Evidence, Implicit Claims, and Warrants

How evidence is marshalled -- and contested -- online and elsewhere

A partial summary of this series so far (see part 1, part 2, part 3). You can skip this if you’ve read the previous posts:

1. Argumentation is a process by which people enhance the reasonableness of a position. This way of looking at discourse is different than what you may be used to. In particular, it is not a process to prove a claim.

2. Reasonableness is an important social constraint, and we invest resources and time in making our beliefs (positions) seem reasonable, both because perceptions of reasonableness impacts our self-perception and constrains social and political response.

3. Generally we believe that one position is more reasonable than others, but we can also believe in a range of reasonable positions.

4. We often argue to convince people to adopt a position, but we sometimes argue with other motives, for instance, to get our position to be seen as reasonable, even if not adopted, or within the scope of consideration.

5. In long-running arguments we search for evidence over time in an effort to enhance or affirm the reasonableness of a position. This is different than proving a claim, even though it is related.

6. The goal of evidence is to make positions regarding claims substantially more reasonable, to remind others of their reasonableness, or to provide others with means to enhance the reasonableness of their positions.

In this post I’ll talk a bit about one model (the Toulmin Model) for how evidence connects to claims. Let’s start with something simple.

Fascistic protest at Drag Queen Story Hour

As I’ve mentioned previously, my goal here is not to talk about misinformation specifically, but about online argument more generally as a route to talking about information quality. The following tweet has a real video and is making a credible point.

Note that as discussed in part one of this series that we see the evidence first, and have to infer the argument. Here’s what the video looks like:

The tweet says these are “verified Nazis”, and there is evidence that is true, but the tweet is also read in the context of a larger argument. While arguments are inferred — and therefore people can differ on how they see the underlying argument — I think in this case it is safe to summarize this argument as something along the lines of “the anti-trans protests are tied up with a rising tide of fascism” where fascism is a shorthand for this type of bigoted, authoritarian vigilantism seeking to push certain classes of people out of public life.

It’s good evidence of that, in my opinion. Why?

Think about the elements of the claim this is supposed to support. First, it’s hard to watch the video and not see the echoes of the KKK or white supremacist skinheads. The Nazi salutes are not subtle. The argument is this is a “rising tide” and note this is not a single individual standing out here — this is a larger group. They are clearly organized. They are intentionally threatening. Also to the point it is “rising” is the fact that we are not used to seeing this sort of thing in recent memory. Yes there are analogues throughout history, but something here feels new and therefore significant.

For those a bit traumatized by the video, I do apologize if it feels weird to dissect this logically. But this is part of my point. We often talk about videos like this as if they are powerful emotional experiences routing around our critical faculties. But when the underlying facts are correct and unmanipulated those emotions are often logical. It’s true that certain visual and verbal references here evoke narratives in non-logical ways. But it’s also true this is powerful evidence.

As we’ve said before, the job of the evidence here is not to prove there is a rising tide of fascism tied up in the anti-trans protests. Rather, it is to make that position substantially more reasonable, to remind us of its reasonableness, or to supply evidence to others wishing to show their friends, families, or constituents that this claim is reasonable.

Warrants

Why is the evidence here good and relevant? The Toulmin Model uses the concept of a “warrant”. The warrant is the often unstated relationship of the evidence to the claim.

In this case we look at the claim and the evidence and write out — as we sort of did above — why the evidence supports the claim.

People have different preferences in how they approach warrants, but for stuff like this I prefer generalizable warrants, that is, warrants that have been applied in the past to a broad variety of similar phenomena. The two warrants I have chosen here:

• Protests are characterized by the actions at them and people who attend them, especially when those actions are planned/organized

• New or intensifying behaviors constitute (sometimes accelerating) trends

...have been used when making claims about protests on both the left and the right to argue points about events as different as the Unite the Right rally and the Summer of 2020 protests against racist police violence. The warrants don’t have to be generalizable, but a good warrant in an argument is something that your audience already accepts as valid. So in practice, warrants are often on the generalizable side.

We then look at the evidence and note the elements here that map. Since we built the warrant and claim around the evidence it is not surprising the warrant maps quite well. The key here is doing this shows us exactly what is important in this video:

• The claim involves fascism, and these people look and act like Neo-Nazis. Since protests are defined in part by the people who participate and what they do there, it seems fair to say there is a fascist element to these protests.

