Computer Theology: Rice University Lecture

The emergence of trust as a computer concept is quite recent. For example, Tim Berners-Lee presented the semantic web in a W3C conference in 2000 with an architecture schema that has been reproduced many times since then (http://www.w3.org/2000/Talks/1206-xml2k-tbl/slide10-0.html) and that we'll summarize as follows:

At the top, the word "trust" reigns supreme. However, even today, we are hard-pressed to google encompassing definitions of trust in computing (this is bound to change as we see more and more studies of trust appearing in software conferences and publications besides our own). That's a gap that we attempted to close, not without some difficulty, as we discovered soon enough that encompassing studies of trust are as lacking in philosophy and social sciences as they are in computer science. So the semantic web has a roof, but we had little understanding of what it was.

What made us progress was the realization that when computers engage in serious business, say in the exchange of money on open networks, they don't rely initially on policy, but rather on evaluations of trust, often based on various forms of cryptography. The most intricate such trust infrastructure is that of public-key cryptography, where a hierarchy of certification authorities vouches for the trustworthiness of a computer or computer application, with a self-certifying top authority. When we realized that policies leading to the actual exchange of money were in fact computer-to-computer interactions dependent on trust, we had the eureka moment of realizing that a recursive organization of trust and policy infrastructures was the organizing principle of great leap societies. Following is an XML description of such (Computer Theology, Power of Prayer, page 375):

What we see is the stacking of subordinating trust and policy infrastructures where there is always a trust infrastructure covering a policy infrastructure; rules of interaction are always subordinate to their trust environment.

Once we had identified the computer construction of trust elaborated by public-key infrastructures, we took this as a starting point to examine human society. We started from the understanding that trust had all the appearances of religion, and that policy was rather like government, operating on values established by  observance of religion. (We later studied the evolution of professed secular societies, but we'll come back to that below). While we had no difficulty associating the concept of certification authorities with that of progressive orders of authority leading to God in monotheistic religions such as Christianity or Islam, we couldn't find such organizations of trust in other religious forms such as Buddhism or Zen. It looked like the latter forms of religion demanded a different kind of trust; a form we came to label "trust by process," thereby  emphasizing that the ritual itself is a leading cause of mutual trust in the way to ecstatic revelation. By contrast, we labeled "trust by causality" the God-led road to ecstatic revelation exemplified for example by monastic orders.

While trust by causality was readily identifiable as parallel to computer network causal cryptographic chains, even to the point of identifying a seminal origin of causality (the root certificate), would we find in computer networks the equivalent of "trust by process?" Actually we did, observing that peer-to-peer negotiations, with each computer maintaining a state of the other, are actually process-driven mechanism that only go forward if the other performs as expected. In fact, we came to understand that this was the fundamental human means of sympathy evaluation illustrated by mirror neurons. We needed to then understand if the rituals led to some form of computer ecstasy, a point we reached when we saw that an operating system in a state of upgrade was disconnected from its applications and therefore from the social part of its interactions; a situation akin to the detachment that is the hallmark of religious ecstatic end states. We will detail that point later in our discussion.

At this point we needed to understand how computers would actually build trust as the web scales up. We made the hypothesis that religion is a central source of trust in human societies up to modern times. Before going there, though, we needed to better understand what human and computer societies are made of, to illustrate the vectors of evolution that would actually lead to development of larger forms of organization.

If trust builds societies, are there ways to measure their evolution; is there a human yardstick that can be used for computers as well? While biological evolution is evaluated by the genome, we looked at the work of Maslow to understand vectors of interactions that would bring about group dynamics. The table of Computer Theology, Fabric of Society, page 146, expresses the needs hierarchy of Maslow [2] (with transcendence at the top level, that Maslow introduced later in his work). We associate the needs with their physiological underpinnings, which we'll use later when we look at the computer equivalent.

In this table, that we call pretergenesis as it expresses social evolution in a hierarchical progression, we see how human needs map to social evolution up to modern days. We can now look at the correspondence of human needs with computer needs (Computer Theology, Mutation, pages 321-340):

Computer Needs – A Reflection of Computer Provisioning

Following MacLean [6], we consider the layering of the brain, phylogenetically (across generations), and ontogenetically (within generations), as a layered construction starting with our sensori-motor apparatus, then our reptilian ancestry, then our mammalian selves, and finally the prefrontal cortex expansion leading to modern humans. This layering parallels and nourishes Maslow's hierarchy of needs. For example, we can consider that motion brings feeding, reaction brings security, emotion brings belonging, and cognition establishes a path to self-actualization.

