Before We Begin
In attempting to introduce Pask’s notion of conversation, and to show that it is both more fundamental and more economical (Occam’s Razor) than older notions of communication, there were a number of loose ends I promised I would return to and tidy up, which I attempt here. This is not pure Pask, or even my interpretation of him: it is my (conversational) response to him, an extension. In particular, I attempt to deal with what I see as problems in conversations which are not, somehow, controlled, and with what the conversational paradigm tells us about ourselves.
How Do We Know What We’re Talking About In A Conversation: And How Do We Get Back On The Rails When We Leave Them (Error Correction)?
I had a long-term disagreement with Pask on this subject, ie the question of how participants in a conversation decide what to talk about (the "Topic" of the conversation, as Pask called it) and of how they should correct the "errors" of understanding (ie, the construction of meanings by the listener that were not recognisable to the speaker as expressed in the listener’s response). He claimed these were not special problems but could be dealt with in the conversation, as he depicted it. I disagreed. While it is conceivable that nothing special need be done, my opinion is that, without some mechanism, error regulation is in effect random and inconceivably clumsy. Participants in a conversation would know they were not in agreement but would not know, other than by hit and miss, how to resolve the situation. There would be no way to direct the next trial after a previous one had failed, no way to reduce the error and home in!
In the first instance, a conversation needs to have some shared focus (ie, a general topic). It can wander from this, of course, but it needs to start off establishing the idea pretty fast that there is, at the very least, to be a conversation (ie, conversation is the topic). Even if the focus is just to have a conversation.
The difficulty lies not so much in the initial choice, but in the error correction mechanism. It does not matter much what initial hunch you make for your topic so long as you can effectively reduce the distance between you and your conversational partner. (The old panel game "Twenty Questions" is a demonstration of how powerful even binary partition can be in helping narrow down differences.)
I propose that error regulation and correction in a conversation is best thought of as occurring through a meta-level. This is where I differ from Pask, who did not believe that such a special mechanism was necessary. (Nor is it, if we have an infinite quantity of time.) If the conversation is taken to operate on one level, then the error regulation must happen on another, meta-level—where it is also possible to hold a conversation. The inclusion of the meta-level allows that not only is an error recognised, but the error can become a topic of the conversation, and therefore be discussed. (A conversation, thus, being held on the level of the meta-level: a conversation about a conversation.) Without the structure of this discussion (used as a means of characterising the error), error regulation is merely by trial and error. And, in spite of Twenty Questions, this can be very time consuming and tiresome, destroying the flow of the conversation.
Note that, in this version of error regulation (correction), the conversation has shifted topics. There is what is being talked about, and above it there is a critical examination. And when there is an error, this allows that error in turn to become the (temporary) topic of the conversation, permitting the nature of (and hence the cure for) the error to be discussed and acted upon. I claim this is not practicable without the meta-level, and that the topic can switch, so that the "commentary" of the meta-level can become the topic of the conversation (with a further critical (meta-)meta-level above it). Thus, what was the meta-level now becomes the level of the conversation, with another meta-level above it. But, meanwhile, below, at what was the original level, this is a substratum where the original conversation is kept (the context within which the error is discussed). And for each conversation there is a substratum, that is, a context against which the conversation takes place: the general topic of the conversation. Three words indicating the similar ideas but, themselves, with different "intentional contexts". (See, for instance, my 1996 paper "Communication without Coding" in Modern Language Notes Vol III no 3.)
Thus, to permit effective error regulation in a conversation, I believe three levels are needed. The level of the conversation, the meta-level, in which monitoring and error regulation take place, and the substratum, which provides the context. At any moment, the conversation may be raised so that what was the meta-level becomes the level of the conversation with, respectively, a new meta-level and substratum above and below, or be lowered so that the substratum becomes the conversation.
Reaching an Agreement
The purpose of conversation (insofar as it is necessary to talk of it as having a purpose), and of the error regulation structure proposed above as an extension to Pask’s formulation, is to permit participation and the building of meanings such that participants can retain their individual meanings yet still communicate. The error regulation structure proposed above facilitates what would be, in Pask’s original formulation, very clumsy.
The reason for having error regulation is that it allows agreement to be reached. Conversations continue until an agreement is reached (they may continue afterwards), or they are abandoned. Abandonment is an extreme form of disagreement: and agreement includes the agreement to disagree. Thus, conversations, while they may freewheel for long periods and from topic to topic are inevitably tied up in time: they start, and they end. (Pask’s later work, as I have mentioned, centred on the form of the conversation when it did not begin or end but was eternal: the structure of our constructed world. He called it the Interaction of Actors Theory.)
Novelty
Novelty is inherent in the conversation. Indeed, it is virtually impossible to avoid. Since no participant is the same as any other, his/her understanding (the meaning he/she constructs) will be different (that is, uniquely personal). Therefore, no one understanding is the same as any other. Thus, in-so-far as it is possible to present one’s understanding so that other conversational participants can construct their own meanings, what one presents will, in principle, always be different to what any other will present: hence, novelty (that which is not in my current understanding) is endemic. (Note that this concept of novelty is local: it is not necessary that what seems novel to me is universally novel, in the way we sometimes require. Novelty, as I use it here, refers to personal novelty.)
