Comment-section jyotisha
Three hundred and eighty replies under one chart post, sorted by hand into seven kinds. What the tally shows about where belief actually sits online.
Two ways in. The gist assumes you have never met any of this before.
The reply left most often under a big astrology post is a stranger's date of birth, time of birth and town, typed into a public box under somebody else's chart. Three or four words usually travel with it: check mine, please see, what about me. Beside them sit the writer's account name and photograph.
I took three hundred and eighty replies under one post and sorted them by hand. Seven kinds covered all of them, and the birth details were the largest by a distance. (The post is not one page. It is what I kept finding across many, put together so that a few hundred readers do not arrive at somebody's real thread; no account is named below and no reply is quoted word for word.) Read the counts as proportions. A different thread moves every figure and leaves the kinds where they are.
The post itself was ordinary. A chart drawn as a square, a headline naming one placement, five or six lines about what people with it are like, and a closing instruction to comment your sign. It had travelled well past the people who follow the account, and that is the condition everything below describes. A quiet post gets a shorter comment section, and a kinder one.
Now the sort.
A hundred and forty-eight of the replies were the birth details described above: a date, a time to the minute where the writer had one, a city, a short request. Many carried the reason in the same breath as the data. A job, a marriage, a court date, a mother in hospital. Nobody had asked for any of it. The post offered no readings and promised no answer. The details went up anyway.
Seventy-one were tags. A friend's name and nothing else, or a name with one word after it. Nothing in a tag is addressed to the person who made the post. It is a message from one reader to another, delivered through a stranger's page because that is cheaper than sending it directly. One tap, and it is the most efficient object in the thread.
Sixty-two were jokes, nearly all at the writer's own expense, and nearly all by people who know the vocabulary well. A joke here is a way of standing near the subject without having to say whether you are in or out.
Thirty-eight were what I have started calling the fear echo, and these I read slowly. Somebody has recognised a word in the post and is asking, in ordinary worried English, whether it means them. I have this one, is it bad, my aunt said the same thing last year. They are not arguing with anyone. They are asking a person they cannot see for a sentence that would let them put the phone down. Whether the frightening term deserves its weight is Nikhil's subject. What belongs here is the container: a real question, asked in a format with no way of printing "not enough to go on".
Twenty-two were sales. A practitioner, or somebody presenting as one, offering a detailed analysis by private message, sometimes with a price in the line. These arrive fastest, and they are not aimed at the post. They are aimed at the hundred and forty-eight people above, who have just published a birth minute and a worry in the same sentence.
Nineteen were counter-readings. Another reader disagrees with the verdict and supplies a different one, usually with more confidence than the post had. Some are competent, and two careful people can differ for dull reasons: a different starting point for the zodiac, a different way of dividing the houses, a different school. A thread has no way to show whether a disagreement is careful or just loud, because both look identical: one paragraph, same typeface. Notice also who they are addressed to. Not the poster. The audience. Under somebody else's post, visible disagreement is about the only free introduction available to a person with the same skills and a smaller account.
Four were corrections. Four. Somebody had checked the one claim in the post that could be checked, found it wrong, and said so plainly with the reason. All four sat near the bottom of the thread, below the jokes, where they remain correct.
The remaining sixteen are unsorted, the drawer every notebook has: a single emoji, a word I could not read, two people continuing an argument from somewhere else.
Where the belief actually sits
The usual question asked about astrology online is whether the people in the replies believe it, as though sincerity and irony were the only two categories worth counting. Sorted, that question is answered by twenty-three replies out of three hundred and eighty. Nineteen counter-readings and four corrections are the only ones arguing about whether the reading is right. Everything else is a use of the post rather than a view about it.
So the thread is not a debate. Picture a queue instead: a few people joke while they wait, one man works the line handing out cards, and nobody at all is behind the counter to serve them. The queue is where the belief is. Not in a stated position, which almost nobody offers, but in the willingness of a hundred and forty-eight people to hand over a birth minute in public and wait. Belief you can see in what people do, rather than in what they say, never announces itself out loud, so there is nothing to argue with.
The visible opinions are a poor guide for a second reason. Everybody who read the post is in the audience; only the people willing to type in public are in the thread. Jakob Nielsen put that imbalance at 90-9-1 in 2006: ninety per cent of an online community reads without contributing, nine per cent contribute occasionally, one per cent write most of what is there. It is a rule of thumb from before any of these apps, offered as an observation rather than a law, and still the most useful sentence I know about a comment section.
Look at what the queue is handing over. A date, a minute and a town are not a description of a person. They are close to a name. Ankit has done the arithmetic on how few others are left standing beside you once all three are known, in your birth details. The comment box adds one thing to his argument: the reply stays, in public, for everybody who arrives at the post next year.
Some of that pile is not the writer's own life either. A brother's details, a fiancé's, a child's, posted by somebody with the screenshot and not the permission, about a person who is not in the thread and will not be told. Raman has written about that request as it arrives in private, in someone else's chart. In public it is the same act with an audience.
Why the correction is at the bottom
Nothing conspiratorial happens to those four replies. The burial is ordinary mechanics, and the mechanics are published.
A correction is always late, because it has to arrive after the thing it corrects. Under an ordering by time it sits at the end, below every joke posted in the first hour. Under an ordering by engagement it needs reactions, and a correction is hard to react to: it closes a subject instead of opening one, and the best answer to a good correction is no answer at all.
Both orderings are real and both are documented by the companies themselves. Facebook's help pages describe three settings for a comment section: most relevant, which shows friends' comments and the most engaging comments first; newest comments; and all comments, including possible spam, which still puts the most relevant first. YouTube's public documentation for developers offers the same pair in plainer words, letting comment threads be requested ordered by time or by relevance. Neither company publishes the recipe, and neither has to. The published fact is enough: what you are reading was sorted before it reached you.
Then there is the fold. A reply written under another reply sits inside a small closed stack that opens only when somebody taps it. A correction written that way is read by the person corrected and by almost nobody else.
The reply worth typing
One line, three parts. It asks about the one thing in the post that has an answer. It says where the answer can be found. Then it stops.
Under a caption naming a sign: is that counted from the seasons or from the fixed stars, since the two counts often give the same person different signs. Under a timing claim: which city is that worked out for. Under a quoted line: which chapter is that from. Each can be settled in about four minutes, by the poster or by anybody reading. None of them tells a stranger what to believe, and none requires the writer to have decided whether jyotisha (ज्योतिष, jyotiṣa) works, which is the argument a thread is built to have and cannot finish.
If you would rather answer one of the thirty-eight frightened people, the honest sentence is short and it is not a reading. A post written for everybody at once cannot be about you in particular, because nobody there has your chart. The question behind the worry is real, and answering it takes one person and one kundli (कुण्डली, kuṇḍalī), unhurried.
One thing not to do, and it is a helpful person's reflex. Do not answer a birth-details reply by quoting it back. A quotation publishes the numbers again, under your name, and outlives the deletion of the original.
Go back to a thread from a few months ago and watch the sorting settle by itself. The jokes have stopped being funny, because they were about a week that is over. The counter-reading is arguing with a post nobody remembers. The corrections are still four from the bottom, still correct. And a date, a time to the minute and the name of a town are exactly where they were, spelled the way their owner typed them, still waiting for somebody to reply.
Twice a week
Essays like this one, Tuesdays and Saturdays. One email per essay, and the archive is never mailed out. Subscriptions open shortly; the RSS feed is live now.
Nothing else arrives, ever.