Mind Your Language

Published in TDAN.com January 2003

“Effective communication depends on a common language” is one of those blinding statements of the obvious for which consultants are paid a small fortune. And so you may find that language hinders
rather than helps communication when you are analysing the business of an organisation. For instance, you know that modelling the “real” world is fundamental to understanding the business and
providing information system solutions to support its needs. The model here provides the means of common communication by the organisation about its business [1], and forms a stable base for use by solution providers.

Yet, in the course of wandering (and wondering) about the modelling world for a number of years, I have found solution providers often overlook or give scant regard to the fact that a model about
its business is an expression of the common business language used by the organisation. This article looks at some details and facets of language and how they may affect business data models in
particular. A business data model is concerned with the kinds of information an organisation needs in order to run and manage its business affairs. It is key to the understanding and management of
what information is available to the organisation.

On the subject of wandering – according to Douglas Adams’ The Hitchhiker’s Guide To The Galaxy, the Guide’s outer wrapping holds the comforting message “DON’T PANIC”, which I often find
appropriate when helping to produce a model. Moreover, as any hitchhiker knows you don’t get to where you want to go by the most obvious route, some parts are interesting and some are boring, and
sometimes you never arrive where you intended at all. Therefore, this article is not structured in a logical, lucid or coherent way, or one that you’ll necessarily always find enthralling [2].

The following table is a simple map, or representation, of the territory I’ll be covering.

The human mind is designed to categorise things that are met in the real world. Given the way the world is put together, dividing space-time into things and then categorising them is a sensible way
of making predictions. If you can slot a new occurrence into a category you know, then you can predict what may happen next based on knowledge of that category. The organism best able to make
predictions about what is going to happen next is more liable to survive. A common, shared understanding of the categories helps ensure the survival of the species. Words are used to label,
delineate and communicate categories, or kinds, of thing [3].

Like the body of people that contribute to it, the organisation [4] best able to make predictions about what is going to happen next is likely to
survive longer. It needs a common, shared understanding of those categories of things it is interested in, so that its human resources may communicate effectively about the business matters that
are encountered in the world and so assist the organisation’s survival.

A common meaning for a word supports both spoken and written methods of communication. Whilst speech is normally used for immediate communication, writing enables you to record matters for use by
yourself or others not directly involved. However, the point about writing is not that it makes it possible to record things, but that it enables them to be done accurately and permanently.
Similarly, if an organisation makes accurate and permanent records of those things in which it is interested, it is better prepared to share this information and carry out its business.

Information technology (IT) is seen to support this process by allowing for data to be kept in a computerised form [5].However, computerisation is
not the only solution for record keeping: manually written records can also prove effective.

It is important to distinguish between the part of the business that concerns “real” things (such as an order being placed) and the part that concerns recording the data (such as writing the
order down or entering the details into an electronic document). The boundary is often blurred in a technology-led society; but, when you take away or disable the technology, some manual backup
system often comes into play and the distinction is clearer. So computerised or written information is only a record of some thing that exists elsewhere. This thing exists in space and time whether
or not a computerised or manual system is present to store and manipulate it for record keeping. In a very real sense then, this is data that exists in its own right and is not transient or system

It is tempting to try to jump straight from the data perceived in the real world to the corresponding records, but you know that before this it is necessary first to represent in a data model the
kinds of business things in which your organisation is interested, and about which it wants to keep records. I will call these representations business concepts as they are an abstraction
or notion of kinds of things [6]. You can then make record-keeping designs based on the business concepts, which you then implement, in reality,
to hold the records of the organisation’s business.

If an organisation’s information system is to be accurate what it records must precisely reflect the real world. You may equate the real world on the map with the Enterprise (or Row 2) view of the
Zachman Framework. However, it is the real world as the organisation perceives it: which is a view or opinion that other observers may not necessarily share. For example, a general observer may
view an ‘order’ in the commercial sense as a request or instruction for the supply of something; yet, your organisation may consider that orders it makes (purchase orders) are distinct from those
placed on it (sales orders [7]). It has separate business processes that deal, for example, with the raising of purchase orders and the receipt of
sales orders.

As I have mentioned, an organisation is interested in business concepts that are categories or kinds of things. Nouns are the part of speech category that typically refer to kinds of things.
However, as you probably know, things are not confined to physical items. When you construe an aspect of time and space as something that can be identified and counted or measured, language allows
you to express that aspect as a noun, whether or not it is a physical thing. Such abstract notions may often be actions.

As a part of speech category, verbs can express a wide range of meanings, such as actions, sensations or states of being. In order to reference actions as things, a noun form of the verb is used.
English often employs the same word as a noun and a verb. It may be, however, that the gerund form of the verb is more apposite when your organisation is interested in the action rather than the
outcome. For example, it may be that the organisation is interested in an order as the action of ordering – perhaps in knowing when it is proposed, accepted, or completed. The actions of proposing,
accepting and completing an order may then be seen by your organisation as simply properties or states of an order.

