3.3.B. The Evolution of Collective Intelligence

From P2P Foundation
Revision as of 17:00, 25 April 2006 by GSimon (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

3.3.B. The Evolution of Collective Intelligence

Related to the above evolution of cooperation is the concept of collective intelligence, which concerns any knowledge of the collective, which goes beyond or transcends the knowledge of its parts . Collective Intelligence is the process whereby groups take charge of their challenges and future evolution, by using the resources of all its members in such a way that a new level emerges which has added qualities.

Jean-Francois Noubel in an online book-in-progress at http://www.thetransitioner.org/ic outlines three stages, arguing that we are in a transition to a fourth. The following is a synthesis of his work.

The first stage is the 'original collective intelligence', which can only exist in small groups, and historically has been typified by the human organization in the tribal era. Seven characteristics define this stage:

  1. an emerging whole that goes beyond its parts
  2. the existence of a 'holoptic' space , which allows the participants to access both horizontal knowledge, of what others are doing, and access to vertical knowledge, i.e. about the emerging totality; to have collective intelligence, all participants must have this access, from their particular angle
  3. a social contract with explicit and implicit social rules about the forms of exchange, common purpose, etc..
  4. a polymorph architecture which allows for ever-changing configurations
  5. a shared 'linked object', which needs to be clear. This can be an object of attraction (the ball in sports), of repulsion (a common enemy), of a created object (future goal, artistic expression).
  6. the existence of a learning organization, where both individuals and the collective can learn from the experience of the parts
  7. a gift economy, in the sense that there is dynamic of giving in exchange for participating in the benefits of the commons

This original stage had two limits: the number of participants, and, the need for spatial proximity.

The second stage is the stage of pyramidal intelligence. As soon as a certain level of complexity is reached, it will transcend the limits in numbers as well as the spatial limits. Cooperation takes on hierarchical formats, with the following characteristics:

  1. division of labor, in which the constituent parts become interchangeable; based on specialized access to information and panoptism, i.e. only a few have centralized access to the totality
  2. authority organizes a asymmetrical information transfer, based on command and control
  3. regulated access to scarce resources, usually through a monetary system
  4. the existence of norms and standards, often privatized, that allow knowledge to be objectified

Pyramidal intelligence exists to obtain 'economies of scale' through repetitive processes that can add value to an undifferentiated mass of raw material. To see what kind of intelligence predominates in an organization, adds Noubel, look at how it produces. If it produces mass products, then, despite eventual token usage of peer to peer processes, it will essentially be based an hierarchy-based pyramidal intelligence.

The third form of collective intelligence is swarming. It exists where 'simple individuals' cooperate in a global project without holoptism, i.e. collective intelligence emerges from their simple interactions. The individual agents are not aware of the whole. This is the mode of organization of social insects , and of market-based societies . The problem is that in the insect world, individuals are expendable for the good of the system, while this is unacceptable in the human world because it negates the full richness of persons. This means that the contemporary enthusiasm for swarm intelligence has to be looked at with caution. It is not a peer to peer process, because its lacks the quality of holoptism, the ability of any part to know the whole. Instead, swarming is characterized by 'stigmergy', i.e. 'environmental mechanisms used to coordinate activities of independent actors' .

Thus, a fourth level of collective intelligence is emerging, which Noubel calls 'global collective intelligence'. Compared to original CI is has the following added characteristics:

  • a 'sufficient' money as opposed to a scarce money (see The Transitioner.org/ic site for more details)
  • open standards that maximize interoperability
  • an information system to regulate symbolic exchange
  • a permanent connection with cyberspace
  • personal development to acquire the capabilities for such cooperation

In this new global collective intelligence, the original limits in numbers and spatial proximity are transcended by creating linkages through cyberspace. In this context, we can see why technological developments are an integral part of this evolution, as it enables this form of networking. What cyberspace does it to create the possibility of groups cooperating despite physical distance, and to coordinate these groups in a network . An important aspect of the new cyberspace-enabled collective intelligence will be the increasingly symbiotic relationship between the countless human minds (one billion at present) and the huge networked intelligent machine we are creating . This noospheric networked intelligence is not an alien construction imposed on us, but something we are collectively creating through our sharing and participation.

