Did the ancestors think of the future?

Meditations on the nature of historical knowledge come in two main varieties. Foremost are those from the philosophers, who want to contribute to theoretical discourse on history as a discipline. Then there are those from the historians, who are concerned to offer reflections on their craft from a practical point of view.

While Ged Martin presents his book as the "reflections of a jobbing historian upon a worthwhile craft", Past Futures is more interesting and important than this modesty suggests.

Martin thinks that history is both an impossible and necessary discipline.

Impossible because causal explanation of what happened in the past is not possible: there is too little evidence, too much complexity, and no means of conducting experiments to demonstrate even the simplest of causal chains. Yet the effort to reconstruct the past remains socially and psychologically necessary.

The alternative to causal explanation is reason-giving explanation - the exploration of the reasons for past action. This approach typifies much historical work, but, as Martin points out, many past decisions were instantaneous, intuitive and irrational - and hence beyond the rational reconstructions of historians.

Thankfully, much human decision-making is rational and recoverable on the basis of surviving evidence, and Martin agrees that decision-making analysis and narrative constitute a viable foundation for the discipline of history.

For Martin, the "why-when" of decisions - why a course of action was adopted at that particular time - is primary. Why-when is critical for understanding the actual content of decisions (the why-what) and also for assessing their consequences.

Martin's focus on why-when questions is part of his general argument that the primary function of historians is to locate events in time, to explain why events happened when they did. In Martin's view, the key to locating events in time is how participants saw their futures.

These "past futures" come in many different shapes and sizes. For some, the future is inevitable; for others, it is determinable. Some people are pessimists, others optimists. Some take a short-term view, others a long-term view. Some care about the future, others do not. Some people act because they think they know what is going to happen next, while others just want to see what happens.

Understanding the type of future that was envisaged by historical actors will take us a long way towards the explanation of why-when.

I doubt that many historians would object to Martin's proposed procedure, and I am sure that they would be enlightened and entertained by his many case studies. Martin was professor of Canadian studies at Edinburgh University, and Past Futures began life as a series of guest lectures at the University of Western Ontario. Much of his illustrative material is drawn from Canadian history, but he also cites imperial history and the history of Ireland.

My favourite example is the case of the Earl of Kildare, who was asked in 1495 why he set fire to Cashel cathedral; he replied that he thought the archbishop was inside. But Martin does not explain the why-when of this particular legend.

More contestable is Martin's attempt to tackle the issue of significance in history. He needs to do this because to locate events properly in time, the past needs signposting - and one way of doing this is to assess the significance of events.

Martin's solution to this thorny issue is counterfactual: events are significant if we can reasonably argue that the past would have been radically different in their absence.

This takes us into the endlessly speculative realm of what might have been, but, Martin says, this is a necessary exercise that provides us with a series of signposts pointing the way from the past to our present.

Moreover, counterfactual reasoning remains central to historical explanation, irrespective of the question of significance.

To understand what happened, we need to explore the alternative past futures of participants, who continually used counterfactual reasoning to assess their options, as we do ourselves.

Ultimately, Martin's conclusion is a traditional, historicist one: the past must be understood in its own terms as well as our own. We cannot talk to the past but we can have a dialogue with it, one that moves back and forth and through time.

Geoffrey Roberts is associate professor of history, University College Cork, Ireland.