Whilst nouns referring to physical things can stand alone as business concepts, those referring to actions show a dependency on other business concepts. This is because of a verb’s role in a

A statement is a sentence whose primary purpose is to convey information. Your organisation uses them to describe its business. Regular sentences can be broken down into one or more specific and
predictable patterns of elements that involve atomic items such as nouns and verbs. Such patterns are known as clauses (or sentences, if you follow the convention of American linguists [8]).

In a declarative clause, or statement, the verb does not exist alone. In the case of a transitive verb there are typically two other statement elements that are necessary for an occurrence to exist
in the real world. For example, an order, whether a sales order or a purchase order, requires two participants – one of which is your organisation. One participant acts as the buyer and the other
as the seller for the ordering arrangement. Your organisation’s role as one of the participants is not always considered. This often happens when an application system is scoped against the
business data model. Because a sales order or purchase order must, by definition, be some kind of thing of interest to your organisation its role in the matter is assumed. This assumption can be a
problem where an organisation consists of more than one legal body and you have to maintain the separate financial positions for legal and taxation purposes.

So, the fundamental link between any two business concepts is one of dependency: for an occurrence of one business concept to exist an occurrence of another business concept must exist [9].

A preposition is the part of speech that is used for a meaningful link between two elements in a statement; most often it shows how they are related in time and space. This is all that is needed in
documenting the fundamental link between one business concept and another. Thus a purchase order may be defined as “a request from one participant (our business) to another agreeing to buy the
goods or services supplied by the other”.

If you use a verb or verb phrase instead of the preposition it implies your organisation may be interested in the particular action itself; that is, it may regard it as a business concept.
Furthermore, it may cause you to want to document the link in both the active and passive voice. For example, instead of documenting the essential link as “a sales order is from a participant”,
you are lead into saying something like: “a sales order is originated by a participant” and “a participant originates a sales order” because of the two real world dependencies for the verb

The organisation’s view of the world is couched in its business language. Therefore, in producing the model, you may consider that you are creating some kind of dictionary [10] for the organisation. This is a reference work – a thesaurus or ‘storehouse’ of knowledge – dealing with the business concepts or terms used in
the vocabulary of a particular organisation. The terms or concepts may be peculiar to the organisation itself, or particular to the business field in which it operates, or extracted from those in
general public use. It does not have to contain words with meanings that are generally known and understood (and with which you describe a business concept).

Some words used by your organisation are terms used to manifest a thing’s behaviour and for that reason are often called “roles”. When used, this kind of word often asserts a converse. Such
pairs of terms are known as relational antonyms. For example, in talking about an order I used the terms “buyer” and “seller”. If I order from you, then I am the buyer and you are the seller.
Each of these terms is referring to one of the two dependencies for an order.

An organisation may, however, favour the term “customer” instead of buyer when considering this dependency for a sales order, and use “supplier” to denote the latter link because it sees this
as the role being played for a purchase order. In these cases, I can describe the terms customer and supplier in my dictionary by reference to the business concepts and dependencies that exist
there already. If you simply document the term as the dependency link phrase (instead of or together with the preposition) this word is not so readily found in a data dictionary, as it is probably
held there as a different kind of record. Roles are an example of things that are referred to generally as business terms.

You may find that different parts of the organisation use the same word to mean different things. Perhaps you’ve been advised to try and reconcile these different views in order to get a single
agreed definition for the organisation. Yet, if marketing use the term customer to mean some body who may buy our goods or services, distribution use it to mean some body to whom goods are to be
delivered, and finance mean some body who has an account for the purchase of goods and services these terms deserve to be documented separately. A dictionary can list more than one sense of a term,
so why doesn’t your data dictionary?

Also, if you want to define business terms precisely you need to record the scope of the definition. That is, you document the organisational context where the term is used in much the same way
that a monolingual dictionary may put “Austral”, “Canad” or “Brit” against a word to describe its coverage. By default the coverage is your whole organisation. As well as homonyms, you can
define synonyms like you do for business terms.

Because business terms are encountered during business modelling you should collect and formally define them at this stage. You use business terms:

  • To build a business dictionary (as already noted).
  • As a basis for information exchange (otherwise how do people know they are talking the same language as the data on which they are enquiring?).
  • To assist in requirements when scoping an application system. An enterprise view of the world involves business compromises for its application systems. These allow the various systems to serve
    the business as a whole and provide for its operational and informational needs.
  • To save repeated use of the same selection criteria when writing process logic.
  • As an aid to improving business communication and control in general.

Words are sometimes used to refer to a type or sort of thing, as when you say “the car is a vehicle” and sometimes to refer to some individual thing of the sort in question, as when you say “the
car is in the garage”. It is important to distinguish between an actual thing and a type of thing. An actual car (possibly identified by its chassis number) is not the same as a car model (given a
name by the manufacturer).