• Part of our warrant protects addresses the few bad apples defense — there are always a few people that turn up to a protest that act horribly, and protests shouldn’t be characterized by that. So the warrant here is one that has been used elsewhere — that characterizing protests by actions is particularly valid when those actions are organized and substantial. And there are a lot of people in this video, and the actions are clearly coordinated.

• Finally, this might be debatable, but my feeling is part of what makes this compelling evidence is it feels new (at least in recent memory). It feels like things are accelerating. And crucial to that acceleration is newness in scale or kind. To give one example, when I lived in NH as a kid we used to go to political events and there would always be three or four Lyndon LaRouche people there, really fringy far-right folks, arguably part of a cult. If someone showed a couple LaRouchies here in NH passing out flyers in front of this saying AIDS was a lie or something anti-Semitic, my reaction might be, well, it’s NH, LaRouchies aren’t new. Not great, but not an acceleration. Here, this is new enough that it is both evidence of the character of the movement and a certain acceleration.

Five possible points of argument failure

For someone that wishes to counter, they have multiple options.

1. They could claim that the video is fabricated.

2. They could claim that the implied warrants aren’t valid, for instance that they disagree that neo-Nazis at a protest make it a neo-Nazi protest (this would be hard obviously, but they might say something along the lines of “If this partially characterizes these protests, why doesn’t looting characterize the 2020 summer protests?”)

3. They could claim that the video is mischaracterized in a way that breaks the connection via the warrant once context is known, for instance, if the video turned out to not be new but from 2015 then the “new” element of the warrant would no longer pertain.

4. They could claim that while all this connects up it merits a much softer position than many are claiming — maybe “it’s possible that” (again, a bit of a hard sell here)

5. And finally, they could say that no matter what this individual evidence may show, the overwhelming evidence from other sources is that overall there is very little fascistic involvement in these protests.

In practice, #1 and #3 are very close. The difference is that #1 would be a class of things where this would not be good evidence of anything, regardless of the warrant — the sort of “fake news” situation we saw in 2014-2018 and that we still see occasionally. In the case of #3, the accusation is that key context is missing that shows the evidence does not match the implied warrant. (In practice, these lines get a bit muddled, but I find the distinction useful for the moment).

Here in the comments we find a (false) allegation of the #1 type. A user replies “this glows”, a reference to a common conspiracy trope that these sorts of events are staged by federal agents.

We see this elsewhere as well, like here:

At a scan, there’s a dozen or so comments like this in the replies and the QTs, probably more. In this case we don’t see people attacking other possible points of failure, possibly because in this case if you accept the evidence as real it’s very difficult to claim it doesn’t relate to the claim, and possibly because “What about the events without Nazis” sounds pretty weak. But note that many feel they do need to respond. Likewise, to claim a failure at point #3 (missing context or irrelevant/overstated evidence) — it’s hard to imagine what that would look like. “Well, if you watch the longer version of the video…” Well, what?

Sometimes when we see stuff like this — people looking at something like this and saying it’s faked — commentators say it’s a sign that facts don’t matter. I’m not sure that’s true. There’s a bit of a paradox here — my sense is that if other parts of the argument were weak people might choose to attack the argument there. If this was a peaceful protest (however bigoted it might be) and there was one person in it doing this we might have a debate over whether the warrant applied to the evidence. Can you characterize a protest by the actions of one person? If the video was old, the reply might be “old video, unrelated to current claim.” If it’s true that replies end up gravitating towards the claim of fake (#1), it might not be because they don’t realize the strength of the argument, but rather that they do. Just a thought at this point, but one worth investigation. (It could be, of course, that calling things fake is just generally a lower effort rebuttal).

Whatever the point of possible failure these people have chosen, one thing is clear — they do feel a need to address the argument.

Narrative… or Argument?

Narratives can be true, of course, and many people would read this exchange not as argument, but as two dueling narratives. In one narrative, there is a rising authoritarian movement in the U.S., driven by bigotry, that is quickly accelerating towards violence. In the other narrative there is a deep state that is creating staged events so that the government can discredit those standing up against perverse practices being shoved down the throats of “real Americans”.

I’m not against using narrative analysis. It can be useful. But something is lost when we see this as primarily narrative and not argument. The trope of the staged event, for example, is an attractive narrative element here not because there is something inherently anti-trans about stories of staged events but because it forms a counterargument (however imperfect) to the initial argument here.

In this case, where the prevailing counterargument is seems confined to “the evidence is faked” the benefits of the argumentation lens may be somewhat less clear (though I think they are clear enough!). Future posts will bring its strengths into clearer relief.