Not a hymn so much as a device: twelve names, numbered, in metre, with a retrieval schedule and a delivery date.
Eight verses. Four of them are a numbered list. One tells you when to say it, one tells you what four kinds of people get from it, one gives a six-month and a twelve-month milestone, and the last tells you to write it out. Read that way it stops looking like poetry and starts looking like engineering.
The Atharvaśīrṣa makes a metaphysical claim and spends fourteen verses defending it. This text makes no claims about reality at all. It contains a list, an index, a schedule, an audience segmentation, a delivery estimate and a distribution clause — and that is the whole of it.
This is not a criticism. It is what makes it the single most widely memorised Gaṇeśa text after the Chālīsā's Hanumān equivalent, and it is why children learn it first. Every design decision in it serves recall: the names are numbered in the accusative (first the curved-trunked, the one-tusked second…), the metre is the plainest in Sanskrit, and the total length fits comfortably inside ninety seconds.
The saṅkaṭa in its title means a narrow place — a tight passage, a strait, hence a crisis. The stotra is what you carry into the narrow place, which is why it had to be short enough to hold in your head with nothing in your hands.
The colophon places it in the Nārada Purāṇa, spoken by Nārada — the sage whose narrative function throughout Purāṇic literature is to carry things between worlds. He is the messenger, the one who tells someone in trouble what to do. Attributing a crisis-stotra to Nārada is not an accident of authorship; it is a statement about genre.
Purāṇic material of this kind was composed for householders rather than specialists: people with obligations, in trouble, without Sanskrit training and often without literacy. Compare the two texts' access clauses and the difference is stark. The Atharvaśīrṣa says do not give this to one who is not a student. This one says the opposite — विद्यार्थी… धनार्थी… पुत्रार्थी… मोक्षार्थी: whoever wants knowledge, money, children, or release. Four wants, no qualifications, no initiation.
Whether it was revealed or composed is, here, unusually easy to hold together. A Purāṇa is by its own account a retelling — that which is ancient, put into words that the present can use. The text does not claim originality. It claims usefulness, and it names its price in months.
Half the text is payload and half is instructions for using the payload. Very little is decoration — there is one epithet in verse 1 and no adjectives anywhere else.
Nine of the twelve name a physical feature. Two name an office. One names a colour. There is not a single name in the list for a virtue — no compassionate, no wise, no merciful. This is a description of a body, and the tradition reads each feature as a capacity.
The text does not say. Anything offered here is inference, so here are the candidates in order of how much weight they will bear.
Japa is traditionally counted on the finger-joints: three phalanges on each of four fingers of the right hand — twelve. One pass of the thumb across the joints is one round of this stotra, one name per joint. The list is the length of the counting instrument it was designed for.
Four names per verse, one per eight-syllable quarter, name plus ordinal in each. Twelve is simply the payload capacity of three verses of the standard metre — no more, no less. See §07.
Twelve is the year. A list of twelve said three times a day for one saṃvatsara — the text's own stated period — has an obvious calendrical shape. Suggestive, and not stated anywhere in the verses.
Charts mapping each name to a rāśi or an energy centre circulate widely. None of it is in this text, in the Nārada Purāṇa around it, or in any commentary before the modern period.
अनुष्टुभ् is the ordinary metre of Sanskrit narrative — four quarters (pāda) of eight syllables each. It is the plainest, most forgiving container in the language, which is why the epics are written in it and why almost every stotra a household knows is in it too.
Watch what this text does with it. Each eight-syllable quarter carries exactly one name plus its ordinal — nothing else fits, and nothing is left over. The metre is not decoration around the content; the content was cut to the size of the slot.
Thirty-two syllables, four key–value pairs, one syllable of padding. If you have ever wondered why these texts survived a thousand years of purely oral transmission with so little corruption: this is why. A missing syllable breaks the metre audibly, and the ordinal tells you exactly which slot is empty. The form is error-detecting.
सन्ध्या is a joining — the seam where one thing becomes another. Dawn, when night becomes day. Noon, when ascent becomes descent. Dusk, when day becomes night. Vedic practice treats these as the hinges of the day, and prescribes recitation precisely at them.
The instruction is therefore not three times a day in the sense of a dosage. It is at each transition. A text about obstruction — about narrow places one has to get through — is scheduled at the points where the day itself changes state.
Practically: it also produces spacing of roughly six, six and twelve hours, which is close to what a modern memory researcher would design if asked to schedule three daily rehearsals of a short list. See §12.
