March 18, 2020


I was in the US Army post-9/11 during the wars in Iraq and Afghanistan. Because of the way 9/11 shaped American thinking about the military (as well as about police and firefighters), I had countless experiences of being thanked by strangers for my service.

If 9/11 caused people to recognize and appreciate the importance of military and first responders to our society, the COVID-19 pandemic should do the same for the workers who--it is not an exaggeration to say--are standing between us and the collapse of society.

These workers include custodial staff, grocery store workers, delivery people, utility crews, maintenance workers, pharmacy techs, LPNs, sanitation workers. I admit I don't know the comprehensive list, because I'm socialized to overlook their contributions.

It's not an exaggeration to say that these are the people who are holding off disaster. They're stocking the shelves that we're panic-buying from, cleaning all the surfaces we're whining about touching, and keeping basic services running. Frankly, any worker connected with utilities should be able to claim that they've saved countless lives. However bad coronavirus is, if the clean water stops flowing to our houses, coronavirus is a just drop in the bucket. We're only days away from dysentery, typhoid, and dehydration that would be vastly more lethal.

Many of these workers are among the most marginalized in western societies. They are often paid a minimum allowable wage--which politicians rountinely oppose raising. In the United States, they are likely to not have good health insurance. Their positions are often insecure and subject to being fired in a downturn. They cannot do their jobs from home. And, most ironically of all, many of them must closely interact with many members of the public, so they are probably among the most vulnerable people in a pandemic.

A valuable outcome of the COVID-19 pandemic would be for it to force us to recognize the essential role that these workers play in society's survival. At a one-on-one level, that recognition should mean thanking these workers.

I'm going to start saying to workers in these positions, "Thank you for keeping us together." The idea behind this phrase is, "We would be falling apart, but you're keeping us together."

Workers in these positions deserve "Thank you for keeping us together" more than I ever deserved "Thank you for your service."

I'm hoping also to capture this with the h/t #KeepingUsTogether

This recognition also has much bigger implications. If we recognize that our society cannot survive without the work these people do, then we should also recognize that society has a practical and moral obligation to support their well-being. In an American context, for instance, we could get over the discourse that we can't afford to pay essential workers a living wage or provide universal health care--and instead recognize that we can't afford not to.

March 17, 2020

Language going viral

I don't mean for this entry to make light of coronavirus. I don't belittle the fear people are experiencing, or the severe economic consequences the virus is causing, or the grief of people who lose loved ones. That's all real.

I'm not actually writing about coronavirus here. Instead, I'm just interested in this entry in the words and phrases that are becoming part of our shared vocabulary for talking about coronavirus.

So this is just a running list of words and phrases. I'm writing about language, not about coronavirus. I'm just curious about tracking when words and phrases emerge, how they take hold, and which ones last beyond the current crisis.


This is a bit of a cop out, but fair to include since coronavirus is actually a family of viruses which includes the common cold, but will now for most of us likely always mean this specific coronavirus. Coronavirus caught on long before the name COVID-19 was bestowed, and COVID-19 is not catchy. The only viable contender for a name would've been if pandolins could've been established definitively as the animal source before coronavirus got set in our vocabulary. Pandolin flu, I think, might've stuck.

Out of an abundance of caution

In my mind, this phrase is the kind of thing government officials say during press briefings to justify actions that are potentially/probably over-reactions. It feels like, during this crisis, the term has generalized, so that we expect to read it in any statement about risk management decisions, even if the statement isn't especially important. I exemplify this with the public statement from my hometown American football team--which was posted at a time when there are no games, no practices, no public interactions, and it's not really clear why fans would need to know that staff are working from home, as we generally would not know they were working at work right now under normal circumstances. The Chiefs could actually all pretty well work from home between the Super Bowl and training camp, as far as I am concerned, global pandemic or not.

Flattening the curve

This one has staying power because it's vaguely statistical in a way that can be applied to anything with rapid growth--and because a genius named Drew Harris came up with a really good graphic to illustrate the concept. It will stick around like regress to the mean has become part of sports talk radio parlance during baseball season. Importantly, I think it's generally unclear to people that the number of cases in the steep curve and the flatter curve are supposed to be the same, so I think this term will be perpetually misused to mean 'decrease the occurrence of' rather than 'decrease the rate of occurrence of.'


There seems to be something of a coming out process (as it were), where someone announces they're beginning social distancing. As an introvert and a nerd, I admit to sometimes feeling a bit annoyed by these proclamations--especially when they come via interviews on TV/radio with people who are struggling with social-distancing or worried about the potential toll of social-distancing. On the introvert side, I've spent a fair portion of my life social-proximiting against my will or fabricating excuses for avoiding social events. On the nerd side, I've spent a fair portion of my life being social-distanced by others. I think this word has a good chance to last--i.e., "I'm too tired to go out this weekend so I'm practicing some social-distancing."


Here is the problem with panic-buying: if you can't find toilet paper for a few days, then you're definitely panic-buying, but not panic-buying in the 'hoarding' sense. But panic-buying because of the hoarding.


Basically see social-distancing. I'll take a smug angle here, though, because, in a sense, I've been preparing for self-isolating all my life. I sometimes imagine that when all of this is over, us real self-isolaters will emerge from our hovels into a sunny day, look around at our happy survivial of the disaster, and then go back inside to finish the book we're reading.

