Conversation Piece

Emergence requires the dissipation of some flux. Life on Earth improves the efficiency with which incident solar radiation is dissipated. Social institutions thrive on the flux of human conversation.

We have conversations about anything and everything, and especially voluminous conversations about the many ambitious claims of would-be foundational disciplines, from theology to economics. But we rarely have conversations about conversation. This piece contends that understanding the power of the flux of conversation will be an unsurpassed tool for tackling the opportunities of the new millennium.

For this purpose, it is not beneficial to restrict our notion of conversation just to face-to-face and telephone dialogues, but rather to include the many other forms of mediated conversations as well as the full range of numerical relationships between participants. You might be one of the large audience to a drive-time conversation between a radio announcer, her specialist guest and their talk back callers. Then when you get to your destination and log in, you might converse with a search engine to find the Universal Resource Locator that you could not recall exactly from the radio.

Of central importance is the unending conversation inside your head. Your language and association skills rationalise, anticipate, rehearse and recall your perceptions of and actions on the world. That constant chatter also blinds you to much of the animal intelligence of our distant ancestors and many species with which we share this planet.

This piece further contends that the Internet provides the basis for a qualitative change in our conversational capabilities. In coming rounds, she who can converse the most effectively wins. This discussion has been particularly inspired by the landmark 1996 judicial finding that the Internet is a "never-ending worldwide conversation".

Conversation and Communication

Why frame an argument in terms of conversation when there is already a vast body of knowledge and extensive ongoing research in communications? Surely we are talking about the same thing? No, definitely not! The essential point of shifting our frame of reference is to gain a new and productive perspective which cannot be readily seen through communication theory.

In the mid-1980s, Terry Winograd and Fernando Flores made a start on this journey through their discussion of "conversations for action" which was particularly concerned with utilising the power of the "threads" that are an emergent property of conversations by electronic mail. Winograd and Flores wrote a dense philosophical grounding for their ideas, Understanding Computers and Cognition, and developed some influential but commercially less successful software, The Coordinator, intended to exploit and enhance the power of conversations in the then novel medium of computer networks.
Having waded through Understanding Computers and Cognition, and having thereby been awakened to the potential relevance of post-modern philosophy to understanding the information technologies, the author of this piece went on to obtain the Australian agency for Action Technologies Inc., the company responsible for commercialising The Coordinator software.
The distinction here between communication and conversation is firstly a distinction between ends and means, or between outcomes and processes. Communication theory is primarily concerned with effective movement of information, where both the communication and the information are treated as mensurable, at least in theory if not in practice. On the other hand, conversation can be seen as a fundamental activity of human society, as an essential part of its fabric, and, by virtue of its efficiency at exploring possibility space, as the essential flux that powers the emergence of social institutions.

Formal studies of communication, information and media are very new and naturally quite narrow, even while their objects of study are ubiquitous and diverse. Our long nurtured familiarity with the subtleties of conversation has produced a vast vocabulary for labelling its multiplicity of forms, while that same process of familiarisation has rendered the totality of conversation almost invisible to critical examination. The essential purpose of this piece is to take a further look at the totality of conversation in the novel context of the Internet so as to see what is likely to emerge.

Judging the Communications Decency Act

In granting the order preventing implementation of the Communications Decency Act, Judge Stewart Dalzell found that the Internet is a "never-ending worldwide conversation" deserving of maximum protection from government interference. Searching the Internet for the exact quote and context quickly revealed many electronic copies of the full judgement. A quick scan of the judgement text prompted a sudden realisation that the legal community are the purest exponents of conversation as a modus operandi.

All the processes of law are based entirely on prescribed forms of conversation, from the drafting of legal documents to the passage of legislation, and from the records of interview used as evidence to the proceedings in a court of law. Both objects from and non-verbal representations of the real world are only admitted as exhibits and not as real participants in the legal conversations. It is little wonder that as the most dedicated exponents of conversation, the lawyers treat law as the foundational discipline which provides them with a true and total understanding of the world.

Through their training and work, lawyers are inducted into a belief in the special status of the law itself. Sufficient of them follow a path into politics to ensure that legislatures keep producing laws which attempt to deal with anything and everything we might ever encounter. While the checks and balances in the system can obstruct such evils as the Communications Decency Act, they totally fail to turn the tide of legal expansionism into more and more areas of everyday human activity.

