A description theory

A new look at how we know what we know

Copyright 2016 Graham Berrisford. A chapter in “the book” at https://bit.ly/2yXGImr. Last updated 26/04/2021 14:13


So far, this book has been about the description of reality that occurs in the modelling of systems. For example, cybernetics is about the storage and transmission of information, in memories and messages, to describe and direct the state of things.


Part four discusses the acquisition, communication and verification of information from a more psycho-biological perspective. It discusses how we know what we know, and describe things by typifying them in memories and messages. This first chapter discusses the nature of description, and its use in cognition. It details the semantics of the epistemological triangle used in other chapters.


Description and reality (repeat) 1

Correlating descriptions to what is described. 1

The semantics of our epistemological triangle. 1

On the evolution of human cognition (thinking) 1

Our constructive epistemology. 1

Relevance to EA?. 1


Description and reality (repeat)

“Knowledge is a biological phenomenon” (Maturana). In other words, there was no description before life. The universe existed for billennia before there was any description of it. Animals evolved to perceive and describe the world because doing that helped them survive and thrive. A description is an abstraction that can be correlated with what it represents, near enough for some purpose.


We can never understand reality directly or fully. We can understand only descriptions or models we make of reality. That does not mean that “perception is reality” (as some relativists or perspectivists say), or that all descriptions or models of a real-world entity or situation are equally true or useful. There are degrees of truth in description. Is it true that I am six foot tall? It is near enough true for most practical purposes. But of course, other descriptions can be more misleading, or completely misleading.

Structures and behaviors (repeat)

We describe reality in terms of structures (continuants or objects that exist in space, and persist for some time) and behaviors (occurrents or processes that happen over time, and change the state of things). This dichotomy is so fundamental to our descriptions of reality that we draw the distinction using many words. This table of word pairs is far from exhaustive.


Structural elements

Behavioral elements


















The distinction may seem clear. However, the chapter on system dynamics discusses how, when a structure/object changes state over time, or a variable changes its value, the succession of state changes can be represented on a graph as a line of behavior.

Typification and instantiation (repeat)

A description typifies what it represents (A J Ayer). A description of one apple applies to every other apple that is near-enough similar. Even if we see only one member of a set (one apple, or one universe), we can, from its description, envisage more members.


A description is a general type that contains descriptive properties shared by each member of a set of near-enough similar things. Not only can one type describe (characterize, represent, typify) many individuals, but also, one individual can be described by (embody, exemplify, exhibit, instantiate) many types.


The chapter on description and typification explores the topic further.

The good regulator theorem

A biological approach to human knowledge naturally gives emphasis to the pragmatist view that theories [descriptions of reality] function as instruments of survival.” Stanford Encyclopedia of Philosophy


Descartes is famously said to have started his philosophy from the premise “I think therefore I am”. Psycho-biologists presume more. They presume that animals, which occupy space and live for a period of time, can perceive phenomena, remember them and communicate about them. Implicit in those premises is the idea that there was no description of reality before life.


The Darwinian view of description is that both memories and messages are biological phenomenon that evolved to help animals survive. Remembering and sharing descriptions helps animals to understand, predict and manipulate things in reality, and so, improves their chance of reproducing.


The good regulator theorem (Ashby and Conant) declares that a good regulator has a description or model of what it regulates. Here, a regulator is an animal or a machine that either has, or has access to, a model of the target that it regulates.


Regulators <monitor and regulate > Targets

Regulators <have and use> Models

Models <represent> Targets


To function and respond to changes in its environment, an animal must “know” what it going on in its world. It needs a model of entities and events its environment if it is to find food and mates, and avoid enemies. The more intelligent the animal, the richer its model. (Similarly, a business needs to know the state of things it seeks to monitor or direct.)


The question is not whether an animal (or a business) has a model of its environment; it is how complete and accurate is the model? To which the answers might be both “very incomplete and somewhat inaccurate” and “remarkably, complete and accurate enough”. Demonstrably, animals remember and share knowledge, like where food can be found. But a description need not be complete or perfect. It need only represent things well enough. A cat remembers a mouse’s features well enough to spot and catch mice. A honey bee remembers the location of some pollen well enough it can direct other bees to that location.