September 25, 2019

Words communities should learn to be ready for university students

Just before the start of the 2019/2020 academic year, I was asked to provide a list of "words university students should learn to be ready for their new community." The request was, of course, just a gimmick to drive traffic to a commercial site. "Local words" exist much more in people's conceptualisations of local dialects than in reality. Linguistically, the interesting differences between dialects don't exist in words but in the domains of phonetics and phonology, morphology and syntax, and pragmatics and discourse. Lexical differences between dialects tend to be pretty shallow, and in most communities local lexicons are given way to regional and national lexicons.

But perhaps most importantly, "traditional speakers of a local dialect" (i.e., the older people whose speech is representative of a local variety) aren't really doing anything innovative with words. By definition (if they exist), they're just using old words. The innovation is happening in the speech of young people. So, if anybody should be learning new words, it's the local community, which is going to have to try to keep up with the language of the young innovative speakers descending into their clubs, pubs, and Ikeas.

Therefore, I surveyed new linguistics students at Warwick to determine their best slang. This is their dictionary.

Word: peng (adj.)
Definition: attractive
Example sentence: They’re really peng. I want to chat them up.

Word: bare (1, adj.)
Definition: a lot
Example sentence: He’s got bare money.

Word: bare (2, adv.)
Definition: very
Example sentence: This song is bare good.

Word: lit (adj.)
Definition: good
Example sentence: That was a lit Wasps fixture.

Word: gucci (adj.)
Definition: good
Example sentence: It’s all gucci in our flat.

Word: bevved (adj.)
Definition: very drunk
Example sentence: He is proper bevved from the club.

Word: rec (n.)
Definition: park
Example sentence: We’re going to the rec for a kick-about.

Word: paigon (n.)
Definition: a person who does nothing; a person who should not be given attention; a person who cannot be trusted
Example sentence: "I do want to use this to say that Theresa May is a paigon and you know what we're doing right now." - Stormzy in speech at GQ Awards

Word: snake (n.)
Definition: a person who betrays people
Example sentence: That snake was bad-mouthing my friends.

Word: clapped (1, adj.)
Definition: ugly
Example sentence: That bloke over there is clapped.

Word: clapped (2, adj.)
Definition: stupid
Example sentence 2: That was a clapped thing to say.

Word: –ed can be added to almost any noun to mean 'drunk'
Example sentence: I got absolutely wallpapered last night.

Word: yeet (1, interjection)
Definition: general exclamation of approval
Example sentence: I love Warwick! Yeet!

Word: yeet (2, verb); pt. yote
Definition: throw
Example sentence: I yote the ball to Mary. She will yeet it back.

March 02, 2019

Why you probably already speak some Cockney

Writing about web page

This was originally published on the University of Warwick Knowledge Centre on 1 March 2019. I have modified it slightly from the form that appeared there, and have changed the title. It is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)

I’m a sociolinguist, which means that I research the ways people use language to do things--and the ways language does things to people. As a rule, I cringe when people imitate accents. Imitations are usually lousy, and invariably include problematic stereotypes—think Dick Van Dyke stepping in time. Bad imitations and rehashed stereotypes reinforce negative attitudes about accents, and these negative attitudes have real consequences for speakers of these accents. In the United States, for instance, researchers have shown that people are denied jobs and housing because of their accents.

Cockney, of course, has a long history of being highly stigmatised. It’s therefore a prime candidate for being badly imitated in ways that harm its speakers. If #SpeakCockneyDay doesn’t get farther than “’ello Guv’na” and “up da apples”, there will be much to cringe over on 3 March.

On the other hand, Andy Green recognises that a dialect is not just a way of speaking, but a central aspect of our personal and cultural identities. That is absolutely worth celebrating--especially when it means spending time with friends or doing things for charity to express innate generosity. If these are part of the culture that’s expressed through Cockney, then Cockney indeed belongs in the DNA of all Englishes.

The good news is that, across Britain, many people can speak Cockney just by talking normally. Rather than disappearing from the streets of London, features of Cockney are spreading into accents all over Britain.

In most cases, wherever these features spread, they still carry negative social stigmas, and those stigmas carry adverse consequences for speakers. If we can agree to celebrate Cockney speech, perhaps we can agree to celebrate Cockney features wherever they show up in English. If so, we can combat negative evaluations of accents and reduce the negative effects of these evaluations on speakers, while also celebrating the connections between our ways of speaking and our ways of living.


Andy has put #SpeakCockneyDay on “the ’fird of the ’fird” as a nod to one of the most salient features of Cockney: the replacement of the sounds traditionally associated with the letters <th> with the sounds of [f] and [v]. Linguists often refer to this feature as “TH-fronting.” TH-fronting is thriving throughout Britain (or friving froughout, if you prefer). Researchers have found TH-fronting across England and Scotland, and a 2016 article in the Mail warned that the "th" sound would disappear in 50 years. TH-fronting is often stigmatised as “yuf speak” or, in the case of the Mail article, blamed on “foreign visitors.” But TH-fronting is a clear case of a traditional Cockney feature spreading into other British Englishes. We should celebrate TH-fronting as one of the ways that Cockney is alive and well.