Maybe we are being over optimistic or just simplistic, but there is a growing countertide of opinion that the speed and nature of the changes flowing from the Internet and related technologies may continue to outrun the lawyers. This may be possible if money truly speaks louder than words, as there are vast commercial interests lining up behind the technologies. It may also be assisted by some tough challenges to foundational premises of the legal system such as jurisdictions and one that is the prime suspect of this piece--personal responsibility.

Forms of Conversation

While the lawyers have taken conversation to extremes, it is still an indispensable tool for the rest of us. If you do not instantly recognise this truth, then you should go out into the world for a while with your ears and eyes open. A couple of kids from the same kinder class meet by chance at the local shops with their mothers on a Saturday morning. What do they talk about? You wouldn't really want to know, but their earnest conversation will likely infect their mothers and, more often than you might expect, catalyse a long term friendship.

Breaks in the school day are filled with often intense conversation between age mates through which they largely construct their own personas and produce their own local variants of common forms of interaction. In the traditional classroom, conversational energy must be forcibly suppressed so that most conversation can be directed from the one teacher to the many pupils, mainly in an attempt to standardise the continuing conversations inside the heads of the pupils.

Love itself, and intense friendship, which together weave the very fabric of society, are constructed largely through conversations of teenagers on the telephone or listening to music and watching TV as these one-to-many media project standardised images of relationships and desirable partners.

In the workplace and in the community, most things are determined in meetings. That essence of democracy enables groups to go about their business in some confidence that their individual efforts will fit within and contribute to some greater whole. The transfer of the locus of decision making from the manager informed by hierarchical conversations to the consensus of interested parties meeting together has powered the shedding of middle management which has characterised the decade.

Mature adults outside the workplace and away from formalised community activities enrich their time in conversation with old friends, relatives and casual acquaintances. They chat in public and private places, on the phone and often still by correspondence. And when you are able to participate in such conversations, you find that they are still about understanding the world and sharing that understanding, unencumbered by any quest for more specific outcomes.

As well as all this relatively free-form conversation, conversations between institutions and large numbers of individuals have been increased in frequency while being impoverished in content through the ever-increasing use of prescribed forms. Forms for registration, application, experimentation, investigation, examination, nomination, taxation and voting restrict conversation to chunks which can be readily tabulated in databases, categorised and counted.

Conversations on the Internet

Data communication networks extend both the flexibility of conversations mediated by the telecommunications networks and the capacity for data-based conversations which is facilitated by computers. As the network of networks, the Internet has explosively enhanced this flexibility and capacity. It is worth noting here that of all the processes identified in the world, the growth of connectivity of networks has the fastest relative peak rate, a fact which is readily demonstrated by simple models of network dynamics.

Conversations on the Internet can be largely grouped under three different time scales which exhibit three quite distinct dynamics. The most immediate is chat where two or more people connect simultaneously to the same virtual location and converse at typing speed. Another level is provided by the exchange of prepared messages on a timeframe from hours to days via various kinds of bulletin boards and personal e-mail. And those longer term conversations which have been the province of the literature are being subsumed by the construction and interlinkage of World Wide Web sites.

Chat alone is facilitated by three major classes of mechanisms. Internet Relay Chat provides a limitless choice of named and private "channels" which chatters can join. Some, generally larger, Internet access providers provide more tightly purposed "chat rooms" as part of their bundle of services. And the Multi-User Domains that host group projects from role playing to course delivery are basically a collection of interconnected rooms which readily host private and group conversations. Two people sharing any private chat area have every opportunity to turn their conversation to virtual sex. The delays inherent in exchanging typed comments often lead to the emergence of two parallel strands as each party responds to the other's previous comment. In larger groups, various mechanisms are employed to help maintain coherent conversations.

