Vācā, phuro, ājñā

One of the four items counted here may not be a word at all. It stands thirty-three times in the 2016 compilation and it is a construction rather than a term, and where it stands a 1968 Gita Press commentary carries ईश्वर उवाच, īśvara uvāca, a plain Sanskrit speech-frame meaning “Īśvara said.” If that is the older reading, the compilation’s form is a corruption of it and there is nothing here to treat. The question is not settled here, and nothing below is built on it. The full statement is in the section headed The phrase that may not be a phrase.

Three words from one paperback

Three words came out of one book, and read on their own they look like the vocabulary of a whole genre.

The book is a paperback. Somebody typed out a very large quantity of Śābara mantras in Hindī and put the result into circulation, and the imprint page names it: Durlabh Śābar Mantra, presented by Tāntrik Bahal, Rājā Pocket Books, Delhi, new edition 2016, 323,692 characters. There is no manuscript basis stated, no apparatus, no note on where any of it came from, no indication whether a given item was copied from a chapbook of the 1920s or heard from somebody’s uncle. By every standard of textual scholarship it is worthless.

You cannot establish a text from it. You cannot date anything in it. You cannot tell which words are old.

You can count it.

Counted, it gave up three recurring words, and the three looked like a set. वाचा, vācā, speech. फुरो, phuro, a verb form meaning let it come true. आज्ञा, ājñā, command or permission. A noun for saying, a verb form for taking effect, a noun for the authority a person speaks under. Taken together they read as the working vocabulary of a whole kind of book.

The difficulty is never with the counting. It is with what a count of one book is a count of.

Chapter 10 counts it and publishes the extraction method, the search strings and the normalisation rules. What no count of one book can supply is the rest: how many books like it are in print, whether they copy each other, and whether this one is typical.

Two of those three questions have partial answers, and they were got by counting seven more books.

Chapter 10 has run the identical script over eight independent printed compilations from six houses in Delhi, Vārāṇasī and Prayāg, with a second scan of this book as an optical control, a further book by the same compiler as a dependent control, and two occult books from the same shelf that are not Śābara compilations as genre controls.

Two of the three words came through that enlargement and rose sharply while doing it.

The third did not come through at all. Ājñā stands denser in this Delhi paperback than in any of the seven books counted beside it, and at its thinnest, in one of the Prayāg volumes, it is seven times thinner than it is here. It is not characteristic of the corpus. It is a habit of one pocket book from one house, and it is the word most of the philosophical argument below would otherwise have rested on.

A word about the word “corpus,” which is used here only where eight witnesses earn it. The population is still unbounded. Eight independent draws from it exist. Those are different statements and this chapter keeps them apart.

Counting punishes carelessness, and chapter 10 records how. This chapter therefore takes its numbers from there and does not re-derive them.

The counts, and a fourth item

वाचा, vācā, sixty-five. फुरो, phuro, fifty-one. आज्ञा, ājñā, forty-two. And a fourth item that is a construction rather than a word, standing thirty-three times, which chapter 10 has struck from its table for that reason and which is taken up below.

Those figures are token counts on a poor scan, folded across the spelling variants this research could identify. Chapter 10 prints the substring count beside each of them, and on vācā and on phuro the two counting rules disagree. The distance between them is the size of the error bar on this instrument. The disclosure is chapter 10’s and the terms are this chapter’s, so it is repeated here.

Now the part one book cannot supply. Counted across all eight compilations, as rates per 100,000 Devanāgarī tokens, the three words do not behave as one group.

Two of the three rise, and they rise by a factor approaching three. The third falls, and the standing witness holds the highest rate in the set.

That third row breaks the set, and it is dealt with in its own section below rather than folded quietly into a sentence.

Take them one at a time, lexicographically, because that is the only thing a word count entitles anyone to do with them. Counting tells you these words are in the book in these quantities. It tells you nothing whatever about how any of them sits in a line, and this chapter will make no claim that it does.

Vācā belongs to vāc: speech, the spoken word, utterance. It is the oldest and most loaded noun for saying that the tradition has, and it carries the largest freight of sacral association with it. Dāsa’s Hindī Śabdasāgara enters वाच् in the revised edition with वाचा, वाणी and वाक्य as its glosses, which is to say that in Hindī the word has stayed inside its Sanskrit range. The वाचा head-word itself could not be isolated in the scan worked from here, so no address is cited for it and none should be invented.

Phuro belongs to the vernacular root phur, to come true, to be fulfilled, to take effect. The form’s grammatical force in this register is optative, or imperative, or something between the two, and let it come true is the nearest plain English. Dāsa gives फुर in the revised edition as सत्य, सच्चा, true and real, from हिं० फुरना, and cites Kabīr and Tulsīdās for it. The head on the page is फुर. Nearby, फुरहरना is derived from Sanskrit प्रस्फुरण, and the original edition gives स्फुरण as the ancestor outright. That closes the question of the word’s descent, and the closing is done below.

Ājñā is Sanskrit for command, authority, permission. Both directions are live in the word. An ājñā is what a superior issues, and it is equally what a superior grants, the leave or sanction under which a subordinate acts. Chapter 10 files the same word the same way. A speaker who says ājñā is not necessarily issuing one. He may be citing one.

Hold that ambiguity. The corpus takes part of it back below.

Read that lexicon against what the popular literature says this material is, and the mismatch is real, though it is narrower than it first looks. These are not the words of a description. They are not opposed to the words of an oath either: chapter 10 shows the standard Hindī dictionary glossing the appeal word and the oath word by each other, so no such opposition exists in the language to build on. What is left is a positive claim about kind. This is the vocabulary of a declaration.