Linguists refer to the sound that occurs in the middle of a word like uh-oh as a “glottal stop.” Imitations of Cockney will invariably include the replacement of /t/ with glottal stops, especially when the /t/ is in the middle of a word like better. Cockney accents may extend this “T-glottaling” to use glottal stops in the place of sounds associated with <p> and <k>, or may replace consonants at the beginning or end of a word with a glottal stop, or may reinforce a consonant by adding a glottal stop to it. As with TH-fronting, T-glottaling has long been documented in a wide variety of British Englishes, and gives every indication of continuing to spread. As with TH-fronting, the spread of T-glottaling represents the continued vibrancy of Cockney in English, and should be celebra’ed.


When Andy wishes readers “’appy March 3rd,” he is demonstrating H-dropping. H-dropping is so widely attested in Englishes that it’s not really accurate to call it a Cockney feature (Donald Trump, for instance, is H-dropping when he pronounces huge as “yuge”). H-dropping is widespread in the UK, documented, for instance, in recent books on the East Midlands and West Midlands. But while it’s not a “Cockney feature,” it is definitely a feature of Cockney. So celebrating H-dropping wherever it occurs is another great way to celebrate Cockney sounds.

These are just a few of the features that have escaped the East End to become part of the future of English. #SpeakCockneyDay is an opportunity to recognise the spread of these features, to celebrate the perpetual vibrancy and innovativeness of English and its speakers, and to appreciate that we can now hear the Bow Bells all over Britain.

September 30, 2018

"filler words" fill the news

Writing about web page

Non-linguists often use the term "filler words" as a convenient catch-all for utterances that, for various reasons, don't meet arbitrary prescriptivist rules for "correct" speech.

The BBC, Independent, and other news outlets recently reported the results of a pseudo-scientific study about "filler words."

"Pseudo-scientific" is overly kind. The study was apparently conducted by a marketing firm on behalf of a website, so it's really just a company taking advantage of the press for free advertising. This is a cursory list of obvious problems with the study.

Perceptions of language ≠ productions of language

The "study" surveys people's perceptions of their usages of linguistic features and presents them as facts about their productions of linguistic features. The absurdity of this should be obvious, since the "study" finds that 23 percent of people "don't realise they're using [filler words]" at the same time that it reports situations and demographics where people realise they're using "filler words." In other words, the "study" says, "People don't know when they say err, and those same people say they say err when they're on a date." You can't study what people actually do by asking them what they do. You must study them doing the thing.

The "study" design doesn't meet basic scientific standards

The "study" apparently showed people two videos of the same person giving talks about Brexit, one factual speech with "filler words" and one false speech without "filler words." As a basic methodological principle, you cannot tell people what an experimental condition is and then examine their responses to it. The method in this "study" of language is analogous to giving a person a trial medication and a placebo, telling them which is which, and then asking them whether they feel like the trial medication or the placebo worked better. That's obviously bad science. Beyond that, we'd need to know how the researchers identified the "filler words" included and whether there were other cues to the speakers' perceived confidence (like body language), not to mention conflating factors like informant demographics and attitudes toward Brexit. And the genre of speech matters, too: people like respond differently to someone using "filler words" in a situation where prepared remarks are expected (like a speech on major international policy) than they do in a spontaneous conversation.

The researchers don't know language

The "study" problematically conflates many different kinds of language under the label "filler words," revealing that they have no knowledge of the tremendous body of science about language. For instance, there are many different versions of the word like:

  1. He was like forty years old.
  2. I was like, "This is not good."
  3. Like, what do you want to do?

In (1), like carries an approximer function, so that the man is 'approximately, but not exactly, forty years old.' In (2), like introduces a quotation that captures the spirit and general idea of what was said, but may not be literally word-for-word. In (3), like functions as a focuser, which points the hearer toward the importance of the clause that immediately follows it.

These are not fresh insights into like. There's a whole Wikipedia page about like with a great bibliography of some of the work linguists have done on it. The fact, though, that this marketing gimmick conflated these very different forms of like under the heading "filler words" reveals the lack of knowledge about language and linguistics.

Indeed, there's real science behind every word that this "study" calls "filler." umm, for instance, is a hard-working discourse particle that helps govern conversation. People say umm in part to signal that they're still talking and that it's not yet someone else's "turn" to talk. Linguists refer to this as "floor management"--managing who "holds the floor" during conversational turn-taking. innit isn't "filler" at all. It is an innovation in British English that functions like a "tag question," but also indicates a speaker's strong feelings about the statement they've just made. It's lumped in as "filler" by the "study," because innit is associated with young people and the speech of young people is socially stigmatized.

"filler words" and future study

I take three points from reading the reports of this "study" in British media:

  1. Media must remain vigilant to being duped into handing out free advertisements to companies on the basis of phony science.
  2. Linguists have much more work to do to share scientific knowledge about language with the public.
  3. There are no "filler words" in language. Every linguistic utterance means something and is doing something to make language work.