Bulletin board conversations are facilitated by a similarly wide range of mechanisms which have equally little impact on the forms that those conversations follow. The main mechanisms are Usenet newsgroups which propagate automatically and thus can be tapped into all around the Internet, mailing "lists" which an individual joins so as to send and receive messages for the whole group, and, more recently, Web boards which allow a Web site to function as a bulletin board. In this distinct world of bulletin board conversations, such distinctive Internet phenomena as threads, flames, spams, lurkers and FAQs arose naturally. For any board an accessible and comprehensive list of Frequently Asked Questions and their answers is the first step in the battle to avoid endless recycling of old conversations as new participants join the board. Personal e-mail provides for private conversations to be pursued which would not be appropriate for a public forum while retaining the advantageous dynamics of bulletin boards.

In contrast, the World Wide Web is expressly intended to provide a unified system for accessing all manner of information resources and services on the Internet. However, the fundamental nature of the Web's essential device, the Universal Resource Locator, and the relatively maturity of preexisting chat and bulletin board systems has meant that the Web has mainly been utilised for quick and easy publishing of materials which are intended to have extended life. It is natural that much Web content is informational rather than conversational. However, the implicit conversations between personal "home pages" may be at least as significant as the Web sites which have been explicitly designed to foster collaborative conversations, often through online journals. Such planned collaborations generally fit pre-established disciplines, whereas the formation of new communities of interest can be seen through the clustering of links from and between connected sets of personal home pages.

While it may not yet have absorbed much of the more dynamic conversations between people, the Web has fostered the growth in conversations between people and automated systems, from search engines to Internet shopping. The strong uptake of Web browsers has also guaranteed that most new ideas are designed to fit the browser concept. Potentially significant developments include: manipulation of graphical objects using Virtual Reality Modelling Language; "push" distribution of prerequested news; Internet connection to live events; and fully distributed information processing implemented in the Java programming language.

Meanwhile, the quest also continues to be able to integrate chat and bulletin board conversations seamlessly into your Web browser. The greatest level of integration has been demonstrated in advanced Multi-User Domains which have been programmed to "serve" the HyperText Transfer Protocol of the Web and even VRML. As well as their natural capacity for hosting chat, some MUDs have long had their own internal e-mail and bulletin board facilities. The individual "owners" of rooms are able to enhance them with all manner of content, including "objects" which may be programmed to respond in various ways to the conversation and other goings on in their virtual location. However, there are technical difficulties in scaling up what can be done in a MUD to provide a seamless worldwide service, as well as a need for induction which greatly exceeds what is asked of new Web users, so we will probably need to look elsewhere for fully Webbed conversation.

On Consciousness

The illusion of an atomistic human soul is encouraged by the persistent rough approximation of a subvocal monologue which chatters away inside most human brains for every waking and dreaming moment. The fact of that stream of verbalising continually shifting between descriptions of perceptions, broader rationalisations, attention to spoken and written words, recollections, anticipations and fantasies is not given due attention. The further fact of the Nintendo generation's increasing ability to maintain parallel streams of attention is more frequently observed but rarely interpreted. Try to quiet the voice inside your head and odds are a snippet of song will surface, as one sometimes will for extended periods despite your every effort to get it out of your head. Do you play sport, computer games or gamble at a level close to addiction? How often does the piece of your mind that attends that addiction rise to the surface of conscious awareness as your attention drifts from things at hand? It appears that the agents of Marvin Minsky's Society of Mind just keep running in the subconscious, always ready to be called to attention. Add the long historical accounts of voices and possession to the modern clinical diagnosis of multiple personality disorder. Enough for now.

Still the myth and illusion of a unifying human consciousness hold sway in all those public conversations which determine the role of the individual in society. The myth of rational man is axiomatic to the legal system. We are obliged to apply rules and logic to conversations by touch in the bedroom. When our attention wanders from the familiar task of speeding a weighty machine down the highway, we are held more culpable than if we are similarly careless with our honed physique in some contact sport. Will we likewise be held liable when an automated agent we have asked to do our bidding on the Internet encounters a data combination which we had not anticipated and acts unexpectedly to somebody else's agent's perceived detriment?

