Information is often given away at no loss to its owners, while recipients gain, because it can be duplicated. For some purposes, it could be paid for in money that behaved in the same way - adding to the recipient's wealth, while not reducing that of the donor. This money would have to be quite unusual; but so is the commodity being traded.

Buying and selling knowledge is full of implicit transactions. One doesn't stop to write a cheque every time one hears industry gossip, however useful, in the corridor. More importantly, one doesn't even need to think about doing so. So much of the knowledge traded in the 'real world' could only be done so informally. Trade is no less valuable for this - the world runs on the chats with senior colleagues, the friendly advice over lunch, the learning and experience gained from working with a client. Formalizing it all would kill it.

That's as far as we can get with knowledge outside the distance-eliminating domain of cyberspace. The most promising aspects of a world with fewer geographical impediments in the flow of information include the translation of the gossip in a corridor to an electronic mailing-list, the ability to be advised by friends never seen over a virtual lunch, or to chat with a senior colleague across continents.

But the limitations of geography had advantages. People now begin to wonder whether they're not being cheated, when the corridor-mailing-list can have thousands of people. The narrow confines of office buildings provide a sense of community, and an assurance that the consumers of the knowledge traded there are worthy of it, being likely producers too. But restricting access to electronic corridors will negate their advantages, of connecting more minds across space.

The main option being considered these days is not just overtly tied to economics - it must be, as that's what the knowledge trade is about - but also tangled with hard cash, with dollars. This is not only unnecessary, much of the time, but also harmful.

With the safe assumption of secure international digital cash transactions in the near future, the hard-currency solution to the electronic corridor problem would be to charge every consumer a small amount for everything - a cent a word, say. This may work with traders in diversity, the indexers of the world to whom you pay for anything you need to find. It works rather less well for the traders in consistency, to whom you would pay much more than a cent a word for continuous, reliable content.

But for those who lurk in the electronic corridors of the infosphere - and all the implicit cooperatives that form between individual contributors to cyberspace's live, people resources - the hard cash model is useless. Among others, there is the question of who takes the money. The Department of Corridors? The Treasurer of Electronic Mailing Lists? It certainly wouldn't be fair to pay just the individual author of the particular piece you read and found useful, as it built upon the unpaid work of others. After all, the informal nature of these gatherings of minds is for a reason. They, or at least the active participants, contribute not for cash but in return for the contributions of others. They only take exception to the one-way consumption by stray, if numerous, visitors.

The alternative to hard cash in these implicit transactions in cyberspace is the equally implicit currency of the real world. The currency that is used almost solely in the trading of knowledge and which, like knowledge, extracts no direct cost to the buyer at the seller's gain. A currency that can be paid equally to corridors full of industry gossip as to colleagues who throng there. Indeed, a currency that is, and is traded like, information itself: reputation.

Reputation may not, at first glance, resemble money. But it is the implicit reward of good products, and their producers. It certainly adds to - or subtracts from - the worth of those who receive it. And it is the ultimate free market currency. Traded as a commodity like other data, reputation is based on, and influences in turn, the average of 'prices' set by individual consumers - what individuals think of a producer. Reputation encourages improved production as much as, or more than, cash does - at least in the informal knowledge trade. And reputation will be a very important aspect of the knowledge economy, with the increasing anomie of cyberspace as e-mail replaces firm handshakes.

There are, of course, problems with transporting the informal reputations outside cyberspace to reputation systems within it. But work does continue on the necessary technologies. This progresses at a slower pace than that on automatic cent-a-word payment systems, because developers don't realize the importance of informality in the information age. This will have to change - the knowledge economy is actually a people economy, and its most common currency should reflect that.