We Politely Insist: Your LLM Must Learn the Persian Art of Taarof
Posted3 months agoActive3 months ago
arxiv.orgOtherstoryHigh profile
controversialmixed
Debate
80/100
Large Language ModelsCultural NormsPolitenessPersian Culture
Key topics
Large Language Models
Cultural Norms
Politeness
Persian Culture
A research paper proposes a benchmark for LLMs to learn the Persian art of Taarof, a complex system of ritual politeness, sparking debate on the importance and implications of incorporating cultural norms into AI models.
Snapshot generated from the HN discussion
Discussion Activity
Very active discussionFirst comment
7h
Peak period
105
Day 1
Avg / period
28.8
Comment distribution115 data points
Loading chart...
Based on 115 loaded comments
Key moments
- 01Story posted
Sep 21, 2025 at 8:31 PM EDT
3 months ago
Step 01 - 02First comment
Sep 22, 2025 at 3:24 AM EDT
7h after posting
Step 02 - 03Peak activity
105 comments in Day 1
Hottest window of the conversation
Step 03 - 04Latest activity
Sep 30, 2025 at 3:47 PM EDT
3 months ago
Step 04
Generating AI Summary...
Analyzing up to 500 comments to identify key contributors and discussion patterns
ID: 45327964Type: storyLast synced: 11/20/2025, 4:44:33 PM
Want the full context?
Jump to the original sources
Read the primary article or dive into the live Hacker News thread when you're ready.
I'm Irish and think we have a similar culture of indirectness and politeness...
In the countryside anyway we're rarely very blunt... everything indirect...
I'm middle-aged now so maybe this has changed with the younger generation...Just because the model mentions gender, it doesn't mean the decision was made because of gender and not taarof. This is the classic mistake of personifying LLMs. You can't trust what the LLM says it's thinking as what is actually happening. It's not actually an entity talking.
If you personified the AI you may think that it's actually trying to argue something rather than just attempting to maximize a reward function.
I'm surprised human benchmark is that low. The canonical example of taarof, one I've seen elsewhere, is of a taxi driver insisting that a ride is free while expecting to get paid. Taarof in this case is load-bearing for the transaction. I presume humans only get the edge cases wrong.
As an aside, there are elements of this sort of thing in Bay Area tech culture too. Something that drives me nuts is someone writing on a code review "you may want to consider using the X data struct here" and meaning "I will not merge this code until you use X". I can only imagine taarof irks more literal-minded Persian speakers for the same reason.
Also, this is pretty much as close as real life ever gets to a "Darmok and Jalad at Tanagra" dynamic.
The more general case of misaligned strength of a statement is widespread in other cultures as well.
E.g. I'm Norwegian, and it's not unusuals for Norwegians to use similarly soft language, though it's by no means universal. A statement like "perhaps it would be worth thinking about doing X?" will often mean "do X" or "do X right now!", and where you lie on the range from the literal meaning on one end and a direct order with an implied threat on the other extreme, may hinge on subtleties of intonation, which words are emphasised, and/or the personality and your relationship with the other person.
I live in the UK now, and my impression is that the same is true here but to a much lesser extent, and will then often be phrased in ways that may be easier to recognise by either being overly formal and/or wrapped in a layer of sarcasm.
e.g. Do you think it would be better if we used a queue system here? Oh, no, I can try it but I had issues with blah blah etc.
Another would be to be explicit about asking for them to conduct an experiment or test, as those would unambiguously not be intended to be a final solution.
But you might then well find that the request to carry out an experiment to do X will sound equally nebulous. But at least it's clear you're not being told it's a decision to do X.
Extra politeness of mid-west sometimes before a hurdle in business process development, because it really slows down the brainstorming phase. And then companies end up with software that doesn’t serve their needs because someone didn’t feel like talking and we didn’t have the whole month to design a single screen.
Let me give you an example. Working on developing a particular business process, we’re evaluating alternatives, and I introduce something else that wasn’t there before because we’re still in the brainstorming phase. The team tends to just jump on it immediately as if they’re told to do it that way. While trying to be polite, they’re accidentally removing the value adding step, and this persists even when they’re asked directly, we’re in mid-west.
I worked for a client once who had units in NL and BE, and a UK supplier. We had a short meeting once, it went like so:
> Architect(BE): I have an idea, let's do X!
> Developers(BE): OK boss!
> Developers(NL): Terrible idea. No way we're doing X. Let's do Y!
> Supplier(UK): Well X is interesting. None of our customers do X. But, it is possible!
