The Problem of Pluto: What Is being Defined?

I wanted to return to the issue of Pluto, which has already been the subject of a number of posts.  The International Astronomical Union (IAU) created a rich array of issues and problems when it undertook a definitional change that resulted in the demotion of Pluto to the class of "dwarf planets".

The topic this time is what exactly did the IAU define?

I was watching a PBS special on the status of Pluto a few days ago.  It included scenes from a diner where the genial Neil deGrasse Tyson was asking customers what they thought about the new status of Pluto.  The reponses varied, but the issue at hand was about whether Pluto was "a planet".  The diners all thought that they were dealing with the general concept signfied by the term "planet".  Yet there is reason to think they were mistaken.

The IAU resolved (see http://www.iau.org/public_press/news/detail/iau0603/) concerning the following:

"The IAU therefore resolves that planets and other bodies in our Solar System, except satellites, be defined into three distinct categories in the following way:"

So what is being defined? Answer: "planets and other bodies in our Solar System, except satellites". 

Not planets in general.  But wait a moment - on the web page referred to, it also says "Resolution 5A is the principal definition for the IAU usage of 'planet' and related terms."  Yet this is not part of the text of Resolution 5A.  It seems to be some extraneous comment of uncertain provenance.  It certainly appears to be in conflict with the text of Resolution 5A, which, again, is only dealing with the situation in the Solar System.

So we have: a lot of people thinking that the IAU defined "planet"; and the text of Resolution 5A which is defining "planets and other bodies in our Solar System, except satellites"; and a statement on the IAU web site saying that Resolution 5A is to be used for planets in general. 

This is contradictory.  The definition is for a "planet in the Solar System" but somehow can be used for a planet not in the Solar System also.  In other words, we can substitute the definition for both A and Not-A. 

Let's try that with the proposition about one of the extrasolar planets:

"51 Pegasi b is a planet that orbits the star 51 Pegasi". 

Substituting the definition presented in Resolution 5A for the term "planet" we get:

"51 Pegasi b is a celestial body that (a) is in orbit around the Sun, (b) has sufficient mass for its self-gravity to overcome rigid body forces so that it assumes a hydrostatic equilibrium (nearly round) shape, and (c) has cleared the neighbourhood around its orbit that orbits the star 51 Pegasi."

So we have a contradiction.  51 Pegasi b apparently orbits both the Sun and 51 Pegasi.

This contradiction arises from the IAU restricting the definition of "planet" to the Solar System, but pretending that it can be used for any planet.  It also shows how Natural Science is dependent on Logic, which is part of Philosophy.  But that is a far more controversial topic.

Three Classes of Identification in a Definition

Stijn commented on my earlier post "What is an Identifying Characteristic?" (http://definitionsinsemantics.blogspot.com/2011/11/what-is-identifying-characteristic.html) raising the point that "identification of a thing is dependent on the application".  He lists out things that identify him, and notes that one cannot always be substituted for another.  E.g. a passport cannot always be substituted for a driving license.  It depends on the application, and each application has rules about what can be used as identification.  Stijn asserts that trying to capture all such rules in a definition will create conflict between the parties representing the applications.  So he advises us to separate a definition from capturing such rules.

There are a lot of topics compressed into this comment, so I am only going to pick one here.  It is the different classes of identification that should be captured in a definition.  I suggest that these are:

• Characteristics of the concept being defined that set it apart from other related concepts.  These are the classic specific differences (differentia)
• Characteristics that can be used to recognize an instance of the concept.  This is what I was trying to highlight in the original post when I stated that an exit row in an airplane could be recognized by a sign saying "no children in this row".  There is no reason for these characteristics to be specific differences.
• Characteristics that can be used to identify an instance of the concept.  This is what Stijn was talking about, saying his passport could be used to identify him.  Identifying an instance is not the same as identifying or recognizing a concept. 

So it turns out that identification is quite complex in definition work.  Simply talking about "identifying characteristics" as I did, does not take this richness into account.

The per-application rules that Stijn mentions add more complexity, but that will require another post.

Role versus Relationship - What Does it Mean for Definitions?

A couple of days ago I was reading some material on a semantics product and came upon the term "role".  We see role used in data modeling where a primary key migrated into a child entity can be assigned a "role name".  This is the name by which the attribute is known in the child entity, and is useful to disambiguate the same attribute migrated for other relationships between the same two entities.

You also hear about "role" in the party model.  Rather than say that Unindicted Broker is a client of Overleveraged Bank, and that Unindicted Broker is a prime broker for Overleveraged Bank, we can say that Unindicted Broker is a party that plays two roles with Overleveraged Bank: (a) client; and (b) prime broker.

I think that there are deeper issues here.  We think of a relationship in a data model as a line between two entities.  We cannot allocate attributes to the relationship as we can to entities.  Our notations, methodologies, and tools will not allows it (at least the commonly used ones).  Furthermore, it is relatively rare to find multiple relationships between the same entities.  When we do find quite different relationships between the same two entities we seem to start thinking of roles.

Now, a relationship is a concept, and therefore must have a description and hopefully a definition.  If a relationship can have a definition, it must have characteristics (qualities, i.e. attributes). This worries me somewhat as relation and quality are two Aristotelian categories and one should not be reducible to the other.  However, I cannot find the theoretical foundation for what I am describing.

A further issue is that what we are calling a relationship such as "Unindicted Broker is a client of Overleveraged Bank" is a generalization about a lot of processes.  Unindicted Broker had to be solicited to be a client, then onboarded as a client, and then assessed in terms of how the relationship would be managed.  All of these processes come under the umbrella of "Unindicted Broker is a client of Overleveraged Bank" but break down to many more detailed entities and relationships. So "Unindicted Broker is a client of Overleveraged Bank" is a generalization, although it is valid.

So where does this leave us?  Not very far I am afraid, but we can begin to see the outlines of the problem.  The term "role" is used in semantics, but it is not clear if it is used technically or informally.  "Role" exists in data modeling, but is for refining names of attributes associated with relations.  And "role" exists in the party model, for "high-level" relationships.  There is some evidence that relations can be broken down into more detailed entities and relations that may serve to describe a role.  Ultimately it does seem that roles can be resolved into sets of entities and relationships at the data model level.  However, at the level of semantics it is not clear how they can be treated as other than relationships.

A role does seem to demand a definition that is greater than what is to be supplied for a "regular" relationship.  A role must be distinct from other roles, or you could argue it should be collapsed with its sibling roles into one role. So at least we can conclude that if we have identified a role we need to provide it with a good enough definition to provide such distinction.  Obviously, there is a lot more to this, but that is enough for now. 

Is A Data Model An Abstraction?

Rob brings up a good point in his comment on The Problem of Abstraction in Definitions of Data (http://definitionsinsemantics.blogspot.com/2011/12/problem-of-abstraction-in-definitions.html).  He notes that what I am describing is not really abstraction but really a number of different things.
Today it seems the term "abstraction" is used in all kinds of situations when talking about data.  For me, it is often difficult to figure out what "abstraction" is supposed to mean in any one of these situations.  I strongly suspect that at least sometimes it does not really mean anything.  Sometimes I suspect it is even used for marketing hype.