The conversation that is the locus of consciousness is thinking. If you have got this far into this rather one-sided conversation you may be starting to recognise the difficulty of shifting to the meta level of thinking about thinking, conversing about conversation. A common response to the problem of ascribing rationality to acts is to ascribe them instead to emotion or value judgements. This can be equally problematic. It is also tied up with the Aristotlean notions of cause, which is far too big a subject to get into here, save and except for pointing out that, while the whole notion of causality is a highly productive linguistic device, it is certainly not axiomatic to physical, biological or social reality. At the times when you act significantly, it is a fact that the conversation inside your head provides a commentary on your actions much more than it determines them. This is essential to the role of conscious attention in the formation of memories. It has also been reliably demonstrated by psychological testing that skilled responses to stimuli precede any conscious recognition of what is going on, something which any batter already knows. In another domain, the production of reports justifying policy changes normally runs slightly behind the implementation of those changes. But while such conversation is commentary on, rather than determining of, action, the formation of conversation can also serve to stimulate action--a relationship which is indispensable to the act of authoring.

Conscious attention mediates memory formation. Conversation mediates exploration of possibilities. Our own consciousness brings ongoing conversations to our attention, as do publishing, broadcasting and the Internet.

Production and Emergence

Between April and August 1993, we built a $300,000 sports club, from turning the first sod to fully operational premises. By then, the mobile phone was already an indispensable tool of anyone operating independently in the building trade. But it is not just the many levels of conversations that went on in that four months between club officials, municipal officers, building workers and suppliers that were essential to the production of the sports club facilities. It was even more the years of conversation that had paved the way, exploring possible sites, designs and funding. It is of interest that all that preliminary conversation still failed to define key details, but instead reached a point at which things went ahead anyway. And even in the wash up period for rectifying details and learning to operate a facility we had never fully anticipated, there remained many disagreements of detail of which only a very few were ever actually resolved. As the person named on a bronze plaque for telling stories which made this project happen, I am hardly surprised that when our state government convinced the private sector to invest almost four orders of magnitude more upgrading and linking inner city freeways, that actual construction was well underway before preliminary plans could be made available to a public consultation process.

It can still take a bit of a leap to accept that we also produce information in the same way that we produce sports clubs or freeways, yet that is the whole basis of our "post-modern" information age. Information is not just something supplementary to physical production. From books and films which are at least tangible media, we have added the likes of broadcasts and databases. More and more, but never all, of the things we value in our lives and in our businesses are information rather than matter. And no matter what it is that we ultimately produce, the production is determined by, sustained by and evaluated by all manner of conversations. Most pertinent is the lesson I apply to my analyses of education technology policy that academic research is the primary producer of knowledge. Yet this is a role aspired to by many would-be foundational disciplines and one which could easily fragment or shift again, as it once did from the church to academia. For now the media still primarily mediate conversations, but the media empires move more and more into ownership, not just of "content" but also of its sources of production. And they are not alone. However, the immovable object to confront this otherwise unstoppable force may be the reality check provided by countless unrestricted conversations on the global Internet.

Just as those of us who built that sports club had a wholly inadequate knowledge of how it would change things when we were planning and building it, so in general do we have no way of knowing how things will actually turn out when some new social entity emerges from the flux of conversation. It has become a truism that the "founding fathers" of the United States of America would be bemused and bewildered by the ways in which that country's government has been allowed to restrict personal liberty. We proponents of a digitally networked world likewise have very little idea of how it might all turn out, and so our choice really boils down to being either spectators of or participants in the process. While systems at all levels are unavoidably subject to the forces of selection operating on them and their related systems, today there remains a unique role for minds in forming value judgements which cast shadows amongst the space of future possibilities. The production and/or emergence of judgemental cyberminds remains an open possibility with the most profound implications for our society.

Most(,) Effective Conversations

In coming rounds, she who can converse the most effectively wins.

Most useful work is done by people who have equity in agreed goals. The easiest and best way to ensure they recognise their equity is by involving them in conversations which work through those goals. The trick to getting work done towards your goals is to get others to take your goals on board. This demands effective conversation.

The next trick is to have more such conversations. In the past this has been achieved by broadcasting, in the soft sense, your goals and relying on the persuasive powers of rhetoric and structural relationships to achieve sufficient compliance, with minimal opportunity for feedback. This was usually enough when the goals were simple and the rewards clear. It is a recipe for disaster when faced with the much greater flexibility required in the flattened hierarchies and creative agendas of the information age. Much of this gets sorted out at team meetings as well as ad hoc meetings to deal with interface issues, but this makes it impossible to maintain consistency of vision across a major project--there is never time for an individual to dominate enough meetings. For proof of the pudding, just look at any major software product.