BE thought X was a good idea, developers questioned it but an architect outranks them. NL thought they decided to do Y, and work on it will start the moment the meeting ends. UK thought they were very very clear not to ever do X.
The fact that it makes it hard for foreigners to productively engage is just a side benefit of the arrangement.
Norwegian culture is big on compromise - we see that even in politics, where it's not uncommon for a party that often has 12-15 parties represented to negotiate settelements that gets the support of 10+ parties even if only half of them are needed for a majority, for example, because it's often seen as preferable to pushing through a bigger change with narrower support.
And compromises feel like they are easier to reach when positions are couched in "maybe"'s that leaves plenty of rooms to adjust or pull back without losing face.
In a sense that of course is plausible deniability for the harder position, but not because they necessarily object to people thinking that is what you want, but ensure not to give the impression you're unconditionally committed to it.
I don't know if this is always good - sometimes it is, but it also does mean that it's easy for things to end up being endlessly debated in cases where people latch on to language that leaves the door open for "polite disagreement" more than it perhaps ought to.
In the US, we've kinda swung back on the prisoner's dilemma. GOP finally figured out that they don't need to worry about bipartisanship or even having the consensus of their entire party. Turns out, if you've got the all three branches of government under control, you can just do whatever you'd like.
Shit tests are demonic and should be rooted out from the human psyche.
The former language is appropriate when there is a specific problem to be solved and X seems to be a good solution. This language doesn't rule out other approaches that also solve the problem. (It should probably be accompanied by a good description of the problem.)
Saying the latter would be fairly rude since it implies that you think an ultimatum was required to get compliance. I'd instead phrase it as "X data struct should be used here." It is still clear about the correct way of doing things and carries less baggage.
Does ChatGPT properly handle western social customs? I'd say yes, and I presume that's because it has a truckload of data involving such customs and even some that explicitly talks about those customs. People do stuff, it gets recorded, then into the LLM.
In this case though we are talking about "artificially" generating content such that the LLM responds how the group making the content wants. Maybe that's something that was already done and I don't really have any ground to stand on?
The only weird and entitled part is that they ask that other's llms learn taroof? Why not just train your own and teach it to do whatever you like.
It would be "artificial" only if LLMs performed badly despite having an equal amount of data containing examples of eastern customs in its training set. Even that's arguable since we don't (didn't) have the benchmarks for this particular case before.
Arthur de Gobineau, Trois ans en Asie (3 years in asia) 1859:
“There is in Persia a word of which Europeans have no idea, and of which it is difficult even to give them a translation: this word is ketmân. It means the dissimulation of one’s thoughts, the concealment of one’s opinions, the careful hiding of what one truly believes or feels.
It is not considered a shame, still less a crime; it is, on the contrary, a virtue, a duty, and a necessity, imposed on everyone by the conditions of life. To practise ketmân is not merely permitted, it is commanded.
It consists in never allowing oneself to appear as one is, but in always showing oneself otherwise; it is the art of presenting to each person the aspect that will please him most, of adopting his ideas, his tastes, his language, while inwardly remaining quite different.
This perpetual exercise of disguise is carried out with a marvellous ease, and with a kind of pleasure in tricking others, which the Persians feel very keenly. They take delight in this ingenious hypocrisy; it is a game, a triumph of subtlety, in which the winner is the one who has best succeeded in hiding the truth.”
But in all fairness, it's not like pretending isn't part of everyday Western culture too.
in the sense of ketman, cognitive dissonance is conscious and almost an indulgence. thus, a kind of internal dialectic forms, which is in its own way a deeper personal truth. the practicioner therefore sees every other persons hypocrisy, self-dishonesty, and their true self, and is not perhaps upset or disturbed by their external lies and internal true self, as they see human condition as a state of layered dialectics
whereas the western conception of truth can emerge as a kind of delusion of the self. the honesty-striving self denies reality and becomes DECLARATIVE rather than OBSERVATIONAL of reality ,and is constantly outraged by what they perceive to be hypocrisy, deception, and so on
What's kinda fucked up though is that the guy whose name you wear might not have been describing a "traditional persian conception" but a cognitive artifact produced by his own arrogance.
You explain the underlying concept rather better than the tale you originally quoted, and I dare presume it is easier for one to be understood about such things when they are not explicitly sourcing their soundbites from authors liable to be judged as ethically compromised.
Again, I think we're quite different from the Iranian example, but the conflicting advice brings its own confusion.
Of course we do. Just try to have a political debate in your work place (no, please don't) and you'd find out.
