Let's talk about open claims, endless argument, and how social media works.

Part of a larger series.

Before I start this, I should warn you that I’m in an unfortunate situation. For many years — a decade and a half in fact — I worked out my thoughts weekly (and sometimes daily) on a blog. The things I am known for — the ideas of federated models for social media, “digital gardens”, web literacy, SIFT — all were hashed out in bite-sized chunks that suited the way my sometimes overactive brain could process.

If I don’t regularly tidy my mental space by writing, my thought becomes too big for me to process, and getting it out is like pushing a firehose of content through a bendy straw. And I haven’t written on a regular basis for almost two years, and now I have accumulated a tangled mess of a whole book’s worth of thoughts that I am trying to get out and organized. Please stick with me as I try to work through that.

Here we go…

Let’s start with claims

I think you all know what a “claim” is. Philosophers go back and forth on this like philosophers go back and forth on everything, but for most people a claim is an assertion about some state of affairs in the world.

Ah, you say, but isn’t that just a tautology? Isn’t an assertion a statement about some state of affairs in the world? And you’d be right. Claims are not only about the world, but they exist in a particular context. Claims are the focal point of an argument. So I might assert that the fact it’s raining means you should skip tennis tonight. I have the focus of that argument on the point about tennis, the claim. You can of course shift the focus (“Who says it’s raining?”) and make that the claim, but the point holds. A claim is an assertion that forms the reason why certain evidence is introduced, and is the assertion that at least one person in the discourse is trying to change your attitude towards.

All this may seem overly complex and a little bit precious. But we are about to get to a fundamental understanding about how social media works. It’s just going to take a few more paragraphs.

Most times you say something — especially if you offer up a “fact” — that thing is interpreted in light of an “open question”. This was formulated in the 1990s in formal pragmatics as the “Question Under Discussion” or QUD. If there isn’t a clear QUD, you will intuit one, just to be able to interpret what’s going on. Let’s take our raining example.

H: I’m heading out to tennis.
A: It’s raining, you know. 
H: Yeah, but I already skipped two days.

When you look at this conversation logically, it doesn’t appear to make sense. I tell someone it’s raining, and they reply with a statement about how adherent they have been to their tennis practice schedule? But of course we process this very naturally. In fact it would be very weird for someone to say:

H: I’m heading out to tennis.
A: It’s raining, you know. 
H: Why are you telling me about the weather?

We even process this when the issue isn’t explicitly raised. For instance, say you are grabbing your tennis racket. I tell you it’s raining. Again, you immediately understand. You understand because you understand the “Question Under Discussion” is you going out for tennis.

None of the above is new or even much debated in linguistics anymore

Open Claims and Ongoing Arguments

Of course, not everything we discuss is as immediate or cognitively salient as a tennis appointment. Consider the following. My wife and I buy a new car. I want to pay more and get a plug-in hybrid. She thinks we should get the cheaper gas-only car — the payment on the hybrid is quite high. We go for the plug-in hybrid, but the truth is I know she is still unsure about that decision. I’ve essentially made a claim, I’ve marshalled evidence (gas is expensive! it’s good for the environment! we get to use those special parking lots!). And where that has gotten my partner is to an attitude regarding the claim that we’ll call “slightly favorable but largely unsure it was the right decision”.

We pay that car bill every month, so that’s not where I want my partner to be. One day several weeks later we are on a long drive and I look down at the mile per gallon display.

“We’re getting 52 miles per gallon!” I say.

“Yeah,” she says, “but with current gas prices how much does that really save?”

Now note this — any money you save is good. Getting 52 miles a gallon is better than getting 40, just like having $100 is better than having $80. So what’s going on here? Roughly, there’s an open Question Under Discussion, and specifically an open claim. The claim is that the car made financial sense to purchase. When I say the mileage we are getting, the issue at hand is not whether we are saving money or losing it, but specifically whether we are saving enough money that the extra expense was justified. We’re looking at the scale of the impact, and the fact I brought this up (assuming I am engaging with this open claim) implies that I think the scale of the savings here is enough that it would — at least partially — rationalize the purchase of the car.

I want to stress I have a harmonious marriage, but if I didn’t you could imagine this playing out over every dimension of that purchase decision. We pull into a special plug-in spot. “This is nice,” I say, “Prime parking!” Yes, says my partner, but it seems like there are a lot of other spaces open. And so on.

This may be contentious — but from the point of view of interrogating arguments it’s completely correct. After all, I am advancing an argument when I say these things, and the context of both scale and purpose matter here. If we’re talking about your hatred of your job and I say, well, yeah, but they do have free ice cream on Fridays in the summer I’m either joking or an idiot. When we introduce evidence — which following Stephen Toulmin’s later work I’ll call “grounds” from here on out — we are making at least two assertions. The first is about the evidence itself — that the evidence is real, and faithfully described. But I am also asserting that the evidence should be seen as a significant support for the claim, whether that claim is stated or unstated.

