Ooh, this looks like part of my Personal Expertise: "Modeling Life on Mars" by Erica Dietlein (@ericadietlein) is my next #SoCIA18 talk!
Begins with some fun paradoelia (sp?) examples - humans seeing patterns where there are none.
Which leads, of course, to the Canals of Mars. Original Italian "canali" can mean Canal or Channel (natural or artificial).
And lo, we see patterns in Mars rocks that maybe look like life-related patterns in rocks on Earth.
So when we talk "Earth Analogues" as models, we mean ~geological features on Earth that resemble features on other Earth-like planets, which we use to enhance our understanding of the other planet's structures/processes.
These can be plenty useful in abiotic cases, but dangerous once you start looking for life. Is Mars an appropriate target for these kinds of analogies at all? Do we really understand what Earth Analogues *are*?
EAs are models from historical sciences (sciences that explain past events using traces left behind, e.g. evolutionary biology, many Earth sciences). All EAs entail this.
Unlike other scientific models (e.g. animal models), these aren't experimentally testable.
In historical science explanations, variables are interdependent & hard to separate, so you have to try to wrap your head around the nature of those dependencies. Unlike constructed analogues (e.g. simulated space missions) or testable models (e.g. animals).
To understand the (any) model, you need to understand the relationship between its variables. To translate: you need to know which part of the model is interesting/relevant, and which parts are like/unlike the thing being modeled!
Forex, why is this Mars rock being compared to this particular Earth rock? Sure, it looks like the comparison-rock, but you need to really explain/understand why that's your comparison.
Otherwise you're just rock-matching, and drawing meaningless parallels. "Mars Rock A looks like Earth Rock B." is not interesting if all you've done is find a visually similar picture.
P.S. The correct word is pareidolia.
I think my "You need to know which part of the model is interesting/relevant, and which parts are like/unlike the thing being modeled" is the key.
It's akin to post-hoc hypotheses in science: whatever data you have, you can find SOME hypothesis to fit them.
It's akin to post-hoc hypotheses in science: whatever data you have, you can find SOME hypothesis to fit them.
Additional worry: expertise makes us want to apply it everywhere. (To a hammer, every problem etc.) Which just makes this problem happen more often.
This talk was not about the kind of analogues I was thinking, but that's ok, it just means it wasn't actually in my expertise-zone!
And now, lunch break. Back around 1:30 Nevada time.
unroll