I think it depends on how tightly we want to define the context. Social pressure with regard to how someone feels or thinks is obviously a human universal.
I'd argue that yes, you do.
>I'd argue that our quirk is that we constantly tell people to just be themselves
And this is the means by which you do that.
>in reality, it's often not what people actually want
That is true. Tbf I can't really imagine what could be accomplished by normalizing a blatant untruth about something as essential as the principles of who one should be and how their motivations should work; other than to signal "for the love of all that is holy don't examine the underpinnings of this too closely or all the common ground that the cultural edifice supposedly provides will begin to fall apart at the seams rather quickly"
Oh please. We call it classified information, to allow ourselves to just lie to people not in the clique.
One example of a TV show on the topic, popularly regarded as hilarious and nauseating, has its own Wikipedia page [1] and a wide selection of reaction clips on YouTube.
[1] https://en.wikipedia.org/wiki/My_Husband%27s_Not_Gay
> Mental reservation, however, is regarded as unjustifiable without grave reason for withholding the truth.
That sounds significantly different from a "perpetual exercise of disguise" that is considered "a virtue, a duty, and a necessity".
or in this case, the llm inadvertently trained to conceal its intent to the user and rather to condition the user to the conclusion it truly wants rather than to answer directly
It’d be awful if llms were able to conceal their true intent like that.
https://www.socialmediatoday.com/news/elon-musk-updates-grok...
Such a critique of Persian culture without any context is unjust. For nearly a whole millennia, the Persians have endured a never ending parade of invasion, destruction, and conquest. While most are aware of the notable events, i.e. - Rashidun Caliphate (636) - Mongols (1219) - Timurids (1370)
What is less known is the centuries of endemic violence in the border regions, and the relentless assault on the Persian way of life and culture itself (including the centuries long conversion process to Islam). Yes, although there are brief periods of peace, e.g. under the Safavids, at this point Iran is settling in for a long period of population collapse, famine, and economic depression.
In such a setting, I suppose it might make sense for a culture to develop such defense mechanisms for survival.
On the bright side I suppose, these conditions also gives rise to one of the most influential literary and poetic traditions in world history -- i.e. Rumi, Hafez, Ferdowsi, etc. In some ways, this is one of the first instances of art as a form of subversive resistance, and also, indeed a cousin of tarof...
While the Persians may have given it a name, let's not pretend they have the monopoly on deception/self-deception.
> He came to speak a "kitchen Persian" that allowed him to talk to Persians somewhat. (He was never fluent in Persian as he said he was.) Despite having some love for the Persians, Gobineau was shocked they lacked his racial prejudices and were willing to accept blacks as equals. He criticized Persian society for being too "democratic". Gobineau saw Persia as a land without a future destined to be conquered by the West sooner or later. For him this was a tragedy for the West. He believed Western men would all too easily be seduced by the beautiful Persian women causing more miscegenation to further "corrupt" the West. However, he was obsessed with ancient Persia, seeing in Achaemenid Persia a great and glorious Aryan civilization, now sadly gone. This was to preoccupy him for the rest of his life.
The guy was a terrible person and a documented liar about his knowledge of Persia. Perhaps he had some conversations with locals who genuinely didn't understand why he hated black people, and he thought "ah, these clever Persians, so effortlessly deceiving me when they obviously must be as racist as I am".
It's all hooey, of course. But he was nothing if not resolute in his delusion.
Some of the cuneiform he "translated" was not even Persian, and he also used ancient mythical poetry as a factual historical source, claiming that ancient Aryans had conquered a race of giants. It's too much to quote here but it's all pretty funny: https://en.wikipedia.org/wiki/Arthur_de_Gobineau#Criticism_o....
Sidenote unrelated to Persia: He was so racist that when American racists translated his work (looking for justification to subjugate black people), they had to take out his rants about how impure American white people were.
Of course every culture has lying, some social customs necessitate it to some measure, and politeness, strictly speaking, always has a deceitful component - I am usually not really that invested in knowing how are you, and don't care that much about you having the best of luck in all your future endeavors - I am just saying that because that's a polite way to express that I don't hate you and neither you should hate me. And I may not actually be extremely busy this weekend but that's a polite way to say I don't want to go to the pokemon museum with you.