A description must correlate in some way to the described phenomenon. An iconic description, like a statue or photograph, mimics features of the described phenomenon, and is recognisable using the basic senses. An indicative description, like the smoke of a fire, points to effects produced by a described phenomenon. A symbolic description encodes some features of the described phenomenon using a code that can recognised by an animal or machine that knows that code.

Correlating descriptions to what is described

This section shows how describers in different domains of knowledge describe things.


Cartographers create maps (descriptions) of territories (described things). The map is not the territory, but the two must be correlated well enough to help map users find things.


Mappers <observe & envisage> Territories

Mappers <create and use> Maps

Maps <represent> Territories


Hmm… Is the territory real? Or merely another description a mental model - we form of the territory as we travel through it? It doesn’t matter to us, as travellers; it only matters that we find the map useful.


Building architects define the structures of buildings (described things) in architectural drawings (descriptions).


Architects <observe and envisage> Buildings

Architects <create and use> Drawings

Drawings <represent> Buildings


Whether the drawings of a building are best called architectural, or blueprints, or detailed designs, is a subjective decision. Other chapters address the topic of system architecture in some depth.



Have you heard that Einstein didn’t shine at school? The truth is, his insights were deeply informed by extensive study of mathematics, which he mastered at a young age.


"Max Talmud says that after he had given the 12-year-old Einstein a geometry textbook, after a short time he “had worked through the whole book. He thereupon devoted himself to higher mathematics... Soon the flight of his mathematical genius was so high I could not follow." His passion for geometry and algebra led the 12-year-old to become convinced nature could be understood as a "mathematical structure". Einstein started teaching himself calculus at 12, and as a 14-year-old he says he had mastered integral and differential calculus". Wikipedia


Physicists <observe and envisage> Nature

Physicists <create and use> Mathematical structures

Mathematical structures <represent> Nature


Throughout his life. Einstein distinguished his mathematical descriptions from the reality of the universe, and declared that it is impossible to directly comprehend the reality.


“There are several kinds of theory in physics. Most of them are constructive…When we say that we understand a group of natural phenomena, we mean that we have found a constructive theory which embraces them.Albert Einstein, The Times (28 Nov 1919)


“As far as the laws of mathematics refer to reality, they are not certain; and as far as they are certain, they do not refer to reality.” Albert Einstein, Sidelights on Relativity (1920), 28.


“Physical concepts are free creations of the human mind, and are not, however it may seem, uniquely determined by the external world. In our endeavour to understand reality we are somewhat like a man trying to understand the mechanism of a closed watch. He sees the face and the moving hands, even hears its ticking, but he has no way of opening the case. If he is ingenious he may form some picture of a mechanism which could be responsible for all the things he observes, but he may never be quite sure his picture is the only one which could explain his observations. He will never be able to compare his picture with the real mechanism and he cannot even imagine the possibility or the meaning of such a comparison.”

Albert Einstein, The Evolution of Physics (1938), 33



Mathematicians define a type (like “point”, “line”, “triangle” and even number) to describe the properties of a thing that instantiates that type.


Type name

Type elaboration

“Even number”

“a number divisible by two”.


A type is a pattern for (one member of) a collection or set. An instance is an exhibition or embodiment of a type in a described thing a set member.


Mathematicians <observe and envisage> Set members

Mathematicians <create and use> Types

Types <are instantiated in> Set members


Two mathematicians had a famous debate about the meaning of descriptive types or axiomatic assertions. The debate might be distilled as follows.


Frege believed mathematics is carried out at the level of thoughts about real things, rather than descriptions of them. He believed descriptions are imperfect representations of our thoughts. We “know” what a geometric entity (like “point”, “line” and “triangle”) is without reference to a description of it, or how it relates to other entities. And every detail of it may be relevant to answering questions about it.


