Many people lie in a casual ‘I’ll recognise it when I see it’ relationship with science. That ambiguity gives room for any interlocutor to add sudden vagaries to their criteria, hurriedly adding and removing things from their definition of science to suit their needs. Science can suddenly need to be given a ‘direct observation’ criteria because someone doesn’t like a particular scientific principle, ignoring the fact that this throws out all of forensics, for example. Having this relationship with science is ― as we shall discuss ― degenerative. And so this is an exploration of one more steady idea of science; not something you can recognise after the fact, but actively hunt for.
The purpose of this post is to take you on a small journey from an entry level concept within the philosophy of science to one of the more nuanced (but less well defined) ideas that are built upon it.
There is a well heard of (but less well understood) demarcation to delineate science, Popperian falsification. And that acts as a good rule of thumb, but there is a lot more to say about defining science than just that. It has its weaknesses and a lot of development has happened based around this early idea. One of those developments (as well as encompassing other elements of the philosophy of science) is Lakatos’ programs of science, and it includes a broader definition of science.
This breadth is alluded to when Sam Harris tries to point out there is no demarcation between science and philosophy. Instead, Harris believes there is a demarcation between rational projects and pseudorational projects, and that science should not be considered an endeavour distinction from the rest of human rationality. That demarcation is whether one is making “valid claims” about reality (Harris, 2014). There is a lot to be said about this broader definition of science, and understanding why and how it applies to modern disciplines is vital for having the more nuanced conversations about the power and limits of science.
A “valid claim” would appear to be one that is supported by specific empirical data, consistent thinking and good reasons more generally. The validity of claims are evaluated in a knowledge-centric context, because reality-centric contexts are impossible to use; knowledge-centricism is as close as we can get and as available to us as reality gets.
When I say ‘specific empirical data’ I allude back to falsification, where there must be a difference in the data one expects if a hypothesis or idea is true or not. But there are problems with falsification.
Falsifying a theory is easy. All one needs is a slightly erroneous conception of the model, or dishonest scrutiny, or an experimental design flaw, or a mistake, or an honest-to-goodness anomalie. Valid falsification of an actual theory is slightly harder. That is why there are a number of philosophies of science that take the spirit of Popperian falsification but reject the strict letter of it.
Deutschian fallibilism (Deutsch, 2011) encompasses all that I have discussed here. It is knowledge-centric and relies on building conjectures out of existing data to make predictions about future data. It’s about creating explanations that hold up against scrutiny, and holding them as provisionally true.
This is an element of the philosophy of science I was to briefly look at, for its initial detraction from what science of often thought as, but its subsequent very powerful very of what science is and the blurred demarcation between science and nonscience. It is Hungarian philosopher Imre Lakatos and his programs of science I want to look at (SisyphusRedeemed, 2017a, 2017b, 2017c)
Lakatos’ programs of science are well illustrated by Hawking and Mlodinow, talking of the fact that a ‘Grand Unified Theory’, or theory of everything, may actually consist of a number of overlapping ‘maps’, a granulated theory of everything. The idea is that many programs can share an intellectual space, so long as they are “progressive programs” (I’ll get to the definition soon). For example, psychology can be populated by many explanatory programs: cognitive psychology, evolutionary psychology and behavioural psychology. The programs do not compete with each other; so long as they function, they are defensible human rationality.
Lakatos’ programs have two key features that allow it to circumvent some of the concerns with falsification, like data not always being reliable and a body of contradictory data may exist merely as a consequence of a large volume of investigations, some proportion of which were anomalous or poorly designed, or smaller and nonessential elements of a theory being attacked which don’t necessarily justify throwing out the whole theory.
This first feature is that the program has a conceptual anatomy: there is an essential core that defines the program, and there are supporting bridges, which are more flexible. In evolution, for example, the core is concepts like descent with modification and common ancestry; the support bridges are things like natural selection. If natural selection were found to be lacking ― say, because microbes are so sensitive to mutation and variation that a great deal of their features were better explained by non-deleterious variation than survival ― or if genetic variation were found to be less important than epigenetic pressure, then those supporting bridges would either be adapted or abandoned, or more bridges added. The core idea ― genetic changes across a population or common descent ― doesn’t need to change. Under Popper’s falsification all of evolution would have to be thrown out, instead of developing and refining the idea.
The second feature is that Lakatos’ programs don’t rise and fall based on their ability to account for existing data. Existing data may well be wrong. It rises and falls based on its ability to make predictions. A good program is one that is “progressive”, which means it can make specific predictions and when a supporting bridge is attacked it is refined in such a way that improves its predictive ability. The program is progressing.
By contrast, a degenerative program is one that loses its ability to make predictions by becoming either increasingly vague or increasingly wrong as it adapts to criticism. Marxism is a good example of this, as one of the supporting bridges of Marxism predicts that the first class revolutions would happen in industrialised states, but the first major one happened in the mostly agrarian Soviet Union. To make sense of this, Marxism had to be adapted and forms of Marxism no longer make reliable predictions; it has degenerated. (That is not to say it can’t be saved, and that class struggles really are the best predictor of socio political situations, but the prediction has to be made before the event.)
One of the consequences of this is that it is not possible to tell what programs are ‘progressive’. It is possible to tell that a program has been progressive, but a current adaptation of the program may have destroyed that; it is only possible to tell that historical programs have been progressive. Lakatos doesn’t accept the idea of an instant recognition of rationality, but instead offers us a retrospective rationality. However, it is possible to say that peer review will slowly strip away any degenerative variation to an otherwise good program.
To answer the title of the post, it is important to understand some of the philosophy of science to recognise that science is not some perfect and discrete entity. Lakatos would, no doubt, agree that the program of science has been progressive and has found defensible facts and models and programs of reality. But it requires scrutiny. It requires attention. Its boundaries are blurred.
(If any one else would like to write about a philosophy of science, I’m interested in hosting a guest writer for that.)
Post-Kuhnian Philosophy of Science: Imre Lakatos (1 of 3) (2017a) [online]. Directed by SisyphusRedeemed. (no place) Youtube. Available from: https://www.youtube.com/watch?v=VExVaR8S_wQ [Accessed 24 March 2017].
Post-Kuhnian Philosophy of Science: Imre Lakatos (2 of 3) (2017b) [online]. Directed by SisyphusRedeemed. (no place) Youtube. Available from: https://www.youtube.com/watch?v=ezJH8MTgTxQ [Accessed 24 March 2017].
Post-Kuhnian Philosophy of Science: Imre Lakatos (3 of 3) (2017c) [online]. Directed by SisyphusRedeemed. (no place) Youtube. Available from: https://www.youtube.com/watch?v=7UBP06tK2g8 [Accessed 24 March 2017].