INTRODUCTION
If you are in the forecasting business, it's best to pick a date in the far future, say the year 2050, not one or five years hence, for your moment of truth. Both predictions will be wrong, but you probably won't be round in 2050, and in any case there's plenty of time for people to forget your forecast. We've all heard of forecasts that went badly wrong-for example a 1940s IBM president predicting that the US would only ever need five computers. (He might well have been correct about needing only five, but a lot more have been sold!) Despite this dismal record of future predictions, attempts at forecasting still flourish, as they must. The US writer Donald Norman has said it well. 'Attempting to predict the future is both foolhardy and essential'.

So we have to try to predict long-term futures. Take as an example transport technology and travel. Firstly, the infrastructure is expensive and long-lived. The City Link tollway in Melbourne illustrates this point. The 21 km tollway took about five years to plan and build. It will operate privately for 35 years, and then be handed over to the state government. So, calculating expected investor returns at the project planning stage involved estimating road traffic up to four decades in the future. Another example can be found in aviation. The Boeing 747 was first introduced in the 1960s and aircraft like these can expect to have a service life of up to 30-35 years. It is estimated that 80-100 years will elapse from the start of initial design of the aircraft to the last one being retired from service. In fact, the major aircraft manufacturing companies Boeing and Airbus Industrie both issue annual 20 year forecasts of passenger travel in order to decide what types and production levels of aircraft to plan for and build.

'There are no past options, there are no future facts'. This assertion by US futures researcher Wendell Bell stresses the sharp split seen between past and future. But often the past lays a heavy hand on the future. Nowhere is this clearer than with cities and their infrastructure-the buildings, roads, train lines etc. Melbourne at the time of the Olympic Games in 1956 looked much the same as today's Melbourne as we prepare for the Commonwealth Games in 2006 (except, of course, that 1956 Melbourne wasn't as large-or in colour on TV). Physical infrastructure has a long useful life, and represents a huge capital investment, so change comes slowly. Also, a lot of us like old buildings. On the other hand, cars only have a 20 year lifespan, and computers perhaps 5 years. Here, change can occur more readily. What this simply means is that some aspects of the future are more easily predicted than others. Once we've built our cities, transport infrastructure, or even aircraft, we want to keep them in operation for decades-or even centuries.

POSSIBLE, PREFERRED AND PROBABLE FUTURES
We can further clarify our thinking by distinguishing between possible, preferred, and probable futures.

Possible futures: The technological progress of the past half-century has seemingly made available an enormous range of possibilities. For the next 45 years to 2050, it might seem that only the physically impossible can be excluded with certainty. The envelope of possibilities seems wider the further we look ahead, probably because we believe that technological progress is ever-increasing our choices. (We tend not to focus on the options that technological progress forecloses.)

You'd think that possible futures would be much easier nail down than probable futures, but such is not always the case. True, some physical science constraints can offer poweful guides as to what is possible for us to do and what is not. We can't build machines that violate the laws of energy conservation, for example. We can't travel faster than the speed of light. But ecological constraints are not easy to specify. Species today are going extinct at 100-1000 times the historical background rate. (Some scientists call it the Sixth Extinction-the fifth saw the dinosaurs off.) Can this continue without fatally undermining the ecological support systems we and other living things depend on? We have little idea-but that doesn't stop us continuing this risky experiment. Political and economic feasibility are also very difficult to define. Example: we can, and have, built hydrogen fuel cell vehicles. But can such vehicles be a real economically-viable alternative to petroleum-fuelled vehicles? I doubt it. The International Energy Agency predict that global oil consumption will be 121 million bbl/day in 2030. Again, I doubt that this is possible.

Some technologies have modest resource needs, and so could expand without fear of physical limits. This freedom from constraints is evident in the glimpses of the far future offered us by computer experts. With them, the future is simply what they believe will be technically possible at that time. We need to be sceptical when considering these future utopias or distopias-we've been there before. All revolutionary technologies-the steam engine, the telephone, the motor vehicle, electricity, nuclear power, and now biotechnology, computers, and nanotechnology-have been accompanied by extravagant predictions. Some forecasters see the end of humans and the emergence of cyborgs-human-machine hybrids. This and other radical forecasts for computers may fail for several reasons. Their promoters may have underestimated the technical difficulties in, for example, implementing Artificial Intelligence in its strong form. Or Moore's Law could fail. Moore's Law is an empirical relationship stating that the number of transistors on a chip doubles every 18 months or so, and has proved to be reliable for half a century. This steady progress has driven advances in IT, and has increased the accuracy of predicting when computer hardware advances will occur. But some experts think that it may not hold for much longer.