The entry for "abstraction" in Baldwin's Dictionary of Philosophy and Psychology describes how abstraction is filtering out of attributes from an instance or a concept to achieve a particular view of the instance or concept.  Rather poetically the entry describes how a child looks at a body of water and becomes fascinated by the lustre caused by the play of sunlight on the surface of the water, to the exclusion of all the other qualities (attibutes) of the water.

This traditional understanding of abstraction as creating a view by filtering out attributes can be used in a special way to create the generalization hierarchies of genus and species (a.k.a. supertype and subtype, or general concept and specific concept).  The particular attributes of a group of specific concepts are left behind and attributes that the concepts have in common remain.  These are used to form the general concepts that include the specific concepts. 

However, abstraction as filtering out of attibutes can generate other perspectives.  Abstraction does not always have to lead to the traditional generalization hierarchy.  I can understand a man's watch as a timepiece, or a piece of jewelery, or as a fashion accessory.

Now, Rob is right in that I was not using "abstraction" in the above senses.  However, I do not have a better term to use for what I was trying to describe.  The main idea I was trying to get across is that one concept system can describe or specify another - such as how a  data model describes a physical database.  The relationship of "description" here is different to every other kind of relationship because the concepts present in the concept system being described have to have some kind of presence in the concept system doing the the describing.  This is not the same class of relationship we see in e.g. "I own a car".

So we somehow have the presence of a concept being described (e.g. a column of a physical database table) in a concept system doing the describing (e.g. an attribute of an entity type in a data model).
Rob terms this "representation" (If I understand his comment correctly).  This has to be right.  However, a representation can often be a picture - a mere image.  Technically, this is called a "phantasm" because it does not have the attributes differentiated from the whole.  Unfortunately, the process of recognizing and separating the attributes from a phantasm is also called abstraction.  It gets more complicated.  we cannot take a photograph of a physical database and produce anything like a data model.  A database has to be conceieved, not imagined.

Obviously, we are getting into a whole lot of other issues here.  I cannot really defend myself against Rob's criticism of my having overloaded (or over-abstracted?) the term "abstraction".  However, I do not have a commonly accepted set of terms that I can use to convey the idea of one concept system describing another.  More of an excuse than a reason, but it will have to do for now.

On Levels of Definitions and the Semantic Web

Stijn made a couple of sharp points in a comment on the post How is a Definition Different from an Explanation? (Part 1) (http://definitionsinsemantics.blogspot.com/2011/11/how-is-definition-different-from.html). 

He notes that there is a need for definitions to be short in certain circumstances, as when a user is scanning through a list.  I think this is a good point.  Users may be more in search mode when they are doing something like this.  They want to know if the definition is close to some target they have in mind.  Obviously, a full definition is not fit for such a purpose.

So we might have three levels of definition: (a) a one-liner, suitable for lists; (b) a one-paragraph, suitable for a quick read with some detail - and display on a screen with scarce real estate; and (c) the full definition, as an authoritative reference.  I have no problem with the last one being very long and including pictures - certainly more than half a page.   

The second point Stijn makes is about the Semantic Web, where there is a need for a "general purpose" description.  I agree with Stijn that there is a problem here.  The definition of a concept must include something about the concept system the concept is located in - such as relationships to proximate concepts. So if one concept can be placed in different concept systems, then the definition will change.  A mortgage loan in a servicing system is not the same as in a loan origination system, and is not the same as in a securitization system.  The Semantic Web may have an unspoken assumption of a single model of reality.  This will cause problems if, as I maintain, one concept can be placed in many concept systems.  It will lead to the frustration Stijn describes.  

Where I disagree with Stijn is his equating explanation with description.  These are quite distinct.  There can certainly be descriptive definitions, and these are very common in natural science.  But explanations are different.

George Orwell on The Advantages of Abbreviation

The issue of the emotive power of terms and definitions is a difficult one.  However, it is never far away, even in data management.  With the recent passing of Kim Jong Il, it seems appropriate to reflect on the political application of terminology.  Below is an excerpt taken from George Orwell's novel "1984" which can be found at the Newspeak Dictionary site www.newspeakdictionary.com.  It deals with how abbreviations can be constructed to achieve certain ends. Admittedly, these are terminological rather than definitional principles, but they are not without interest and certainly have their place. 

So far as it could be contrived, everything that had or might have political significance of any kind was fitted into the B vocabulary. The name of every organization, or body of people, or doctrine, or country, or institution, or public building, was invariably cut down into the familiar shape; that is, a single easily pronounced word with the smallest number of syllables that would preserve the original derivation. In the Ministry of Truth, for example, the Records Department, in which Winston Smith worked, was called Recdep, the Fiction Department was called Ficdep, the Teleprogrammes Department was called Teledep, and so on. This was not done solely with the object of saving time. Even in the early decades of the twentieth century, telescoped words and phrases had been one of the characteristic features of political language; and it had been noticed that the tendency to use abbreviations of this kind was most marked in totalitarian countries and totalitarian organizations. Examples were such words as Nazi, Gestapo, Comintern, Inprecorr, Agitprop. In the beginning the practice had been adopted as it were instinctively, but in Newspeak it was used with a conscious purpose. It was perceived that in thus abbreviating a name one narrowed and subtly altered its meaning, by cutting out most of the associations that would otherwise cling to it.

The words Communist International, for instance, call up a composite picture of universal human brotherhood, red flags, barricades, Karl Marx, and the Paris Commune. The word Comintern, on the other hand, suggests merely a tightly-knit organization and a well-defined body of doctrine. It refers to something almost as easily recognized, and as limited in purpose, as a chair or a table. Comintern is a word that can be uttered almost without taking thought, whereas Communist International is a phrase over which one is obliged to linger at least momentarily. In the same way, the associations called up by a word like Minitrue are fewer and more controllable than those called up by Ministry of Truth. This accounted not only for the habit of abbreviating whenever possible, but also for the almost exaggerated care that was taken to make every word easily pronounceable.

In Newspeak, euphony outweighed every consideration other than exactitude of meaning. Regularity of grammar was always sacrificed to it when it seemed necessary. And rightly so, since what was required, above all for political purposes, was short clipped words of unmistakable meaning which could be uttered rapidly and which roused the minimum of echoes in the speaker's mind. 

When it comes to Newspeak, my favorite word is Prolefeed, which is defined as Rubbishy "entertainment" and spurious news which The Party hands out to the masses.  But I'm going to have to cut this post short tonight - it's time for Dancing With The Stars.

The Problem of Abstraction in Definitions of Data Objects

I think there is a major problem in not being able to understand and work with different levels of abstraction.  By "abstraction" in this sense I mean one concept system that somehow describes or defines (not merely relates to) another concept system.  I think this is a big problem for definitions in data models.

Let us take an example in a retail business such as mortgage banking: Customer Name.  Customer Name exists in the business.  They use it all the time.  Maybe it is sometimes called Borrower Name, but the concept is the same.  This is the Level 1 abstraction.

Now let us think of data values in a column in a table that holds Customer Name.  These data values are stored as a code of 1's and 0's.  Of course these bits are rendered into something we can read.  However, this is not the same as the Customer Name in the business.  I worked for a place where they prefixed the name of anyone who had recently left with "ZZZ".  So we could have "ZZZ_John Smith" as a data value, but the business would call him "John Smith" still.  The data value is the Level 2 abstraction.

