On the 18th of June, a political Twitter account tweeted the following:
Serving Labour MP Admits to Selling Drugs
And then there was a link, but we’ll come back to this later.
This caught my eye for a few reasons, but perhaps the most obvious is that it was so surprising, I went to fact-check it, and was gobsmacked at what it actually meant.
So what’s the problem with this tweet? Well, perhaps the most reasonable interpretation is that the person in question – this serving Labour MP – is a drug dealer, and therefore a criminal.
In fact, Taiwo Owatemi is an NHS cancer pharmacist, and on the 14th of June, the Register of Members’ Interests was updated to declare that “From 5 June 2021 until further notice” Owatemi would be a “Locum Pharmacist for Tesco, Shire Park, Kestrel Way” and that she would “work shifts on an ad hoc basis as required” (TWFY).
So did Owatemi “admit to selling drugs”?
That’s because there are at least two distinct interpretations to “admits to selling drugs” and whilst “trained and registered pharmacist legally dispensing medication” is one, another is “criminal drug-dealer selling illegal substances”. And I would argue quite strongly that the average person who reads the tweet is more likely to arrive at the latter. How do we know this? Well, we could test it on people and ask them what they thought it meant, but luckily, corpus linguistics can show us much more quickly what we tend to mean when we say things like this.
Admissions of guilt
The first element that’s worth a quick look is “admits to”.
The OED lists one denotation of this word (2f) as “To acknowledge a fact or personal quality; to confess to responsibility for an action, esp. a crime.” But in the spirit of fairness, it also has denotations such as 2a: “To accept as valid or lawful; to acknowledge.”
So when we look at real life data, what kinds of things do we admit to? Well, I turned to the BNC and did a quick corpus search for admit[s,] to (which finds both admit to and admits to) and randomised the results and then looked over the top 25. One was a metaphorical use to describe the route of a path (example 2) but after that, eleven of them (3, 4, 7, 10, 13, 14, 15, 16, 17, 24, 25) all have some sort of negative or problematic aspect, ranging from emotions such as fears, “grudgingness”, and disappointments through to outright criminal acts such as theft, travelling without a ticket, and assault.
If we accept – as most linguists do – that words, and especially clusters of words, prime us for what is coming next, then we can infer that there’s a very reasonable chance that “admits to” will prime the reader to expect something bad, if not criminal, to possibly follow.
So onto the next two words…
There are only ten hits in the BNC for “selling drugs” and you’ll be astonished to know that every single one of them is the criminal kind. No pharmacists dispensing here.
But there’s more going on in this little phrase.
We have another very useful word that would have absolutely changed the complexion of this tweet had it been used.
And though technically synonyms, these words carry drastically different connotational baggage, as can be illustrated with a corpus linguistics function called collocates. A collocate is a word that hangs out (co-locates) close to the word you’re interested in (such as drugs) more often than chance alone would dictate. So for instance, the is going to occur near to drugs often, but that’s purely because our syntax works that was. In other words, it doesn’t pass the “more than chance alone would dictate” rule. By contrast, a word like guns doesn’t have to hang out close to drugs. Something else could, like cars. If I need the word the, I don’t have much choice because of the way grammar works, but if I need another object – say, a noun – to make a sentence, I can pick from thousands. I could even use vehicle instead of car. Or hatchback. In other words, if particular groups of lexical words – nouns, verbs, adjectives, adverbs – are showing up unusually frequently around your word of interest, these start to tell you the kinds of concepts it hangs out with. The attitudes and stereotypes it’s imbued with. Its connotations.
What, then, are the top 15 noun collocates that appear on the right hand side of drugs versus medicine…
|Drugs (R1-R3, NN1)||Medicine (R1-R3, NN1)|
Another point here – drug and drugs don’t work in quite the same way as medicine and medicines. I can admit to selling medicine or medicines or drugs but I can’t admit to selling drug. For completeness, I repeated the same search with drug (singular) versus medicines (plural) and three of the top four drug results were variations of the word trafficking, with misusers in fifth. By contrast, the top five results for medicines were over-the-counter, herbal, dispensed, pills, and prescribed. It’s essentially a similar story to the table above.