Preferred futures: In our day-to-day lives we do not usually forecast; assuming that the relevant factors are under our control, we survey the possibilities and select our preferred option. We don't try to predict what show we'll watch on telly tonight, we look at the TV Guide and decide what we're going to watch. Often, governments can simply plan for the future, since they have greater control over events than we individuals. Government planners can collect and analyse past data, project these trends into the future, and then calculate what numbers (for air pollution emissions, or incidence of new AIDS cases, for example), result from this projection or baseline scenario. If the projected numbers seem too high, policies__obviously different from the the existing ones__can then be put in place to achieve the preferred future targets. But even governments like Australia's can't plan everything; they have little control over the international price of oil, for example.

In brief, the more in control of events you are, the more you can plan and the less you have to predict. You become future makers, rather than future takers, as they say. But even here, you have to foresee the consequences of the new policies. Planning always involves prediction. And here we face the problem of unintended consequences. Some years ago, activists Tom Hayden and wife Jane Fonda successfully lobbied for rent control in Santa Monica, California. But the landlords now had a greater selection of applicants. Not surprisingly, they chose young professionals (who were willing to redecorate apartments with their own money), rather than single moms-the opposite of what the Haydens presumably wanted. Not all causal chains are as simple as flicking on a light switch.

An obvious question arises with these preferred futures: 'Preferred by whom?' Different groups want different things-this won't be news for most of you. But surely there are some things we all want? We all earnestly want the Earth to be able to continue to support human and other life for many generations to come, right? Alas, no. 'Dispensationalists'-who may number in the many millions and are an important part of the US Christian Right-seem to believe that the last days are upon us. President Regan's first Secretary of the Interior, James Watt, was at least up-front about his beliefs: 'God gave us these things to use. After the last tree is felled, Christ will come back.' This testimony in Congress got him fired-fortunately. The present thinking seems to be that either the time left is too short for major problems to occur, or that God will provide. But-alas-all indications are that God favours a hands-off management approach to human affairs.

Where we can get agreement from all interested parties (e.g. for unleaded petrol or other air pollution controls) and enact the necessary legislation, the future is rendered more certain. For example, EU motor vehicle manufacturers have known years in advance what emissions limits will legally apply to all diesel vehicles in, say, 2010. But where there is no such consensus, as is the case for climate change in Australia and the US, predicting the future environment for business decisions, for example, becomes more difficult. And who can speak for the generations to come-assuming we don't foreclose that option?

Probable futures: Forecasting the probable future is what most people usually mean by prediction. Consider astronomy, where probable futures are known with high certainty-judged by the success of our past astronomical predictions. The paths of the planets are easy to predict because we have scientific laws-actually just regularities based on past observations-that accurately describe their motion. But an implicit additional requirement is that humans can't easily influence their motion. What would happen to our prediction of the tidal influence of the moon if humans started mining the moon's surface and sending the materials to earth? Forecasting would then not only involve astronomical calculations, but a schedule of lunar mining operations.

Hell is other people, said Sartre. Although it's often difficult to predict human actions, we are helped by the fact that we humans lead fairly regular lives. Our actions tend to be cyclic on a daily, weekly and annual basis. The daily and annual cycles are based on astronomical cycles, but the seven day week is a purely human invention. (During the French Revolution, 10-day weeks were briefly tried, but proved unpopular.) Travel patterns depend upon human regularities in work hours, business opening hours, and hours of sleep. If these change in the future, so will travel patterns. And yes, next week-and next year-Melbourne traffic will still drive on the left. Similarly, groups such as the Lions Club, U3A, or the AFL might publish a schedule of their meetings or games for the forthcoming year. (It might simply be that the group meets on the first Tuesday of each month.). Social scientist Daniel Bell in 1967 made a correct prediction for the US in the year 2000, namely, that they would have a presidential election that year. (There was-well, sort of.)