Such novelty may be glossed over. But it may also be pounced upon and exploited, even exaggerated, to create novelty of a dramatic and more-than-personal kind.
Words with Meanings: Language as Short Cut
I have insisted that words have no meanings: meanings lie in the "we-who-have-and-talk-about-meanings", not in the media of communication. Nor does it matter what I mean to say, (my own meaning): the meaning that each participant in an act of communication makes is his/her own. But above, for instance in discussing novelty, I was using words as if they do have meanings and as if I can specify which meaning (where there are several), and/or propose modifications of such meanings.
At one level, I believe this is irrefutable in terms of both logic and experience (which are not necessarily coincident). We know our experiences are different, and we know how hard it can be to understand even close friends in a familiar conversation. At another, we have dictionaries: and we believe in them (as I write, I have used a dictionary twice already today to help me check words in some texts I am editing).
My explanation for this is rather straightforward. In spirit, it is close to Luhmann.
There is clearly a way in which we treat words (and any other devices we use as media of communication, and the units thereof) as if they do have meanings although they do not. How can this contradiction be resolved?
It seems to me that, in negociating an agreement with a conversational partner, even with the error regulation structure I have proposed, we are involved in a very long-winded business. And it’s very time and effort consuming when we have to go through the same business with every partner and on every occasion. Humans might be characterised as repetition generators and as simplifiers: that is, we find patterns which simplify the random continuum, and we note these patterns again and again. That is a major part of how we learn (from experience). These processes allow us to construct the world as regular, and we do it with flair, frequency and success!
It seems that this is a reasonable account of (some of) what we do. Then why not do it in conversation? If I wish to present again what I have presented before, perhaps to the same conversational partner, why not present it the same way? Thus, repetition of presentation in the conversation can become fixed: and that fixity can be tried with other partners. Indeed, the trying can be most insistent. And when other partners agree to a similar presentation, it does indeed appear as if the means of presentation has a particular meaning: which we take to be no longer as if, but actually the case. Observing more regularities (on the meta-level, as well) allows us to guess new presentations that may work right off. And so on.
When we do this, we agree a sort of social fixity in presentation in the conversation, which makes meaning appear to be in what we have fixed—in the words, for instance. What we end up with is language as a shortcut, a collection of social conventions which become treated as if they were meanings resident in words etc. Meaning, in this sense, is a social construction. And, in this sense only, we can talk of meaning in language, and develop social studies such as history and etymology. Of course, within these conventions we still negociate, but we do not have to negociate everything on every occasion with every conversational partner, from scratch. Thus, we have a social system: to join it we have to accept all that system’s many rules: a down-side, for it involves a loss of choice and a prediction of labels and meanings.
But there is more to the down-side of this fixity. Firstly, it restricts us in how we may make presentions: it does me little good to present an ovoid form emanating from a chicken as a "bolt", when everyone else calls it an "egg". This works at a much deeper level, however, for each of us has to sign on to join the already existing (ie, prior) system. So, as children, we have to learn not just how we want to construct the world, but also how this prior social system will permit us to: for conversing is a (the?) major part of our experience. We have to pay our dues to join in what is a very determining system. We can’t see this in our mother tongue (how appropriate that expression is). But we know it full well when, as adults, we try to join in another system (that is, another language).
(In the extreme, we actually try to program words so that they have unique and "unambiguous" meanings. We do this for situations that move very fast, where team efforts are required and where decisions are not made locally but centrally. We insist there must be no ambiguity. The clearest example is the military. A recruit’s initial training is involved in breaking his concept of himself as independent. It is mind-destroying, and would seem to be so on purpose. The aim is to produce a fighting machine in which each soldier becomes a cog. Each soldier is trained to reduce his individuality and distinction. Communication is no longer by conversation, in which he is a participant and in which the integrity of his meanings is respected and valued. One "benefit" is that he will then do exactly what he is told to, and that orders will be followed unambiguously. Each order—each word—will unambiguously command a particular response from the soldier. In destroying the "ambiguity" of words, in giving them exact "meanings", we destroy the quality that makes us human, intelligent and interesting: our distinction and our meaning. It could be said that the turning of communication from a conversation into a code destroys humanity. It is certainly a feature of fascist regimes that language becomes both horribly distorted, and also highly codified, coded and encoded. Of course, this doesn’t actually really remove our need and our responsibility for making our own meanings, but it surely seems to.)
Secondly, it lets us delude ourselves as to where meaning lies—that meaning lies in the "words" and not with us. This is the source of our treating language like a code, and it reinforces our belief in linearity, cause and effect and coding. It is an outstanding example of how we use the shortcut of language without remembering its origin and where its value lies. Thus, we forget that we are speaking as if, and we even come to believe that word f is bound by encodment to meaning x.
Meaninglessness. And Art.