September 22, 2018

The Linguistics of Pirate Talk

Writing about web page

This was originally published on the University of Warwick Knowledge Centre on 18 Sept. 2018. I have modified it slightly from the form that appeared there, and have changed the title. It is licensed under a Creative Commons Attribution 4.0 International License (CC BY 4.0)

International Talk Like A Pirate Day is, of course, silly. No one in history has ever, based on their adopting a sea-going profession, talked like Robert Newton's Long John Silver in Tresaure Island. The people we think of when we talk about “pirates” would’ve talked mostly like the people they grew up around, just like the rest of us do. Many of them wouldn’t have spoken English. And real pirates (historical or present-day), aren’t especially good people to imitate. Pirates were and are brutal criminals. Nothing kitschy about that. So there’s not much to be said for the “pirates” part of International Talk Like a Pirate Day.

But the “talk” part is interesting.

The ideas we have about Pirate Talk have much more to do with facts about language than facts about pirates. So, pretending that pirates really existed in the Treasure Island sense, here are four rules to follow on Talk Like A Pirate Day, and the linguistic facts behind those rules.

Rule 1. Say ye instead of you

(1a) Pirate Talk: Ye are a bunch of landlubbers.

(1b) Standard English: You are a bunch of landlubbers.

In Old English, ye was a second-person plural pronoun, distinct from the second-person singular pronoun thou. In Middle English, ye disappeared in most dialects of Modern English, though a few dialects retained it for some time. Some Irish English speakers today distinguish between you for second-person singular and ye for second-person plural.

If pirates say ye, then, they may be following a grammatical rule that distinguishes between addressing one person and addressing multiple people. To play it safe, celebrants of Talk Like a Pirate Day should probably use ye only when addressing groups, or risk lashes.

This rule actually makes the pronoun system of Pirate Talk more complex than the pronoun system of standard English, which now has only you. Pirate Talk would join many “non-standard” varieties of English that have supplemented the language by adding a second-person plural pronoun: yall in the American South and in African American English, yous in Scotland and Ireland, and you lot in Britain. In every case, the non-standard (often socially stigmatised) variety follows a more complex and precise set of grammatical rules than standard English.

Rule 2. Say be rather than am, are, and is

(2a) Pirate Talk: The scurvy dog be walking the plank.

(2b) Standard English: The scurvy dog is walking the plank.

Many dialects of English allow the plain form of the verb be to occur where standard English requires the verb to be inflected to am, are, or is. This feature is closely associated with African American English in the United States, and there are examples from Irish English. (See Stan Carey’s Sentence First blog).

The rules for this feature vary across dialects, but there’s a general characteristic that the inflected verb is in (2b) describes an event that is taking place currently, while the plain verb be in (2a) indicates an event that occurs continually or repeatedly. Linguists sometimes refer to this feature as “habitual-be.”

Dialects that allow habitual-be address a shortcoming of standard English. While many languages mark a grammatical distinction between one-off and habitual events, standard English does not.

So, if pirates are following the same rule as speakers of a number of other dialects of English, Pirate Talk might include a habitual-be rule in sentences like (2a). In doing so, pirates are adding grammatical complexity and semantic nuance where standard English lacks it.

If that’s the case, it should be noted that a pirate would probably never actually say sentence (2a), since a scurvy dog could walk a plank only once. On the other hand, we’d be on safe ground with a sentence like, “The scurvy crew be sailing the seven seas,” which could happen habitually.

Rule 3. Drop your “g”s

(3a) Pirate Talk: The captain is keelhaulin’ scallywags.

(3b) Standard English: The captain is keelhauling scallywags.

Linguists represent the sounds at the ends of the progressive verb keelhaulin’ and keelhauling with the symbols [n] and [ŋ], respectively. Both sounds are formed by diverting air through the nasal passage. But the [n] sound is formed by raising the tip of the tongue to the alveolar ridge just above the upper teeth to divert air, while the [ŋ] sound is formed by raising the body of the tongue to the velum in the back of the mouth.

In standard English, there’s actually no [g] sound at all in the ending spelled -ing. So really the rule for Talk Like A Pirate Day should be “use alveolar [n] in place of velar [ŋ]".

But nearly all English speakers alternate between these two pronunciations. This alternation has been part of English for roughly a millennium. Today, English speakers follow a complex set of grammatical and social rules in choosing to say either [n] or [ŋ]. For instance, speakers are more likely to use [n] with a progressive verb (“He’s keelhauling scallywags”) than with a gerund (‘the keelhauling of scallywags.”) English speakers also usually associate the [ŋ] pronunciation with formal, careful, and “proper” speech, while [n] is associated with casual, relaxed, and “solidarity” speech.

Like most English speakers, pirates won’t know they’re following sophisticated grammatical rules when they vary between alveolar [n] and velar [ŋ]. But there is good reason to think that pirates would be conscious of the social rules. After all, if you’re trying to embody pirateness, it’s desirable to avoid linguistic variants that may expose non-buccaneering qualities like formality, carefulness, and properness. The [n] of Pirate Talk could be an important strategy for pirates to construct their pirate identities.

Rule 4. Say arrrrr

In a blog entry on Language Log, Mark Liberman credits Roger Depledge with attributing the ubiquitous pirate 'arrr' to actor Robert Newton in the 1950 movie version of Treasure Island. Ben T. Smith further cites British re-enactment group, Bonaventure, for noting that Newton was from Dorset and that the real Blackbeard was from Bristol, so Newton might have tried to use a dialect associated with Southwest England.