Hilbert said that, even if we do know other details of an entity (say, its color), they are entirely irrelevant to understanding geometry, which merely defines relations between entities. Geometric descriptions are holistic in the sense that they define how geometric entities are related, rather than what those entities are or contain; they can be anything (large or small) that is relatable as described. Mathematics is carried out at the level of descriptions, regardless of what the entities are in reality.


According to the Stanford Encyclopedia of Philosophy, Hilbert is now regarded as the winner of the debate. Geometry does not address the whole of a thing (Frege); rather, it addresses only those features of it that are describable by geometry (Hilbert).



Ashby’s cybernetic is about creating an abstract system (a description) to represent the behavior of a real machine (a described thing).


Observers <observe & envisage> Machines

Observers <create and use> Abstract systems

Abstract systems <represent> Machines


A real machine can be only be called a system in so far as it is correlated with an abstract system. The remainder of the machine is either beyond our ken or not described as a system.

Enterprise and business architecture

Enterprise architecture does not address the whole of an enterprise; rather, it addresses selected business operations. It is concerned with business activity systems in which business roles and processes create and use business data.


Architects <observe and envisage> Business operations

Architects <create and use> Abstract systems

Abstract systems <represent> Business operations


Architecting is a process that encodes, decodes and compares descriptions of baseline and target systems. At design time, general entity types like customer, event types like order and payment, and process types like billing are defined. These types are related to each other in logical data structures and process structures. E.g. order value = order amount * unit price. At run time, physical systems consume inputs that carry information about entities and events, record those entities and events with attribute values, and use remembered information to decide how to respond to events. In all these ways, an enterprise can be seen as imitating what comes naturally to animals.

The semantics of our epistemological triangle

Epistemology is about what we know of reality, through observation, description, testing, reasoning and learning from others. The constructivist epistemologist distinguishes descriptions (constructed) from described things (observed or envisaged).


We have no direct knowledge of things in reality. We know them only via descriptions of them that we construct (in the mind, in speech, writing, mathematics, whatever), which are (ultimately) associated with our sensations of the world.


We (describers) can only understand physical reality in so far as it is correlated with a description we have access to which we can correlate well enough with real-world phenomena for practical use.


A describer is any animal or machine that can observe or envisage something, and create or use (encode or decode) a description of it. A phenomenon is any aspect or part of reality that a describer can observe or envisage. A description is a representation (in mind, speech, writing or other form) that is correlatable with a phenomenon.


The three concepts are relatable thus:




<create and use>          <represents>

Describers     <observe and envisage>    Phenomena


Note that each pair of concepts is related in a many-to-many way. One describer can create several descriptions of the same thing. Conversely, several describers can contribute to one recorded description of the same thing. It may well be that none of those describers knows whole description, which can be larger, more complex, consistent and informative than any human memory can be.


The view of description and reality above seems obvious to some and obscure to others. The semantics of the triangle are explored below.


Descriptions <represent> phenomena

Descriptions embrace every kind of model an animal or machine can make of the world, be it a mental or documented model, a memory or message, in speech or writing, a 2D picture or 3D model, in the brain or in stone. Here, we mostly discuss descriptions expressed symbolically, and verbally.


Phenomena include everything that can be observed or envisaged, including descriptions and describers.


Two descriptions of the same phenomena may be compatible or in conflict. Is light rightly described as waves or particles? Physicists do not say either model is the “true” model, they say only that each can be useful.


Describers <create and use> descriptions

Describers are animals or machines that can encode and decode descriptive models of phenomena. Descriptions are given meanings in the acts of encoding or decoding.


We speak of the describer in the triangle as an animal or a machine. Strictly, it that actor’s cognitive processes its ability to observe and envisage, and create and use descriptions at the apex of the triangle.


In short, to create a description is to encode a model that represents some feature(s) of a phenomenon. To use a description is to decode it, then use it to respond to or manipulate whatever is described.


