Ideally you would consult a list of priorities and decide accordingly, taking into account national strengths, weaknesses and needs. But the chances of finding such a list of priorities are slim. Most governments have never asked for such a list and scientists have certainly never volunteered one. There are signs, however, that this is changing.

As always, we can learn from mistakes. The recent trend towards high-powered government-wide co-ordinating councils, for instance, is not the way forward. Such committees have a poor record on both sides of the Atlantic: the National Science and Technology Council in the US rarely meets, while the Council for Science and Technology in the UK has been all but invisible since it was set up in 1993.

Within the UK, the two research councils most closely involved with physics adopt very different approaches. The Particle Physics and Astronomy Research Council is taking the more strategic approach and has drawn up a “road map” of scientific opportunities that, hopefully, can be reconciled with the available budget. With a broader range of subjects to support, the Engineering and Physical Sciences Research Council has developed an algorithmic approach: two panels, each made up of business executives and senior academics, judge physics, chemistry and six other programmes according to a range of criteria. The council then decides the level of funding for each, guided by the panels’ votes. The system ensures that priorities are set, but the method is anything but transparent.

The National Science Board (NSB) in the US has revisited the subject of priorities in a recent working paper, Government Funding of Scientific Research, that concludes: “We are aware of no examples of the scientific community agreeing on the relative priorities for investment across scientific fields. Although many scientists consider the task both undesirable and undoable, the NSB believes that this difficult task will become increasingly important and must be faced over the next few years.”

The NSB report champions the findings of a little-noticed 1993 study by the National Research Council. That report, National Goals for a New Era, is remarkably clear. It advocates that the US should be among the world leaders in all areas of science and that it should maintain clear leadership in some major areas of science. It calls for field-by-field peer assessments by expert panels and even considers, hypothetically, reducing funds for areas in which the US leads the world but in which world leadership is not important. But no other country is as wealthy as the US, so being among the world leaders in all areas is not an option elsewhere.

In a rare example of explicit priority setting, Austria recently asked the European Science Foundation (ESF) for an assessment of two proposals for a transnational research facility – a pulsed neutron source and a crystal-growth facility. The ESF panel, which did not contain any Austrian representatives, did not mince its words: the crystal-growth facility is a non-starter and the neutron-source proposal will need a lot of work if Austria hopes to attract financial support from neighbouring countries.

But without the advantage of size as in the US, or a straightforward question as in Austria, how can priorities be set across science? There are no easy or obvious answers but somewhere, or somehow, astrophysics is being weighed up against zoology in decision-making processes around the world. These are important questions that need to be addressed in public and in detail in every country that takes science seriously. A panel of wise men and women from all disciplines has obvious flaws – the very make-up of the panel will inevitably colour its outlook – but in the absence of anything better, such a root-and-branch (and atom and molecule) review of all science would be an improvement on what is happening now.