Regional accents in southwest England retained the “r”-sound for a long time after accents associated with London began deleting “r” at the end of words and before consonants (turning the London arrr into plain old ah, and "me hearty" into something like "me hotty"). So, for a long time, a pirate’s arrr could have marked them as being from Devon, Cornwall, and other r-pronouncing places.

Because many Englishes (outside of North America, at least) have followed the London pattern to become r-less, nowadays arrr probably can’t differentiate between regional accents of British English. However, arrr is still a good shorthand way to indicate a person is using Pirate Talk. When people begin to consciously associate a language variety with a specific group of people--a process that sociolinguists refer to as “enregisterment”--it’s typical for a particular linguistic feature to take on social meaning as an indicator of the variety. This has happened for bostin’ in Brummie English and 'ello govna in old stereotypes of Cockney.

Pirate Talk has undergone enregisterment. And the “r”-sound, which was once a way to mark regional dialects of English, now functions as a shorthand way to indicate that a person is participating in that enregistered variety.

“Proper” Pirate Talk

Of course Pirate Talk is not a real variety of English. But if it were, it’s clear that speakers of Pirate Talk, like all speakers of all language varieties, would be unconsciously following an extremely sophisticated set of grammatical rules. Furthermore, like speakers of most stigmatized dialects of English, Pirate Talk would often add grammatical and communicative complexity and precision where standard English lacks it.

Pirate Talk would also be doing a lot of social work, as its speakers would constantly choose among linguistic variants as part of a range of practices to actively construct identity. All native speakers of all language varieties constantly navigate competing linguistic variants in this way to shape the personae they present to the world.

In this sense, while Talk Like a Pirate Day is silly, we all talk like “pirates” every day. Our remarkable linguistic abilities, both grammatically and socially, are definitely worth celebrating.

September 02, 2018

The Real News about Fake News


The term fake news has two meanings. When mainstream media like BBC, the Times, and CNN say fake news, they mean ‘fabricated stories on social media.’ When politicians associated with the political right say fake news, they ‘mean mainstream media like BBC, The Times, and CNN, which are biased in favour of liberalism.’ When mainstream media talk about fake news, they reinforce this second meaning, and de-legitimate themselves as an instrument to protect democracy.


What mainstream media mean when they say fake news

  1. A Nexis search on 12 August 2018 found 21 unique stories with the term fake news.
  2. Three of these 21 were quotes attributed to Donald Trump, 2 were attributed to MP Judith Collins, and 1 to an attendee at a rally. One was in a story by Russian state media, which probably shouldn’t count as mainstream media. The remaining 14 were straightforward instances of news stories discussing fake news.
  3. The remaining 14 stories used fake news to refer to fabricated stories in online media, especially in India, where such stories contributed to lynchings. Other stories used fake news in a more general sense for fabricated news, as in Conor Brady’s editorial in The Times:
There can come a point at which a convergence of populist pressures, the attenuation of resources and the gathering of existential fears will combine to render the watchdogs toothless, opening the way for the purveyors of rumour, untruth and fake news.

What right-wing politicians mean when they say fake news

  1. Donald Trump first used the term fake news on Twitter on 10 December 2016. As of 12 August 2018, he had used it in 259 tweets.
  2. 40 of those 259 tweets name CNN. Most of these do not respond to anything that was reported on CNN, but simply invoke the name:
  • Trump invokes CNN as a metonym. CNN is symbolic for all mainstream media, which (in this meaning of fake news) report news according the liberal biases of reporters, rather than reporting facts. This is indicative of a strategy by politicians associated with the political right to cast all mainstream media as biased in favour of the political left.

What happens when mainstream media report on fake news

  1. When a news story uses the term fake news, it reinforces the term as being part of the public vocabulary. So, when news stories talk about fake news, they legitimate the label, fake news.
  2. When the mainstream media and right-wing politicians use the same word, it seems like they are talking about the same thing. In other words, when mainstream media use the term fake news to describe fabricated news and right-wing politicians use the term to describe the mainstream media, it appears that they are agreeing with each other.
  3. This means that each time mainstream media media like BBC, the Times, and CNN talk about "fake news," they confirm the realness of fake news.
  4. By reinforcing the realness of fake news, mainstream media legitimate the claims of Trump and other politicians associated with the political right that fake news is a large problem. Therefore, the mainstream media provide a foundation for claims that mainstream media are fake news. By attempting to differentiate themselves from fabricated information, mainstream media are perpetuating politicians’ attempts to de-legitimate them.


Fake news is a problem. But the bigger problem is that fake news has two meanings, and the mainstream media are helping politicians de-legitimate the mainstream media by reinforcing the politically motivated meaning of fake news. The mainstream media are thereby supporting their own dissolution as an instrument to protect democracy.

