Introduction: Considering Communication1
I should like, in this issue, to consider communication, the second element in Wiener's definition of Cybernetics (although I prefer the term characterisation: definition is trying to be too objective, too prescriptive). I shall consider this from the point of view of a cybernetician, NOT a semiotician (which I do not pretend to be). Semioticians might like to develop or censor this approach through their own expertise. I shall be interested to hear, and to enter into debate.
There is a sense in which control without communication is impossible. In the conventional view, somehow the indicators - the messages - of the actual and desired states of a system must be transmitted and received within that system, so that the controller (as classically determined) can compare the actual with the desired, determine any difference and what to do about that difference, and return messages in the opposite direction so that the behaviour of the controlled system may be modified to suit the wishes of the controller. When this is the view taken, communication can't be very interesting (except in the technology of encryption) as we shall see: the messages should just be passed along accurately; for that is the sole function of this form of communication.
This is essentially the flavour of Shannon/Weaver's notion of communication2: the notion of communication as the transmission of a message, transferred unaltered3 from one actor-location (the clearly defined sender) to another (the clearly defined receiver) through a channel of communication via a pair of transceivers at each end of a channel, by means of some unambiguous and determined encodement. This is the condition that I used to refer to when I insisted that communication, being a sine qua non for control, really was of very little interest (and which gave me a permission not to think about it). Of course that simplification was wrong.
Why was it wrong? Because what is interesting in communication is precisely what happens when we consider communication in other ways. When we think of communication not as simply the direct and unambiguous transmission of control acts and intentions, but as adding or facilitating more than this. When communication becomes able to be other than just coding.
Coding and Communication
Before considering alternatives to communication-by-coding, we should ask what we get from the coding notion of communication.
Of course, to a large extent what we get is what we were looking for in the first place. That is: no ambiguity, simple encodement, point to point transmission, clear sender-receiver roles and consequent action. Thus, we have the possibility of command: we are able to assume to control by the issuing of commands or orders precisely because there is no ambiguity (and because we assume that messages are acted upon correctly upon their decoded receipt). We can indeed send messages and expect them to be understood (there is no ambiguity) and to be acted on in the correct manner: the word entirely predicates the action.
This model is ideally suited to those situations in which control/command/predictability is seen as central, in which human and other agents are required not to think, not to interpret, not to understand, but simply to execute. That is, in a world without meaning(s), a world in which we deny those characteristics we (generally) most value as the most human!
Such worlds are quite familiar. We live in automation and with the amazing results that have come from the application of control principles. But we also let this approach dominate in the everyday human world. The clearest example is the military, where orders are passed and obeyed, and where any questioning is done at great prejudice to the questioner's career! Indeed, at one level, the whole of a military training is designed to remove individuality and humanness so that the soldier becomes a cog in the fighting machine. If you don't do as commanded, you are treated as an idiot/criminal. Recently, it seems, other human institutions, including, even, some of our Universities, have begun to subscribe to the same model, which is, of course, astonishingly contrary!
There are many other examples, usually where we require that what we do is completely without ambiguity because, in our model, we find situations in which we believe we cannot trust ourselves, or because we believe there is a direct cause-effect relationship outside our human agency4. For instance, in traffic management and flight control (where we act in a speed-frame that is well outside our normal perceptual and unaided/self-generated range), and in machine to machine communication (again involving us working to limits that we, unaided and unamplified, cannot manage). We spend a lot of time and effort, under these circumstances, in drilling a uniformity of understanding and action into our operators, and refining a lack of ambiguity into our messages (i.e., in command).
It is clear, too, that many of our legislators frequently act as if they believe that the world is of this nature ("objective"), which is peculiar, since the Law, as we have evolved it, is clearly not a set of unambiguous instructions - that is why we require judges to give us interpretations and refer to precedent, hierarchy of intention etc. And how often do we hear politicians insist "I have made it perfectly clear...", which can only be the case if we communicate through code rather than language. I.e. when we näively consider that language is entirely lexicographical, that words have proper (and attached) meanings - rather than that we give meanings to the utterances we perceive others to have made.
