What Evidence Ease of Imagination?

Will Wilkinson responds to Ross Douthat:

In this landscape, it’s easy to imagine the middle-class anxiety that the political scientist Jacob Hacker termed “office-park populism” defining the domestic debate over the next 20 years, and easy to imagine a Democratic majority that capitalizes on the opportunity.

The phrase “easy to imagine” has all the virtues of theft over honest toil. It is “easy to imagine” that the Kaiser won the Great War and that I’m writing in German (and a pith helmet). Likewise, it is easy to imagine Jacob Hacker’s now-largely-discredited thesis of income volatility and  our current cyclical financial worries defining domestic politics in a generation, but why would we bother to imagine it? Let’s imagine instead the centrality of the coming “robot gap” in American politics.

Do we over or underestimate the probability of scenarios that are "easy to imagine"?

Comments

[Roy Haddad]

I learned about this one in Psychology 101.

Availability bias:

The availability heuristic is a rule of thumb, heuristic, or cognitive bias, where people base their prediction of the frequency of an event or the proportion within a population based on how easily an example can be brought to mind.

[Robin Hanson]

Roy, a heuristic is not necessarily a bias. I was asking if there is a bias here.

[James Bach]

The availability bias, yes, and also other biases touch on this, such as representativeness bias ("an effect probably looks like its cause"), post hoc ergo propter hoc ("this must be happening because that other thing which happened right before it"), prominence bias (something that grabs our attention or touches our emotions is given more weight as evidence), and assimilation bias ("that new evidence just confirms what I already believe, especially if I kind of squint and hold it sideways...").

Also, the Occam's Razor heuristic seems like an example of easy to imagine.

To be easy to imagine is to be matchable to a model already held in our minds. It's not easy for some people to imagine that I could pour water into a glass, turn the glass over, and nothing pours out. But when I show them how Aquagel works (it turns water to a wax-like gel in a few seconds), suddenly it becomes easy to imagine.

-- James

[Hopefully Anonymous]

"easy to imagine" is up there with verbal tropes like "Clearly" and "obviously" -I think we'd be better off without them, like we would be without an appendix.

"I think", "I suspect", "I intuit" are more honest, because they're closer to how Eliezer suggests we use "magic" as a placeholder. Those acknowledge that bias could be playing a large role, whereas I think "easy to imagine", etc. obfuscate that.

[Roy Haddad]

It's a bias as well, because this heuristic is what many people use in place of better processes for judging likelihood (especially while in the grip of fear). Aren't many biases just over-applied heuristics?

[Tom Breton]

Perhaps the ease of imagination heuristic is being relegated to the trash heap too soon.

It certainly has its flaws, and is a rich source of bias and false belief.
On the other hand, it is a powerful mental tool for exploring ideas, especially ideas that interact with that hard-to-capture context that we call reality. It is powerful for anticipating problems and opportunities inherent in an idea which might not be superficially apparent.

We shouldn't be too quick to discard such a powerful and useful tool.

To use it well, it seems to me we must keep in mind these things, and perhaps others that don't occur to me at the moment:

• 
  Argument from not being able to imagine something is argument by failed search, with all the qualifying conditions that imposes.
  


• 
  Argument from ease of imagination, whether absolute or relative ease, is subject to biases from the sources that feed our imagination, particularly the more vivid sources. For example, do we find it easier to imagine something because we saw something similar from a fictional source, such as TV? Or a source that selects what it presents, such as TV news? If so, our confidence should be no higher than our confidence in the unbiasedness of that. Usually this is a fatal limitation. If we can't identify the relevant sources, we should be especially careful.