Jun 24 2008
Prove, Shmoove
Over at ScienceDaily yesterday I encountered this headline and article: New Discovery Proves ‘Selfish Gene’ Exists
As you might guess by now, the word nerd in me went critical all over some of its sorry letters. As I have mentioned before, the desire for precision is part of doing science. As is expressing, explaining, and arguing over data, concepts and theories.
My beef about the title is twofold. We’ll tackle the lesser matter first.
A “selfish” gene. Okay, the editor/writer put it in half-quotes. Good. They had to. A few paragraphs into the article we encounter this:
In studying genomes, the word ’selfish’ does not refer to the human-describing adjective of self-centered behavior but rather to the blind tendency of genes wanting to continue their existence into the next generation. Ironically, this ’selfish’ tendency can appear anything but selfish when the gene does move ahead for selfless and even self-sacrificing reasons.
Genes “wanting”? We obviously need more words, better words, more accurate words. So many of our words originated in a context of social relations. When we haphazardly use these same words to describe non-social phenomena, we can mislead, particularly because our brains are primed to find social implications, even where none may exist. We are a hyper-social species. That’s what we do.
The more philosophically important beef over words I have with the title is “prove.” Science is not mathematics. Sure, math can be an integral part of it, but it doesn’t work the same way. With math you can develop “proofs” of concepts and theorems. A correct result is a proof of concept/theorem.
With science that is NOT the way things work. What data and research findings do is provide support for a hypothesis; they bolster theories. One experiment, one study proves nothing. Even a hundred studies with results in line with predictions prove nothing. A proof is absolute. While math can cleanly identify, define and isolate all of its variables, and subjective biases rarely enter the picture, that isn’t the case with science.
Rather than providing proof, scientific findings support hypotheses; they boost our confidence in a theory; the increase the probability that our understanding of how things work is accurate.
An example: Two research teams working on the subject of whether or not taking vitamin x is good for your health. Thanks to its protocol/methods, team one finds an increase in its measure of health. Team two, perhaps using different methods and measures, arrives at a null result. Has team one proved the vitamin works, while team two has meanwhile proved it doesn’t?
If you want absolutes, you’ve got to look outside of science. Like to religion, which will provide you with absolute answers.
In terms of science, the best we can do is accumulate a mountain of evidence that brings us great confidence that the hypotheses, theories and models of the universe we hold are accurate.
While it is always possible to increase one’s accuracy, once something is proven, game over.
Another example from a news release about a scientific study on the diversity of dog traits:
In the cover story of tomorrow’s edition of the science journal Genetics, research reveals locations in a dog’s DNA that contain genes that scientists believe contribute to differences in body and skull shape, weight, fur color and length — and possibly even behavior, trainability and longevity.
This was a strong bit of scientific research. The quantity of data was very good: 13,000 samples of dog DNA. The quality was also good: DNA analysis is quite precise with little problems in measurement, at least relative to the social sciences.
Can we conclude that this one, strong study, “proved” something about dog diversity, or even that evolution is a fact?
No we can’t. Sure, it increases the confidence we have in a hypothesis and perhaps theory; it increase the probability that our model is accurate. Good data – yes. Solid evidence – yes. Final proof? No.
[...] To read more of that post, click here: Prove, Shmoove [...]