This is the world that Nicholas Rescher discusses in his (very powerful) "Logic of Commands", where he sets out the conditions in which commands may be issued and enacted. What is interesting about this is that Gordon Pask demonstrated that you need only vary one of Rescher's conditions (the predication that what is commanded will be executed) to convert the Logic of Commands into a Logic of Questions (where the predication is, at its strongest, that what would have been in Rescher's case the command, just might be carried out: that is, where it is accepted that there is doubt as to the outcome). Thus, Pask showed us that it does not take much of a relaxation in the conditions of our descriptions to open up a world of much greater interest and variety than the literal one-to-one world of command, i.e., a world of question! Opening it up means that we are no longer looking for the a deterministic outcome from an act of communication.
Coding-as-communication gives us a picture of a mechanical world. It may be complex, a wondrous clock-like mechanism, but it is mechanistic nonetheless. And it is a world that obeys: where control intentions are, in principle, translated without flaw into control actions (except when there is noise) just as we expect in our scientific/mechanistic view of the world.
This is a world in which communication is transparent, merely carrying the unambiguous message exactly and correctly. Its role is secondary, and I don't find it very interesting. It serves the purposes of control, and may be forgotten, to almost all extents and purposes. Here, communication is the means for the exercise of control, yet is (virtually, and from the point of view of the controller) without quality in itself. What qualities it has are the domain of the specialist and only appear when there is some sort of silly inadequacy (e.g. the channel capacity is too small). It never gives us room for novelty, for the development of ideas, for involvement and passion. It never objects, it is not a medium, but a sort of super-ether. And it denies meaning: for meaning, like observation, belongs in and to us. Just as Heinz von Foerster has pointed out that objectivity is the pretence that there can be observations without an observer, coding is the pretence that meanings can be communicated without "mean-ers" to construct the meanings.
Even so, the control model of communication gives the agents that own the lines of communication great power. They (may) appear neutral, but their ability to pass the instructions means that they determine which control messages will be transmitted. They have the ultimate form of censorship.
And it also presumes that everything that is worth encoding may be encoded: that is, that what is to be said is intended as a control action and that all such control actions can be encoded (rather as a Turing machine can compute all that is computable). The notion is tautological in spirit, and it misses a vast area in which we know (from our experience) that we do communicate. The area of those things that cannot be encoded, because they cannot be said (of their own nature), or they are imprecise, or they appear in and through the acts of communication (between individuals) themselves rather than existing ready to be encoded. Or when the communication channel adds something of its own (is not transparent). Or whatever else. (Thus, this presumption is also a form of censorship).
It is at this point - the point where communication is seen as possible without the unambiguous and encoded transmission of messages - that communication becomes interesting: and it is here, too, that second order cybernetics occurs, and where one of its "fundamental" models, Gordon Pask's "Conversation Theory", arises.
For we can ask, what about communication that is not intended primarily to facilitate control? What about communication about communication, communication for its own sake, communication in which it is not assumed that we pass messages unambiguously and with certain decoding to those who are to receive? What about communication as we so often experience it, as our normal experiencing?
Meaning as Conversation and Communication
The crucial concept behind Wiener's and Shannon/Weaver's original cybernetic notion of communication is coding. It is coding that allows they type of communication that they discuss, to exist and to function.
But there is another form of communication that we know of as conversation.
What conversation gives us is meaning. And it gives us communication in its own right, that is, as something which is not transparent but has its own positive qualities.
Consider, for a moment, what happens when there is error in a (first order cybernetic) system - i.e. it does not perform as we would hope: there is some perturbation, some unpredicted condition, some chaotic event, some failing in our model or in the control system which leads to error - that fundamental which is the raison-d'être of cybernetics.
We can see the error through comparison of the system's state as is and as is intended, through the transmission of messages about that state. But we already have a new value: the value that the discrepancy is, indeed, an error. I.e., we have given the discrepancy a meaning that we tag with the name error. Through the mechanism of feedback we may work (with grand intention) towards reducing and annihilating that error: but note that the judgement that we have an error causes us to change in our behaviour so that we change the behaviour we observe in the system and our consequent evaluation.