Donald Norman gives the following advice when trying to make predictions: examine the same time period in the past that you wish to forecast in the future. In general, this is sound advice. But the long-term trends we see for smaller household sizes, increased longevity, rising real incomes and so on, can only continue if the relevant conditions that prevailed over the previous 50 years remain unchanged. Here's an analogy that might help. Imagine a smooth steel ball set in motion on a vast, horizontal, frictionless surface toward a far-distant energy-absorbing barrier. We know that the ball's uniform motion must end at the barrier, even if the collision won't occur for another 50 years. The point I'm labouring to make here is that just because a trend-decreasing household size, for instance-has been with us for 50 years or more is no guarantee of its future continuation. Indeed, I believe that there are several barriers, or constraints, which must soon halt this long-term linear progression of global growth in incomes and resource use, trends that we presently think of as inevitable and irreversible.

Potential serious constraints include oil depletion, global climate change, fresh water scarcity, as well as loss of biodiversity, and risk of global epidemics. Projecting decades years ahead has always been an uncertain exercise, but this uncertainty looks like increasing. The probable future must of course be drawn from the range of possible futures, and as we have seen, deciding what is possible is not always easy. But forecasting must also take account of values, and the way in which these might change over time.

APPROACHES TO THE FUTURE
Prophecies-an unorthodox approach to future prediction: There is no shortage of long-range projections for planet earth and its people. I've even seen some-appropriately in The Futurist journal-that look ahead a 1000 years or more, giving a century-by-century description of what's going to happen. (It's all going to be wonderful.) The most unorthodox predictions are those provided by religious fundamentalists. For what it's worth, a 2002 poll found that most Americans believe that the prophecies (if that is what they really are) in the Book of Revelation are going to come true. (This last book of the New Testament, incidentally, has an unhealthy preoccupation with both violence and the number seven.)

There are many other prophecies of this type. They include those by the Oracles of Delphi, Cassandra of doomed Troy (inventor of the phrase 'I told you so, but would you listen.?), various astrologers, visionaries like Mother Shipton in 16th century Yorkshire, Edgar Cayce in 20th century America, and the infuriatingly vague prophecies of Nostradamus. (He may have had a good excuse: he was probably a Protestant sympathiser in France, at a time when directly expressing such views is what today is coyly called a 'career-limiting move'.) A set of prophecies ascribed to the Irish medieval scholar and bishop St. Malachy give brief descriptions about the identities of 112 future popes. The present Benedict XVI is the 111th pope described; the last is to be 'Petrus Romanus' (Peter the Roman). Benedict's Latin tag is 'gloria olivae' (glory of the olive): one tongue-in-cheek commentator saw this as predicting the end of the age of oil! These predictions, of course, owe nothing to recent past history, but a lot to events important in the lifetimes of these prophets. So do more conventional predictions I'll now discuss, but a forecast for the year 2050 made in 2005 is, I think, likely to be give us greater insights than one made in 100 AD.

More conventional (but boring!) approaches: Other, seemingly more secular, explorations of the future can be roughly divided into three categories. The most common could be termed a modified 'business-as-usual' approach, and include most official forecasts. We shouldn't dismiss these, as a study of the past 50-100 years shows that Australians have, for example, experienced a steady rise in per capita incomes, car ownership and travel, energy use/capita, educational level, and longevity. Our large cities have grown steadily larger, while household size and urban densities have continued to fall. Given these uninterrupted trends, we can easily justify predicting more of the same for the year 2050. The only uncertainty countenanced is whether future rates of change will be higher or lower than in the past. The State Government's plan/projection for Melbourne in 2030 is of this type.

The other two approaches both acknowledge that the future will in some way entail a profound break with the past-they have this in common with the prophecies I've just discussed. The first, which is closer to my own position, argues that because of a range of environmental challenges, including finite natural and mineral resources, and the finite capacity of the ecosphere to absorb waste products, limits to both resource consumption and economic growth will occur, requiring major lifestyle changes. But we need to be very careful here. It seems to be the conceit of people in every age-including ours-that their particular era is approaching some turning point in history, that a decisive break with the past is presently under way, or soon will occur.

