Cheap talk

Sapir-Whorf politeness, virtue signalling etc

February 1, 2020 — April 11, 2023

Is talk cheap? If so, why are we so concerned with who pays for it?

A miscellaney.

Figure 1

1 How much does talk cost?

Cheap talk is not free. For one thing, every breath and every heartbeat is precious in this sweet time we have sandwiched between two oblivions. But for another, learning dialects takes time.

2 “Virtue signalling”

This term is slung around in common dialogue (as opposed to more precise use in the original set up). I’m trying to find a core idea in that usage. How about this?

I think that as much as we would like positive political change X to happen by rational argument or scientific experiment, or whatever, a lot of it will happen by conversations that could be subtitled we are on the same team and people on our team do X so let us do X together. A lot of progressive politics is about enlarging the franchise, so that previously excluded outsiders become insiders. This is why powerful progressive messages have content about in-groups. We are in this together, we are stronger together and so on. Expanding the vote beyond the landed gentry, and then to women, ending slavery, allowing naturalisation of immigrants, all these have expanded the size of the group that might plausibly be us and enabled us to work together on things in a broader sense of mutual interest.

If, while persuading someone that I am on their side, I can also cultivate genuine empathy and understanding of their position, and we can truly be on each others’ sides, all the better; perhaps then we are in a mutual virtuous circle that will truly represent our collective interests. That sounds nice. If I have a belief that my political program is of broad benefit to society I probably want as many people as possible to be on my team, and presumably I should be doing all I can to cultivate the sense of being on the same team.

I probably want to be wary about making my political program a point of difference in group identity. We are in the same group because we like X. Let us do X together and also while we are there complain about people who do not like X and are therefore not on our team, which is to say, pig-headed and morally suspect.

In this case, making my political agenda part of our group identity, rather than an addition to our group identity, we are no longer working with the grain of tribalism, but rather encouraging others to ally with us or oppose us because of who they already are. Now, the success of the notionally-generally-positive policy is governed not by who of us can be persuaded into it, but rather by whether we can dominate them, which may or may not coincide with the quality of the policy. Many other things can go wrong at that point; what else can make its way onto the group platform that I am required to support against the other groups?

Good. OK, so is making peoples’ political stances part of their group identity the best possible interpretation of the term virtue signalling as used in internet comment threads?

That would be interesting.

Of course it is easy for me to glibly say “it is probably healthier to detach political agendas and group membership”, but this is hard in practice because in fact many agendas do naturally attach to groups, with identifiable differentiating characteristics. Queer liberation is likely to be proportionally less salient for the heterosexually invested, for example, and people who have it good under the current system might prefer to think there is nothing wrong with it rather than admit new people to the in-group who might have some notes.

Handling that, in practice, is hard. I am open to suggestions.

3 Goodharting courtesy

If courtesy is the target instead of the measure, it ceases to be a good measure.

4 Prescriptive language use

Figure 2

George Packer , in In The Moral Case Against Equity Language (“What’s a “justice-involved person”?”):

Katherine Boo’s Behind the Beautiful Forevers (Boo 2012) is a nonfiction masterpiece that tells the story of Mumbai slum dwellers with the intimacy of a novel. The book was published in 2012, before the new language emerged:

The One Leg’s given name was Sita. She had fair skin, usually an asset, but the runt leg had smacked down her bride price. Her Hindu parents had taken the single offer they got: poor, unattractive, hard-working, Muslim, old— “half-dead, but who else wanted her,” as her mother had once said with a frown.

Translated into equity language, this passage might read:

Sita was a person living with a disability. Because she lived in a system that centered whiteness while producing inequities among racial and ethnic groups, her physical appearance conferred an unearned set of privileges and benefits, but her disability lowered her status to potential partners. Her parents, who were Hindu persons, accepted a marriage proposal from a member of a community with limited financial resources, a person whose physical appearance was defined as being different from the traits of the dominant group and resulted in his being set apart for unequal treatment, a person who was considered in the dominant discourse to be “hardworking, a Muslim person, an older person. In referring to him, Sita’s mother used language that is considered harmful by representatives of historically marginalized communities.

But the people in Behind the Beautiful Forevers [Katherine Boo, 2012] know they’re poor; they can’t afford to wrap themselves in soft sheets of euphemism. Equity language doesn’t fool anyone who lives with real afflictions. It’s meant to spare only the feelings of those who use it.

Dave Karpf argues George Packer is kind of right about equity language guides.

Does inclusive language increase inclusion? A salient case study for me is “gender-fair language” and there are some further musings there, although they are not particularly deep either.