August 04, 2018

Rhetorical Fallacies: False Cause

Actions have consequences. People do things and, very often, those things cause other things to happen in the world. This sort of intuitively obvious cause-effect relationship is “causation.” One of the ways that we measure causation is by calculating the extent to which two sets of data are related. Health professionals can look at incidence of lung cancer, and count how often it is that someone who is diagnosed with lung cancer is also a smoker. More specifically, if health professionals structured the study this way, smoking and lung cancer would both be “binomial” variables, because they would have “yes” or “no” answers. We would likely test whether the correlation between lung cancer and smoking was statistically significant with a Chi-Square test. It’s also, course, possible to establish correlations between “continuous” variables—e.g., the number of cigarettes a person smokes and the mass of tumors. This could establish a more linear relationship, like as a person smokes more cigarettes, the mass of their tumors increases. In either case, we’d use this correlation as evidence of causation: smoking cigarettes causes, or at least increases the incidence of, lung cancer.

However, in an equally obvious sense, two things can happen independent of each other. For instance, I taught in an English department at a university in Missouri from 2014 to 2017. During those years, the football team won two national chamionships and the basketball team won one. It is possible that the Powerpoint slides I showed to freshmen about argumentation and sentence structure percolated up to the varsity teams and equipped them with critical thinking and reasoning skills that enabled them to outplay their opponents. In which case, to the fans of Northwest Missouri State University football and basketball, you’re welcome. However, I confess that it’s more likely that I had nothing whatsoever to do with these championships. Furthermore, even though across the whole history of the university, there’s a fairly strong correlation between my being on faculty and national championships (in the case of basketball, the university only won a national championship while I was teaching in the English department!), there’s still no causation. Regardless of statistics, we must use a rational perspective on events to determine (and show) that two events which are correlated are also connected. Otherwise, we risk confusing correlation for causation.

In rhetorical analysis, confusing correlation for causation may be called a “false cause” fallacy. An arguer (either intentionally or unintentionally) interprets two events that are correlated as being causally connected. President Trump, potentially committed this fallacy in his January 2, 2018 tweet, where he suggested that he could claim responsibility for 2017 aviation fatality rates:

Merrit Kennedy reported later that day that there were actually two fatal accidents involving turbo-prop planes in 2017, so there is potentially some equivocation in Trimp’s tweet--there were no fatalities on commercial passenger jets. But, the claim that 2017 was the safest year on record is valid.

The issue is whether Trump’s being “very strict on commercial aviation” in the time “since taking office” caused the safe year. In order to show that the correlation between Trump being president and the absence of air traffic fatalities is connected in a causal relationship. To make this claim, we would need to know in what ways Trump has been “very strict on commercial aviation.” Ideally, we could point to a specific event. For instance, if Trump had signed an executive order directing the Federal Aviation Administration to increase penalties on airlines for pilots who fly fatigued, then Trump could claim that greater enforcement of safety regulations had flights to be safer.

It’s not clear that any action taken by Trump during 2017 could have contributed to a increase in passenger air safety. Alternatively, Trump could have a stronger claim to causation if the event of his inauguration corresponsed to a change in airline safety numbers. However, Kennedy quotes an Associated Press report that, at the time of Trump’s tweet, in the United States there had been no air fatalities since 2013 when an Asiana Airlines flight crashed in San Francisco. So, airline safety during Trump’s presidency appears to be a continuation of airline safety during Barrack Obama’s presidency, rather than a consequence of any action by Trump. It seems likely, then, that Trump has committed a false cause fallacy by claiming credit for airline safety as a result of his being “very strict.”

On the other hand, Trump may be indexing some more complex psychological frames. It’s possible that he means for “very strict” and for references to airlines to subconsciously recall his travel ban. In this frame, his tweet about air statistics would actually mean something like:

I signed Executive Order 13780 on March 6, 2017, which banned entry by immigrants under various conditions from Chad, Iran, Libya, North Korea, Somalia, Syria, Venezuela, and Yemen. People traveling on planes from these countries would have committed terrorist attacks by exploding or crashing these planes, thereby causing passenger air fatalities. Because of my Executive Order, these people didn’t fly on planes, and so they did not commit terrorist attacks, and so there were no air fatalities.

I think this may have been what Trump actually meant by the real tweet on Jan. 2, and may have been something like what it encoded for his supporters. While there are clearly other reasoning issues in the hypothetical tweet, it would at least do better to establish that an action caused a consequence, so that correlation is connected to causation. The actual real-world tweet of Jan. 2 doesn’t do this, and thereby commits a false cause fallacy.

July 21, 2018

Rhetorical Fallacies: Straw Man

In good argumentation, a person who is making an argument will summarize or restate a claims of their opponents and then respond to them. (In great argumentation, an arguer will summarize and restate their opponent’s position to make sure that they are genuinely talking about the same things, and potentially work toward consensus. But that’s a separate point.) In a straw man argument, an arguer summarizes or restates a claim that their opponent does not actually hold. They attribute that false claim to their opponents and then respond to that. The straw man argument can be an extreme version of an opponent’s claim, a misleading interpretation of an opponent’s claim, or even just something the arguer has made up. The person making the strawman argument is able to “win” against the false claim because it’s not a real claim. The opponent doesn’t really hold the position. Very often, no rational person holds the position. So the person making the strawman argument has defeated an argument, but it’s not actually an argument that existed in the world prior to them making up the argument just to defeat it.

The idea behind the name “straw man,” by the way, is something like, “you create a fake person, filled with straw, and then fight it.” It’s an easy fight to win because it’s not a real person.

