Kṛṣṇa returns to Dvārakā and is embraced by his mothers. The text names seven of them, with Devakī foremost. Debroy's endnote does something a less careful translator would skip: it checks the number against what the text has established elsewhere, and reports that the check fails.
Vasudeva had several wives. The seven mothers can be interpreted as Devaki's co-wives. But in that event, since Vasudeva had eighteen wives, the number seven does not fit. 'Seven mothers' is interpreted as the biological mother, the preceptor's wife, a brahmana's wife, the king's wife, a cow, the nursemaid and the earth.1
The arithmetic is stated plainly. Seven does not equal eighteen. The obvious reading — that these are Vasudeva's wives, Kṛṣṇa's stepmothers — produces a number that contradicts information the text supplies elsewhere.
What Debroy supplies in place of the failed reading is worth examining, because it doesn't fix the arithmetic so much as change what's being counted.
The substitute list — biological mother, preceptor's wife, a brahmana's wife, the king's wife, a cow, the nursemaid, the earth — isn't a list of Vasudeva's wives at all. It's a conventional enumeration of the seven categories of person or entity a man is traditionally supposed to regard as a mother. Two of the seven aren't even human.
So the discrepancy doesn't get resolved by finding the correct seven among eighteen. It gets resolved by concluding that the passage isn't counting wives at all, but invoking a standard formula about maternal categories that happens to also number seven. The arithmetic problem disappears because the arithmetic was never the point.
This is the specific pattern worth naming, because it recurs in this volume's apparatus. The note performs three moves in sequence:
First, it identifies the natural reading — these are Devakī's co-wives, the women who raised Kṛṣṇa.
Second, it tests that reading against other information in the text and reports the failure explicitly: eighteen wives, seven mothers, the numbers don't reconcile.
Third, rather than leaving the discrepancy open, it supplies an alternative interpretation under which no discrepancy exists — because the passage was invoking a different category entirely.
The third move is where the interesting thing happens. The note doesn't say "this is an inconsistency in the text." It says "under this other reading, there's no inconsistency." Both are honest responses to the same observation, but they leave the reader in very different positions: one flags a possible textual problem, the other resolves it by reinterpreting what the text meant.
Worth distinguishing this from the volume's more significant unflagged numerical problem — seven body sheaths in one chapter, eight in another, glossed locally each time without any note observing the difference.
Here, Debroy does notice, does report the failed arithmetic, and does offer a resolution. That's the careful behavior. The sheath case is the opposite: two contradictory numbers, each annotated as though the other didn't exist, no note anywhere connecting them.
The contrast matters because it shows the apparatus is genuinely capable of catching numerical inconsistency when it looks — which makes the cases where it doesn't look more informative than they'd otherwise be. The seven-mothers note demonstrates the checking, and its existence establishes that the absence of such checking elsewhere reflects what happened to be noticed rather than a general policy of leaving numbers unexamined.
The seven traditional mother-categories are worth a second look for what they take for granted. Biological mother, preceptor's wife, brahmana's wife, king's wife, cow, nursemaid, earth — this is a list that assumes a specific social world, in which a man's relationship to authority figures' wives, to livestock, and to the land itself are all describable in the same vocabulary as his relationship to the woman who bore him.
That the list numbers seven and can therefore stand in for a numerically problematic passage is convenient. What it actually tells a reader is that "mother" in this text's vocabulary is a considerably wider category than modern usage assumes — closer to a relationship of dependence and nurture than to a biological fact, which is why a cow and the earth can appear on it without strain.
There's something in the passage the note leaves alone, and it's worth raising because it bears on which reading is more likely. The text describes the seven mothers' physical response to seeing Kṛṣṇa: "Because of their affection, milk began to flow from their breasts and the mothers placed their son on their laps."
That detail sits awkwardly with the traditional-categories reading. A brahmana's wife, a king's wife, the earth and a cow are not a group one would naturally describe as lactating in unison at the sight of a returning young man — the image assumes women who nursed him, which points back toward the co-wives reading the arithmetic was supposed to have ruled out.
This doesn't overturn Debroy's resolution, and the note may simply be addressing the number rather than the whole passage. But it does mean the two readings aren't cleanly separable: the arithmetic favors one, the imagery favors the other, and the note resolves the first without engaging the second.