Following Lakoff and Johnson [7], we consider that cognition revealed by language is metaphoric. When we compare:

(1) A flock of birds is above the lake

(2) Rice graduates are above the fray

we use our embedded geometric engine that positions birds in relation with the lake to express a relationship of graduates to the rest of humanity, at least as they see it before meeting the realities of life. Now going from geometry to emotion we can compare:

(3) The fire warms our hands

(4) Our love warms our hearts

where we now build up concepts from physical properties beyond geometry. In Computer Theology, In Search of Enlightenment, page 283, we have formalized those metaphorical constructions as follows, illustrating that computers are now equipped with the cognitive apparatus needed:

<Rule>
  <Premise>
      <Union>
        <Determination>
          <Cause> A </Cause>
          <Effect> X </Effect>
          <Property> P </Property>
        </Determination>
        <Determination>
          <Cause> B </Cause>
          <Effect> X </Effect>
        </Determination>
      </Union>
  </Premise>
  <Conclusion>
    <Determination>
      <Cause> B </Cause>
      <Effect> X </Effect>
      <Property> P </Property>
    </Determination>
  <Conclusion>
</Rule>

 While we can derive metaphoric understanding, we are as yet incapable of destroying on a grand scale our own capabilities to make place for new insights in the restructuring  of our knowledge propped up by self-actualization.

In order to understand that last tidbit, we turned to another metaphor, that of theater. We paid particular attention to three steps in the evolution of theater since antiquity.  Greek theater was  explicitly formal, to the point that laws were written against departing from the text. Italians introduced much later commedia dell'arte, a type of performance where the scenario was quite stable but much freedom was left to the actors in expressing it. A new form of creativity was born out of a ritual making place for communication with the audience for reconstructive knowledge typical of comedy. Finally, modern times saw the coming of improvisation, where the scenario itself is in play. We follow Stanislavski [8] in recognizing the preparation of the actor as a personal mix of ecstasy for innovation, as well as ritual allowing other participants, actors and audience, to welcome the new perspectives offered by the performance.

To better understand the forces behind that progression, we turned to observing computers. Whenever an operating system is in place with applications, it is usually possible to update certain applications while the computer is running. In most cases, the application itself is suspended pending installation of the new version; in effect, previous knowledge of the computer is deleted and new knowledge comes in. A more intricate form of provisioning is a modification of the operating system itself. While minor modifications may leave some applications and sensori-motor capabilities of the computer unscathed, major modifications are of a very different nature. We are all familiar with the blank state that a computer, be it an iPhone or a Mac, undergoes when a new operating system is installed. The computer stops functioning, applications also, and often, applications themselves need to undergo major changes to adapt to the new situation. Data may have to be restructured, and the new computer comes back to life with a personality change.

Within the limit that such metaphors carry, we realize immediately that for the human brain to update its knowledge, similar constraints apply. Some knowledge will accrue during the day while we may notice discrepancies in our apprenticeship, but then a good night's sleep may bring us to a new state of understanding that will help us pass that exam next week. But more drastically, we can see that changes that affect deeper constructs of our brain may demand the same disconnection from our sensori-motor capabilities as well as from our operating functions, that we've observed with an operating system change. That state of detachment and subsequent illumination has been thoroughly described for religions. Grace and nirvana, for exemple, are different versions of ecstasy. Combined with rituals structuring both the ecstatic experience and its outcome, humans have found a mechanism for updating their common understanding and practices in an environment of trust allowing progressively larger communities to develop. At this point, we'll conclude by observing that trust induced by ecstasy and rituals has been largely documented as facilitated by mind-altering drugs that give a physiological basis to the whole process that may yet be emulated by computers.

And what about the "great leap forward?" If we are correct and the convergence of the Internet (child of Arpanet) and of the Web (child of hypertext) has generated human-like societies of computers, maybe we've learned something about the emergence of human societies themselves. One hypothesis, formulated to me by Tom Zimmerman, in the audience today, is that the human great leap forward of 50,000 years ago may be the product of such a convergence. Maybe the convergence of ecstasy and ritual, constituents of transcendent trust?

Epilogue

Thank you for your attention, and we should have time for some welcome discussion.

Bertrand du Castel is a Schlumberger Fellow, and co-Chairman of the W3C Oil & Gas Semantic Web initiative. A former Chairman of the Java Card Forum Technical Committee and vice-Chairman of POSC (now Synergistics), he received in 2005 the Visionary Award from Card Technology magazine for his role in bringing the Java Card from inception to billions. With publications in artificial intelligence, linguistics, logic, and software engineering, Bertrand has a PhD in Computer Science from the University of Paris and an Engineer Diploma from Ecole Polytechnique, France. He is based in Austin, Texas.