President Trump released a string of tweets on Jan. 12, 2018 to defend himself from a claim that he had described African nations as “shithole countries,” questioned why the United States would want to accept immigrants from Haiti, and said the U.S. show desire more immigrants from Norway. Amidst tweets where Trump claimed his language was “tough” but not derogatory, he tweeted:

The Democrats’ position on DACA negotiations is, of course, irrelevant to the question of whether on not Trump made disparaging remarks about African nations and immigrants from majority-black countries, and described a desire for immigrants from majority-white countries. The opposition of “our Country” to DACA is also highly loaded. Trump is indicating that the people most directly affected by DACA are not part of “our country.” These people, who were brought to the United States as young children and who have lived in the US for most of their lives, are literally part of our country. Indeed, the point of DACA is to offer a pathway to American citizenship to these people, which would formally make them part of our country. But the opposition of “our Country” to DACA casts people who benefit from DACA expressly as not part of “our Country.” Trump furthermore claims that the people he includes in DACA are benefitting at the cost of people he includes in “our Country,” potentially inciting conflict between “our Country” and (primarily Latino) immigrants.

The strawman, though, occurs in Trump’s claim “Democrats want to stop paying our troops and government workers in order to give a sweetheart deal, not a fair deal, for DACA.” This would be a valid claim if and only if Democratic negotiators had proposed to eliminate funding for the military and all Federal employees (presumably even members of Congress?). Of course, no one in Federal-level mainstream American politics--Democrat, Republican, or otherwise--has ever made any such proposal. To do so would be absurd in every way imaginable. At a purely symbolic level, no legitimate politician could take such a position because the military is so beloved in the American electorate. But it would also be disastrous economically because of the huge role the military-industrial complex plays in the American economy, and administratively because of the gargantuan amount of work done by Federal employees to manage Social Security, justice, food safety, air traffic, interstate commerce, and on and on. It would be even more ridiculous for Democratic members of Congress to negotiate on behalf of a very small number of people (who cannot, at present, vote), at the cost of the voters in their constituencies who would lose the economic impact of military service, military contracts, and military installations, and who would lose all the recourse to Federal services that they currently enjoy.

So, yes, Trump is right. The Democrats would be wrong to demand that the US stop paying the military and Federal workforce in order to make a sweetheart deal for DACA. But the Democrats aren’t demanding that. No one is. It’s a stupid position. So Trump is right against a stupid argument he fabricated, not an actual argument being made by his opponents.

With more careful argumentation, we can get to a valid basis for Trump’s argument. Democrats, at the time of the relevant negotiations, had threatened to vote against increasing the Federal debt ceiling if DACA was not re-authorized. In 2013, Republicans refused to increase the Federal debt ceiling as a means to defund the Affordable Care Act. This resulted in the Federal government furloughing non-essential employees for more than two weeks, including many military servicemembers. This was politically and economically problematic. It was unpopular at the time, and resulted in a downgrade of the Federal government’s rating for its worthiness to borrow money. So, here is how Trump could have constructed a valid argument about the Democrats’ position:

The Federal government needs to borrow money to continue to operate. In order to do so, Congress must pass (and I must sign) a bill to allow the Federal debt ceiling to be raised. Otherwise we cannot borrow more money. If the Federal government cannot borrow money, it will not be able to pay workers. Therefore, the government will shut down until the debt ceiling is raised. In the past, Republicans have used debt ceiling negotiations as a way to make demands for their own legislative agenda. For instance, in 2013, Repulicans in Congress refused to raise the debt ceiling for a brief period as a way to try to defund the Affordable Care Act. This resulted in a Federal shutdown, which I supported. Here’s some of my tweets from 2013:
I tweeted that last one on Sept. 20, 2013, but I've deleted it from my Twitter account. Huh. Anyway, in 2013, the Republican decision not to raise the debt ceiling, which I supported resulted in a government shutdown. As part of this, Federal workers and military servicemembers were furloughed. So we literally stopped paying our troops (until they were given backpay later on). I also claimed the United States would benefit from a shutdown earlier in 2017:
I tweeted that last one on May 2, 2017, but I've deleted it from my Twitter account. Huh. Anyway, Democrats now want Deferred Action for Childhood Arrivals to be re-authorized. They are threatening to vote against raising the debt ceiling as a negotiating point if DACA is not re-authorized. If the debt ceiling is not raised, the government would again shutdown until it is raised. This would likely result in military servicemembers and other Federal employees being temporarily furloughed. This negotiating strategy was wrong when the Republicans used it (and I supported it in 2013 and 2017), and it’s wrong for the Democrats to use now. #DisentangleDebtCeilingNegotiationsFromOtherIssues

Clearly, this valid argument would take several tweets to express. It would, however, avoid the straw man argumentative fallacy, and provide a legitimate space to discuss substantive issues. Also, Trump would have to not delete his past tweets when they conflict with his current actions and positions.

Though, to be fair, it would still be a nonsequitor to the context in which the straw man occurred--Trump’s slurs against majority-black countries and expressed preference for majority-white countries--and the opposition of “our Country” to DACA would still be problematic as an incitement to race war.