You're working from a source — a document, a dataset, a historical account — and two numbers it supplies can't both be right. The temptation is either to pick the more plausible one and move on, or to conclude the source is unreliable and discount it generally.
This note models a third move worth trying first: check whether the two numbers are actually counting the same thing. Debroy's resolution doesn't establish which number is correct, because on his reading they were never in competition — one counted wives, the other invoked a formula about maternal categories. A surprising proportion of apparent numerical contradictions in complex sources dissolve this way, not because someone made an arithmetic error but because a term shifted category between two passages without announcement.
The discipline is to test the category before testing the arithmetic. If the categories match and the numbers still conflict, you have a genuine inconsistency worth flagging. If the categories don't match, the inconsistency was an artifact of assuming they did.
One clarification worth making explicit, since the whole note turns on it: the figure of eighteen wives isn't supplied in this passage. Debroy brings it in from elsewhere — background about Vasudeva's household that the immediate scene doesn't state.
That's what makes the check possible at all, and it's the part of the note most easily overlooked. The contradiction only becomes visible to someone holding information from two different places at once. A reader encountering "seven mothers, of whom Devakī was the foremost" in isolation has no reason to suspect a problem; the number is unremarkable on its own terms. The apparatus is doing work here that reading alone cannot do — supplying the second figure that turns an ordinary sentence into a discrepancy.
This is the general shape of what a good annotation layer provides: not merely explanation of what's on the page, but the surrounding information that lets a reader notice when what's on the page doesn't cohere with what's elsewhere.
The strongest evidence that this note reflects genuine checking rather than routine annotation is that it reports a failure. A translator simply glossing "seven mothers" would supply the traditional list without mentioning that the co-wives reading produces a contradiction. Reporting the failure is what demonstrates the check occurred.
Tension — the resolution is plausible but not proven. Debroy offers the traditional-categories reading as an interpretation ("'Seven mothers' is interpreted as"), not as an established fact about what this passage means. The arithmetic problem is resolved under that reading and would return under the co-wives reading, and the note doesn't establish which the Bhāgavata's author intended. Preserved.
Open question. How many similar numerical checks would a systematic pass over this volume's enumerations turn up — and how many would resolve as cleanly as this one does?
The volume's apparatus contains at least two other flagged arithmetic problems: endnote 1021 observes that Abhijit "is usually taken to be one muhurta, not two" against a text that says two, and endnotes 887–888 wrestle openly with a list that counts to nine where the text claims ten. Together with this one, that's three separate instances of the apparatus noticing a numerical problem and saying so.
Set against the unflagged sheath discrepancy — seven versus eight, glossed locally twice with no connecting note — the picture is of an apparatus that checks arithmetic genuinely but not systematically. Some counts get tested; others get annotated as they stand. Which receive scrutiny appears to depend on whether the discrepancy surfaces within a single passage (checkable in place) or across chapters far apart (requiring a comparison nobody ran).
With Legge's Translator Discipline — Inter-Translator Disagreements. That page examines a translator's practice of documenting his own uncertainty on the page — the footnote that says "I cannot tell" rather than silently resolving a difficulty and presenting the result as settled.
The seven-mothers note belongs to that same discipline but adds a step Legge's practice doesn't require: it doesn't merely report uncertainty, it reports a specific test that failed. "Eighteen wives, seven mothers, the number does not fit" is a falsification, not a hedge — the note tells you exactly what was checked and exactly how the check came out.
What the pairing produces: disclosed uncertainty is more useful to a reader when it names the test rather than only the doubt. "I cannot tell" tells a reader to be careful. "I tried reading it this way and the arithmetic contradicted the text's own figures" tells a reader what to be careful about, and hands them the means to evaluate the translator's reasoning rather than merely accepting the flag. The second is a higher standard of transparency and a genuinely harder one to meet, since it requires the mediator to have done the check in the first place rather than simply registering that something feels uncertain.
Sharpest implication. A translator who reports a failed arithmetic check is telling you something more valuable than the resolution he supplies: that he was checking. The resolution may or may not be right, but the demonstrated habit of testing readings against the text's other claims is what lets a reader calibrate how much to trust the passages where no such note appears — and, more usefully, notice where the habit apparently didn't operate.
Generative questions.