The technological optimists, or cornucopians, on the other hand, assume that one or more radical new technologies will decisively alter society and the economy for the better. The obvious candidate is information technology, but biotechnology and nanotechnology are also increasingly discussed. The cornucopians usually see endless technological change, but no social change beyond that needed for us to adjust to the new toys. What's more, many computer enthusiasts see the new technology as effortlessly ending wars, poverty, ignorance, disease, and environmental/resource problems. Are they correct? Just take a look around you! As the 1960s protest song 'The Eve of Destruction' argued:

Some also argue that we are moving to a 'weightless economy', where physical inputs of materials and energy progressively decline per unit of output. The end point as, Nobel Laureate economist Robert Solow once notoriously argued, is that we can largely get by without nature. Again, an effortless solution to our problems. Or as the Vice-President of the American Petroleum Institute simply put it 'There are no limits to growth'. There may be limits to resources use, but not to human ingenuity. The modified business-as-usual forecasters also share this belief in continued technological progress, but in a more muted and less explicit form.

SOME PREDICTIONS FOR 2050
So far you could be excused for thinking that all I've done is elaborate on the difficulty-or impossibility-of forecasting. Here I'll give some specific indications of what I think urban Australia will look like in 2050. I'll cover both the physical and social aspects. Remember, I'll be mostly wrong-my excuse is that being right isn't the main point of future forecasts. It's to get us thinking about where we're heading, where we want to be in the future, and what the likely consequences are of various possible actions. (You're obviously not convinced!) I'll concede that in some cases, for instance if you're buying horse-racing or stock market tips, future certainty is exactly what you want. But the problem is that predictions of vital importance to us are often subject to self-altering effects. Examples? Warnings of failure of a small bank can be self-fulfilling and lead to its failure if nervous depositors withdraw their money. And prominent stock market tipsters can even successfully talk up certain stocks. On the other hand, warnings of a high incidence of AIDS cases can be self-negating and lead to measures being taken to avoid this outcome.

Future cities and their transport: Do large cities have a future in anything like their present form? We have to seriously ask this question-cities have been abandoned in the past. Think of the ancient cities of Mesopotamia, of the Indus valley, or the jungle-encrusted ruins of Ankor Wat in Kampuchea, or those in Mexico. Rome is once more a large city of over four million, but its growth has been far from continuous. At its height, imperial Rome had a population of between half and one million, but after the collapse of the empire in the 5th century, it fell to less than 100,000. Rome's survival has of course been helped by its position as the headquarters of a major religious franchise.

Even New Orleans has been temporarily abandoned, because of too much water. Like Cassandra, the scientist's predictions for New Orleans were accurate-but ignored. The future of sizable cities like Sana'a, capital of the Yemen, and Quetta, in Pakistan, are also in doubt-because of too little water. Some towns near the Chernobyl nuclear disaster have been abandoned. However, the UN sees cities continuing to prosper, with further growth in the share of the world population in cities. In particular, they see steady growth in the urban population of Australia out to 2030. So do official Australian forecasts, which predict an even higher share of Australian population in our large cities in 2050 than is the case today.

David Orr, a US environmental academic, is having none of this. He sees the large cities of today as an 'aberration'. Accordingly, he thinks that 'long before 2030 the trend toward ever-larger cities and an increasing ratio of urban-to-rural dwellers is likely to have reversed'. He argues that this reversal will occur not only because of the resource and ecological costs of large cities, and their susceptibility to diseases such as AIDS and tuberculosis, but because of 'the unmanageability of all cities beyond a certain size'. A very American view, perhaps. Here, I'll simply assume that no major catastrophies occur that would render this discussion pointless. No collisions with 10 km diameter asteroids, no Black Plagues. Sea level rises to the year 2050 are likely to be manageable for Australian cities.

