Response to Husain Sarkar

Question: Taking into account sections II and III of Chapter I as well as the prisoner’s dilemma, how would you as an individual scientist structure your society of scientists?


In practice the crux of this question turns not only on the structure I choose for my group of scientists but also on the capacities of a single individual scientist amongst his fellow scientists. I assume we are to overlook this practical problem.

In theory the answer to the question depends on the approach: one approach would be to give my (rather naïve) particular appreciation for what constitutes the best structure for a society of scientists. Another approach, the one I am inclined toward, is to do a study of many individual structures and combinations of those structures against one another in some principled manner in order to discover the optimal structure for a group of scientists. My preference rests with my inclination that what I fancy as the best structure of a society will in all likelihood differ from one of my colleague’s favorite structure, and so going about things in that route will bring me to a political impasse. That is, the former approach is subjective in the worst sense.

The problem with the latter approach lies in analyzing structures. The intuitive method would be to lay out each structure in some list, choose one and compare it against the very next structure on the list. Whichever structure turns out to be better is next compared to the third structure on the list, then the victor of the this comparison is compared against the fourth structure, and so on until we have finished the list. One potential problem with this approach lies in the assumption that there are finitely many structures (and hence combinations of these structures) in principle (since this approach is theoretical). If there are infinitely many structures then our work is never done: for any finite list of structures our analysis will give us a victorious structure (or group of structures), but we must then compare it to one of the infinitely many structures that did not appear on our list, and again ad infinitum, in order to arrive at the best structure. So there’s one hitch. Another, as I alluded to above, is the problem of choosing between two structures that yield equal scientific results. This problem may never arise (note: the previous problem cannot arise in its principled form; it is more-or-less a specter over the entire method), but it remains a problem for the theoretician.

The big assumption of this approach, though, is that one can compare scientific structures, which according to Kuhn is impossible. This assumption is a playing field for further problems: what is to be compared in scientific structures, results or principles (or norms if you prefer)? If we answer “results” then how are we to valuate the results of any member of a set of theories? Here the answer is always “against some standard”, but implicit in this response lies a presupposed standard. I will leave this regress here, but I trust one will note that we need not. Alternatively if we answer “principles” then we are left with the same problem as we encountered when we responded “results”: what is the standard to which we evaluate principles? Again, this is necessary if we are to compare structures. That given two structures an individual scientist can respond to the question “Which structure is best?” with “theory a” or “theory b” does not get us out of this theoretical impasse. Unless there is an underlying principle or method for determining which response the scientist should make we are left with the problem of comparing scientific structures theoretically.

So the answer to the question for my part has to be that at the moment I am completely unable of choosing a structure for scientists in a non-ad hoc way. So let me choose the option I am most fond of right off of the top of my head: structure the group in accord with the principle that each scientist should practice in their own specialized field only as much as their scientific principles that are reconcilable across scientific domains allow, and to otherwise work with one another to resolve conflicts between principles that are irreconcilable over scientific domains. This principle intends for a homogenous scientific edifice.