I would be mildly surprised if using gender-neutral job descriptions had in isolation a substantial effect on the friendliness of a job for a person of a given gender, because I was taught as a linguistics undergraduate to treat the Sapir-Whorf hypothesis with skepticism. However, I could overcorrecting; see Sczesny, Formanowicz, and Moser (2016) for a review of research in this area.1

But maybe language-to-change-individual-behaviour directly is not the aim, language-to-change-organisational-behaviour is. Perhaps using gender inclusive language, as a slightly tedious thing to orchestrate, is an effective signal of institutional commitment to non-discrimination. That is certainly why I attempt to use inclusive language despite my skepticism about the direct method of action. If that is table stakes for more substantive action, it does seems affordable to me personally, because actually I would like to address my own and organisational latent sexism and am prepared to pay for it to operate in an organisation which was good at that.

But maybe these small gestures are not affordable in practice, because frequently discussing such linguistic manoeuvres leads to a pernicious and potentially counterproductive type of invasive argument which I think of as a tokenism versus table stakes debate.

The Equity Trainees is poignant.

5 Simulacra

See Simulacra.

6 A thousand slights

I refer to this as microstressors, which is an older and better name for the (real) phenomenon than microaggressions.

7 Letting rip on that joke

Figure 3

What is the role of humour? Irony? Is edginess blameful or necessary? What does it even do?

For sure humour requires context.

Tressie Mcmillan Cottom discusses context in this form:

That means we all now have to be thoughtful. We all have to consider, oh, wait a minute. Is that what we say in this room? We all have to reconsider what the norms are, and that was the promise of like expanding the discourse, and that’s exactly what we’ve gotten. And if that means that I’m not sure about letting it rip on a joke, that’s probably a pretty good thing.

See also Laughing to Transgress:

I also think that if they had known I was gay, they wouldn’t have joked around, or been embarrassed after, and then it would have been a Big Thing. The way it went, a concern about a person they shared a house with could be brought up and made legible. I joined the “one of the guys” dynamic, by understanding its rules, context and intent. I can get the same effect now just by saying something funny and insensitive about gay people behind closed doors, to show that it’s not a personal landmine at all.

It’s also entirely in line with existing LGBT culture. Drag queens in particular have used sass and mockery as a shield, but no mockery is complete without self-mockery.

Different again, Robin Hanson argues against irony for being outgroup-exclusionary. I don’t think blanket discouraging irony is plausible or desirable, but… there is an insight in there. It is important to remember that indicators of in-group membership, such as irony, are shibboleths, not indicators of any intrinsic quality. I like this idea as a trial balloon for some better ideas that do not conflate local culture with hostility to wider culture, or assume the plausibilty of operating without in-groups. I am not convinced that it shows a plausible kind of group organisation, where future human only operate in simple, maximally unambiguous declarative statements without any attempt to do team bonding by demonstrating awareness of shared context and understanding, which is essentially what irony is for.

Martin Sustrik considers irony in a totalitarian state :

That brings me directly to the topic of subversive jokes. Nobody had told you directly that they want to rebel against the communists. Well, few did, but those were mostly in jail. Everyone else had to find a more subtle way to communicating the message. And jokes turn out to be the optimal medium for doing that.

Yet another caveat: Not everyone who told you a “subversive” joke was planning to overthrow the regime. Telling the joke could mean a range of things from being violently hostile to the system to being generally in favour of the system but signaling your insider status by making it clear that you don’t buy the official propaganda.

All in all, it seems that spreading of common knowledge under oppressive system is a subtle and delicate matter. It’s a system where everyone balances close to the line where propaganda becomes indistinguishable from irony.

But in the end, everybody must participate in the game lest they inadvertently wander into harm’s way and by participating in it and by seeing others participate in it everybody gets at least some level of insight about the state of common knowledge.

Also, he tells this joke which is good.

A drunk Russian man in front of the Kremlin shouts, over and over again: “Down with the mustached dictator!” A policeman arrests him and brings him to Stalin. Stalin asks: “Who did you mean?” The man, suddenly sobered up, answers: “I was shouting against Hitler, of course.” Stalin turns to the policeman: “And what did you mean?”

Figure 4: The ill-fated monument to civic virtue

8 Common knowledge

Nerds love this one as an example of when apparently empty talk changes much.

9 Fiction

Does narrative fiction sublimate our urges for what it depicts, or encourage those urges?

I first became aware of this debate in the context of porn: Do violent pornographic scenarios normalise and therefore encourage violent behaviour, or do they allow people with troublesome violent urges to sublimate them into fiction? But it works for everything. Are rom-coms harmless recreation for people living in an unromantic world, or do they setup for unrealistic expectation of healthy relationship dynamics in vulnerable people?

My trashy fiction is harmless escapism, someone else’s trashy fiction is brainwashing propaganda.

10 Rituals and niceties

TBD (Stein et al. 2021)

10.1 Acknowledgement of country

Recently formalisation and standardisation of Acknowledgement of Country in Australia is one case of apparently cheap talk. It can be contentious. I think investigating why is an interesting case study.

Finding the traditional owners of land can be hard.

11 Incoming

Here are some think-pieces taking assorted positions.

