Mind the Gap: The Pitfalls of Hasty Generalizations
Includes the third musical short on appeals to ignorance. Also in this issue, why hyperbole is unhelpful, and how real world objects influence the images we create.
Issue No. 11
This issue brings you the third episode of Goat & Bear. It’s on the appeal to ignorance fallacy and is my favorite one yet. Really silly all around. We’ll have one more issue later in the month, and then I’ll be away for the month of December. To celebrate making it to the end of the year, I’ll be running a giveaway for several signed books. More on how to sign up for that next issue. Be well.
Paid subscribers, you can now listen to issue no. 10 on the motte and bailey fallacy. I’ll send out the audio version of issue 11 tomorrow.
1 Reasoning
Hasty generalizations
Generalizing is the act of making a leap from a specific experience to a broad statement of reality. We might reach for that instinct when classifying some thing or some one (should I trust this person), when learning about the world in order to survive it (should I touch this prickly thing), or when making new discoveries that rely on prior assumptions holding (will heat always flow from the hotter thing to the colder thing).
It can be a useful instinct, as pointed out in this passage for instance, which I happened to be reading the other day:
One of the social skills that we all develop is an ability to estimate the capabilities of individual people with whom we interact … we use cues from how a person performs some particular task to estimate how well they might perform some different task. We are able to generalize from observing performance at one task to a guess at competence over a much bigger set of tasks. We understand intuitively how to generalize from the performance level of the person to their competence in related areas.
When in a foreign city we ask a stranger on the street for directions and they reply in the language we spoke to them with confidence and with directions that seem to make sense, we think it worth pushing our luck and asking them about what is the local system for paying when you want to take a bus somewhere in that city.
If our teenage child is able to configure their new game machine to talk to the household wifi we suspect that if sufficiently motivated they will be able to help us get our new tablet computer on to the same network.
—Rodney Brooks, personal blog by way of IEEE Spectrum
When the generalization has a basis in facts, as in the examples above, or as is the case with scientific discoveries that go through clinical trials, the generalization holds until proven otherwise. If it doesn’t have a basis in facts, however—if it’s premature, if it relies on too few or on no data points—then it becomes a crutch. A shortcut that might make us feel good in the moment, but which subsequently doesn’t get us any closer to better understanding the world around us. We end up in a state of perpetual floundering, where we create brittle models of the world, ones that don’t stand up to any sort of scrutiny.
I want to share two types of generalizations that fall into that category. We’ll start with the appeal to ignorance (argumentum ad ignorantiam)—a type of argument that claims something is true because there’s no evidence to say it’s false. For instance, I see a flash of light in the night sky and determine that aliens definitely exist and that’s them there up in the sky. If someone presses me, I respond with, What’s to say it’s not aliens? A general claim about aliens existing and visiting Earth is inferred from one observation, without evidence.
Similarly, say I jolt up in the middle of the night to the sound of creaking floorboards, and determine that it must mean ghosts exist and the house is haunted. If someone presses me, I respond with, Do you know for sure that it’s not ghosts? A general claim about ghosts existing and being present in the house is inferred from one experience, without evidence.
The pattern here is that I shift the burden of proof away from me. I’m the one who’s supposed to present evidence for a claim I’m making. But instead, I appeal to an unknown—a gap—to make a determination about something definitely being true. Not that there’s some reasonable evidence for and some against my claim, and on balance, I’ve decided to take one position or the other. In the case of the appeal to ignorance, there’s no evidence at all. And I use that lack of evidence to make a leap to a general claim. The conclusion may well be due to wishful thinking.
Argumentum ad ignorantiam … Another way that men ordinarily use to drive others and force them to submit to their judgments, and receive their opinion in debate, is to require the adversary to admit what they allege as a proof, or to assign a better. And this I call argumentum ad ignorantiam.
—John Locke, An Essay Concerning Human Understanding (1690)
The composition fallacy is another way we might generalize from inadequate evidence. The generalization in this case takes a quality that applies to an individual part of a larger thing, and applies that quality to the whole thing. As in, what’s true of a part is assumed to also be true of the whole. And an inference then follows.
A sports team might be made up of individually talented players (the part), but that doesn’t mean the team (the whole) will win the season. A team of engineers might be made up of rockstars (the part), but that doesn’t mean the team will work well together, finish projects on time, and deliver groundbreaking products.
One example that I personally fall into is with foods that have a basis in something healthy, but where the ship on healthiness has long sailed due to a plethora of other ingredients being added in. A berry like açaí for instance (a part) may be good for you. An açaí bowl that has that berry, and then also has chocolate chips and cookies and fudge sauce (the whole), maybe not so much.