David Weinberger has recently summarized the history of knowledge exchange for the Release 1.0. newsletter, showing how digitalization has freed categorization from the shackles it had in the physical world . He notes how humanity first started to separate things (shoes in shoe boxes, etc..), then, we the advent of the alphabet, it started to separate the information about things, from the things itself, putting the books in library shelves, and the data in card catalogs. The information would inevitably be classified in a hierarchy of knowledge, a tree structure. One way to know the world, one way to access knowledge. In the 1930's, an Indian librarian named Shiyalin Ranganathan, 'decentralized' knowledge categorization. An object has different facets, and the user can determine which facet is the most important for him/her. Thus, the catalog will organize the information hierarchically, but flexible, starting with one facet, then another, following the specifications of the user, as long as the programmer has prefigured these choices. On the web is now emerging a bottom-up approach, which does not necessitate any prior hierarchical categorization. Users will add tags , and different users or user groups will use different groups of tags, each reflecting their personal or group ontologies, thereby illuminating different aspects of the object. This peer to peer categorization methods are called folksonomies . In the sphere of abundance that is the Internet, it is near impossible to continue using hierarchical and well-designed metadata systems, due to the sheer volume of data and the large numbers of users who would have to be disciplined, so bottom-up tagging makes a lot of sense . In the peer to peer era, knowledge is liberated from preconceived and forced categorizations. Many authors have examined how our categories of classification are also instruments of power, and noted how different social formations overturned previous forms of categorizing the world . With the P2P classification schemes we see for the first time a recognition of multiperspectival worldviews. Knowledge is a distributed network, following a peer to peer logic. Notice how computers themselves have followed a logic from linear calculation to parallel and distributed computing, and from the mainframe/dumb terminal (centralization), via the client server (decentralization) model, to the internet filesharing ('the network is the computer') model. In computer programming the shift has been from linear ad procedural software production methods, to object-oriented programming, conceived as autonomous objects.

What is certain is that the emergence of networked media involves a new epistemology, a new way of relating to the truth. In one of the articles in a collection on ‘Subjectivation du Net’ in the journal Multitudes, issue 21, Jean-Louis Weisberg problematises the new epistemology. The explodion in the number of interconnected humans and machines, who are individually and collectively posting information and sharing knowledge , and linking multiple media formats through hyperlinks and RSS feeds, prefigures entirely new ways of knowing and learning. It is linked to the growing distrust of the older forms of mass media and even representational democracy, that we are witnessing.

The objective dimension of truth, implied in mass media, where specialized reporters verify facts to establish one narrative of truth, is making place for a truth that is emerging out of continuous intersubjective confrontation. Networks users, experiment, modelise and communally discuss events, acquiring the possibility of a much greater intimacy with the discussed event or process, a view of reality which is enriched by the multiple interpretations. But what is needed is a common meta-framework, so that the space is opened up for families of interpretations to compete. Another important related shift that is occurring relates to how we are learning and visions about the learning process. Behaviorism, cognitivism and constructivism, the 3 main learning theories, locate learning <within the person>, even as the latter admits that learning is socially constructed. Emerging connectivist learning theory on the other hand, acknowledges that learning can take place 'outside' the individual, through his connections. Since reality is ever shifting and changing, the individual ambition to know everything is a lost cause, paucity of knowledge has been replaced by an abundance of knowledge. Crucial skills are know the ability to know 'where', in our field of connections, the actionable knowledge is located, to be able to evaluate 'what' has to be learned in a context of abundance, and to negotiate and integrate a variety of opinions on any given subject.

Finally, theories about how individuals learn, still the focus of connectivist theory, must be coupled with a study of the new peer to peer knowledge dynamics, especially of the communal validation of truth which occurs within peer groups, and which replaces institutional mediation .