I attended a security meeting a few days ago, run by an international NGO whose sole purpose is to monitor the security situation and share that information with other NGOs. This NGO collates all of this information—mostly based on what NGOs share with them—and sends a list of them out weekly in spreadsheet format. That is the core product. (There are biweekly and quarterly reports that try to track trends in incident numbers in various regions of the country, but it’s fairly basic data processing.)

This is to say, the weekly reports I get from this organization are “just the facts.” That doesn’t make them infallible: the standard epistemological caveats apply. But if somebody’s detonated an IED, then we get a report that somebody detonated an IED: with little or no speculation as to motives, or the identity of the perpetrators. The reports are, I imagine, as close as we can get to objective news reporting.

In the notes I typed up for myself after the security roundtable, I wrote, “It’s difficult to create a narrative based on a lot of discrete data points.” That reflected my observation that the security NGO truly struggled to place the facts that they had collected into any sort of narrative context. It wasn’t for want of trying, but just about all they are able to say were things like, “Incidents are up X% in this province from last year,” or “Incidents are down Y% from three months ago.” And, tedious as those sorts of observations can be, I admire their intellectual humility in choosing not to create a narrative out of the facts for which they did not have evidence.

On the other hand, the security reports are utterly unreadable. In fact, until recently I was just ignoring them, because the barrage of facts made no sense to me. That’s because they had no narrative.

Humans process facts primarily through narratives. Without a narrative, it’s very difficult to hold a large number of facts together. I was able to place my spreadsheet-of-facts into a narrative context by getting the data into Google Earth, grouped by week. All of a sudden, my brain could work with the data: the events were distributed spatially and with time. With some furious clicking I could create little animations of how (badly) things were going, and where. Even if my simple data visualization lacked some crucial narrative elements—I still don’t know who the actors are, and I don’t have a clear sense of their goals, beyond seeing where they choose to operate—it provides enough narrative scaffolding for me to make sense of the data.

If I were going to invest in a news media company, I would not invest in one that sent around weekly spreadsheets of events. News is only comprehensible to us because the facts of the situation are presented in the context of a narrative. A more nefarious way to express the idea of the last sentence is to say that successful news must be placed into a narrative of the media company’s choosing. So in this sense, “media bias” is unavoidable, not just with respect to the selection of facts, but also with respect to the narrative context in which the facts are placed. (This is probably where confirmation bias comes from.) So, for instance, atrocities committed by military personnel from various countries can be presented either as isolated atrocities, or as part of a larger narrative an injustice. In the American media, Syria and Russia tend to get the latter treatment. America got the gentler treatment in the earlier Bush years and in the Obama years. Or, if you can remember the news from 2015, there was a tremendous spate of stories about gun violence and shootings—despite the fact that gun violence was decreasing at that time.

In all eras, stories from media outlets have reflected the narratives that those outlets wish to promulgate (for instance, during World War II). It seems to me that in the 1990s, the growth of cable news added a market element in: consumers could choose who they wanted to get their news from. And then of course in the last fifteen years or so, social media has encouraged the formation of echo chambers. Fake news flows inevitably out of this situation. Once the echo chamber has been formed, all that is needed to satisfy our itching ears is a tiny narrative to fit into the larger one. The factuality of the incident is less important than the narrative resonances that it evokes.