
In late October 2022 one of the larger left-perspective rumors of the U.S. general election emerged, a claim that at least one Mormon church was putting together an illegal ballot filling party, where voters would be “instructed” to vote in a certain way.
It centered on a single report of a story purportedly relayed by a friend, in a configuration that folklorists call a FOAF (friend-of-a-friend) story. These are stories relayed through a peer network where the person who had the experience is too socially distant to have their authority verified, but is socially close enough that the story feels real. (Think “My aunt had a friend that used tanning beds a lot, and she went to the doctor for some pain. The doctor is looking at her MRI and is like this can’t be right – your organs look LIKE THEY HAVE BEEN COOKED.” )
One of the ways FOAFs propagate is that when they are passed on the new teller of the story portrays the speaker as their aunt, or a friend of their friend. As the story moves through a population, each person makes the story seem just a bit closer to them than is the case, so the story can pass through a thousand people, and will still reach you as a story that feels just close enough to your social circle to feel real, and just a bit too far out to verify.
Online, the dynamics of spread are different, but the dynamics of credibility are similar. We don’t know what happened here with the LDS story – whether this speaker truly had a friend relay this to her, or whether (as is more often the case) she heard a story indirectly and then gave it a more direct attribution to make it more compelling. But either way the result for the reader is the same.
Types of evidence used for backing
Argumentation theorists have divided up “backing” (evidence used to support grounds or warrants) in various ways. When looking at evidence on the web I have found this modification of my own to be useful. I structure it based on from whence the credibility of the evidence derives.1 For what it’s worth, I am still fine-tuning this as I blog, but it’s been stable for a while now.
For instance, any personal testimony derives credibility in part through the closeness to an event. A good rebuke to an testimony you don’t trust might be “You weren’t even there” or “Are you sure about what you saw?” When that testimony is reported in a newspaper, however, the reporting derives credibility not because the reporter was there, but because their process of verification is solid. It would be odd to reply to reporting on a plane crash with an accusation that the reporter had not actually seen the plane go down. Likewise, an analyst derives credibility from their ability to apply expertise to the analysis of data or events.
Take the case of a car accident. The two drivers of the cars involved both have their testimony of what happened, and maybe a third person witnessed it from a distance. We judge those accounts by such things as closeness to the event, ability to interpret it as a witness, and trustworthiness in recounting it. To the mix we add a traffic photo of one of the participants running a red light; this is documentation. We judge the documentation’s credibility by whether it is real and unaltered. There are statistics – one driver had five previous speeding tickets, the other driver was going 5% over the speed limit when they hit the brakes. The accident investigator uses forensic analysis, applying expertise and professional experience to judge the evidence and render opinions that count more than the average person. A reporter – whether a professional one or a friend relaying news – has their reporting judged by how careful they are with the facts, and how fair they are historically with relaying the whole context, often sewing together testimony, analysis, statistics, and documentation into an informative whole.
Even something like a research paper is a mix of these things. A paper might draw on a mixture of common knowledge and documentation (in the form of citation) in its introduction, in what we would consider reporting. The questions here are similar in some ways to what we might look for in news — do the authors summarize the existing research fairly, is the description unbiased? The methods section of a paper is, in fact, personal testimony. The experiment may produce statistics that can be evaluated directly by the reader, followed by analysis which relies, in part, of one trusting in the acumen and technique of the authors.
Note that these categories only deal with credibility, not context or relevance. An accurate statistic can be miscontextualized, and photos misrepresented. Questions of credibility are narrower than argument evaluation and deal with the raw inputs into argument.
The problem of unsourced testimony
But back to our story. In cases where the person relaying an account has some approach to verifying that account, we could see our Tucson LDS rumor as reporting. If the Arizona Sun-Times reported, for instance, that a woman had said these ballot parties were happening, you’d count that as reporting. The fact that the Sun-Times has a set of standards on how to evaluate and verify the plausibility of these stories would put the burden of credibility on them. Even if a paper with a poor record of credibility reported it, again, we could judge the credibility by attributes of the reporter.
Here we don’t have that – Tucsonbelle is simply relaying an personal testimony about personal testimony. We can map this like so:
For our purposes here we’ve added in some backing for the warrant, the US Code that prohibits this kind of coercion. We don’t find this mentioned in the text or the responses, but we add it here for our own reference to show that if the testimony was credible this is about as solid as an election theft argument can get. The actions described would be a clear violation of law.
The problem here is simply that there is no way to substantiate the event.
In the end, this claim ends up being unsubstantiated, and given the scale of the accusation and the lack of any confirming evidence, it is almost certainly false. Notably, it took nearly a day to debunk the claim. This is a contributing factor to one of the most frustrating aspects of rumor – it takes much longer (seven times longer in one study) to debunk a false claim than to corroborate a true one. And if you look at this example you see one reason why. Without a corroborating source – or even a specific church – the only way to fact-check this is to call contacts in multiple LDS churches, to see if anyone has heard of this thing at all. If the event really did happen, one reliable person testifying to it could confirm it. But if it didn’t happen, the reporter has to keep calling until they feel they have covered all ground.
The evidence schema laid out here is derived in part from Rieke’s more limited schema. Rieke breaks evidence into three categories: instances, statistics, and testimony. Testimony is further broken down into testimony of opinion (“this statistician believes the election was stolen”) and testimony of fact (“this statistician said that there were more voters than voters in these three counties”). In some ways this is unhelpful, since in debate about elections most “instances” are also testimony, and most testimony of fact is an instance. Here, for example, the related story about the LDS seems much more parallel to someone sharing a picture and saying it was a “ballot party” than it does to an expert sharing a general opinion on whether the election was stolen. So I introduce the idea of breaking instances into personal testimony, documentation, and statistics (primary sources), making testimony about analysis and noting that reporting gains its credibility through different methods than direct expertise. Finally, for things that are simply references as known (part of Stalnaker’s “common ground”) we notate these as common knowledge.
Thanks for writing, really interesting!
I suppose there is an selection dynamic, within all kinds of evidence, for bullshit to evolve towards the appearance of being-possible-to-verify without actually being possible to verify. So the FOAF story adapts as it travels the social network, including enough details and to appear to be always socially close enough that you could go back to the source (but you can't); and for bullshit academic papers (say) the claims are backed by citations, to papers that appear to be in journals you could check (but you can't, because they are fake, or don't support what is claimed or whatever); and so on and so forth for all evidence types