Now let us think of the column itself that stores Customer Name, irrespective of whatever it contains.  This is the container used for the data.  It is merely a container, and anything can be put into it - just in the same way as the old peanut jelly jar I have on my desk is used to hold pens.  The column has certain characteristics, like the maximum length of text it can hold.  This is the Level 3 abstraction.

Now let us think of the data model that describes the column that will hold Customer Name.  In this, Customer Name is an attribute.  We worry about what naming convention to give it.  And behold!  Our data modeling tool asks us to enter a definition for Customer Name!  Yet, we are now at Level 4 of abstraction.

Let's summarize.  The concept system of the data model (Level 4) is a design for the concept system of the container of the data (Level 3) which will store the concept system of data values (Level 2) which we hope will satisfy the concept system of the information needs of our users (Level 1).

So tell me again what the definition entered in the data model is referring to?  Which of the four concept systems?.  Suppose it is stated as "an attribute that holds customer name" - I have seen this kind of thing quite often.  Well, an attribute is something in a data model (Level 4), and a thing that holds data is a container (Level 3).  

It would seem that the ideal thing would be to understand the Level 1 abstraction - the business information.  However, the chances of getting a good definition of this when you are at Level 4 would seem to be a challenge.  There are too many layers of abstraction in the way.  This, I think, is why semantics are so important.  They deal with business information as is, and do not have to worry about other concept systems.

Should A Technical Term Have Only One Definition?

There are far more concepts than there are terms to describe them.  This leads to the use of a single term to signify more than one concept.  Such terms are known as homonyms.  For instance, the term "table" is used in conversations in business about whether or not to discuss a topic.  Someone may say "Let's table that".  Unfortunately, some people think this means "Let's take that topic off the table", while others think it means "Let's put that on the table".  The differences in interpretation are geographic, with the British thinking it means one thing and the Americans another.  It makes for pretty interesting conference calls on transatlantic projects.  I have managed to forget which side thinks of it which way.

So homonyms exist, and we have to deal with them.  But what about technicial terms?  Technical terms are specific to very specialized domains.  It might be thought that the narrowness of the domain would itself guarantee that a technical term would have only one definition.  But there is no guarantee of this.  An example I often come across is "data model" in the realm of data management.  To some this means an artifact for the design of a database produced by utilizing a standard symbology.  To others it means the actual design of a physical database.  The first means an artifact produced by a tool like ERwin.  The second means the underlying design of an actual physical database, and certainly not a design artifact.

This is very confusing, and can cause a lot of problems in communication.  Working with technical language is very difficult to begin with.  Having technical terms with more than one definition makes things much worse.

If it is possible to set up a technical vocabulary, or to reform a technical vocabulary, then it is possible to ensure that one technical term has one definition.  This is part of the work that terminologists do.  

Yet, even if a technical vocabulary is set up by terminologists, they cannot control its usage.  If a technical term starts to have marketing value it will be used to signify things that may be the same as the original concept, or a different concept, or no concept at all.  We call this "hype".  If a general community cannot fully understand the concept signified by a technical term, they will use the term to mean something they think they understand, as in the example of "data model" above.

To answer the question originally posed, a technical term should only have one definition, at least within a particular technical domain.  However, we have no way to ensure a technical term will always signify the same concept.  

So what is the conclusion?  If we want to master technical terms we need to understand what they are intended to signify, and be alert to detecting when they are used to signify other concepts.  A big piece of this is having a glossary of terms with adequate techncial definitions.

From Vice to Virtue - The Changing Definition of "Sophistication"

Today, to be called "sophisticated" is considered a compliment.  In fact, it seems to be a virtue to be aspired to.  Here is an example from: http://www.iexaminer.org/editorial/definition-sophistication-obsolete: 

"...I resumed thinking. What exactly is sophistication? The things that we deem to epitomize sophistication—going to the symphony, ballet, dressing up, sipping fine champagne and delighting in witty conversation while daintily snacking on tiny foods, etc.—are they even relevant anymore?" 

This was not always so.  In the not too distant past, to be called sophisticated was to be insulted.  Here is the definition of "sophistication" taken from Samuel Johnson's Dictionary of the English Language(3rd edition, 1766) at http://openlibrary.org/books/OL7124798M/A_dictionary_of_the_English_language : 

SOPHISTICATION: Adulteration; not genuineness

Sophistication is the act of counterfeiting or adulterating any thing with what is not so good for the sake of unlawful gain...Quincy.

The drugs and simples sold in the shops generally are adulterated by the fraudulent avarice of the sellers, especially if the preciousness may make their sophistication very beneficial...Boyle 

There is no hint of anything good about sophistication in Johnson's entry.  Now contrast this with the definitions provided by Wordnet, the online dicitonary maintained by Princeton University.
(http://wordnetweb.princeton.edu)

S: (n) edification, sophistication (uplifting enlightenment)
S: (n) sophism, sophistry, sophistication (a deliberately invalid argument displaying ingenuity in reasoning in the hope of deceiving someone)
S: (n) sophistication (being expert or having knowledge of some technical subject) "understanding affine transformations requires considerable mathematical sophistication"
S: (n) sophistication, worldliness, mundaneness, mundanity (the quality or character of being intellectually sophisticated and worldly through cultivation or experience or disillusionment)
S: (n) sophistication (falsification by the use of sophistry; misleading by means of specious fallacies) "he practiced the art of sophistication upon reason" 

It seems that Wordnet is giving us both the old definition of sophistication as a vice, and the new sense, where it is a virtue.  However, I have only heard the term used as a compliment, except on a very few occasions.  I think that very few people know that it once had a negative meaning, even if they are familiar with related terms like "sophistry", which are still generally understood to be negative.

This is a pretty extreme example of how a term can come to signify a concept that is in direct opposition to the concept it originally signified.  I am not able to trace how this happened, but I am sure somebody knows.  Perhaps it is in the book whose appalingly written review can be found at http://www.timeshighereducation.co.uk/story.asp?storyCode=413279&sectioncode=26 

A lesson that can be derived here is that when analysts encounter a term used in a technical context, they should not expect it to signify the same concept as in common usage.  If "sophistication" can change in this extreme manner, so can anything.

Clear, Obscure, Distinct, and Confused Ideas and How They Relate to Definitions

I wanted to capture the meaning of these terms because they are very important in dealing with definitions.  The terms have a formal place in logic, and are often encountered in the traditional literature.  Yet it is also fair to say that we probably all use these terms (or their synonyms) quite frequently in analytical work.  Having a good idea of what they actually mean makes them, I think, more useful tools for us.

To get understandable definitions, I have used two sources: [1] C.S. Peirce's essay "How to Make Our Ideas Clear" (1879); and [2] Leibnitz's tract "Reflections Touching Knowledge, Truth, and Ideas" (1684).  This is because each source, in my opinion, is understandable for only two of the terms.  

Here we go.