To summarise all these results into a fairly obvious point, drug(s) are overwhelmingly linked with crime and crime-adjacent concepts – illegal and addictive substances, violence, gambling, and the police. By contrast, medicine(s) are overwhelmingly linked to fields of, well, medicine, other sciences, places we put medicines, and places we learn about or apply them. As synonyms go, this is a pretty stark contrast.
(Incidentally, I used to set my first years exactly this task many years ago. They had to pick a pair of synonyms and use corpus linguistics to test the notion of perfect synonymy, and all those years ago, one student did indeed do drugs versus medicines. One of the best assignments I read compared toilets to lavatories, and you’ll be thrilled to know that toilets are also linked more heavily with crime and drugs than lavatories are. But I have digressed quite a long way here.)
Remember my name…
One thing that I didn’t recognise in my first pass over this tweet and that was helpfully pointed out to me was the intertextual link to popular culture – all the worse because I gave the series in question a try and made it through the whole of the first season.
Anyway, the black and green graphic in the tweet represents the first letter of each word as though they are elements from the periodic table (we’ll overlook the fact that M is just entirely made up at this point since Magnesium is Mg), and it turns out that this is a fairly clear intertextual reference to the textual and cover-art of Breaking Bad.
And what is Breaking Bad, I hear absolutely no one asking? Well, IMDb describes it thus:
A high school chemistry teacher diagnosed with inoperable lung cancer turns to manufacturing and selling methamphetamine in order to secure his family’s future.
So, an upstanding member of the community (a bit like, say, an MP) starts illegal drug-dealing.
Where does that leave us? We now have a Twitter account that looks as though it publishes news, and it has tweeted what appears to be a headline claiming that an MP “admits to selling drugs” alongside an image that references a TV series in which the main character illegally sells drugs. There is nothing to indicate that it’s a joke or satire or any sort of prank, so, for all of the reasons given above, the ordinary reader would be absolutely forgiven for thinking that this tweet claims that the MP is a criminal drug dealer. The likelihood that someone would naturally conclude, instead, that she’s a pharmacist, is pretty low.
But there was one more thing in the tweet.
That link I mentioned.
The missing links
Rather than taking us to what we might expect – an exposé or investigation into criminal activity – the link in the tweet takes us to a page that reads:
Guido can reveal a Labour MP has admitted to actively selling drugs – and doing so since her election in 2019. Taiwo Owatemi quietly let slip the announcement earlier this week, revealing she’s been operating primarily in the Kestrel Way area of Welwyn Garden City.
[Screenshot of the Changes to the Register of Members’ Interests for 14th June 2021 for Taiwo Owatemi]
Guido’s sure she’s a valued team member at the Tesco Pharmacy…
NB. I dislike calling this content an “article” because that seems to elevate it to something that is conveying newsworthy or useful information, and I can’t see how it’s doing anything remotely resembling news, but that’s what its implied status seems to be, so I’m going with the unhappy alternative of “article”.
That aside, there’s more linguistic slithering here – “quietly let slip”, “operating primarily in” and so on, but the “article” clearly ends on the screenshot of Owatemi’s pharmaceutical role, and the words, “Tesco Pharmacy”. In fact, therein lies the single stark difference between the tweet and the “article”. And, by extension, this is where we arrive at the figurative link that takes us from the corpus linguistics that helped to contextualise why someone might interpret the tweet the way that they did to the forensic linguistics that looks at whether that interpretation could be seen as defamatory or not.
