• Question: How much research do you have to do before putting forward a theory?

    Asked by anonymous to Steve, Emma, Christian, Bose, Alice on 9 Mar 2016.
    • Photo: Steve Marsden

      Steve Marsden answered on 9 Mar 2016:


      In particle physics (and I imagine other fields of science as well) scientists are divided into experimentalists and theorists. As the names suggest, the theorists propose new models, which the experimentalists then go out and attempt to (dis)prove them.

      A particle physics experimentalist will typically work on a couple of analyses at any one time, with each analysis lasting about a year.

      I’m an experimentalist, so I hope theorists will forgive me if I get the process slightly wrong… A theorist during their PhD will often spend a few years working on one theory proposed by their supervisor. During this time they will be performing sanity checks on the model. For example, making sure that the theory gives the correct values for known parameters, and sensible values for any unknown parameters. (For example, if a theory says that the probability for something happening is greater that 100%, it’s probably broken). After their PhD, a theorist may continue working on other theories, maybe proposing fixes to theories that are known to be broken.

      A professor who has spent decades in the field may have a sudden stroke of inspiration and write a paper documenting their theory in a matter of weeks. They may then pass it to a PhD student to study the properties of the model, completing the cycle.

      Once a theory has been studied for a while by theorists, it may be picked up by experimentalists who may then start performing experiments to check the model’s predictions. Some times this whole process from first hypothesis to discovery takes a few years, however in the case of the Higgs boson, it was almost 50 years between it first being proposed, and it being shown to be true.

    • Photo: Alice Harpole

      Alice Harpole answered on 9 Mar 2016:


      It depends on the theory! Generally before you put forward a new theory you need to show several things: that it is consistent with observations/experiments, that it agrees with related theories and how it improves on the previous theory. You can show these things in different ways: in my work, I build computer simulations to show what my theory predicts will happen. To show it improves on the previous theory, you can make new predictions of things yet to be observed. It’s then the experimentalists’ job to try and measure these new predictions and prove your new theory right or wrong.

      All these things can take a long time. Some new theories are thought of in a matter of minutes in a ‘eureka!’ moment, and their details worked out in a few weeks; others take many months or years to fully develop.

Comments