• Clear: A clear idea is one which is so apprehended that it will be recognized wherever it is met with, and so that no other will be mistaken for it. [1]
• Obscure: A notion is obscure when it is not sufficient to enable us to recognize the thing which it represents: - when, for example, I remember some flower of animal which I have formerly seen but this remembrance is not sufficient to enable me to recognize its image or to discriminate it from others which resemble it. [2]

So ideas can be clear or obscure.  If an idea is obscure, we can go no further.  However, if an idea is clear, then it can be either distinct or confused. 

• Distinct: A distinct idea is one that contains nothing that is not clear.  This is technical language; by the contents [from the word "contains"] of an idea logicians understand whatever is contained in its definition.  So that an idea is distinctly apprehended according to them when we can give a precise definition of it in abstract terms. [1]
• Confused:  It [an idea] is confused when we are not able to enumerate marks sufficient to discriminate the thing from others [i.e. form a definition of it], although it may in reality have such marks and requisites into which its notion may be resolved...Thus [for example] we cannot explain to the blind what red is [even though it is clear idea]...In like manner we see that painters and other artists discern well enough what is well or ill done; but often are not able to give a reason for their judgement, and reply to those that inquire what it is that displeases them in the work, that there is something, they know not what, wanting. [2] 

Leibnitz took up the what Descartes had earlier said about clear and distinct ideas, but which Descartes and his followers had never been able to clarify.

I suppose we can summarize by saying an idea is clear when we can recognize it the next time we encounter it, and distinct when we have a good enough definition of it such that we can distinguish it from all other concepts. 

We can see that "clear" and "obscure" apply to ideas, but "distinct" and "confused" really only apply to definitions.

One Term, Many Meanings - Why Are We Surprised?

David Eddy kindly supplied me with the following military tale:

A true story heard around the Pentagon goes like this:

One reason the services have trouble operating jointly is that they don't speak the same language.

"secure a building" has been found to have the following meanings...

• Navy would turn off the lights and lock the doors.
• Army would occupy the building so no one could enter. 
• Marines would assault the building, capture it, and defend it with suppressive fire and close combat. 
• Air Force, on the other hand, would take out a three-year lease with an option to buy.

I think that we can all appreciate the humor in this, but must recognize that there is something deep and important about it.  But what is the moral in this tale?

The story shows that "secure a building" means different things to four different groups.  In each case the term refers to a different concept.  And in each case the concept is clearly defined.  The concepts are all very distinct - there is no chance of confusing them.

However, the four groups are all part of the same overall organization - the Armed Forces of the United States.  It is a common assumption that one organization is a monolithic semantic community.   The reality is that enterprises are often mosaics of different subcultures, who each see the enterprise through a different ontology.  At least, this is my observation.  I would like to find some research material on it, rather than anecdotes like the one quoted above.

The view that that enterprises are mosaics of subcultures also goes against the idea that there is a single data model - a "single reality", or a "single version of the truth" - that must underlie every enterprise.

Saying that the services "don't speak the same language" is a telling statement.  Rather than suggest that each service has its own view of the world - its own ontology - the fundamental difference is attributed to language.  This brings us back to the idea of the primacy of language over conception, and Wittgenstein's notion that language is the mirror of reality.  I believe these views are invalid, and that we need to "make our ideas clear" as Peirce said, and that language can as easily trick us as inform us. 

Perhaps the lesson for an analyst is not to be surprised when the same term is used to mean different things in the same enterprise.  Indeed, the analyst should be on guard for it when technical or unusual terms are used.  Homonyms can also indicate the existence of different concept systems, or ontologies.

Dodd-Frank "Swap" Definition: Industry Reaction (Part 1)

In an earlier post I attempted to analyze the definition of "swap" in the Dodd-Frank Act.  I have come across some articles about the definition which are worth looking at to see if they teach us anything about definitions in general.

The first is an article from Risk.net, an online magazine about financial risk management, entitled "US power bodies call for clarity on Dodd-Frank “swap” definition" (http://www.risk.net/energy-risk/news/2096674/power-bodies-clarity-dodd-frank-swap-definition).  The article was published on 2011-07-26.  It presents the opinion of the National Rural Electric Cooperative Association that Dodd-Frank has defined "swap" too broadly.  The definition would include transactions "long used to manage electric grid reliability" - essentially putting a rural electric cooperative in the same category as Goldman Sachs.  These transactions are used to optimize generation resources to ensure grid reliability.  The industry wants all these types of transactions exempted from the Dodd-Frank definition of "swap".  Ominously, however, the FERC's (Federal Energy Regulatory Commission)  general counsel, Michael Bardee, also stated that “a detailed listing of all such excluded commercial and merchandising agreements commonly used in the electric and natural gas industries is not feasible”. 

Here are my take-aways from this story:

• I would guess that the individuals responsible for the definition of "swap" were not aware of these electric energy transactions, based on the fact that there is no response justifying the inclusion of such transactions in the definition of "swap".  This illustrates the difficulty of trying to produce a definition without complete understanding of the ontology covered by the definition.  How does one obtain such a complete understanding?  I am not sure I have the answer to that.
  
• The individuals responsible for the definition of "swap" were certainly not aware of the consequences for participants in certain subsectors of the energy industry.  Thus we can clearly see that poor definitions play an important role in "the law of unintended consequences".  How can effects be predicted accurately at the point in time that a definition is made?  It would seem an even more complete understanding of the ontology covered is needed.  But how to achieve this is something I do not know how to do. 
  
• A legislative definition should be right from the outset.  That is quite different to other kinds of definition, which can be gradually improved over time (e.g. those of science).  A legislative definition, such as provided for "swap" is part of a more general act of creation - the overall Dodd-Frank Act in this case.  I would argue that if the conceptual system created is to be stable, the definitions of the concepts involved must be complete and coherent.  If this is not done, then the system will be unstable, and will create problems until such time as it is modified (or crashes).  Regrettably, we cannot model this kind of system to see what behavior is exhibited, and then optimize the definitions to achieve the intended results.
• Mr. Bardee's comment about the impossibility of creating an exhaustive list of types of contract to exclude is worrying because he is dealing with a subset of the general problem.  It is unclear why such a list cannot be produced.  Maybe different types of contract come and go frequently over time.  However, if this kind of difficulty exists in this narrow area, there is pretty much no hope for an adequate definition at the higher level of "swap".  Perhaps an approach would be to try to establish completeness in a sample of small subareas under a definition.  If completeness cannot be achieved, then it is pointless to continue with the higher level definition.   

An Outline of ISO 704:2009 - Terminology Work - Principles and Methods

ISO 704 is a standard that has a good deal of relevance for anyone interested in definitions.  I am providing an outline of it here, so its scope can be appreciated.  Substantive discussion of the content of the standard will be provided in future posts.  

The standard has been produced by ISO Technical Committee (TC) 37, which is called "Terminology and other language and content resources".  Actually the standard has been produced by Subcommittee SC1 of TC 37, which is known as "Principles and Methods".  ISO 704 existed in a previous version, that was published in 2000. 

The present version of the standard is 74 pages long and can be obtained from http://www.iso.org.