By itself, the tweet contains no clear indication that it’s anything other than a serious claim about an MP engaging in criminal activity. But at a best guess, the “article” seems like it’s aiming for something that might approximate satire. I don’t find it especially satirical, as you can probably gather, but I’m not the Final Universal Arbiter of Satire and things I find funny may not amuse anyone else at all, so quite whether I think it counts as satire or not isn’t really the issue. I would be extremely surprised if the “article” isn’t comprehensively defensible as satire, and much as I find its content and particular subject grotesque, I suspect action against it as defamation would probably fail. Bear in mind that I’m obviously not a lawyer, however, so I could have my mind changed on that point. The article, then, would presumably have satire as its first defence.
But the tweet? That’s something else. And the question now is, should the tweet and article be seen as two separate publications, or as one and the same publication?
On the one hand, if we argue that a tweet is its own publication in its own right, we might do so because it’s unreasonable to ask a reader to take a further step, click another link, go to a different platform that could be controversial, full of viruses, banned, unavailable on our device, not accessible to screenreaders, or otherwise problematic in its content or availability, and then consume however much content there to contextualise and understand the tweet. If that were the accepted position, then this tweet could even be actionable, though I obviously couldn’t speculate on how worthwhile or successful any such step might be.
By contrast, do we accept that short-form messages like tweets can’t always tell a whole story, that by their very nature, headlines only ever give us a partial view of the story, and that the link is very clear? If the position is that both together constitute one publication, then I doubt that the tweet could be actionable. However, if we do take that view, what does this mean for the ephemeral nature of the internet? What if the external site goes down, or the index structure shifts so that the link breaks? Can the link be guaranteed for the same lifespan as the tweet so that they always remain connected? Does it even need to be? Would a certain duration be acceptable before the “article” might be allowed to conveniently disappear, leaving only the tweet behind with no way to ever know whether it was serious or satirical? And I can’t even imagine how this would intersect with issues like paywalls.
Can satire leap through the link?
Lastly, this whole case is complicated by the way that the average reader interacts with content like headlines and previews on platforms like Twitter. There’s a lot of messy research on this and I’m not convinced by it all, but generally it appears that of roughly every five people who read a headline, only one may go on to read the article itself. Some research quotes far lower figures. Whatever the case, people happily share without reading, and through this, highly inflammatory but inaccurate headlines get huge amounts of traffic without reaping much in the way of consequences. (As it happens, too, most people who do go on to read the article still seem to skim most of it anyway, but in this case that isn’t entirely relevant because the “article” is two sentences and a picture – to skim that you’d have to read maybe three words.)
As the above suggests, plenty of tabloid outlets and malicious actors have already discovered the effectiveness of grabby, outrageous claims linked to much more mundane, if not flat-out contradictory content elsewhere, since they know that a reasonable proportion of the readership will share the tweeted headline and its linked article without ever reading it. For exactly this reason, Twitter now sometimes prompts users, asking them if they’d like to read the article first before sharing it. And if we look at this particular case, we have a clear issue: plenty of users will have read the tweet and taken it as a serious standalone publication. Fewer will have clicked the link to read more and discovered that it’s supposed to be something like satire.
And for the record, I absolutely do not think that those users are to blame. It is simply impossible to click on every single link, and then stop and vet every single story that one scrolls past in a day on a platform like Twitter. People have quite a lot else to do than check whether a news site’s headline aligns with their copy, and even if they had the time to read it all, how would they know that the claims in the copy were correct? The research involved in thoroughly fact-checking even one story can be enormous, especially if it is big, developing, or contentious. And remember, the person arriving at this particular tweet didn’t know that they were only going to find fifty-five words waiting for them at the other side. The problems for most people are time poverty and information overload.
So what happens if someone discovers the potent cocktail of mixing saturation with satire? Malicious actors have already weaponised this general habit of not checking the linked articles, and through it they’ve been able to propagate massive disinformation campaigns through headlines alone, but what if the linked content is supposedly satire? Should that satire be able to leap back through the hyperlink and act as a defence for non-satirical-looking content hosted on a different platform?
I don’t know the answers to these questions, but I would be very interested to see how a satire-once-removed defence would play out in court.