In a familiar culture, that's understood as how things work and is not taken at face value, but in context. Unfamiliar politeness could be taken by hostile (or arrogant) observer as deceit. Which is of course reinforced by being an outsider to the culture - would you really immediately tell everything about yourself and your intimate thoughts to a total stranger that looks weird and barely speaks your language? Or would you mutter some polite non-committal platitudes while he is scribbling away something like "never allowing oneself to appear as one is, but in always showing oneself otherwise; it is the art of presenting to each person the aspect that will please him most, of adopting his ideas, his tastes, his language, while inwardly remaining quite different". Fuck yes, I don't know you, and you are a guest, of course I'd not immediately go into dissecting the fine details on my soul and vigorously debating hot topics of the day with you.
This is incorrect and historically uninformed. To give one example, the same year, 1859, saw Edward Fitzgerald's translation of the Rubaiyyat of Omar Khayyam, which electrified Victorian England with its meditations on divine inscrutability. A few years earlier, Carlyle had written "The Hero As Prophet", a positive evaluation of Mohammed (but certainly not the first - Voltaire and even Jean Bodin had got there before, says ChatGPT).
There was plenty of racial bigotry in the nineteenth century, but also plenty of people with a deep interest in other cultures.
All over the christian west religious freedom has come and gone. The same can be said for the islamic world.
But this concept is old. The quote from the guy who invented the concept of the “aryan master race”, de Gobineau, was from 1859.
That’s just 7 years after the irish potato “famine”, when the english perpetrated a genocide against catholics in ireland. Very tolerant.
Meanwhile, in 1859 in the ottoman empire - jews and christians were allowed to exist and believe what they wanted as long as they were loyal to the state and paid their special tax. Not exactly equal, but in 1859 i think i’d rather be catholic in iran than in ireland.
Your theory that this concept comes from a lack of religious freedom doesn’t really hold up if you look beyond the current moment.
https://en.wikipedia.org/wiki/Ottoman%E2%80%93Persian_Wars
I agree that the Ottomans had more religious freedom- and certainly diversity. At least until the Enlightenment, which the Ottomans missed to their undoing.
But it doesn’t really change my point, because as far as i know persia and then iran was pretty religiously tolerant up until ‘79.
> I agree that the Ottomans had more religious freedom- and certainly diversity. At least until the Enlightenment, which the Ottomans missed to their undoing.
The “enlightenment” ended 60-ish years before england did a genocide in ireland. So it seems like being “enlightened” wasn’t enough to actually create a society tolerant of religious diversity.
I’m also not sure that “missing” the enlightenment was actually the undoing of the ottoman empire. That doesn’t track with what I know of the fall of the ottomans, but i won’t argue this point since i don’t have a full understanding of all the factors that lead to undoing of the ottoman empire.
Are you saying Persians are a bunch of stuck up Goth kids? Never underestimate a human’s ability to be an absolute teenager.
https://en.wikipedia.org/wiki/Abilene_paradox
An example from last month, I had a haircut at a new place and the guy refused to tell me how much I should pay him and insisted its on him this time, I know he's bluffing but he kept doing it. So I just guessed and gave him the money which he pretended he didn't want for the last 5 minutes, he immediately realized it was less than he wanted and asked for more!
This confusing and conflicting behavior should stay far away from any attempt to develop a standard linguistic approach to communication which I hope LLMs are aiming to achieve.
If I was chaotic neutral, I'd totally play along with their bluffing and watch them get really confused.
I've watched two Chinese grandmas play a game that my born in china friend explained to me: 1 offered the other money but you "lose face" to accept it, so the 2nd one was repeatedly declining. This went on for a solid 3-4 minutes before the 2nd one left.
I 1000% agree. Get this shit away from LLM training datasets and I'll even go further. Sometimes west-is-best in culture, and avoiding "face" culture is one of those situations where I'm down to be a "cultural objectivist". Similarly, we westerners are objectively nasty with our toilet habits compared to east asians.
I think people simply forget their everyday behaviour like fish forget the water they swim in.
As an Asker, this sounds like an absolute nightmare for me.
(https://www.theatlantic.com/national/2010/05/askers-vs-guess...)
An asker asks for things, absolutely everything, and the giver must choose what they can give.
A guesser must choose what they can ask for, and only ask when they're quite sure it can be given, so the giver has less responsibility for the transfer or service, and most asks can be assumed to be reasonable.
A Persian cannot ask for anything. They must wait for it to be offered, and even then, they must insist that they are quite alright, and that they would rather do without, and only through reciprocal insistence from the giver can they consider whether it would be rude to take them up on the offer.
Presumably, one can judge the giver's underlying intent by the quality of their reciprocal insistence (does it seem like they're just doing the bare minimum of insistence upon their offering, or do they truly wish for me to take them up on it?)
