The Hindu Business Line : Double loop learning

WE HEAR a lot about organisational learning; in this context it is worth asking if we realise the very important distinction made by Chris Argyris between two very different kinds of learning: The single loop and the double loop learning. The first is the more familiar, best illustrated by the thermostat or a trip switch. It is set to recognise a deviation from the expected state, i.e., change in the temperature (or pressure, volume of flow or whatever) in the system. If that is beyond permissible limits, the machine shuts down or switches back in, as required. This is how air-conditioners and refrigerators work.

Left to themselves, managers would ideally like all business systems to work this way. Clear decision rules of the form "if this happens, then do that" would make life a lot easier and dictate what needs to be done. If the bidders for the airport modernisation projects did not get, say, 80 per cent on certain pre-agreed criteria for technical evaluation, well, then, they were ruled out.

No one would try to rework the weights and scores to sneak someone in. How simple life would be if indeed administrations, public or private, worked like that.

Back in the 1960s, working as area sales managers, my colleagues and I were taught that the most shocking offences a company stockist could commit were to tamper with the pricing (make an undue profit margin) or bounce cheques. On the latter, there was some leeway provided for genuine crisis and unintentional errors, so in a lifetime the stockist could do this twice; the third time, he was out, and no appeals or excuses entertained thereafter. The system simply rejected him.

Few rules in organisations today actually do work like this — and if they did, almost anyone could run a department, unit or sales area. This was also one of the reasons why the government, for example, which is notorious for being rule-bound and procedure-ridden, could afford to post very young officers in their twenties into the districts as collectors. Their most important guidelines were rules and precedent established over the years. The second kind of learning is, in some sense, real learning as it applies to human and social systems. Let me explain: If all the action that was required of the manager or an IAS officer were to follow specified directions, then there is no scope for exercise of judgement or intelligent appreciation of the situation.

And this is where double loop learning comes in: It asks the question, "Do we need this rule? Must we have the self-correcting mechanism work like a thermostat or is there a way to avoid the occurrence of the problem altogether? Can we learn some things about why the problem arises in the first place and take some preventive steps?

In other words, should we do something else altogether?" Asking such questions is so uncomfortable and inconvenient that managers tend to avoid it. Yet breakthroughs can only occur if someone takes up the unpopular role and asks whether (to go back to the earlier example) the stockists can be dealt with in a different way so that the termination of their stockistship does not automatically happen.

Soon, one finds that there are not enough good distributors to go round and one might well have to put up with some degree of flexible financial standards in return for, perhaps, more effective selling. In other words, double loop learning does not take the system as given; it does not hesitate to ask for the fundamental justification for a decision rule — and whether, in the light of events or over time, the assumptions of an earlier era haven't become obsolete.

From such scrutiny does the prospect of real managerial learning emerge. If we want to encourage such double loop learning, top management must keep the natural overweening tendencies of bureaucracies at bay.

(Feedback can be sent to srchander23@netscape.net)