I was having an interesting discussion over dinner the other day with my aunt and cousins, which began as a relatively minor complaint about the propensity of Agents of SHIELD screenwriters (yes, I only just discovered this show) to use real science words in absurd ways, rather than making things up. At some point, the conversation had morphed into something about the general habit of filmmakers to publish misleading science as if it were plausible. (I found myself attempting -- but failing -- to communicate a point better made by Eliezer Yudkowsky in his post Science as Attire.) Some of us were of the opinion that this was a pretty bad thing that should probably stop; others didn't see much harm in it, so long as it was in works that were clearly fiction (false-science documentaries another matter entirely.)

My aunt, in the latter group,

"It's fiction, and it's art. If you're watching it as an audience and as a scientist, then it means one thing to you, but if you're just watching it as an audience, the science doesn't really matter, unless it's somehow important to the plot. Besides, anyone who's getting their science education from movies should really educate themselves better."

My one cousin and I had spent some time pushing against this head-on, without much success, when my other cousin tried a different tactic:

"You two, you're framing the stakes wrong. Mom, what would you say about a movie, clearly fiction, which had some offhand scenes involving domestic violence in a light that made it seem acceptable, or even normal? I mean, it's clearly fiction, and the screenwriter isn't an expert or adviser on marital relations or anything; it was just a thing she wrote into the movie. Besides, anyone who gets their social education from movies should really educate themselves better."

And very quickly the conversation (at least until we finished the drive home) moved to the topics of what knowledge we should reasonably expect filmmakers to have, and how much diligence was due on their part, in order to stop from casually disseminating problematic messages in ignorance. We eventually got back to science when I found a chance to reference Why I Hope Congress Never Watches Blackhat:

It wasn't until this week -- Tuesday evening, to be exact -- that my anxiety over the timing of the movie set in. That's when the White House released its legislative proposal to "reform" US computer crime policy in reaction to the Sony breach. President Obama plans to formally announce it at the State of the Union next Tuesday, but the details are public now. And many are troubling.

The general thrust of the proposal is to broaden the reach of the Computer Fraud and Abuse Act, and boost penalties for violations. The White House proposal will quadruple the maximum possible sentence for some crimes from five years to 20. And where under current law some hacks are misdemeanors -- specifically a first-time offense that doesn’t involve credit cards or more than $5,000 in information -- those crimes will now be felonies...

More disturbingly, the proposal includes sweeping language that directly impairs legitimate security work. It makes it newly illegal to "traffic" in any "means of access" into a computer if you have reason to know that someone will use it illegally. Releasing or using hacking code is a staple of cyber security work. Researchers publish it to demonstrate and describe the vulnerabilities they find, and professional white hats use it to audit their customers' networks. Like many security tools, bad guys can use the software too, and they do. But a sober computer crime proposal doesn’t ban tools that benefit thousands of people because one of them is a criminal.

Obama has struggled and failed to get similar CFAA changes through Congress in the past, but this time he has the Sony hack behind him -- and now Blackhat. If it’s farfetched to think lawmakers will be swayed by a work of Hollywood fiction, consider that it’s happened before. Congress passed the original CFAA in 1984 in direct response to the seminal hacker flick Wargames... The result was a law that -- after several revisions -- led to cases like the Lori Drew and Andrew Auernheimer misfires: People charged for lying in their social networking profiles or conspiring to access an unpublished URL. (...)

Oops that was a lot of blockquote.

Anyway, this post has an object-level point and a meta-level point. The object-level point, not a new one, is that sloppy science in film, when it's not clearly fake, occasionally is harmful to society, and that filmmakers really should be more careful with their huge megaphones.

The meta-level point is that sometimes, when you're having difficulty convincing a conversational partner that X is bad because in situation A it causes Y in a particular way, it's worth asking yourself if you might make your point more clearly if you liken it to situation B, in which X causes Z in a similar manner. Obviously, it's not an un-sticking strategy for conversational impasses in general, but rather, one tool in the box of "things to make conversations go more smoothly" -- and not one that I had realized could be so useful before I saw it used so effectively.

tl;dr Sometimes, when you're trying to make headway in a theoretical conversation, raising the stakes doesn't work -- it just takes you into increasingly-ridiculous-seeming hypotheticals that inject additional vectors of doubt into the conversation. Changing the stakes sideways instead gives you a way to make the conversation concrete and important to your conversational partner -- without driving it into the realm of fantasy.