Here’s the thing: Occam’s Razor might help you decide on which explanation of the bump in the night you accept as “true” (your mental construction of the phenomenon), but it doesn’t actually have any value in determining what actually happened (the animal or the tree branch). But that scenario has almost nothing to do with the Scientific Method. You could use it to help you scientifically understand your “bump in the night” scenario, but not in the way you’re suggesting (or most people would use it).
To understand your “bump in the night” from a scientific perspective, you’d need to do the following:
- Describe the phenomenon.
- Propose hypotheses for what may have caused the phenomenon.
- Test those hypotheses.
- Summarize the results empirically.
- Re-test those hypotheses.
- Draw conclusions from repeated testing.
- Propose generalized principles (that can be tested) to explain similar phenomena.
Your example of the bump in the night doesn’t work, because you’re talking about a single case, and “ruling one out” isn’t an example of either Occam’s Razor or finding an “explanation.” Since you didn’t observe the event, any amount of “ruling it out” is just playing with mental models. It may help you to feel as if you have an explanation, but from a science perspective, it doesn’t help at all.
However, if your “bump in the night” is a phenomenon that continues to happen (say, around 11:00pm on nights in the Spring and Fall) over time, then you might find a use for the principle of Occam’s Razor. In that case, it can help you determine which hypothesis you should test first (because it is simplest and “cheapest” in terms of resources). So, consider the above, and then add these hypotheses to explain the repeated bump in the night:
A. A raccoon is playing on your roof.
B. A tree branch is brushing the side of your house.
C. Aliens are periodically using mind-waves to disturb your sleep in some kind of experiment, and your pitiful human intellect construes that as a “bump in the night” because it’s the only way you can process the information given your limited experience.
In this case, you can absolutely use Occam’s razor to help you scientifically assess the bump-in-the-night phenomenon by picking the simplest hypothesis or set of hypotheses to systematically test first. So, drop C and let’s focus on A and B. Set up a system of game-cameras with enough sensitivity and coverage to record both animal and branch movement. Analyze data over six months to a year, and see if there is a correlation between animal sightings or tree branches being swayed by storms. Follow the rest of the process and draw conclusions. If neither animal behavior or tree branches are correlated with your bumps-in-the-night, then go on to test the next hypothesis (Aliens fucking with you).
That’s it. Occam’s Razor has no explanatory value in itself, because the “simpler explanation is better” rule doesn’t correlate to empirical reality at all. No single event can be explained in such a way. But, Occam’s Razor is fine for determining how to systematically explore a system and generate and test hypotheses. If you have two hypotheses and either seem comparably plausible from a mechanistic perspective, choose to test the simpler hypothesis first.
If, at the end of the day, you have equal evidence to support two hypotheses, which is unlikely in empirical, quantitative science, and further tests don’t reveal way to distinguish which is more applicable, then Occam’s Razor still doesn’t help you determine a scientifically valid explanation. The only valid scientific explanation is to state the results of the experiments: neither proposed hypothesis is sufficient to explain the phenomenon on its own. Occam’s Razor has no explanatory power at all, but it might help you sleep better at night if you need an explanation.
As for hummingbirds and genetics, you’re kind of missing the point, I suspect. Do hummingbirds that have 700% of a “normal” human’s BG levels (as opposed to “normal” hummingbirds at 600%) develop diabetic complications? Forget hummingbirds (unless you’re a veterinarian or can’t find a better experimental model) altogether, and focus on relevant questions. Do diabetic complications (peripheral neuropathy, retinopathy, etc.) develop in non-diabetic human populations? Do such complications occur in those with certain “genetic profiles” but no overt symptoms or diagnoses of diabetes mellitus? Do complications occur in the same proportions in populations of diabetics with different levels of BG control?
There are good scientific protocols for disentangling how different factors (like genetics, BG control, chemical exposure, exogenous treatment, or phase of the moon) contribute to something like the development of diabetic complications. The reason we don’t have good models (which would like something like “the probability of developing retinopathy P(Ret) = A x GeneSeq + B x GlucCon + C x Medication + Error”) isn’t because of failing to apply Occam’s Razor. Rather, it’s because we don’t have sufficient data and resources to meaningfully construct and test such models of the development of diabetic complications. It will be done, eventually, but it goes back to what I said above: we (scientists) need more time, more data, new techniques, and more resources to adequately understand such complex systems well enough to inform treatment decisions.
There is a practical takeaway to all of this, which you may or may not find relevant: there exists no consensus treatment recommendations to prevent diabetic complications. As such, the best we (as individuals) can do is support further research and make decisions based on what we do know. At this point, that is not a whole lot, but we can do something about the few factors of diabetes we can control. I.e., we can
- do our best to minimize frequency and duration of hyper- and hypo-glycemic episodes;
- we can get regular examinations from eye-doctors, heart specialists, and neurologists (so that early onset of complications can be detected and treated, of not prevented);
- and we can determine the level of effort and resources we can devote to (1) and (2) above based on emotional, financial, and temporal considerations.
I can’t change my genes, and I can’t make myself non-diabetic. I can’t go back and prevent whatever damage my pre-diagnosis period of hyperglycemia did. I can take my medicine and adjust the levels to try to achieve as “close to normal” levels as possible while minimizing hypoglycemic episodes; I can go to my eyedoctor and dentist for yearly exams; I can take my GP’s advice on whether I need to consult a cardiac or neurological specialist; I can eat low-carb and exercise a lot (sorry, I know you think only monkeys should exercise), which helps in both minimizing frequency and duration of hyperglycemic events but also minimizing variance in BG; and I’ve determined that I can afford to do all those things, because the effort, time, and money are worth it to me in the chance that it will minimize my complications. Maybe it will, maybe it won’t, but I’ve made an educated guess that it will.