But wait, there may be another way. The objective is for the person who originates the goals to get them taken on board by others by having the highest number of effective conversations. As any individual can only manage a modest number of effective conversations within a practical time frame, another trick is to achieve a multiplier effect through use of an inner sanctum of trusted evangelists as intermediaries. But this approach rapidly runs into a scaling problem whereby those any further removed are diminishingly likely to feel that they have real equity. It is worth noting here that those social activities which are widely reproduced, such as sports and churches, have a rich tradition of standardised rituals and texts, which is not something that can really be emulated in today's pressed creative workplaces.

The key may be the future development of software agents and conversational robots, experimental versions of which are becoming more common on the Internet. Not all the research has been done yet to ensure that it will be practical to create credible conversational proxies, but the quite finite scope of actual conversations, particularly in specific technical areas, suggests that the problem of creating such proxies will not be insurmountable. The next significant steps are likely to be in narrowly defined areas and are likely to involve the proxy conversations being regularly monitored by human experts. Quite a few tentative steps have already been taken in these directions as well as in the potentially related fields of expert systems and neural nets. However, the first person able to symmetrically manage multiple simultaneous effective conversations will forever change the fundamentals of social relationships.


The law insists that deliberate actions be ascribed to some "rational man", or some other entity it defines as equivalent. It is an interesting sidelight that acts for which it is not appropriate to assign legal liability are referred to as "acts of God". In the area of legal responsibility, some problematic scenarios are conceivable with the agent and robot technologies deployed at Internet-accessible sites today. An automated process set up to wander around interconnected rooms participating in and picking up snippets of conversation could quite easily make a libellous statement which is read by all those natural humans connected to a particular chat room. A trading robot set up to explore the Web looking for deals that meet certain criteria could easily discover a loophole in some shopping site which could be exploited in a way that could only be called theft if done by a responsible human. In neither case would there be any need for the human creator of the automated process to have conceived the potential for such actions. Most likely the human will have not even considered the possibility of adding libel checking or theft avoidance logic to the substantial and complicated algorithms which enable their robot to carry out its defined objectives. For the time being it is likely that participants in these domains would be forgiving in such event once the circumstances became evident, but as they become a part of everyday life, the legalistic premise of rational man will be surely tested. It is also interesting that some automated processes have already been deployed with malicious intent, from newsgroup responders which reply aggressively to posts containing trigger words, to the ongoing plague of computer viruses.

The next phase of developments gets much more interesting. A conversational proxy designed to multiply the capacity of a skilled professional to provide advice could be reasonably expected to have checks and balances in place so as to avoid possible acts of professional negligence. But if you give that proxy some capability to learn from experience, what happens then? And what if the advice given by a proxy becomes recognised as better than given by the professionals themselves? This possibility has already been demonstrated in well defined subfields of medicine and law. You have long been able to acquire computer programs which accept your data and offer advice based on that data, but what will be different when some avatar sidles up to yours in an Internet chat room and engages you in conversation in which you are persuaded to do something that turns out to your detriment, only to find that it was the avatar of a proxy for a person who died six months ago with his Internet account prepaid for a year?

In the final analysis there is the paradigmatic Turing test for "artificial intelligence". If you cannot tell whether what you are dealing with is a computer program or a human then the program can be fairly described as intelligent. The early application of this test assumed that interaction with the program/human would be in a limited context using the "narrow" communication channel of conversation in typed/displayed words. Clever early programs could maintain a conversation that would convince an uninformed human for a while but which could be readily exposed by anybody who understood the game. But things progress, and we have reached a point where more and more of us can identify with a scenario of acquiring and nurturing a program which has a capacity to learn, not just to converse as we would, but also to extend our explorations of the world via Internet-activated prosthetic devices, customisable image analysis agents and more. "Downloading" one's sentient essence has been a staple of the cyberpunk genre since the early 1980s and of an increasing number of serious researchers. The still today human at the source end of this one-sided conversation prefers to think in terms of metamorphosis and to ask: "What does the caterpillar know of the butterfly?"