Our theory leads us to expect that we can build interorganisational information systems using logical database technology, enabling interoperation among organisations that share institutional facts. The sharing of institutional facts is represented by the participants’ commitment to a common ontology. This ontology can be either transcendent or immanent. The question now is: given that we can interoperate where can we then go?

One possibility is to recognise that once an interoperating community is established, it can generate a large number of institutional facts. These institutional facts can be interpreted by any of the players who share the common ontology. These ontologies or institutional fact schemas constitute the atomic behavioural units, but do not necessarily determine behaviour. The rules of chess determine what constitutes a chess game, but there are lots of different games.

So we can use techniques like data mining that depend for their atomic data on the exact classification/logical equality nature of institutional facts, but which can find emergent patterns in the multiplicity of instances. These emergent patterns can be used as an immanent ontology for strategic or tactical decision making, for example advertising campaigns to encourage or discourage behaviour patterns, or as evidence of undesirable behaviour to be subjected to further investigation (e.g. fraud, money laundering).

Where the interoperating community consists of many small players, there may be an advantage to each player giving up its exclusive access to the institutional facts it creates in favour of a community-wide pool to which all players have common access. This is common, for example, in real estate where individual sales reports and auction success rates can be published for a whole city market area, enabling each player to see trends to which they can respond in their own fashion.

The information spaces opened up in this way give great scope for the development of interoperating autonomous intelligent agents. Each agent can develop its own immanent ontology, which it uses to govern the strategies and tactics it uses to interoperate with others to perform speech acts using the common transcendent ontology. The theory of this paper predicts that a research and development program along these lines would be likely to be productive.