Describers <observe and envisage> phenomena

To observe or envisage things is to create and use descriptions of them.


Both describers and descriptions are also physical phenomena. In observing a concrete house, a describer is creating and using a description of that house, which exists in a physical form, and can be translated from memory to message and back again. Similarly, in envisaging a purely fantastic thing, such as a flying elephant, a describer is creating and using a description of it, in some physical form.


It is difficult for describers to observe the largest known prime number, because it is so long; still, it exists in the memory of a computer. And describers can envisage the next prime number beyond today’s largest known prime, as a future instance of the abstract type the describer has in mind or in writing. In short, to observe or envisage things is to create, use and compare descriptions of them.

On the evolution of human cognition (thinking)

Cognition involves acquiring knowledge (through the senses and thinking) and capturing it in descriptions of the world for later use. In our triangular epistemology, we could replace the describer (animal or machine) by cognition, thus.


Cognition <observes and envisages> Phenomena

Cognition <creates and uses> Descriptions

Descriptions <represent> Phenomena


The psycho-biological view of thinking is called “cognitive embodiment”, meaning that the brain is inseparable from the rest of the nervous system. Thinking is not done in the brain’s grey matter alone. Below, distilled from the next chapter, are some staging posts in the evolution of human cognition.



In molecular memory, organisms sense and respond to molecular structures.

In primitive animals, sense-response thinking is an end-to-end input-to-output process (cf. a value chain in business architecture). It begins in perception by a bodily sensor, where the state of things is first described by encoding a neural message; and ends with a direction to the body’s motors and organs, where a message is decoded and used in actions to change the state of things.



In neural memory, animals remember things they have perceived through their senses.

In higher animals, with memories, perception is a mix of observation (sensing) and envisaging (guessing) at what there is out in reality. Both observation and envisaging are processes that create (encode) and use (decode) descriptions in memory.



In social interaction, animals first used fixed format messages, like alarm calls.

Animals cannot not only inherit and remember knowledge, and learn from experience, but also communicate knowledge. Even very primitive animals signal mating intentions to each other. Other early social acts were related to marking territory, signalling danger and locating food. E.g. Cats spray scent to mark their territory and other cats smell that scent. By 100 million years ago, some animals had evolved to cooperate in groups by communicating descriptions of things to their fellows.



In consciousness, animals compare descriptions of the past, the present and possible futures

Every remembered description of the past serves as a type that defines (one member of) a set that may contain more members in future. Envisaging the future involves creating and playing with descriptions of possible futures. Consciousness enables an animal to compare descriptions of past, present and possible future phenomena.


Speech and symbolic language

In speech, humans encoded complex messages in sounds.

In humans, memories are translated into and out of verbal messages for communication. Ashby observed that in thought and communication “coding is ubiquitous”. The multiple translation steps involved in social communication are illustrated in the next chapter.



In writing, humans recorded ever more complex descriptions in a persistent and shareable form.

The proposal here is that thinking is a process that encodes, decodes and compares descriptions. Moreover, in humans at least, this is process is not entirely internal. I observe my own thinking process is partly externalised via writing. It involves a circular feedback loop in which I encode some half-baked thoughts in written words, then read back what has been written, test it for consistency and coherence, and correct or clarify it. Writing makes it possible to document a description of the world that is more complete, consistent and coherent than any I can hold in mind.


Moving data from the brain, to the hand, to the keyboard, to the screen, to the eyes and back to the brain, is a succession of processes that translate data encoded in one form to data encoded in another. In short, cognition is observing and envisaging - is creating and using descriptions - is a process that encodes, decodes and compares descriptions, both internal and external.



In science, humans refined how to form theories, predict outcomes and test them.


Artificial intelligence

In machine learning, human-created machines abstract descriptions from phenomena.

AI software can now abstract types/patterns from things it observes as sharing features.


Human and computer actors process data in different forms and ways. A human’s memory, being encapsulated in a brain, is inaccessible to others (unless translated into a message). By contrast, in software, memories and messages are merely varieties of digital data storage; both can read/written by many software components.