The abstract of the standard states:

ISO 704:2009 establishes the basic principles and methods for preparing and compiling terminologies both inside and outside the framework of standardization, and describes the links between objects, concepts, and their terminological representations. It also establishes general principles governing the formation of designations and the formulation of definitions. Full and complete understanding of these principles requires some background knowledge of terminology work. The principles are general in nature and this document is applicable to terminology work in scientific, technological, industrial, administrative and other fields of knowledge.

ISO 704:2009 does not stipulate procedures for the layout of international terminology standards, which are treated in ISO 10241. 

The standard begins with an overview.  This goes straight into the subject matter, and does not provide any philosophical or logical background.  Notation used in the standard is introduced.

The next major section is "Terms and Definitions". It is rather brief, but has several points of interest.  It is followed by a section on Concepts.  What concepts are is discussed, along with "general concepts" and "individual concepts".  Characteristics of concepts are then reviewed, and their importance highlighted - especially for terminological work.  An example of terminological analysis is presented, where a computer mouse is discussed.  Different classes of characteristics are outlined.  The next subsection within "Concepts" is about "Concept Relations".  This subsection is particularly important as it describes different types of concept systems.  It is quite a long subsection, and will be of interest to both data modelers and ontologists.

The next section is on definitions.  After a discussion of intensional definitions, a treatment of definitional writing is provided.  There is a good deal of detail in this long subsection.  It is followed by a subsection on "Supplementary information to the definition", essentially metadata for the definition.  Deficient definitions are discussed next.

The next section is on "designations".  These appear to be identical to logical signs - the things that signify concepts.  A number of very interesting points are discussed.  Obviously, this area is very close to terminology, but still has a lot of relevance to definitions.  Towards the end of the section, the "formation of terms and appellations" is discussed.  This might be of interest to data modelers, who often engage in debates about "naming conventions".

An annex is next, dealing with "Other types of definitions" - that is, other than the intensional type of definition, which was dealt with in the main body.  Here we get some of the usual suspects in the ontology of definitions.

A second annex deals with "Examples of term-formation methods".  As might be expected this is concerned with creating terms. 

The final annex is "Categories of appellations".  Although not claiming to be a classification, a list of major types of appellation is provided, and each is discussed in detail.  Again, this is closer to pure terminology than definitions, but is of considerable interest.

Data-Centricity vs. Concept-Centricity - Supposition vs. Definition

Definitions of concepts are extremely valuable, but are they sufficient for data management?  There seems to be a need for something more - for "data centric definitions".  By this I am referring the definition of a table, or column, or other data object in a database.  A concept definition - a full description of a business concept - is undoubtedly needed for a data object, but there is a strong case that more is required to form an adequate "data-centric definition".  I do not really like calling what I am discussing here "data-centric definitions", because a definition has traditionally been thought of as a property of a concept, not of a data object.  However, the term "data-centric definition" at least focuses us well on the area of interest, so I will keep it for now.

Let us first try to discover if there is anything in logic that relates to the topic we are discussing.  And indeed there is. Definition has been part of logic since its inception, and continues to be an area of interest, even in modern logic.  However, an idea called "supposition" goes beyond definition and seems to have something to say about data-centric definitions.  "Supposition" appears to be a purely Medieval idea, apparently not appearing in either ancient or modern logic. 

Supposition works like this.  Imagine I can successfully come up with a definition of "employee" for my enterprise.  The definition, as such is describing the concept of employee.  The definition, as such, has nothing to say about any actual employees.  It is only when I begin to use the term "employee" in propositions that I have to consider exactly what set of instances (actual employees) I am referring to.  The definition alone cannot help me with this.  

If I make a database table called "employee" I have to ensure that I am conforming to the definition of the concept "employee".  But the table may be designed to capture a specific set of employees.  Perhaps the table is only intended to store US employees, or Canadian employees.  Perhaps it is only intended to store current employees, or perhaps only past employees.  None of this information about the nature of the table can appear in the definition of the concept "employee".  This information is the supposition of "employee" - better yet, the supposition of "employee" in a data-centric context.  To work effectively with my employee database table, I need the supposition as much as the definition.

So what is supposition?  Here I will paraphrase George Hayward Joyce's discussion in "Principles of  Logic".  The supposition of a term in a proposition must be understood because even a univocal term (a term which signifies only one concept) can be construed is various ways.  For the logicians there were three main forms of supposition:

(1) Collective and Distributive Use.  When anything is affirmed or denied of a plural subject, the predicate may apply: (a) to individuals (instances); or (b) to the individuals taken as a group.  (a) is called the distributive use (suppositio distributiva) and (b) is called the collective use (suppositio collectiva).  E.g. for (a) "The employees attended a town hall meeting" applies only to the individuals that came to the event.  E.g. for (b) "The employees in our enterprise sign an employment contract" applies to all employees, without having to think about any instances. 

(2) Real and Logical Use.  Is the term being used as it applies to: (a) the real order, or (b) as it is at the conceptual level?  E.g. for (a) "Employee A is sitting at his desk right now" - this refers to reality.  E.g. for (b) "I am working on the definition of 'employee' right now" - this refers to my dealing with the concept of "employee", and not any individuals who are employees.  In logician's Latin, (a) is called suppositio realis and (b) is called suppositio logica.  

(3) Material Supposition (suppositio materialis).  This one is much closer to our ideas about metadata.  It happens when I am referring to the sign I am discussing.  E.g. if I say "'Employee' is a word consisting of eight letters", I am not dealing with the concept, nor any individuals, but only discussing the sign.  Aspects of naming conventions in databases intended to keep the names of data objects reasonably short would seem to fall into this category.

I suspect that the distinction between supposition and definition may answer some of the problems brought up under the heading of "context" by data managers.  It seems to me that "context" is an overloaded and overabstracted term that refers to a number of quite different issues.

I hope this proves that there is a theoretical basis for "data-centric definitions" and that supposition provides part of it.  We have a lot to follow up on regarding "data-centric definitions" and will be returning to it in  the future.      

Definition as Content and Container

We often use terms without fully thinking out how we are using them - particularly if abstraction is involved.  The other day I was looking at a data modeling tool, and asked a colleague "What did you put into the definition?".  It suddenly occurred to me that I was not talking about the content of the definition, but about the definition field into which I could type the content.

After thinking about it some more, I realized that the definition field is a container which we are free to use as we want (at least in the tool I was working with).  So an immediate question is, how do we want to structure what goes into this field (e.g. with section headings) and what metadata about the definition do we want to put in (e.g. what person last updated it)?  

Of course, the more sophisticated tools have these separate elements of the definition more explicitly segregated.  The "definition" is a set of smaller containers inside a bigger container. Perhaps this might mean that a different set of questions arises, such as the order in which we develop content in this set of smaller containers.  Or perhaps such questions depend on the actual set of containers that are exposed to use in each of these tools.  Either way, I think questions arise.

It is good to be concerned about how to develop good definitional content, quite apart from what this content is stored in.  However, the fact that some kind of container exists (whether simple or complex) means that we have to be concerned with the management of the container also.  Since both container and content are signified by the term "definition" there is ample scope for confusion here.  

I do not think this is being pedantic.  Information Knowledge Management is difficult because it is dealing with things that are inherently abstract and immaterial.  It means we really have to think our way through it, and it is easy to make mistakes.  Keeping in mind that there is a distinction between container and content may help us to avoid a few of these mistakes.