Verse 6 is the four puruṣārthas — the classical aims of a human life — in the order a life tends to meet them, minus one. Dharma is absent; vidyā, knowledge, stands in its place. That substitution tells you who was expected to be reciting.
Note the last one refuses the obvious word. The seeker of mokṣa does not obtain mokṣa; he obtains गति — a going, a course, a state moved toward. Three concrete nouns and then, where you expect the biggest noun of all, a direction. The text will promise you a son. It will not promise you liberation.
This is the only one of the four texts on this site that quotes a delivery date. Put the numbers together and the commitment is visible:
Thirteen and a half hours, spread so thin that no single day costs more than four and a half minutes. Compare the Atharvaśīrṣa's thousand repetitions — 117 hours, realistically a group undertaking. This text was built for one person with no spare time.
नात्र संशयः is a stock phrase — it appears hundreds of times across Purāṇic literature, and it functions roughly as and that's that. It is a rhetorical seal, not evidence, and reading it as a guarantee misreads the genre.
What can be said without any special pleading: six months of a three-times-daily practice will reliably change the person doing it — in attention, in the shape of the day, in what is available to recall under pressure. Whether it delivers a son or a salary is not something this page can adjudicate, and not something the text can either.
Worth noticing: the stated interval is long enough that almost any life will have changed by the end of it, in some direction. Promises with a six-month horizon are unusually hard to falsify. That is a fact about the form, and it need not be cynical to say so.
The Atharvaśīrṣa ends: having caused eight brāhmaṇas to grasp it properly. This one ends: having written it, present it to eight brāhmaṇas. Same number, same position in the text, different verb — one oral, one written.
Whatever else is going on, this is a replication instruction with a reward attached, placed at the end of the payload, and it is the reason both texts are still here. A recitation that ends by asking for eight copies has a reproduction rate of eight. Nothing else in either text does more work for its survival.
Read historically, brāhmaṇa here means the literate professional class who could actually store and reproduce a text — a restriction that excluded most people, and there is no point pretending otherwise. Read practically today, the operative content is: this does not stay with you.
Two findings in cognitive psychology are about as solid as the field gets. The first is the testing effect: retrieving material from memory strengthens it far more than re-reading it does — demonstrated repeatedly since Roediger and Karpicke's 2006 experiments and replicated across materials and populations. The second is distributed practice: the same total study time produces markedly better long-term retention when spread across sessions rather than massed, a result confirmed in Cepeda and colleagues' large 2006 synthesis of the verbal-recall literature.
Now describe this stotra in those terms. A twelve-item list. Indexed, so each item can be retrieved by position. Encoded in a metre that makes an omission audible. Recited from memory, not read — that is retrieval practice, not review. Three times a day, at fixed intervals. For six months. It is, structurally, exactly what a modern protocol would prescribe.
One caution that matters. The evidence supports the claim this schedule is an excellent way to make twelve names permanently and instantly available to you. It does not support the claim and therefore the six-month result arrives. Those are different sentences, and only the first one has data behind it.
A finite, ordered sequence gives panic something to do that has an end. You know where you are and how far is left. The thumb moving joint to joint keeps the body in it too.
Twelve names for the same being is twelve angles on one thing. Whatever the trouble is, the practice trains the move: give it a name, then give it another, and it stops being weather and starts being a shape.
Under real distress, working memory narrows and long sequences collapse. Ninety seconds of eight verses is inside the budget. Design constraint, not modesty.
Name seven is sovereign of obstruction, not remover of it. The stotra's implicit posture is not make this go away but you are the one who put this here; let me through. That is a markedly different relationship to a difficulty, and it is available whether or not anyone is listening.
Paste a YouTube link to the recording you use; it is saved on this device. Tell me which one you have settled on and I will time the line highlighting to it. Until then, Follow along in §04 paces the verses evenly, and the tap button records real timings you can hand back to me.
The Saṅkaṭanāśana Gaṇeśa Stotra as transmitted with the Nārada Purāṇa and printed in standard stotra collections. Eight anuṣṭubh verses; the division is universal.
Common variants: verse 2 is printed both as वक्रतुण्डं च एकदन्तं and with sandhi as वक्रतुण्डं चैकदन्तं — the second is metrically tighter, the first is how most people say it. Verse 5's प्रभो is sometimes प्रभुः. If your household's version differs, tell me and this page will follow yours.
Translations, glosses and readings written for this page. Where an interpretation is inference rather than something the text states, the page says so explicitly.