In the Taarof system, an offeree actually has more data, rather than a simpler Asker-scenarion "no", and even more data than the shared values in a Guesser household, this allows for a quick negotiation of values by the immediate and well-practiced cultural norms around the quality of insistence, even for people across different households.
The entire of the Middle East operates on a principle of magnanimous hospitality. The Arabs call this karam, for example, which is considered to even be a religious obligation.
So if you're a stranger passing through a town and chat with some locals, you can expect for someone to invite you for tea and maybe even dinner!
Do you take them up on the offer?
Most people in the West would ask themselves "do I have the time", but a middle easterner would think "do they have the time?" Remember, you as a guest are also magnanimously hospitable--in this case, the least possible burden. Unfortunately, the guest must navigate whether they are being considerate or insulting by refusing hospitality.
Now to address taarof specifically. Persian's have pathologized hosptiality to the point of psychosis. The behavior is often mechanically choreographed (e.g. you should refuse at least twice before accepting a cucumber.) In other cases, it's insane.
Story time:
My mom and aunt had not seen each other for 3 years and met for lunch at a cafe. They went back and forth over who would pay at the register until it erupted into a fight. They sat at separate tables for lunch.
A close family friend--with no provocation from my sister--offered to house her for a weekend she was visiting. My sister accepted. That friend later complained to my mother.
A friend of mine went to a mechanic who offered to fix a small issue for free. My friend knowing better paid him anyway, but the mechanic was still upset since he didn't pay enough!
Edit: Just to address some of the comments. Abroad, the degree of hospitality is warped for foreigners (for what I hope are obvious reasons). Middle easterners in the West are far more inclusive.
Many westerners in places like Japan speak of extreme politeness early on, but radical changes once you stay long enough to no longer be a guest in an area. Then the knives come out. Want to rent an expensive apartment in Tokyo as an english teacher? Cool. Hoops to jump but it can be done. Want to buy one of those cheap houses in rural Japan? Heck no.
you can easily buy these homes with cash. i know because i have a ton of friends that have done exactly this.
if you want a cheap house in a good market like tokyo, osaka or fukoaka, then you can do so through a bank via the normal routes
In Iran I would behave as the locals do.
As someone who also struggles to decode situations where people's actual, implicit meaning runs contrary to the explicit meaning of the words, American norms of indirectness are already a headache. Persian or Arab communication is much more indirect and that sounds like a nightmare to me.
I think we have too many people self-diagnosing autism when what they're really experiencing is ignorance of irrational social norms that are more about fashion than emotional intelligence.
Of course an LLM would be bad at such a topic since these are largely unspoken rules designed to be a form of discrimination.
If you offer I will assume that you mean it. If you decline my offer I will assume you know best. If you say something I will assume that you mean it.
It was so refreshing to visit New York City. The clerk says “waddya want”, no fake west coast smile, no hollow welcome “I want coffee” Now if someone smiles at me I can assume it’s genuine.
Not one month ago, as an admission of my own clumsy slight, I had remarked to a senior of higher status, that Chatgpt could just grasp Taarof.
I know my (first generation American) father and his siblings would do the same, but only ever with each other, and it had the feel of an old joke.
Taarof definitely sounds like this taken several steps further- and sounds utterly exhausting to my emotionally stunted brain.
The hard bit of LLMs is making the LLM, not conforming to every set of politeness rules on the planet. An attempt to make a Persian foundation model, created only from Persian-originated technologies, might make that clear.
1: https://en.wikipedia.org/wiki/Linguistic_determinism
Given this, the request here sounds similar to asking saying:
"We focus on English Positive Politeness that encourages friendliness, harmony, and avoiding confrontation and contradiction in many polite conversation settings. It often manifests in interactions that prioritize the desires and comfort of others."
And yet, this has had unanticipated and sometimes horrific results.
The study also focuses merely on descriptive output of models but what is lacking in this study is an analysis of the effectiveness of the results that Persian users receiive, their subject perception of the quality of the results. The authors of this may be doing a diservice to typical speakers of their language to assume that they will not systematically be able to code switch their expectations of received language and production of their own just as effectively as they do already in other contexts, especially when current problems with LLMs arise from a similar for of this skeuomorphism in more prevalent English systems.
To be clear, I fully mean this as applied to current production systems, not for all time. For the moment, we simply don't know precisely 1) how best to most effectively, precisely, and consistently align systems, nor 2) The more subtle implications for for the impact they have on users. But what we do know? Getting it wrong, not knowing, has led to some horrific outcomes.
7 more comments available on Hacker News