Why Do Definitions - Why Not Just Use Wikipedia?

Today I spoke to the New York ERwin Modeling User Group (NYEMUG) on "Creating Great Definitions".  One question came up, which was why bother crafting definitions at all - why not simply rely on Wikipedia for them?  I suppose it could be any external source, and not necessarily Wikipedia (e.g. the Enterprise Data Management's Semantics Repository).  The way it would work conceptually would be to associate a link to Wikipedia with any term.

It might be thought that Wikipedia only deals with common terms, and not specialized technical terms.  However, there are a good number of technical terms that are present in Wikipedia.

My first reaction was both "yes" and "no".  "Yes" because it is simply obvious there is considerable value in Wikipedia, but "no" because Wikipedia does not understand the enterprise I work in, which ultimately supplies an enormous amount of context that influences definitions.   

I will have to think about this topic some more before I can fully answer it.  But, I at least wanted to capture a few initial ideas.  I take it for granted that Wikipedia has value, so I need to think why this value might be so limited that we still need to do definitions.   

(a) Wikipedia cannot provide details of a concept which is defined by a common relation of the instances included in the concept to another concept (usually a special way of managing these instances) that is unique to the enterprise (or part of the enterprise).  So Wikipedia cannot tell us what a Financial Asset is for our enterprise, because what we include in "Financial Asset" depends on our business model.    

(b) Wikipedia cannot deal with definitions modified on a per-context basis.  Speech communities are contexts.  So, for instance, Wikipedia cannot tell us what our Marketing department considers a Customer to be, versus what our Accounts Receivables department considers a Customer to be.  Wikipedia does not know the speech communities in our enterprise.

(c) The structure of Wikipedia entries differs on a per page basis.  There is no consistent set of sections to a definition, nor metadata for a definition.  This might inhibit use in an enterprise.

(d) Concepts are arranged in concept systems.  A definition shows how a concept is placed in a concept system.  One concept can occur in many concept systems, and the definition is modified in each case.

(e) Wikipedia can be wrong.  I have found errors, but very few - so I am not sure how valid this objection is.

Some of the above points may or may not overlap.  They are also preliminary, so subject to revision. 

That is all for now - I will have to return to this one.  However, the question raised in the title is one that must be answered.

How is a Definition Different from an Explanation? (Part 2)

In Part I of this series (http://definitionsinsemantics.blogspot.com/2011/11/how-is-definition-different-from.html), we explored explanation defined as "bringing a mind to an understanding of a topic".  

There is, however, another form of explanation, of which Aristotle said "We believe ourselves to know a thing when we are acquainted with its cause" (Posterior Analytics, II C. II para 1).

One way to provide this kind of explanation is by arguing from the cause to the effect.  Traditionally this involved using syllogisms where the cause was in the major premiss and the effect was in the conclusion.  E.g. All ellipses show a pattern of positions X; the orbit of Mars shows a pattern of positions X; therefore the orbit of Mars is an ellipse.

This form of explanation is very satifying, and it might seem natural to try to incorporate it into definitions.  However, there are reasons not to do so.  First, the above form requires putting the concept to be defined into propositions, and then putting the propositions into a syllogism (an argument).  And then putting all of that into the definition.  But if all this is part of the definition, and the definition is supposed to substitute for the term being defined, we would seem to end up with an infinite regression.  Of course, this is always a danger when the term to be defined is in the definition, but having an argument in the definition seems to assure this as the term has to be included in the argument.  At least it seems that way to me - although I cannot find any description of my opinion in the literature, and I will willingly defer to others who can prove me wrong.

Secondly, putting argument into a definition exposes the definition to much greater chance of error.  The propositions may be false, and the argumentation may be invalid.  And why would we put the concept into propositions and arguments before we have a completed definition - surely that is jumping the gun.

My provisional conclusion, therefore, is that causal explanations should not be placed in definitions.  

BUT, there is another consideration.  In modern enterprises, definitions are containers as well as content.  If an explanation has to be provided, and there is nowhere else to put it, then it should be put into the definition (as container, not content).  The distinction between container and content is not found (at least by me) in traditional logic.  Yet it is a most important consideration.  I suppose we need another blog on definition as container vs. content.

If we are forced to put a causal explanation in the definition (container) then at least get the concept defined fully (content) before any explanation is provided, so the explanation is not part of the true definition (content).

Dodd-Frank "Swap" - A Definitional Disaster?

The Dodd-Frank Act is intended to reform the financial system in order to reduce the chance of any future systemic failure.  Obviously, it is very important, and one of the most important parts of Dodd-Frank revolves around swaps.  It may be recalled that lack of understanding about Credit Default Swaps (CDS) was a big part of the financial crisis that began in 2008 - and specifically caused AIG to fail.  CDS, however, are only one species of swap.  Prior to 2008 there had been little regulation of swaps.

Before we go further, there is one other piece of background for those unfamiliar with the sausage-making process of US financial regulation.  An Act of Congress is just the beginning.  Agencies of the US government must take the Act and turn it in to rules - usually many rules - and then enforce them.  This means that if there is a problem in the Act, there can be difficulties across many rules.

Back to swaps.  The Dodd-Frank definition of "swap" is given below.  It is a pretty mind-numbing read.  This definition clearly shows that there is no simple definition of a swap.  The definition has a long list of concepts that are included, and a long list of concepts that are excluded (Paragraph B).

What we have is a collection of "things" that the government wants to be regulated in an identical manner - as "swaps".  However, the only consistency is the way these "things" are to be regulated.  There is no consistency - no common attributes intrinsic to these things - that distinguishes them from other things.  If there was a set of common characteristics, then these would presumably have been listed.  Instead, we get enumeration of members of the class (a practice frowned upon by traditional logicians). This is seen especially in the items labelled "(I)" through "(XXII)".  These are part of the definition, not a list of illustrations.

The definition even abstracts from enumeration of members when it says:  "(iv) that is an agreement, contract, or transaction that is, or in the future becomes, commonly known to the trade as a swap".  

I was shocked when I first read this.  How can a definition simply point back to common usage of a term?  The government is throwing the burden of definition back on the speech community!  Even then, what is meant by "the trade" and "commonly known" is puzzling (I cannot find definitions for these terms in the Act).  Also, if some bright investment banker figures out a new product that he or she brands as something other than a swap, then they can presumably escape regulation.  However, there is an "escape hatch" - the Commodities Futures Trading Commission (CFTC) is delegated the authority to broaden the definition, which should prevent this kind of problem.  But suppose something gets called a "swap" that is nothing like a "real swap".  That sort of thing happens all the time in the evolution of language.  There is no way the definition can be modified to exclude it as the Act is written. 

There is a lot more to be said on the Dodd-Frank definition of "swap" and we will cover more topics in future posts.  This post is merely an appetizer to what promises to be a true feast of lessons that can be drawn about definitions.

Here is the defintion of "swap" - only subparagraphs (A) and (B) of the text - subparagraphs (C) through (F) are not part of the general definition.  The full text of the Act can be found at: www.sec.gov/about/laws/wallstreetreform-cpa.pdf - please refer to Section 721 and scroll down until you see "SWAPS".