Yet another view is put forward by some authors terribly impressed by the rapid advances in Information Technology over the past 2-3 decades. They believe that the new technology will make present cities obsolete, just as electric power made the steam engine obsolete. And that advances in IT will make much travel, including urban travel, redundant. Mitchell, an expatriate Australian urban planning specialist at MIT, is one who has developed this idea in detail. He uses the term 'demobilization' for the substitution of work, shopping, and other trips by networked computers. But arguments to the effect that IT will radically reduce urban travel needs have now been made for almost three decades. Actual results so far have been very disappointing.

Let's look at global climate change and Australian/global oil depletion. Many technical fix solutions to these two now-pressing problems have been proposed. Present favorites include carbon sequestration, hydrogen fuel cell vehicles (with the hydrogen ultimately being derived from renewable energy), liquid fuels derived from biomass, and huge increases in vehicle fuel efficiency, as in the Rocky Mountains Institute concept vehicle, the Hypercar. These fixes, the technological optimists argue, will permanently answer any resource/environment challenges to the car. With the so-called 'green car', the historical end point of urban transport will be reached ('.and they lived happily ever after.')

If, as I believe, these two problems are serious, and not capable of speedy technical resolution, Australia's present urban transport system is likely to change dramatically. Travel change can happen fast: in 1947, about 80 % of vehicular travel in Melbourne was by public transport. By 1964, less than one-third was. I think some features of the public transport era will reappear, with alternatives to the car having a far greater role than today. If motoring costs rise, or if motoring is restricted, (petrol rationing again?) urban residents will adapt in a number of creative ways. In the short term, they can increase vehicle occupancy rates and satisfy their entertainment, recreation, or shopping needs more locally. They can also use public transport to travel to existing destinations that are more accessible by public transport.

It doesn't have to be catastrophic. Our large cities, particularly Sydney and Melbourne, are fortunate in having extensive fixed-rail electrified networks, which is some insurance against oil shocks. Eventually, what I call the 'logic' of urban travel could change once more. Once we realise that most trips won't be made by car any more, both destinations and even the timing of our activities will change to reflect the new transport realities. For some time we'll still use cars for trips unsuitable for other modes because of location, time of travel, or the need to carry goods. But the present automatic choice of car travel for nearly all our trips will no longer prevail. Exactly how travel will change depends on government policy. If, during the transition, transport fuels are subject to general rationing-as presently with water in Melbourne-what car travel remains will be spread more evenly over the urban population, with the car used selectively by all. But if transport fuels are rationed by price, car travel will mainly be restricted to higher income urban groups-we'll get a nicer class of driver. Nevertheless, this is just for the transition-by 2050, I predict far fewer people, rich or poor, will be driving cars.

I don't foresee much change to cities themselves, for reasons I outlined earlier. City populations will continue to increase in the short-term, but growth could slow or even reverse in the longer-term. Sustainable agriculture is likely to be a labour-intensive enterprise, so we could see some drift back to rural areas. The uses we make of various buildings in our cities could change, and corner-stores could once more reappear. Far fewer flights will leave our airports each week.

Energy: Not only will we be using the car and motor fuels a lot less, we'll each have to use less energy in general. We can easily improve what I call 'social efficiency'-that is, we can get more useful services from a given amount of energy. Two examples. A TV set provides little social benefit if no one is watching, so improving social efficiency means getting more total viewer-hours per hour the set is on. For transport we can increase seat occupancy rates for all modes. So by changing our practices we can greatly reduce our energy needs. When you think of it, social efficiency is all we can change in the short term. Given time, we can also further improve the technical efficiency of our devices. We'll have to get by with a lot less energy, as running our economy on renewable energy won't be easy or cheap. And in 2050 we'll see many more more rain water tanks in cities, and a lot more re-use of 'grey' water for our gardens and parks-and toilets.

But all these energy savings in transport or in the home will not get us far if the saved energy is simply used elsewhere in the economy, and CO2 emissions don't fall. Were William Wordsworth around today, the lines in his poem 'The Solitary Reaper':

would not have been inspired by the haunting song the highland lass sang as she reaped, but by the 'Suburban Leaf Blower'. We've all witnessed-and heard!-one of these loathesome devices chasing half a dozen leaves up and down a driveway. Leaf blowers show how good a growth economy like ours is at finding extra uses for any energy we manage to save elsewhere. So we may need to simply ban some energy-using practices. Another alternative: in the U.K., they're even considering energy rationing. All residents would get a 'carbon card' entitling them to emit a certain amount of CO2. They could sell some of their quota if they want to.