Your organisation needs to decide whether or not it needs to manage both actual and type, or only one of them. For example, a tailor needs to manage garment types (such as trousers, jacket, tie)
and also needs to manage a particular garment being made for a particular person. In this case the tailor needs both “garment” and “garment type”. A clothes supermarket may not need to manage
an individual garment, and thus only needs “garment type”.

Note that a garment could exist without having a type allocated to it (a prototype, for example). Also, a new garment type could be envisaged for which there were no actual garments yet made. This
means that the actual thing and the type are both independent business concepts. It is common to find a further business concept that allows you to know which type(s) the actual belongs to. In the
above example this business concept is acknowledging a garment as a garment type. This concept can be useful in business situations where (a) the typing of an item may change over time, such as
recognising that an upgrade to its specification changes a car’s model (it’s still the same car as far as the owner’s concerned), or (b) where different classification systems may apply at the
same time, or (c) in both cases.

The basic structure of a business concept (or word) definition has been known since the time of Aristotle, who distinguished two factors: a general category to which a word belongs, and the
specific feature(s) which distinguish that word from related words. So, if your organisation is interested in people (and what organisation isn’t?) it may simply describe “person” as “a human
being” [11]. Hyponymy is an important idea used in trying to categorise the world and an important sense connection for a dictionary.

However, not all members of the same category must necessarily have the same features. Most vegetables are green, but carrots are not. Most birds fly, but penguins don’t. These kinds of categories
are more like family resemblances, where some members of a family have the same traits and features but not all do. It may therefore be necessary to provide examples of the kinds of things that are
both included and, more importantly, excluded from the scope of the business concept.

At first it appears that I can recognise “Peter” in some taxonomic-like hierarchy such as “Lop-eared English grey”, “rabbit”, “mammal”, “animal” and “living being”. However, if I am
interested in an animal as a business concept (being any living being except a human being), then I may be interested in two separate business concepts when I refer to its kind:

  • Animal Type allows me to recognise the kind of breed (“Lop-eared English grey”, “English grey”). These may be considered as specialisations or sub-types.
  • Animal Category allows to recognise the kind of animal (“rabbit”, “mammal”). These may be considered as generalisations or super-types.

These three business concepts (actual, type and category) are independent and allows my organisation to acknowledge different combinations of them, if required for business purposes. So, I can
recognise that:

  • Peter is a Lop-eared English grey.
  • Peter is a rabbit.
  • A Lop-eared English grey is a rabbit.

Phew! I’ve managed to get through an article about data modelling without using words such as ‘entity’, ‘object’ or ‘class’. Although, properly speaking a part of speech is known as a word
class, and one of the elements of a sentence is called an object I tried to avoid using these terms.

So far you may expect the representation of the real world to consider:

  • Business concepts – A kind of thing in which the organisation itself is interested, and about which it wants to keep records.
  • Dependencies – A fundamental link between any two of those kinds of thing.
  • Business terms – A word or phrase that is meaningful to the business and is defined against the business concepts, dependencies or other artefacts in the business model.


  1. Anybody who thinks that you create a model without involving the business users from the start may face an uphill task to get the model accepted. It may be seen purely as an artefact of
    technical design rather than a model of the organisation’s business.
  2. You may consider this as a pathetic excuse for a rambling, randomly written article.
  3. As well as categories, words are often used to label things themselves; for example, ‘rabbit’ may be used to label a type of ‘rodent’, and ‘Peter’ used to label an occurrence of a
  4. Throughout this article I use the word ‘organisation’ to cover any planned and controlled group of people, whether or not it is a commercial enterprise, and ‘business’ to cover the affairs
    they undertake.
  5. The term Information Technology (IT) illustrates itself that the language used for computing has changed over time. Do you remember Information Systems (IS), or even Data Processing (DP)? Do
    you find that these terms evoke different meanings?
  6. Also, this avoids using such contentious words as entity type or object class.
  7. Strangely, whilst most organisations use the term ‘purchase order’ for those orders it places, they use the term ‘order’ instead of ‘sales orders’ for those which are placed on it.
  8. If you thought modellers had different words for the same kind of things, then be assured that linguists have the same kinds of problem too. Perhaps some of the mystique of a discipline comes
    from the hidden power suggested by new terminology.
  9. A business concept cannot be dependent on itself. This leads to a chicken and egg situation.
  10. Indeed, early repositories used to support record design were called data dictionaries. You may still find people like myself who have been around data administration for some time still
    referring to them as such.
  11. Inevitably all classifications and definitions in this manner lead to some basic idea of “being”.

Share this post

Mike Vaughan

Mike Vaughan

After leaving school to work in the money mines of retail banking, Mike Vaughan soon gave it up for the lure of labour in information technology (or data processing as it was then known). He couldn't find anything better and has now been there for years. Handsome, charming, witty and talented are none of the adjectives that can be applied to him. He lives in Hampshire, England. He may be contacted by email addressed to datahead@ymail.com.

scroll to top