Another example I recall from several years back is that of a young, electric car manufacturer, who used to highlight in their early days that many of their components were from Mercedes—an established manufacturer with brand equity. An association that aimed to imply quality. But the quality of individual components (the part) isn’t the same as the quality of the car (the whole). A car could have the best wipers on the market, but a subpar driving experience.
We covered how to spot language that prematurely generalizes in issue no. 2 (as in, phrases like “people are saying”, which in fact refer to one person who said something). The best way to spot other uses of hasty generalizations is to be on the lookout for gaps that are either used as the basis for a conclusion or that we find ourselves leaping over to make some inference. In those cases, stop and ask for the evidence. It’s annoying, at first, but it becomes second nature with time.
2 Language
Why hyperbole is unhelpful
Hyperbolic language exaggerates, embellishes, and amplifies in order to make the thing being reported on more exciting. You’re more likely to pay attention to the TV screen if you see Breaking News or Key Alert flash across it. Similarly, you’re more likely to pay attention to a headline or to a lede that mentions emotive words like: slams, blasts, lashes out, destroys, decimates, flays, slices, or dices.
Here’s a headline picked completely at random yesterday, from a respectable outlet:
Rivals blast [so and so’s] pitch to make social media users “verified by their name”
The article furthermore includes the line,
[So and so’s] presidential rivals … slammed her Tuesday comments.
Reading through the article, one finds that two political rivals disagreed with the person on a particular position during a political debate. Far from acts of blasting and slamming.
From earlier in the week, a video analysis of a recent political exchange carried the title,
Reporter BODIES guest
Bodies, in this case, being slang for utterly destroys. Upon watching the exchange, the anchor merely took a strong position on a topic and the guest answered, albeit inadequately. No one was tackled in the process. No suplexes were recorded.
Other, somewhat humorous, examples I’ve seen include,
Renters Break Leases and Flee City.
Natural … threats seem to multiply by the hour.
Zuckerberg Dumped on by Facebook Employees Over Handling of Militia Groups.
A sub-headline before the last US presidential elections read:
If [so and so] wins [the election], it will be seen as the moment when the destiny of a man and his nation converged.
A presidential election shouldn’t be portrayed as a new Marvel movie dropping. It’s a forum, in which a new public servant replaces an old public servant. Policies ought to be central to it, not characters. Language can make the whole atmosphere surrounding the event feel like that of a boxing match, with characters—caricatures even—participating in it. I recall another headline during the last election. I believe it was from the editorial board of a paper of record. It read,
In the midst of unrelenting chaos, [the candidate] is offering an anxious, exhausted nation something beyond policy or ideology.
As one political commentator I used to listen to at the time pointed out, policy and ideology are precisely what constituents expect from their politicians. Unless the something beyond is actually divine or metaphysical. Even if it’s simply character and values, which the paper is likely referring to, it had better manifest as policy. Otherwise it’s not of much use.
The danger with news that’s made more exciting by way of hyperbolic language is that it can eventually come at the cost of accuracy. Accurate news isn’t always exciting. Oftentimes, it’s banal. Or downright depressing. We read it to stay informed. To learn about our fellow countrymen and women. But when hyperbole overtakes the general discourse, it ends up normalizing a charged, polarized political culture, paving the way for opportunists and populists to gain a foothold in the mainstream.
Be on the lookout for words that get you to click on a link, or to watch a video. Words that make you feel like you’re in an episode of Game of Thrones, with two sides battling it out (I’ve never watched Game of Thrones, but it felt apt to reference it here).
3 Images
How real world objects influence the images we create
I came across the following graphic in a book I’ve had on my shelf for years. And it made me pause, mainly because of how anachronistic it looked to my 21st century eyes. It reminded me of the extent to which real-world objects can prime us without us realizing it.
In this case, we have a graphic of the survivable body temperatures for seven mammals. Note, however, how the graphic is vertically oriented, on account of thermometers being mercury-based back in and around 1980 when this book was published. The form of a ubiquitous real-world object led to a design constraint.
If I had to draw that graphic today, I’d naturally lean toward a horizontal orientation. The mercury-based thermometer has largely been replaced with digital ones, at least for the average consumer. And so, as a reader, I no longer have that mental model, where I might expect a graphic that’s about temperatures to have its range extend bottom to top.
A horizontal graphic does have fewer constraints that can be advantageous in this case. For instance, we can now spell out the labels next to the bars, instead of tagging the bars with letters, as in the original graphic. We can now also position the minimum and maximum values for each mammal’s range either side of its bar. We end up with something that’s still compact and easily digestible on first glance.
(I’m not entirely clear if the sort order is significant here, so that may be worth a rethink. Narrowest to widest ranges? Closest to farthest from humans? By lowest? By highest? Alphabetically?)
‘No one is useless in this world,’ retorted the Secretary, ‘who lightens the burden of it for any one else.’
—Charles Dickens (Our Mutual Friend)
Until next time,
—Ali
