In the previous blog
post, I introduced a poser to show an easy way to establish basic
scientific principles. Here, I want to say that despite there being good and
bad ways to make inferences, the idea of there being a single ‘scientific
method’ is problematic, as the complexity of scientific inquiry cannot be
encapsulated by one method or definition. That is to say, the idea that any
human mental processes can encapsulate a singular method for science is a
faulty one.
Last time, we applied modus ponens (affirming the antecedent) and modus tollens (denying the consequent) to explain why Team B’s method is more reliable and why modus tollens can be effective in some cases. Now, consider the affirming the consequent method:
If P, then Q.
Q.
Therefore, P.
While this logical form is a fallacy, science sometimes uses it to form hypotheses - yet it's precarious without corroboration. For example:
If a carbon atom has 4 valence electrons, then
it can bond with up to four other atoms at the same time.
A carbon atom can bond with up to four other
atoms at the same time.
Therefore, a carbon atom has 4 valence
electrons.
While it is a fact that a carbon atom can bond with up to four other atoms at the same time because has 4 valence electrons, this kind of hypothesis would not qualify as a scientific theory were it stated as a one-off isolated hypothesis. To see why, let’s use a comparable example:
If it is midnight, then my watch will say it's
midnight.
My watch says it's midnight.
Therefore, it must be midnight.
In isolation, this kind of fallacy of thinking would be exactly true of science too – one claim by one scientist is inadequate to the task of a reliable inference; we need corroboration from repeated sources, because if stated in isolation there may well be something (as yet unknown) that falsifies the proposition that “a carbon atom can bond with up to four other atoms at the same time because has 4 valence electrons”, just as seeing my watch has stopped gives me reason to doubt whether it really is midnight.
But equally, suppose my watch says it’s midnight, and also my neighbour’s watch says it’s midnight – I have a much better reason to think it’s midnight, as two watches in 2 houses side-by-side stopping at midnight is less likely. The more verification I get when I look at other people’s watches and see they say midnight, the stronger the corroboration, and the stronger the justification for my belief.
That is what science is all about. Science uses arguments at the singular level, but they become stronger with further corroboration. We infer to the best possible explanation, where a theory best explains the facts in front of us. Of course, if scientific theories are inferences to the best explanation, then the best explanations will also have good predictive success rates too. But context must be established in order for this to happen. For example, the earth is both spherical and flat when seen from the right context. When I place a spirit level on my concrete driveway, I see it as flat, but if I view the earth from space then it is spherical. It would be a fallacy to measure my driveway as flat and infer that the whole earth is also flat.
The strength of science is found in inferences based on repeated corroboration, but there is no single scientific method from which we obtain this. While we often use the term 'scientific method' for ease (by which we mean testable, repeatable, verifiable, and predictable), it is difficult to justifiably claim science to be amenable to some kind of singular ‘scientific method’. It does little good to simply say that the value of science is in an observation being testable, repeatable, verifiable and predictable, because science is only one particular lens of reality, which by itself has no singular ‘method’.
Even if we ignore for a moment the fact that science is limited to only a scientific lens of reality, there are many philosophical questions attached to the process of an observation being testable, repeatable, verifiable and predictable. How does one confirm that one’s observations are sufficiently free from psychological bias to be balanced? How does one decide what is testable? At what level does repetition confirm the validation of an observation? How can prediction be informative without a philosophical framework to police our concepts? What links our methods of experiment with the complexity of data? How do we account for the fact that different levels of reality are attached to different levels of physical behaviour in nature? How does the overarching narrative of our interpretation of reality align with the lenses of reality to which science is amenable?
It’s not that these questions can’t be answered; it’s that there is no singular ‘method’ by which they are all answered, as humans have such a complex nexus of perceptions and conceptions, and require many lenses of interpretation with which to understand the world.
Having built the foundations in parts 1 and 2, we are now ready for a meatier part 3, which is up next: Science & Climate Change Part III: Understanding the Limits of Climate Models in Risk Assessment