On Wednesday former Defence Minister and ex ANC chairman, Mosiuoa Lekota, announced the start of a process leading to the formation of a breakaway party from the African National Congress. Although Lekota qualified his announcement, saying "This is probably the parting of the ways," it is hard to see how he and his allies can turn back from the path they have embarked upon.

The conventional wisdom about the new party's prospects is that the ANC has little to fear. This view was summed up by an analysis piece run on Reuters which stated: "Despite the discontent, most analysts believe Lekota, an Mbeki loyalist, will either crash and burn, like ANC dissidents in the past, or at most make a relatively minor dent in the party's hegemony." Exhibit number one of the futility of challenging the ANC, within a democratic context, is the (relative) failure of Bantu Holomisa's United Democratic Movement to mount a meaningful challenge to the ruling party.

It is certainly possible that the new initiative will collapse in a heap, particularly if its launch is bungled and it fails to bring on board a substantial proportion of the losers at Polokwane. But this does not mean it does not have the potential to dramatically reshape the political landscape.

Many commentators seem to underestimate or misunderstand Lekota. He may have ended up in the Mbeki camp by the time Polokwane came around, but he did not start out there. He was elected to the position of ANC chairman at the ANC's 50^th national conference in Mafikeng, December 1997, by campaigning against and defeating the preferred Mbeki-aligned candidate Steve Tshwete.

According to a Mail & Guardian report of the time he had, apparently, rebuffed pressure to step down "quietly arguing that he was the only non-Africanist in the running for one of the top six positions." With the support of leftwing allies he went on to defeat Tshwete by 1775 votes to 859. It is worth noting that until they turned disloyal Baleka Mbete, Jacob Zuma, and Kgalema Motlanthe were all far more closely identified with Mbeki than Lekota ever was.

The significance of Holomisa's movement into opposition to the ANC is also misinterpreted. At the time he was ejected the liberation movement was united, disciplined, and led by Nelson Mandela. Nonetheless, immediately after its formation the UDM enjoyed a surge of popularity. By July 1998 the UDM enjoyed the support of 6.5% of those polled by Markdata, while the ANC had dropped to 57% (non-choices excluded.) This is the only time, since 1994, that the ANC has ever received less than sixty percent in that series of opinion polls. (See Table below)

The problem new parties have - the Independent Democrats being another example - is that they don't have the organisation on the ground to translate that initial enthusiasm into enduring support. By 1999 much of the UDM's initially support had dissipated. It was, nonetheless, able to win over half-a-million votes (3.42%). In the Eastern Cape alone the UDM won 281,748 votes.

The potential of the new party has to be some order of magnitude greater than that of the UDM given the current internal dissension within the ANC; its loss of moral authority; and, the apparent willingness of certain oligarchs to fund the new organisation. It should also be able to bring with it sections of a pre-existing party organisation; particularly in the Eastern Cape where many Xhosa-speaking ANC supporters feel they lost ownership of the movement at Polokwane. It is possible the new party would be able to draw support from the emerging black middle class, particularly in Gauteng.

Its chances of making serious inroads in Limpopo and the North West - both provinces which backed Mbeki at Polokwane - is more open to doubt. Still, a breakaway could do serious harm to the ANC, however hegemonic its support may currently appear. If the last two elections are any guide the new party would have to take between 25% and 28% of the ANC's electoral support to deprive it of an absolute majority. This is undoubtedly a tall order - it would require between 2,5m and 3m ANC supporters to switch allegiance - but it may remain within the very outer realm of possibility.

What is likely to happen is that the new party will rapidly acquire some kind of ethnic character. South African politics is communally based. And in these circumstances, as the American political scientist Lucian Pye once noted, "Attempts to organize parties in terms of particular political principles or limited policy objectives generally result either in failure or in the adoption of a broad ethic which soon obscures the initial objective. Usually political parties represent some sub-society or simply the personality of a particularly influential individual."

Inkatha, the Independent Democrats, and the Democratic Alliance are all parties which initially had more universal aspirations, but which ended up drawing their main support from one or other sub-group in South African society. The ID, for instance, is now really a party of the Coloured working class. A breakaway party from the ANC is likely to end up a similar place. The difference being that an exodus of one particular group from the ANC could well disturb the remaining ethnic balance in that organisation.

A last point to make is that what is hindering the formation of a new party is South Africa's lack of a constituency based electoral system. This means that MPs cannot cross the floor to the new party, nor do they have a real constituency to bring with them. But once election time comes around South Africa's pure proportional representation system will work to the advantage of a breakaway party. The difference between a "splinter" and a "split" is much narrower within a PR system where every vote counts.