In contrast, those means of presentation that have lasted a long time and have had continuing value over centuries and for millions of people in myriad cultures, must be taken to be at least highly ambiguous, or, as I choose to put it, meaningless. A great work of art can have no meaning. It must be an explicitly empty vessel into which we can invest whatever meanings we individually construct. This represents a timely return to the prime notion in a conversation, before society and language got into the act and created the delusory shortcut: that the means of presentation has no meaning. Listen to the interpretations of great works of art, and you will find vast variety much of which, in a logical world, would be mutually exclusive. Great art returns to us our responsibility to create meaning for ourselves. This is why John Cage was correct when he said that, as a composer, it was not up to him to impose his will or interpretation on his music, for he had no more right or rightness than anyone else. It is this which makes, for instance, authorised recordings of music, so inherently sterile.
What we say has no meaning (other than to us). We do not communicate what we mean. Our partners in a conversation make their own meanings. We can only try to make and re-present to our conversational partners our meaning made from what we imagine they mean. We can never share meanings, nor can we transmit them. We are alone.
And yet, there is a strange experience that we may have, when we loose ourselves in the presence of art (and other phenomena, for instance crowds, sex, ritual). When we do not distinguish the "I" any more, and when, therefore, the question of meaning etc disappears. Then we seem to share, we are united somehow in a whole, perhaps because we are all where there is no "I".
What Does the Conversation Tell us about Ourselves?
There is a final argument I wish to present in favour of the conversation. In order for us to take part in a conversation we have to show some of the characteristics we believe to be most profoundly and attractively human.
For instance, I cannot take part in a conversation with you, with any hope of negociating an agreement, unless I show open-mindedness and generosity: open-mindedness because I must be prepared to listen to what you say, and generosity because I must accept that you put things differently and that I cannot insist on having things all my own way. (If I do insist on this, I end up with no conversation. We have all taken part in these sorts of "conversations": often when we give but our conversational partner only takes, but also sometimes as the one taking and not giving.)
Furthermore, we have to take responsibility for our own meanings, and hence our own actions. This is not only because there are no meanings but the meanings we (each of us) make. It is also because when the form is circular and I am in that circle, there is no stop point but the one I choose: and the only one I can choose is where I am.
From this comes the ability to recognise error and magnanimously accept that we are wrong. This is one of the essentials for us to be able to learn. Error is a very positive phenomenon: the ability of cybernetics to accept error for what it is, and not try to get rid of or minimise it (ie, not to demonise it) is one of its great contributions. As also is the need to consider others as reciprocal and find reciprocity in them: to see others in our image.
I contend (see above) that the use of the encoded message approach to communication and to what we do and can communicate reduces these qualities to insignificance, offering us, instead, correctness, conformity and narrow-mindedness; blame instead of responsibility; meanness and carping criticism instead of generosity of spirit; monologue instead of dialogue, conversation, and novel inspiration and joy. And it allows us to denigrate others rather than to treat them as our equals (and, hence, eventually, to be denegrated by them in turn).
I am reminded of this every time I think of old relatives—trapped in the world of senile dementia. There is no communication here, for conversation is no longer possible. There is no interaction: there are just, possibly, private codes. There is no give and take. There is just an outpouring: from me, just as much as from them, for I do not know how to transcend this breakdown. Private codes uttered without connection. A Beckettian voice from a fixed mouth ploughing through the darkness. Endlessly screaming, "Not I".
I have learnt where my humanity lies. Thank you, Gordon.
Notes
1. Until recently, I was a staff member at the University of Portsmouth in the UK. However, I have accepted early retirement, and am starting the challenge of spending the projected last third of my life trying to learn to do what is good for me, rather than what is demanded of me. I feel I leave education at a good time: before it is entirely destroyed. This does not, however, mean I am leaving the world of cybernetics, or am no longer interested in academic and teaching work: I am open to offers! Please note my details: 52, Lawrence Road, Southsea, Hants PO5 1NY, UK. Tel +44 1705 737779; fax +44 1705 796617; email: ranulph@glanville.co.uk.
2. Gordon Pask was a generous teacher. He did not require agreement and he did not stand on his dignity. Rather, he looked to his students to develop themselves through the discipline of cybernetics. He looked to the imitation of his best standards. Some have been angry that I would disagree with my professor. My professor was not among them. Pask revelled in such disagreement and he and I would, on occasion, develop and orchestrate disagreements in order to open/stir things up a bit. He was interested in education, not training, as befits the visionary of the uses of computers in learning.
3. The notion of negociation in a conversation is probably due to Laurie Thomas. Conversation Theory was developed, in good part, through seminars held at Brunel University (as well as in Pask’s research company, System Research). As well as Pask’s own seminars, he shared a seminar series with Thomas, in which many of the ideas here propounded were teased out and fought over. It is not always clear who contributed what. (The contributions of individual students are even less clear.) See, for instance, Thomas, L and Harri-Augstein, S (1993) "Learning Conversations", London, Routledge.
4. This is what breaks my heart about my favourite writer, Samuel Beckett, who set up a censorious orthodoxy to mind, mindlessly, his works.
Return to the content of this issue
Return to the Cybernetics and Human Knowing Homepage