Future society and its beliefs:
Some politician (I forget who) once stated that human beings can be counted on to do the right thing-after all other avenues had been explored. Sooner or later, but I think well before 2050, it will be universally acknowledged that the earth has only finite resources and pollution-absorbing capabilities. And that human ingenuity, the supposed guarantor of perpetual global economic progress, can only take us so far. The mindset that we Westeners have had since the Industrial Revolution, one now increasingly shared by most of the world's population, will change. Rakesh Kapoor, an Indian futurist, explains the reason for our quasi-religious belief in endless technological progress:

I also think that the nation-states of the world in 2050 will necessarily manage their affairs in a far more co-ordinated manner than is the case today. They'll have to. There's no point in some countries cutting back on greenhouse gas or ozone-depleting CFC emissions if other countries boost their emissions. In fact one proposal (Plan B) calls for all countries to converge on the same low per capita CO₂ emissions by 2050. After all, why would countries like China agree to accept lower per capita emissions than the OECD countries? As a result, it's probable that there will be increasing economic equality at the nation state level. Greater equality inside nation-states is less certain-but it's the preferred option for most of humanity, and one I think with some chance of success. What's also clear is that the closer co-ordination between countries can't be built around existing bodies like the WTO, whose guiding philosophy clashes with the need for ecological sustainability

CONCLUSION: RESPONDING TO FUTURE UNCERTAINTY
Prediction is difficult, especially about the future, as Mark Twain once so wisely observed. It is not the mere existence of global challenges like climate change that pose a problem for forecasters-even if it does for the rest of us. It's the uncertainty of their severity and effects, and of our decision-makers' policy responses-if any-that cause the prediction problem. The barrier to the motion of the steel ball in the analogy I gave earlier, will drastically alter a long-term trend, but its effects can be predetermined, hence predicted.

In trying to get a clearer picture of the year 2050, we face a paradox. On the one hand, I've implicitly argued that we'll get a better view because we are approaching some limits-I've mentioned annual oil production and cumulative greenhouse gas emissions. These, I have argued, act to constrain our possible futures, in principle making forecasting easier. If the world had today's technology but only 100 million people, this would not be the case-our options would be greater. For oil, this constraint is surely valid. On the other hand, recent research in a number of scientific fields, including ecology and climatology, suggests that large shifts in regional vegetation or climate-related factors can occur quite suddenly. For example, starting about 9000 years ago, vegetation in the Sahara followed a gradual but fluctuating decline over several thousand years. But then about 5500 years ago it collapsed suddenly to near today's low levels. What this and other examples show is that as critical thresholds are approached, sudden shifts to alternative states can occur. So, as human environmental impacts intensify on ecological or climate systems, instability can make the future of the world's physical and biological systems less certain. Could it be that similar considerations also apply to our social world?

In general, we try to forecast because we believe that at least some elements of the future is predictable. If so, it's useful to separate out the fairly predictable from the inherently uncertain. That's what I've tried to do by separating out predictions about our future cities and their transport, and the beliefs and politics of the future. But, given the above discussion, we urgently need to continuously monitor both our physical and social worlds for extreme, unpredictable, events. In areas of great future uncertainty, we'll need to adopt strategies which are less dependent on forecasts. Instead, we must take actions to avoid catastrophes where this is possible, and plan to blunt their likely impact where avoidance is not possible. This is not a new idea: it's why we have fire brigades. Responding to global climate change or global epidemics requires elements of both. In the face of rising uncertainty, we'll need more robust infrastructures, economies and social systems. And don't forget the words of US futurist Edward Cornish, echoing Karl Marx: 'The main point in thinking about the future is to change it.' Whatever happens, 2050 is likely to be an interesting year.

ADDED IN POSTSCRIPT
Before the talk, I handed a sealed envelope to the Chair, who opened it and read it out after talk: 'I predict that you will give polite applause after my talk'. At last! An accurate prediction!