretical (uninstantiated) concept existed prior to this time.  Then the concept came into being, and shortly thereafter came instances of the concept.  Prior to this MBS was a Not Previously Existing Concept.

So New Concept, Known Concept, and  Not Previously Existing Concept  are all relative (involve a relation with) at least one Semantic Community. 

Practical Implications

The above still needs a good deal of work, but it is possible to see the outlines of practical implications.  I am very interested in the work of a Business Analyst (BA) and a Data Analyst (DA - which I will subsume within BA for the moment).  Here are some brief thoughts.

  • On a global project, say a Customer Reference Data project, a BA needs to know if the Speech Communities he is dealing with are actually one Semantic Community.  The presumption is likely to be that Customer Reference Data is covered by one Semantic Community, but this is not necessarily the case.  There may be differences in concepts and relations associated with some of the Speech Communities.  I think the best way to figure this out is to construct a conceptual model for each Speech Community using their terms for concepts, definitions, and relations - and then to compare the conceptual models.
  • A BA would be unwise to think they are a member of a Speech Community in the area they are analyzing, if they are new to the area.  Even if the BA has dealt with the same people before, or the BA thinks they are a member of the related Semantic Community, there is no guarantee they are a member of the Speech Community.  Each concept system has its own vocabulary for the Speech Community, and strict terminological analysis is required.  I think too many BA's make presuppositions about this kind of thing and get into trouble.  

I think there are more practical implications, but blogging demands brevity, and this post is already too long.

1 comment:

  1. Philosophically inclined people wanting to separate concepts from terms typically base or relate their thinking to the Semiotic Triangle or Pierce’s Triangle ideas. Sometimes also called a triangle of meaning, it distinguishes a concept that exists in a human mind (a thought) from how it is referred to and a symbol that evokes it. A symbol is typically explained as a pictorial depiction. The important aspect of this theory is that it is entirely about human cognition. There can be no name or identity for a concept as it only exists as a thought in a human mind.

    The challenge to applying this thinking to information modeling is that in the information modeling we must ultimately commit everything to (electronic) paper. Thus, every concept must have an identity and a name. As a result, a separate model for concepts and term where terms themselves have relationships, etc. can be an over-complication that does not deliver practical value. SKOS takes a simpler and, what I believe to be, a more practical approach where each concept has a globally unique identity, a preferred label that is unique for a given human language (such as English, German, etc.) within a scope of a particular “concept scheme”, any number of alternative labels and whatever other attributes and properties that are deemed necessary. A metadata about labels can be captured without giving them identity of their own. Besides the language part, such metadata is typically not just about the label itself, but about its relationship to the concept i.e., who said that this is a preferred label for this concept and when. All the relationships are between concepts, not between the labels.