The first episode of the hit podcast Serial began with a bit in which host Sarah Koenig asked a few teenagers to try to remember what they had done on a random day six weeks before. None of them could do so with any degree of certainty, except for one kid who happened to know it had been the day of a state test. The immediate point of the exercise was to suggest how difficult it would have been for the teenagers caught up in the 1999 murder case at the heart of Serial to have answered police officers’ questions six weeks after the events in question. But, by extension, I also took the exercise to function as a warning signal about how much more unlikely it would be that anyone could remember what exactly they were doing on a particular day when they were teenagers 15 years ago.
I thought that opening was a promising sign that the podcast would take a sophisticated approach to memory. As highlighted in a recent NYT op-ed, many of our “common sense” beliefs about memory don’t square with what scientists know about how and what people remember. For one thing, people tend to think their recollections are true if they’re really confident they’re true. Maybe, but it’s complicated. When tested in a lab, for “true” memories, “greater confidence was associated with greater accuracy.” But “for false memories, higher confidence was associated with lower accuracy.” The problem, of course, is that, outside of a lab, there’s often no way to know which category your memory falls into, in which case your confidence isn’t a reliable guide. People also tend to insist that their recollections of highly emotional or traumatic events — where were you when Kennedy was shot, or the towers fell, etc. — are fixed in lacquer. Not necessarily: “Studies find that even our ‘flashbulb memories’ of emotionally charged events can be distorted and inaccurate, but we cling to them with the greatest of confidence.”
Anyway, you wouldn’t know any of this from Serial. Thus far, my hopes about the podcast haven’t been realized — after that first episode, sophisticated grappling with the nature of memory has been largely discarded in favor of increasingly naive parsing of minutiae. Take for example Episode 9, in which Koenig presents a phone call from Laura, a high school acquaintance of the podcast’s main subjects. “There were never any pay phones at Best Buy,” Laura says. The podcast continues: “[Sarah Koenig:] Laura says she knows this because she used to go to this Best Buy a lot … [Laura:] And I used to, you know, steal CDs from there all the time, so I was pretty aware of what was around.” Both Koenig, and the commentators on a Slate podcast about this episode of the Serial podcast, present this information as though it’s some sort of holy writ, because teenage shoplifters (or the shorter synonym of teenage shoplifters: teenagers) have superhuman powers of observation and recall, or something. This struck me as odd. There may or may not have been pay phones at this particular Best Buy in 1999. Whether one person says she doesn’t remember any pay phones 15 years later doesn’t settle that question one way or the other.
Moreover, how much does it matter, and for what purposes does it matter, whether or not there were pay phones at Best Buy? The state’s key witness told a story involving an important post-murder phone call from a Best Buy pay phone. So if the goal is to poke holes in the case the state presented at trial, then it matters whether other people think there was or wasn’t a pay phone there. It matters for evaluating the overall investigation if the police didn’t try to corroborate this fact. These kinds of questions all might have been useful fodder for cross-examination at trial. But if the goal is to figure out what actually happened on that day in 1999, it doesn’t strike me as that important — or I guess I should say, it’s one detail among many, it could be important, but it’s not some core insight to the entire case in the way it’s been presented. The state’s witness might have been lying about everything, he might have been lying about a few things, he might have simply misremembered where exactly this particular call was placed from.
Over Thanksgiving I talked to a lot of friends and family about Serial. The general sense I’ve gotten is that many listeners find value in these inquiries into details like whether the car could or couldn’t have made it to x place in y number of minutes — many listeners, that is, except for lawyers. The lawyers I’ve talked to mostly seem to agree that this fixation on details is beside the point. There’s never a criminal case where, at the end of the investigation, everyone involved knows 100% for sure what happened, every notation in the police timeline lines up exactly with a corresponding snippet of eyewitness testimony, and there’s independent forensic corroboration for everything that everyone says. There are just degrees of certainty, degrees to which the state’s case doesn’t add up, degrees to which you can say a reasonable jury should or shouldn’t have convicted. People say a lot of things to police (and to reporters, for that matter), for a lot of reasons.
Now cut to the broadly similar online discussions revolving around this Rolling Stone article alleging a university gang rape. When I read the Rolling Stone article, I took the reporter’s point to be, This woman brought allegations to a university official that, if true, would be horrific crimes, and therefore they should have been investigated more thoroughly at the time. That set of claims doesn’t depend on whether the allegations were true. It depends on whether it is true that the woman brought the allegations to the university at the time.
Other reporters are now raising questions about whether the allegations were presented sufficiently provisionally in the article as one first-person narrative, rather than corroborated facts, and about whether Rolling Stone made adequate efforts to contact the parties allegedly involved (some examples are here and here). These are useful questions to ask, as might be more general questions about how far reporters should go to corroborate allegations independently. (As noted above, I don’t think that what I took to be this particular story’s gist depends much on the truth of the allegations themselves, but I suppose reasonable people could disagree about that and also about whether the article took too much narrative license, devoted too much space to these allegations if it wasn’t going to do more to corroborate them, and/or was clear enough in specifying what it was and wasn’t doing.)
Meanwhile, the hordes of pseudonymous online commenters are now discussing amongst themselves whether or not the allegations could or could not have actually happened in every single particular precisely in the way described in the article, based, I suppose, on their encyclopedic knowledge of the Likelihood of All Things. These strike me as much less useful questions to ask. The assumption being made all around, by both defenders and critics of the story, seems to be that if a traumatic episode really did occur, then anyone involved would have pristine recall of all its particulars two years later, unfiltered through cultural tropes and unchanged through various retellings and the passage of time. As discussed above, this is just not a good assumption to make based on what we know about memory.
Anyway, who knows how either of these stories will shake out in the end. But I bring them together here because I wonder if the detail fixation shared by Serial listeners and Rolling Stone readers (many of whom, of course, are probably the same people), as well as the model of memory at work in both works of journalism and in readers’ responses to both works, might tell us something about cultural expectations about truth and how and where to find it. It seems that our bias is to place more confidence in memories, and to place more importance upon isolated details in memories, than is warranted. After all, it’s not as though ordinarily, when defense lawyers do their job of challenging every detail of the state’s case, or when journalists adhere to every “best practice” in reporting and fact-checking, the result is pure apollonian understanding of What Happened. The result is just a better understanding of how confident we can be that we know what happened, allowing that a few details here and there may not add up. Thus, for instance, when Judith Shulevitz at The New Republic writes, “What we don’t know is whether every detail of Jackie’s story, as told to Rolling Stone, is true,” I am not sure who she is arguing with, even if Shulevitz otherwise makes some useful points.
None of this is to deny that details matter, to deny that some details matter more than others for evaluating the overall credibility of an account, or to deny that we have plenty of tools for discerning the truth of individual details to a degree that is usually good enough for our purposes. In a court of law, it matters enormously what details can be proven to a jury and to what degree of certainty. But that’s because consequences ride on such proofs. It’s not because anyone in the legal system is under any delusion that crystalline insight can be found into every component of a past occurrence. To the contrary, the recognition that truth is an aspiration is built into the legal system’s probabilistic vocabulary — “more likely than not,” “beyond a reasonable doubt.” And details matter enormously in journalism. When journalists don’t fact-check, or combine sources into composites, or fabricate anecdotes (or entire stories), or embellish with invented details for literary flair, they generate stories that can’t be trusted and, more generally, undermine readers’ trust that journalism is the result of a culturally agreed-upon set of procedures. But those procedures are worth maintaining because, like all knowledge-generating procedures that humans have yet invented, they are tools for heightening our certainty, not because they are divining rods for uncovering some buried capital-T Truth.