Category: Science

I spent last weekend at the Social & Cultural Issues in Astrobiology 2016 conference at Clemson University, South Carolina. A small academic conference (~30 people) discussing nonscientific issues surrounding astrobiology and space exploration, and an absolutely amazing chance to spend two days thinking through the future among some of the world’s foremost thinkers and researchers on the topic.

If you poke around the website, you can find abstracts on all 26 talks: two one-hour keynotes and 24 half-hour talks on everything from ethics to gene theory. Starting on the 3rd talk¹, I livetweeted my notes and comments. For an overview of some of the top concerns and opinions around space travel, take a look through the Storifies below!

Session 1: Philosophy of Science

Session 2: Philosophy of Biology

Session 3: Connections and Patterns

Session 4: The Importance of Perspective

Day 2 Keynote

Session 5: Education & Outreach

Session 6: Concepts of Life

Session 7: Applied Ethics

Session 8: Ethical Theory

An essay entitled “The Empty Brain” has been bouncing around the internet over the past week, claiming that “Your brain does not process information, retrieve knowledge or store memories. In short: your brain is not a computer.”

Now, rumor has it some of you may not hang out around neuroscientists all day? In that case, let me fill your internet with an expert reaction.

This essay overstates many points, and presents many issues in imprecise or misleading ways. But, fundamentally, it is correct.

Let’s look at some of Dr. Epstein’s key points:

• We don’t have representations

It depends how abstract you want to get. Certainly, something in your brain reflects each skill and prediction and bit of information you have. I call this a “neural representation,” but that’s a technical term, not a literal claim. This “representation” isn’t a discrete thing, an arrangement of data; it’s some fluid time-share assemblage of synaptic weights and connections and more. Our “neural representations” are nothing like computer representations.

Later in the piece, Epstein uses the example of baseball outfielders to explain action without internal models. However, a pure organism-world interaction doesn’t explain human behavior: we correct our movements before we receive sensory feedback! We need internal models of body and world characteristics (e.g. how your joints will interact when you perform a complex movement), but that doesn’t require information processing.

Use one neural network to model another? That sounds hard, doesn’t it? Well, that’s why the cerebellum has as many neurons as the rest of your brain combined.

• Brains don’t perform algorithms

Brains can perform algorithms – after all, you can perform an algorithm, and surely your brain made that happen. But this is a bad way to describe what brains are doing. No part of the brain is operating algorithmically; we have a big mess of automatic processes and linguistic consciousness scaffolded atop basic sensory-motor functions.

• Brains don’t store memories

A wild overstatement. But we certainly don’t store memories like computers, nor even like we introspectively think we do. We store certain key details and associations, and re-invent all the details when we recall it. Moreover, as with “representations,” this “storage” is distributed widely across the brain. Nothing like conscious memory gets stored in single neurons.

When Epstein says “no image of the dollar bill has in any sense been ‘stored’ in Jinny’s brain,” this is obviously false. Any sense is a low bar, and it has certainly been stored in some sense. But only a vague, metaphorical, un-computer-like sense.

• The uniqueness problem

“There is no reason to believe that any two of us are changed the same way by the same experience.” Yup. Brains are hard.

• The IP metaphor has produced few, if any, insights.

Here, Epstein is dead wrong (i.e., “wildly overstating his case for dramatic effect”). Almost everything we’ve learned about the brain in the last 50 years has come about from the IP metaphor. It’s been incredibly fruitful. But we can do so much better!

• We are organisms, not computers

This, reader, is the magic. This is why the information processing metaphor causes problems. The computer metaphor can help (has helped!) you understand the brain, but if you frame brain function in the computer metaphor, you will miss critical features of the brain’s nature.

The brain is not a Turing machine, capable of any computation. The brain is an evolved structure, honed to produce adaptive action. Higher-level capacities like cognition, intelligence, and memory are built out of sensorimotor capabilities. The very act of perception is inseparable from knowledge of our body’s action capabilities.

• Relevance for writers

The human brain is what it is because of its evolutionary history. All the shiny beautiful new things are built atop of (or, rather, out of) basic functions. If some divine power created the brain from scratch, that scaffolding wouldn’t be necessary. There’s no telling what that brain would look like, but it wouldn’t make the same compromises and tradeoffs you’d find in a brain developed through evolutionary time.

This may not bother most readers, but if divine powers created humans from scratch and stardust three thousand years ago, those non-evolved people would behave nothing like the humans of Earth.

For another, funnier take on the ubiquity and consequences of nonsense science in the media, I present a segment from the latest episode of Last Week Tonight with John Oliver.

I would watch those TODD talks all week…

Every now and then, people trot out a scientific study from 2011 called “Story spoilers don’t spoil stories,” which claimed that spoilers generally improved readers’ enjoyment of stories. This study got lots of media attention. Unfortunately, it’s probably wrong.

This is actually the tip of an iceberg. Scientific journals are flooded with studies whose conclusions or results are wrong. There are many reasons why this is true. Some of them are malfeasance, such as data falsification, researcher biases, and “p-hacking”³. But most false results don’t arise from misbehavior. I don’t think the spoilers paper was biased or botched. To understand why this study was probably wrong, we need to get out our sonar and unveil the shape of that iceberg.

As noted since 2005, there’s a hidden structural problem that most researchers ignore: the question of “how likely is your hypothesis?” Because nothing is absolute and certain in our messy world, mainstream statistics are designed to admit a small rate of false positives⁴. In fields like psychology, findings are worth publishing if your data have a  ≤5% chance without the effect you’re looking for. “If I get heads five times in a row, that’s enough data to conclude that the coin is weighted.” But here’s the rub: what if you had a hundred coins, and only one of them is weighted? “Five heads in a row” occurs 3% of the time, so 3/99 fair coins will pass your threshold, in addition to the 1 weighted coin. You’ve now judged 4 coins to be weighted, and 75% of those judgments are wrong.

Issues in scientific culture compound this. Particularly publication bias, where only the most exciting and novel results get into high-prestige, high-visibility journals. Think about this: “exciting and novel” means “unexpected and unlikely.” As a result, higher-prestige, mass-media-worthy research is especially likely to be wrong.

The Good Spoilers Paper was published in 2011 in the journal Psychological Science, one of the top journals in psychology. However, a recent study of the big-name psychology literature showed that few of its studies held up. They re-ran studies published in the year 2008, so they didn’t replicate the Good Spoilers Paper directly. However, of the social-psychology papers in Psychological Science, only 29% (7/24) of the experiments produced the same conclusions when re-run!⁵ In other words, a sample of similar research only confirmed the results for 1/3 of studies.

Between publication bias and all the myriad ways to get a false positive, odds are that if a study has counter-intuitive results, and it appeared in a high-profile psychology journal in the last ~decade, it’s probably wrong.⁶ Remember: “counter-intuitive” means “there is a lot of counter-evidence.”⁷ If most of the evidence points one way, and a little bit of evidence points the other way, the outlier is probably a statistical fluke.

Moreover, some people have tried to extend the findings of the Good Spoilers Paper, using more complex measures of enjoyment. Lo and behold, they found that unspoiled stories were more fun, suspenseful, moving, and enjoyable.

There may even be specific reasons why spoilers are bad for reasons that the original experiment would’ve missed. But others have explained that well already. My little soapbox is here to tell you not to believe the research that says spoilers increase enjoyment, because science is messy.