July 15, 2018

Cognitive bias and bad policing

Writing about web page

Mia Irizarry posted a video on June 14, 2018, which shows her being verbally assaulted and physically threatened by Timothy Trybus for wearing a Puerto Rico shirt. The video also shows Officer Patrick Connor standing close by and ignoring Irizarry’s requests for help.

As of July 11, Trybus had been charged with hate crimes, and Connor had resigned his position. These seem like promising first steps toward justice.

I’m drawing on this Irizarry’s video to think generally and theoretically about roles that language and cognition might play in policing.

I need to acknowledge and apologize at the outset of this writing that I am extrapolating away from Irizarry’s experience and trauma. That’s gross and unfair to her. In the real world, the most important thing that can come out of this experience is justice for Irizarry. Her video, though, reveals larger insights that may point toward ways to reduce unjust policing, and I want to pursue those insights. But ultimately, I’m pursuing something theoretical inspired by this video, and not commenting on the actual events that are shown in the video.

The specific question I want explore is—operating under the assumption that Connor was a moral, well-intentioned, professional police officer—why did Connor fail to intervene on Irizarry’s behalf?

I think issues related to language and cognition may be part of the answer to this.

On the cognition side, I suggest it’s possible that the expectations Connor brought to the situation blinded him to events that were happening immediately in front of him. Language in the video can reveal some of this cognitive blindness.

Police were called to the scene because another man (he’s wearing a black shirt in the video and appears to be Trybus’s friend) was choking a woman. I suggest that Connor understood all events in the context of solving the problem of the man in black choking the woman, and as a result failed to recognize anything that didn’t fit into that context.

This is evident in several moments:

  • At the beginning of the video, Connor has positioned himself next to the man in black in a position to control his movements.
  • At 5:30, the man in black seems to move toward someone shouting at him in the distance, and Connor shifts to block him. This shows that Connor is capable of controlling a dangerous person, when he’s motivated to do so.
  • At 9:56, Connor separates Trybus from the man in black. This move is presumably intended to segregate the Trybus from the man in black so that they can’t coordinate their stories (and not to safeguard Irizarry).
  • At 23:35, Connor tells Irizarry, "When I was called here, it was between these people."
  • At 29:59, after taking information from Irizarry about the Trybus’s attack on her, Connor returns to trying to collect information about the man in black choking a woman: “Did you see any of the stuff between these guys?”

Connor is not alone is showing this bias.

  • At 16:58, an officer interviewing Irizarry asks twice about whether it was Irizarry’s party that called the police, reflecing the need to close out the call that the police received (rather, e.g., than to help any people in need on the scene).
  • At 17:37, the officer tells Irizarry that they had been called about the man in black--"We got the call it was a male choking a female"--reflecting a continued fixation on the incident of the call.
  • At 21:43, an officer approaches Irizarry and again asks about girl who was choked by man in black.

The expectations police bring to the situation also reduce their ability to perceive Trybus as a threat.

  • At 21:43, the officer apologizes, "We know these guys."
  • At 23:44, Connor tells Irizarry that he knows Trybus and "At no time was he going to attack you."
  • Perhaps most significant of all, at 16:41, following Irizarry's account of events, the officer clarifies "But he didn't touch you? Just made you--" The officer’s question reveals a bias toward minimizing the threat posed by Trybus to Irizarry. Note the difference between her question and the more neutral, “Did he touch you?” The officer’s wording shows that she presumes Trybus only made Irizarry uncomfortable and did not actually assault her.

In short, bad policing may have happened here in part because of the expectations officers brought to the situation. Connor’s case is especially egregious, but it seems clear that there was broader potential for police cognitive bias to cause blindness to the violence being done to Irizarry.

It’s well established that people--including highly competent professionals--can fail to recognize seemingly obvious information when they are focused on complex tasks. It’s possible that Connor really did fail to see that Irizarry was being attacked just feet away from him. This could’ve happened because he was so fixated on solving one problem, that he was cognitively blind to anything that wasn’t directly part of that problem.

To be clear, I don’t mean this as a defense of Connor. But I think there are insights for policing in this observation. Bad policing probably often happens because of the way police perceive situations based on the expectations they bring to them via prior information.

In Irizarry’s video, language reveals some of these expectations. It may benefit police and society to look for ways that language might also help combat cognitive bias.

I assume, for instance, that Connor was dispatched to the park to stop the man in black from choking a woman. Of course, if a person is being attacked, that is unquestionably the most important intervention. But if Connor had also been dispatched to “help anyone else there who needs help,” would he have been less cognitively blind to the threat to Irizarry? What if “and provide any other help you can” was a standard mantra for the police department? What if officers were expected to report formally not only on how they responded to calls, but also on any other good samaritan work they did on a site. What if, once Connor radioed that the immediate crisis of the man in black chocking a woman was under control, dispatch standardly radioed back, “Does anyone else need help?”

I don’t know that the answers to these questions would really be better policing. But, if we take it for granted that Connor was a moral, well-intentioned, professional police officer, his being blinded to Trybus’s attack on Irizarry through cognitive bias the only explanation I can find for his failure to intervene. If that’s true, police departments would be well served to look for ways to reduce the effects of cognitive bias. Language may play a role in this.

Search this blog

Blog archive

RSS2.0 Atom
Not signed in
Sign in

Powered by BlogBuilder