So here is our first connection to the Internet. Not everything on the Internet is like this. A lot of the internet is still “here, I wrote a long argument about subject x.” But for a lot of shorter content, like tweets and TikTok videos, like WhatsApp messages and social sharing, this is what’s happening. Like the discussion about the car, a claim like “The 2020 election was stolen” is a long running claim to which a lot of things get attached. So is “Masks were a farce and accomplished nothing” or alternatively “The abandonment of mask mandates is a health disaster”. These claims are in the ether, and whenever someone socially shares something, importantly, the thing they share is unlikely to be the prime claim. It is usually evidence, often supporting some unstated instrumental claim bolstering an unstated primary claim.

Why ignoring this fact is bad

So what, you might say. You say tweeting an article makes a claim. I say it at most makes an instrumental claim supporting an implicit, larger, open claim. What’s the difference?

A lot. I can’t discuss everything here, but consider just a couple of things.

Headline studies. A lot of work in misinformation studies — some of the most celebrated work, in fact — has been around headline studies. In these studies participants are presented with headlines and asked to judge whether they are true and whether they would share them. Some of this makes sense in a world that looks like 2014-2018, where much misinformation was of the sort “FBI Officer who testified against Clinton found shot and burned in bed with wife, police rule accidental fire”, or “Hundreds of migrants caught crossing border with suicide vests.”

But this is a subset of cases where the evidence is fully fabricated. When evidence is fully fabricated, treating the evidence as a claim in its own right sort of works. Most evidence is not like this however. Most successful disinformation campaigns, for example are based on something that has some truth to it — but that something is put in service of a claim that it doesn’t support.

So, for example, one of the biggest stories of this election cycle was the claim that 240,000 “unverified” ballots were sent out in Pennsylvania. Is that true, or is that false? Seen in one light, it is true. Many ballots in the system that were sent out had been marked with an unverified status. But the situation here is like the ongoing argument about the car — yes, you did get that gas mileage, but that fact does not support the implied claim. In this case, the implied claim is that the election is insecure (all these unverified ballots after all!). But the details of the claim argue against that — being marked unverified in the system is a security measure. It is a way of forcing someone who has received a ballot to go through additional verification before that ballot is counted. As evidence, it argues against the larger claim.

Sometimes we shorthand this and say — well, in this case the claim is false. But we should avoid being confused by the shorthand. The problem is not that the evidence is false, the problem is that the evidence, once properly understood, does not support the broader claim it was introduced to serve.

We sometimes talk about this in a narrow sense — “Oh, well, yeah, that’s a tactic called false context” or something of that nature. And it’s good to know tactics. But the tactic is a way of subverting something, not the thing subverted. Putting these all into lists of tactics is like adding epicycles to your Ptolemaic model of claims. The problem is you’ve got the wrong thing at the center of your model.

Moderation. We can look at moderation as well. We’d like to imagine moderation as if we have a world of five paragraph essays to look at and adjudicate. But that is not the case. What we often see is something like this:

Is this “true” or is it “false”? It’s kind of a nonsensical question right? Again, the post has to be evaluated in terms of the endless argument that it is made in service of. In this case this is made just before the 2017 elections, which were the next day. In absence of some other relevant argument to attach it to, a reader attuned to typical arguments about hijacked elections would likely take this as evidence in support of a broader claim that went something like “Our election is being messed with”.

Some people might reply — well, it’s technically true. And when looked at in the something like the Toulmin model of argument that means something like “the evidence is not fabricated”. But to say something is “technically true” is to admit that it is positioned in a larger structure, and within that larger structure is is doing something that is — let’s say — likely to get you further from the truth than closer to it.

We’ve known this for a while in information literacy, but we have lost sight of it in other areas

There are pieces of this that are not new at all. My go-to work on this is Toulmin’s 1958 work The Uses of Argument, hardly some ground-breaking work. Likewise, from very early days, influenced quite a bit by Sam Wineburg’s work (and their excellent work with authentic assessments), on assessments around claims I have generally asked students whether given posts, stats, and other material represent good evidence of some larger claim. This is not revolutionary, it’s normal practice in education. When we present a student with a bar graph we don’t ask “Is this true?” Instead we ask “Does this chart support X?” If we are critiquing a graph, we ask “For a person trying to determine X, is this chart deceptive?” We don’t talk about the structure of argument, but it is there, underneath everything we do.

Why then, when it comes to work around misinformation is so much research and response myopic? I know I haven’t shown the full range of the ways in which the Ptolemaic System of misinformation studies is so often centered around the claim, and not the argument. It’s an unfortunate result of me not writing blog posts for two years — I have some much to get down on the page and without my weekly practice of writing it is all jumbled up. Over the next weeks I hope to slowly clear the backlog of my thoughts on this and pound it into shape — thanks for listening while I work this out.