Note also that software can read a memory without creating one. By contrast, to read a human a memory is also to create a new one. (Which is one reason why human memory is less reliable, as in “false memory syndrome”.) Nevertheless, both human and computer actors can be reasonably be thought of as writing and reading creating and using - memories and messages.

Our constructive epistemology

The proposal here is that to observe or envisage things is to create, use and compare descriptions of them. Our position on description and reality has been set out in this chapter, and can be represented at an overview level in this triangle




<create and use>          <represents>

Describers     <observe and envisage>    Phenomena


There is an essential difference between our triangle and comparable triangles you can find in semiotics and philosophy. Most other triangles separate descriptions into two kinds: internal models (descriptions in the mind) and external models (descriptions in speech or writing). By contrast, our triangle separates cognition processes (observing, envisaging, remembering, recalling, writing and reading) from all the descriptive structures those processes create and use (in the mind, in speech, on the page, wherever). So, a description in the mind is at the apex, not the left.


Philosophy has moved over many centuries from a position in which descriptions perfectly capture the real nature of things (Plato), to positions in which descriptions are, at best, approximations to things. Some propose descriptions are convenient fictions to organize sense data. Others propose all descriptions are untethered from external reality, and equally justified. To decide where you sit in the mess of countless philosophical positions leads you into the endless morass of philosophical debate, which you can find in the Stanford Encyclopedia of Philosophy.


This book may be read as endorsing three philosophical positions. Instrumentalists say models are instruments of prediction. Pragmatists say models are concepts or artifacts used in producing scientific knowledge. Constructive empiricists say models are symbolic representations of empirical phenomena.


Some say these three positions are “anti-realist”, meaning they deny that models give a true description of reality.  Here, it seems futile sophistry to deny any of the following


a)     Reality does exist.

b)     A description can be true empirically (enough to be useful) or logically (a consequence that follows from some axiomatic assertion)

c)     We can share descriptions and so share some knowledge of reality.


Our main interest is in symbolic descriptions, which are encoded using symbols (or signs) and in how animals and machines create and use them to describe things. In our epistemological triangle, we could replace description by symbolic representation.


Chapters after this explore the evolution of cognition, the use of types to describe things, how we share knowledge and verify truth, and challenge alternative views of description and reality and philosophical positions.

Relevance to EA?

Musical composition

Composers (describers) define the notes and instruments required for symphony performances (described things) in symphony scores (descriptions).


Composers <envisage> Performances

Composers <create> Symphony scores

Symphony scores <represent> Performances


The composer does not address the whole of the social entity that is an orchestra. The composer addresses only the roles that orchestra members play in sounding musical notes, and the kinds of instrument they need to play their roles. All the documented role, note and resource types add up to a complex type we call the symphony score.



Symphony scores

<create and use>                          <represent>

Classical composers <observe and envisage> Symphony performances


At run time, conductors use a score to drive an orchestra, beat by beat, to perform the symphony in a discrete event-driven way. Thus, supervised by a conductor who keeps things in order, an orchestra realizes the abstract symphony score in a performance.

Business architecture

Similarly, the designer of a business activity system cannot address the whole of the social entity that is a business. The designer addresses the roles that business actors play in a performance of the business system, and the kinds of resource they need to play those roles.


The designer describes entity types like customer, event types like order and payment, process types like billing, and rule types like order value = order amount * unit price. All these entity, event, process, data and rule types add up to a complex type we may call an abstract activity system.


Business activity systems

Abstract activity systems

<create and use>                     <represent>

Business architects <observe and envisage> Physical operations


At run time, supervised by managers who keep thing in order, business actors realize the abstract activity system in a physical activity system that consumes inputs, creates business data and uses it to decide how to respond to events.


In cybernetic terms, a business activity system of this kind can be seen as a machine that imitates what comes naturally to animals. It detects events, interprets their significance in the light of memories retained, and responds by acting in a way that serves the interests or aims of the business, machine or animal.