‘‘(47) SWAP.—
‘‘(A) IN GENERAL.—Except as provided in subparagraph (B), the term ‘swap’ means any agreement, contract, or transaction—
‘‘(i) that is a put, call, cap, floor, collar, or similar option of any kind that is for the purchase or sale, or based on the value, of 1 or more interest or other rates, currencies, commodities, securities, instruments of indebtedness, indices, quantitative measures, or other financial or economic interests or property of any kind;
‘‘(ii) that provides for any purchase, sale, payment, or delivery (other than a dividend on an equity security) that is dependent on the occurrence, nonoccurrence, or the extent of the occurrence of an event or contingency associated with a potential financial, economic, or commercial consequence;
‘‘(iii) that provides on an executory basis for the exchange, on a fixed or contingent basis, of 1 or more payments based on the value or level of 1 or more interest or other rates, currencies, commodities, securities, instruments of indebtedness, indices, quantitative measures, or other financial or economic interests or property of any kind, or any interest therein or based on the value thereof, and that transfers, as between the parties to the transaction, in whole or in part, the financial risk associated with a future change in any such value or level without also conveying a current or future direct or indirect ownership interest in an asset (including any enterprise or investment pool) or liability that incorporates the financial risk so transferred, including any agreement, contract, or transaction commonly known as—
‘‘(I) an interest rate swap;
‘‘(II) a rate floor;
‘‘(III) a rate cap;
‘‘(IV) a rate collar;
‘‘(V) a cross-currency rate swap;
‘‘(VI) a basis swap;
‘‘(VII) a currency swap;
‘‘(VIII) a foreign exchange swap;
‘‘(IX) a total return swap;
‘‘(X) an equity index swap;
‘‘(XI) an equity swap;
‘‘(XII) a debt index swap;
‘‘(XIII) a debt swap;
‘‘(XIV) a credit spread;
‘‘(XV) a credit default swap;
‘‘(XVI) a credit swap;
‘‘(XVII) a weather swap;
‘‘(XVIII) an energy swap;
‘‘(XIX) a metal swap;
‘‘(XX) an agricultural swap;
‘‘(XXI) an emissions swap; and
‘‘(XXII) a commodity swap;
‘‘(iv) that is an agreement, contract, or transaction that is, or in the future becomes, commonly known to the trade as a swap; ‘‘(v) including any security-based swap agreement which meets the definition of ‘swap agreement’ as defined in section 206A of the Gramm-Leach-Bliley Act (15 U.S.C. 78c note) of which a material term is based on the price, yield, value, or volatility of any security or any group or index of securities, or any interest therein; or
‘‘(vi) that is any combination or permutation of, or option on, any agreement, contract, or transaction described in any of clauses (i) through (v).

‘‘(B) EXCLUSIONS.—The term ‘swap’ does not include—
‘‘(i) any contract of sale of a commodity for future delivery (or option on such a contract), leverage contract authorized under section 19, security futures product, or agreement, contract, or transaction described in section 2(c)(2)(C)(i) or section 2(c)(2)(D)(i);
‘‘(ii) any sale of a nonfinancial commodity or security for deferred shipment or delivery, so long as the transaction is intended to be physically settled;
‘‘(iii) any put, call, straddle, option, or privilege on any security, certificate of deposit, or group or index of securities, including any interest therein or based
on the value thereof, that is subject to—
‘‘(I) the Securities Act of 1933 (15 U.S.C. 77a et seq.); and
‘‘(II) the Securities Exchange Act of 1934 (15 U.S.C. 78a et seq.);
‘‘(iv) any put, call, straddle, option, or privilege relating to a foreign currency entered into on a national securities exchange registered pursuant to section 6(a) of the Securities Exchange Act of 1934 (15 U.S.C. 78f(a));
‘‘(v) any agreement, contract, or transaction providing for the purchase or sale of 1 or more securities on a fixed basis that is subject to—
‘‘(I) the Securities Act of 1933 (15 U.S.C. 77a et seq.); and
‘‘(II) the Securities Exchange Act of 1934 (15 U.S.C. 78a et seq.);
‘‘(vi) any agreement, contract, or transaction providing for the purchase or sale of 1 or more securities on a contingent basis that is subject to the Securities
Act of 1933 (15 U.S.C. 77a et seq.) and the Securities Exchange Act of 1934 (15 U.S.C. 78a et seq.), unless the agreement, contract, or transaction predicates the purchase or sale on the occurrence of a bona fide contingency that might reasonably be expected to affect or be affected by the  creditworthiness of a party other than a party to the agreement, contract, or transaction;
‘‘(vii) any note, bond, or evidence of indebtedness that is a security, as defined in section 2(a)(1) of the Securities Act of 1933 (15 U.S.C. 77b(a)(1));
‘‘(viii) any agreement, contract, or transaction that is—
‘‘(I) based on a security; and
‘‘(II) entered into directly or through an underwriter (as defined in section 2(a)(11) of the Securities Act of 1933 (15 U.S.C. 77b(a)(11)) by the issuer
of such security for the purposes of raising capital, unless the agreement, contract, or transaction is entered into to manage a risk associated with capital raising;
‘‘(ix) any agreement, contract, or transaction a counterparty of which is a Federal Reserve bank, the Federal Government, or a Federal agency that is
expressly backed by the full faith and credit of the United States; and
‘‘(x) any security-based swap, other than a securitybased swap as described in subparagraph (D). 
...
(b) AUTHORITY TO DEFINE TERMS.—The Commodity Futures Trading Commission may adopt a rule to define—
(1) the term ‘‘commercial risk’’; and
(2) any other term included in an amendment to the Commodity Exchange Act (7 U.S.C. 1 et seq.) made by this subtitle.
(c) MODIFICATION OF DEFINITIONS.—To include transactions and entities that have been structured to evade this subtitle (or an amendment made by this subtitle), the Commodity Futures Trading Commission shall adopt a rule to further define the terms ‘‘swap’’, ‘‘swap dealer’’, ‘‘major swap participant’’, and ‘‘eligible contract participant’’.

F.A. Hayek on Definitions - An Anti-Scientistic Stance

F.A. Hayek is well known for his battles against Socialism and Keynesianism.  Less well known is his critique of the misapplication of the techniques and language of natural science to the social sciences - an attack on Positivism made eloquently in his 1952 book "The Counter-Revolution of  Science - Studies on The Abuse of Reason" (ISBN 0-913966-66-5).  Hayek termed this misapplication of the natural sciences "Scientism".  In Chapter 3 of the book the Nobel laureate makes some interesting observations on definitions.  Hayek writes:
 

'Take the concept of a "tool" or "instrument", or of any particular tool such as a hammer or barometer.  It is easily seen that these concepts cannot be interpreted to refer to "objective facts", that is, to things irrespective of what people think about them.  Careful logical analysis of these concepts will show that they all express relationships between several (at least three) terms, of which one is the acting or thinking person, the second one the desired or imagined effect, and the third a thing in the ordinary sense.  If the reader will attempt a definition he will soon find that he cannot give one without using some term such as "suitable for" or "intended for" or some other expression referring to the use for which it is designed by somebody.  And a definition which is to comprise all instances of the class will not contain any reference to its substance, or shape, or other physical attribute.  An ordinary hammer and a steamhammer, or an aneroid barometer and a mercury barometer, have nothing in common except the purpose for which men think they can be used.