Indeed, our decision that we have error, giving rise to the quest for cause, leads to us wishing to indicate the agent of error, to give it form, pattern, purpose, etc.
Thus, even in the simple case of communication error and correction, there is not only the circularity of controlling (as discussed in earlier articles), but also there is at least the potential for evaluation and meaning.
We can design systems in which error is looked for yet in which there is no meaning (mechanical and entirely determined systems): yet it is we who have determined that this is how the world our machine will inhabit is, and how it should act and react.
Which is to say that there is a constructive base in all this, a base that is essential to all that we do (and is also a foundational starting-point for second order cybernetics, i.e. the cybernetics of observing systems and systems that make meanings, systems in which there is participation).
A constructive base clearly undermines the notion of coding as the universal means of communication. Indeed, coding becomes, at very best, merely a special - and highly restricted - case of something much more general (as Newton's mechanics are taken as merely a special - and highly restricted - case of Einstein's). For, in a universe where construction (rather than reception) is the characteristic act, each construction requires a constructor (or, as George Kelly would have it, a construer) and that constructor must be different from every other, building his own constructions.
But, under such circumstances, how could we communicate with others? We have only our own constructions. Can we share at all? Are there others5? (Our experience seems to be that there are others, in some form, and that we can share with them, so we will not explore this question further here. But how we might communicate is important, for unless we answer that, we cannot account for the type of communication we are trying to explore.)
This is where conversation comes in. The mechanism which allows us to communicate without having to encode (and all that goes with it), which not only allows but requires that we construct and are separate, is the conversation. This will be explored in detail (and after all this build-up) in the next issue.
We will discover that conversation is an embodiment of feedback. Feedback is not strapped on, it is integral. Conversation is circular. Conversation, existing without coding, requires that we make our own meanings in our own understandings. Conversation leads of necessity to (re–)interpretation, to construction; and conversation is the result of (re-)interpretation and construction. Conversation is at the heart of communication when we accede that (meaningful) communication does not happen by coding: yet conversation allows coding as a special (and very impoverished) case. And conversation is at the heart of cybernetics.
For Conversation is a (perhaps the) form for communication when communication is considered in its own right rather than as a subsidiary requirement for control to occur, when communication is seen as communication per se. The fact that there is such a means of communication and that encoded communication is not the only way communication may happen in cybernetic systems is important, not just because it is a generator of the arguments for second order cybernetics (as well as being, in the case of the conversation, a resultant form) but also because this indicates that communication is a serious and primary contributor in the Wiener characterisation of cybernetics, meriting serious consideration.
Notes
1 I would rather be able to assert, and have accepted, that meaning requires us to construct it, each of us individually, and meaning cannot, therefore, be communicated by transmission, nor can it be encoded, for it seems to obvious. And that the sharing of meaning is the central concern of communication. But I am afraid I cannot, for it is not at all obvious to everyone. So I present some reflections on coding and the Shannon/Weaver model of communication - their weaknesses and limitations. These reflections. I hope, will at least give credibility to what I shall attempt in the next issue: to discuss the apparent sharing of meaning (hence, communication) through the device of the conversation.
2 Which seems to me to take a lot from Turing's work on (de-)coding, as well as, of course, having its roots directly in Newton's work on Thermodynamics.
3 Unless, of course, there is noise in the system!
4 In case I sound negative or dismissive about this, I am not. I am as keen as the next person to survive when I am in a aeroplane, for instance.
5 I should like to bring to the reader's attention Heinz von Foerster's argument (made in "On Constructing a Reality") against solipsism: first, that if I invent you and then I invent you inventing me, I have no way of knowing that it was not you who invented me to invent you: and, secondly, that if you think you have the only right and I think I have only my own right, you may well insist you are right and I cannot disprove you, but I will desappear and you will have nothing left except your own cloud! From this he asserts that, while it is not possible to disprove your position, it leaves no future or possibilities and is ethically belittling.
Return to the content of this issue
Return to the Cybernetics and Human Knowing Homepage