Timothy Burke, in Academia: Disentanglement is a banger:

Why else are people who identify themselves as the ‘sensible left’ or the ‘rational middle’, etc. so ardently against what they slang as ‘wokeness’? Because they think that you can change a political orientation* by trying to make it embarrassing to be ‘woke’. Anybody who is working social media in the neo-Skinnerite way that we all do, trying to behaviorally condition other social media users into what we consider optimal practices, shares some part of the proposition that the words we use and the frames of reference we adopt have sociopolitical consequences that go beyond the semantic content of what was said in any given exchange. To some extent, much of the left is stuck in the space of believing that language, representation and knowledge have power as a result of the fading prospects of conventionalized mass mobilization and mass action after 1975 but also an increasing inability to believe that conventional persuasion of ostensibly rational subjects can or does happen—a viewpoint that has been increasingly validated by cognitive scientists and by anybody who is an attentive observer of social media.

This entire perspective is why most college students leave college with at least some sense that the words you use have power that goes far beyond their ostensible semantics. And again, this is something that not only the American right believes but in fact has frequently demonstrated the validity of. Look at the recent weaponized use of words and phrases like “grooming”, “critical race theory”, “the Deep State” and so on. That said, one of the major problems with progressive politics in the era of social media has been that it is like a kid with a hammer thinking everything is a nail: the proposition that choices in language and representation create politically meaningful sociopolitical changes is over-applied to situations where that makes very little difference and is over-attributed for changes that have actually happened.

Freddie deBoer, Planet of Cops declaims on this theme. In between the invective there is an attempt to tackle process, rather than intention or outcome, which we do have a hard time disentangling in online arguments about who is being sanctimonious and who is being a troll, and who is dog whistling and all the other things that the opponents of a side will explain it as.

A argument that process is prominent in governing speech is a popular theme for many pundits; at this point we have wandered perhaps into community governance.

Ben Burgis, discusses symbolic versus substantive battles.

Given that there are 535 members of Congress, I suppose it’s possible that there are one or two who regularly attend plays in 2020… Perhaps those one or two are even right-wing Republicans who oppose measures like ending the sale of surplus military gear to civilian police departments. The idea that such congressmen would see that the director of some D.C. theater where they attended a really interesting revival of Eugene O’Neill’s Under the Palms last year put out a #BlackLivesMatter statement and in response decided to change their votes strikes me as fairly improbable.

See also his Cancel Culture Is a Symptom of Capitalism’s Rot interview, wherein he is pretty amusing on this topic.

In the test case of COVID-19 communication we can see the dynamics of costly lies:

To the purely political actor, the implausible lie is better. If the lie is implausible, then those repeating it have sent a costly signal of loyalty, and cut ties with lower levels. You don’t have to worry they repeated the statement because it happens to match the physical world, or that they will refuse to repeat the next one if it fails to match.

12 Enforcement of norms

13 References

Boo. 2012. Behind the Beautiful Forevers.
Dobbin, and Kalev. 2021. The Civil Rights Revolution at Work: What Went Wrong.” Annual Review of Sociology.
Farrell. 2012. The Consequences of the Internet for Politics.” Annual Review of Political Science.
Feltovich, Harbaugh, and To. 2002. Too Cool for School? Signalling and Countersignalling.” RAND Journal of Economics.
Jerolmack, and Khan. 2014. Talk Is Cheap: Ethnography and the Attitudinal Fallacy.” Sociological Methods & Research.
Lönnqvist, Szabó, and Kelemen. 2021. ‘The New State That We Are Building’: Authoritarianism and System-Justification in an Illiberal Democracy.” Frontiers in Psychology.
Maerz. 2019. Simulating Pluralism: The Language of Democracy in Hegemonic Authoritarianism.” Political Research Exchange.
Pascale. 2019. The Weaponization of Language: Discourses of Rising Right-Wing Authoritarianism.” Current Sociology.
Richards. 2001. Coordination and Shared Mental Models.” American Journal of Political Science.
Richerson, and Boyd. 2001. The Evolution of Subjective Commitment to Groups: A Tribal Instincts Hypothesis.” Evolution and the Capacity for Commitment.
Sczesny, Formanowicz, and Moser. 2016. Can Gender-Fair Language Reduce Gender Stereotyping and Discrimination? Frontiers in Psychology.
Stein, Schroeder, Hobson, et al. 2021. When Alterations Are Violations: Moral Outrage and Punishment in Response to (Even Minor) Alterations to Rituals. Journal of Personality and Social Psychology.


  1. The Sapir-Whorf hypothesis, IMO, is better described as a family of hypotheses, on the theme of “the language we use affects how easy it is to think some things.” Some of them are dual to Lexical hypothesis, which supposes that language adapts to the kind of things that we think, i.e. the direction of causation is reversed. Neither is in much scientific favour in linguistics AFAICS.↩︎