It must not be objected that these are merely instances of abstractions to arrive at generic terms just as those used in the physical sciences.  The point is that they are abstractions from all physical attributes of the things in question and that their definitions must run entirely in terms of mental attitudes of men toward these things.  The significant difference between the two views of the things stands out clearly if we think, for example, of the problem of the archaeologist trying to determine whether what looks like a stone implement is in truth an "artifact", made by man, or merely a chance product of nature.  There is no way of deciding this but by trying to understand the working of the mind of prehistoric man, of attempting to understand how he would have made such an implement.  If we are not more aware that this is what we actually do in such cases and that we necessarily rely on our own knowledge of a human mind, this is so mainly because of the impossibility of conceiving of an observer who does not possess a human mind and interprets what he sees in terms of the working of his own mind."

There is much in the above passage to think about.  For me, the critical point is "...they are abstractions from all physical attributes of the things in question and that their definitions must run entirely in terms of mental attitudes of men toward these things".  This implies that definitions cannot always be based on qualities, but sometimes need to be based on relationships involving a human mind.  We will return to the implications of Hayek's viewpoint in future posts.

One last thing.  Few writers have had the courage to criticize any aspect of natural science.  R.G. Collingwood was one, calling "pseudo-sciences" what Hayek calls "Scientism".  I do not think Hayek was aware of Collingwood, but the above passage very closely resembles ideas Collingwood expressed.

Is the IAU Definition of "Planet" A Quality Definition?

In this post we continue to learn lessons from the International Astronomical Union's definition of "planet" in 2006 (http://www.iau.org/public_press/news/detail/iau0603/).  The question tackled here is whether the IAU's definition of "planet" is a quality definition.  After close examination, it seems it is not.

Here is the definition:

'A "planet" [1] is a celestial body that (a) is in orbit around the Sun, (b) has sufficient mass for its self-gravity to overcome rigid body forces so that it assumes a hydrostatic equilibrium (nearly round) shape, and (c) has cleared the neighbourhood around its orbit.

[1] The eight planets are: Mercury, Venus, Earth, Mars, Jupiter, Saturn, Uranus, and Neptune.' 

And here is the analysis of the definition:

(a) The definition is not actually of "planet" but of "planets in our Solar System".  This can be mined out of the text of Resolution 5A, which states: 

"RESOLUTION 5A The IAU therefore resolves that planets and other bodies in our Solar System, except satellites, be defined into three distinct categories in the following way..." 

So the IAU did not define "planet" at all, but merely "planets in our Solar System".  We will need to explore this in a further post, but it is clearly a source of confusion, and hence the definition is of poor quality (definitions are not supposed to cause confusion).

(b) The superordinate genus identified in the definition is "celestial body".  If I look up "celestial body" in Wordnet (http://wordnet.princeton.edu) I get "natural objects visible in the sky".  So, celestial bodies must include planets, stars, comets, asteroids, nebulae, galaxies, and so on.  As such, the genus seems too remote for a quality definition - it is little better than "thing".  There seems to be a strong possibility that it could be divided into genera that are superordinate to "planet", but subordinate to "celestial body".  What are they?  That is not my job - I am not an astronomer.  But I can tell you that a more proximate superordinate genus is required for this to be a quality definition.

(c) The definition contains the phrase "is in orbit around the Sun".  This clearly shows that only our Solar System" is being considered, as we saw above.  Guess what - I could tell from the term "planets in our Solar System" (the definiendum) that the planets would be orbiting the Sun.  That is part of the definition (or description) of "our Solar System".  The term "Sun" should not have been used in the definition.  It is an essential characteristic of "our Solar System", not "planet".  Another point that shows we do not have a quality definition.

(d) If the IAU chooses to define "planet in our Solar System" it is obliged to define "extrasolar planet" (the coordinate species) and "planet" (the proximate superordinate genus).  At the very least these should have been referenced in Resolution 5A.  There is no such reference in 5A.  Again, an indication of a poor quality definition.

(e) What does "(c) has cleared the neighbourhood around its orbit" actually mean?  What is "neighbourhood" in the celestial context?  It is easily understood as a general (human, non-astronomical) term, but that cannot possibly apply here.  What is it?  Suppose I make the presumption that "cleared" means "to have removed matter".  I really do not have any right to do so, but suppose I do.  Well, the Earth has not absorbed the Moon or ejected the Moon from its proximity.  So is the Earth not a planet?  This is a big failure, because definitions are supposed to make things clear.  Again, we have a poor quality definition.

In defense of the IAU, it is quite difficult to produce definitions in natural science.  There is usually no alternative to them being other than descriptive (as opposed to essential or causal).  However, the IAU could have done better.

Thus, we see that we have a poor quality definition of "planet".  Sorry, I meant to say "planet in our Solar System".

The Problems of Pluto - 2: What is a Definition Authority?

In a previous post (http://definitionsinsemantics.blogspot.com/2011/11/evolution-of-definitions-problem-of.html) the topic of the IAU's 2006 redefinition of the term "planet" was discussed (see http://www.iau.org/public_press/news/detail/iau0603/). 

There are many topics relevant to definitions surrounding this event - one of them is the problem of authority.  The International Astronomical Union (IAU) is the organization that came up with the new definition of "planet".  Does this make it an authority?  It does if we define "authority" as "a source of a definition that publishes, and administers, and supports this definition". 

However, the term "authority" carries emotional content also.  "Authority" can imply (a) a role of active enforcement; (b) an obligation of obedience upon the body politic.  These implications exist because "authority" is a term that defines many concepts, and some of them have to do with enforcement and legislation. Now, some entities that produce definitions also have enforcement roles and legislative roles, but that is not as any direct consequence of producing definitions.  

Additionally, the IAU is presumably a collection of scientists, and scientists have an almost priestly aura in contemporary Western civilization.  To challenge any scientist is presumably to challenge science.  To challenge a scientific consensus is presumably an even worse sin.  Just having scientists as an authority seems to add to the intimidating implications when we use the term "authority" when speaking of definitions.

I think the basic stance on this issue must come from the logicians' warning about appeal to authority - the fallacy of argumentum ad verecundiam.  First, let us remember that we are only dealing with definitions here, not argumentation such as deduction or induction.  The idea is presented above that an authority is a special type of source of a definition.  To be an authority, I suggest, means that you are doing more than just creating a definition.  You are publishing it so it is available to some form of wider audience.  You are administering it, so e.g. there is an orderly way to change it.  And you support it - as an authority you stand ready to be consulted on it.

None of this means the definition is of high quality.  An authority can do all of the above and create poor quality definitions.  An authority can also cause terminological havoc outside of the narrow field of definitions.  That it probably why I cannot stop thinking of Pluto as a planet.

We must be very careful, therefore, to understand what an authority is for a definition, and what to expect of an authority. Above all we should not be intimidated.  Remember, just because a source of a definition claims to be an authority, or is styled as such, does not mean they automatically produce high quality definitions.