Thursday, 7 January 2016

Further towards a Grand Unified Theory of bullshit

Sadly, according to Gordon Pennycook, "we ... live in the age of misinformation." Happily, he thinks we can learn to reliably spot misinformation for what it is:
 ...I have done empirical research on bullshit, and the results are clear. My collaborators and I recently published a paper investigating what we referred to as pseudo-profound bullshit. To understand how we investigated bullshit empirically, consider the following examples:
The invisible is beyond new timelessness.
As you self-actualise, you will enter into infinite empathy that transcends understanding.
These statements are, definitively, bullshit. I can say this directly because they were generated using two websites: and the New Age Bullshit Generator. Both select buzzwords at random and use them to form sentences. They have no intended meaning and use vagueness to mask their vacuity. They are bullshit. 

Across four studies and with more than 800 participants, we found that people consistently rate blatant bullshit such as this as at least somewhat profound. More importantly, this tendency – which we referred to as bullshit receptivity – was more common among people who performed worse on a variety of cognitive ability- and thinking-style tests, and who held religious and paranormal beliefs. Put differently, more logical, analytical and skeptical people were less likely to rate bullshit as profound, just as you might expect.

Importantly, we also included motivational quotations that were written in plain language and that had clear meaning (eg, ‘A river cuts through a rock, not because of its power but its persistence’). Surprisingly, more than 20 per cent of our participants rated the sentences that consist of random buzzwords as more profound than the sentences with clear meaning. These people had particularly faulty bullshit detectors. They also scored lower on our thinking-style test, indicating that they tend to be particularly intuitive and non-reflective decision-makers.
In other words, the kind of imprecise, overblown language George Orwell warned us about seventy years ago is what should trigger your bullshit alarm:
Only I don't think it's quite that simple. Meaningless polysyllabic jargon and obscure buzzwords account for a fair proportion of the world's bullshit pile, to be sure, but they're not always  reliable indicators. They don't, for example, account for the world's most efficient ruminant feces extrusion nozzle (pictured below):

Although there are similarities between the Chopra and Trump brands of misinformation ("For the bullshitter, it doesn’t really matter if he is right or wrong. What matters is that you’re paying attention"), there are also big differences. Where Chopra soothes the punters with his warm bubble bath of impenetrable but impressively positive-sounding platitudes, Trump grabs their attention by doing more or less the opposite:
Trump's relentless spray of short, punchy, monosyllables peppered with emotive trigger words has a completely different look and feel to Chopra's wordy pseudo-profundity, but underneath the branding, it's the same old pile of disingenuous nonsense.

So it looks to me as though you can't reliably detect all forms of bullshit using a one-dimensional Orwellian language filter (long, complex, obscure, abstract = misinformation, short, clear, concrete, easily understandable = real information). 

Imprecise and misleading language, it turns out, can come in simple, common-sense disguise, deliberately avoiding the big words and and complex abstractions that Orwell and now Pennycook warned us about. There's no one-size-fits-all template for bullshit, because effective bullshit is all about marketing and branding. And what do marketers do? They segment. They tailor the message to the audience. They will always try to get round the rational brain and anchor their product to vague but emotionally satisfying hot words, but the word cloud and syllable count will look very different depending on whether you're fishing for the dollars of New Age Californians or trying to get blue-collar tabloid readers to vote for the 1% and against their own interests.

So there are probably at least three varieties of bullshit. First, pseudo-profound, as identified by Pennycook. Pseudo-intelligent is probably a better name for the sort Orwell identified - where you don't quite understand what the journalist, politician or manager is going on about, but assume that they* must know what they're talking about, using all those fancy words. And finally, there's pseudo-straightforward, where a barrage of short, simple words, many of them emotionally loaded, are repetitively firehosed at target audiences without the time or inclination to think about details like whether or not any of those words make sense.