The word that did not travel

The disconfirmation is printed where it falls.

Read in one book, ājñā looks like one of three words that characterise the material. Two of the three are. It is not.

This Delhi paperback carries ājñā more densely than any of the eight independent compilations chapter 10 counted. The Banaras compilation, three and a half times its size, runs lower. So does the Vārāṇasī book of 2012, and the other Delhi pocket book, and every volume of the Prayāg series, the thinnest of them seven times thinner than this paperback. Only the imperfect Manoj copy runs higher, and one imperfect copy does not make a trend. Chapter 10’s table carries the rate for every book in the set.

Set that against the other two words. Vācā nearly triples on the way to the largest witness. Phuro nearly triples. Ājñā falls by a quarter, and keeps falling as the sample widens.

So the honest form of the finding is narrower than three words in a row. Vācā and phuro are features of the printed popular Śābara corpus, on eight witnesses, at rates one to two orders of magnitude above the genre controls. Ājñā is a feature of Durlabh Śābar Mantra.

That matters more than a single row of a table, because ājñā is the word doing the philosophical work in the second half of this chapter. The double sense, command and permission, is what invites the question of whose command is being cited in a book with no office behind it. That question is worth asking of the book that carries the word densely. It cannot be asked of the corpus on the strength of a word the corpus does not use at anything like the same rate.

The argument below therefore does not stand on ājñā. It stands on the thing the eight witnesses actually agree about: the density of assertion, the density of appeal, and the absence of a credential vocabulary from every one of them.

Two cautions on the disconfirmation itself, because a reversal deserves the same scepticism as a finding. Ājñā is a common Sanskrit and Hindī word with a life entirely outside this material, and its rate in any given compilation may be tracking how much Sanskrit that compiler kept. And the count folds three spellings and no more. Neither of those rescues the corpus claim. Both mean the reason for the fall is unexplained, and it is left unexplained rather than furnished with a story.

The phrase that may not be a phrase

The fourth item, at thirty-three occurrences, is the hardest of the four to handle honestly, because there is a rival reading and it comes from a better book.

No English rendering of the compilation’s form is given here, and no grammatical analysis of it, because an analysis names the elements one by one and this treatise does not print constructions from this material even in pieces. What can be said without printing anything is that the sequence does not parse cleanly as Sanskrit, and that its failure to parse is what raised the question this section is about.

Mānas-pīyūṣa, the standard Hindī compendium of Rāmcaritmānas commentary, edited by Añjanīnandan Śaraṇ and published by the Gita Press at Gorakhpur in 1968, prints in its Bālkāṇḍ volume a set of specimen Śābara mantras, sourced there to the ṭīkā of Paṇḍit Rāmeśvar Bhaṭṭ. This treatise does not reproduce them and does not describe their shape. What it reports is one lexical fact from them. Where the 2016 compilation has the sequence it counts thirty-three times, the specimen printed in the Gita Press commentary has ईश्वर उवाच: īśvara uvāca, “Īśvara said,” which is one of the commonest frame-formulae in the whole of Sanskrit literature and needs no explanation at all. The commentary’s form is printed here because it is an ordinary Sanskrit tag that occurs nowhere in the compilation. The compilation’s own form is not printed anywhere in these chapters, because it is a construction and this treatise prints terms and counts.

The 2016 file returns ईश्वर उवाच zero times.

If the commentary’s reading is the older one, then the thirty-three occurrences counted here are a vernacularisation or a mishearing of a Sanskrit speech-frame, and what looks like a member of the performative lexicon is a corrupted attribution tag. That is the more economical explanation of the two, and this chapter cannot rule it out. Settling it needs the manuscript and print history of the formula, which would mean a critical edition of Śābara material, and no such edition exists.

So the phrase is named, counted, and left unparsed, and it is left out of the argument below entirely. Nothing in the rest of this chapter depends on it. Flagging a difficulty is not the same as declining to build on it, and the item is left out rather than flagged.

This is the sharpest open question in chapters 10 to 13. What would close it: the formula traced through dated printings and, if any can be found, manuscripts, with the two forms collated.

A saying that does

There is a precise philosophical vocabulary for a declaration of this kind, and using it is worth the small awkwardness of importing it.

Ordinary sentences describe. They can be true or false, and you check them against the world. A second class of sentences does not describe anything: it performs an act, and the act is complete in the saying. A person who says the words of a promise has thereby promised. A person who says the words of a naming has thereby named. There is nothing to fact-check, because nothing was reported. Something was done. The English philosopher J. L. Austin gave this class of utterance its standard name, and performative is his word.

That frame comes from outside this treatise’s evidence base entirely. No source in the dossier supplies it, no page of Austin has been consulted, and nothing here should be read as reporting a claim of his. The vocabulary is borrowed because the alternative is to keep calling this material “magic words,” which explains nothing and never has. The borrowing is visible on purpose.

With that in hand, what the counting turned up becomes describable without any recourse to structure. A book whose commonest recurring nouns are a word for speech and a word for authority, and whose commonest recurring verb form means let it come true, is not a book of reports. Nothing in that lexicon is checkable against the world. It is a lexicon for saying that intends to do.

And on Austin’s own terms it has a problem at exactly one point.

The institution that is not there

A performative works because something outside the speaker backs it. That is the awkward and interesting part of the whole category. The words of a marriage are effective in the mouth of a registrar and empty in the mouth of a passer-by, and the difference is not in the words. The umpire’s call stands because there is a game with rules and a person appointed to make calls inside it. Strip the institution away and the identical sentence produces noise.

So: what institution backs a book like this one?

There is no priesthood of it. There is no ordination, no register of qualified speakers, no examining body, no seat. Chapter 6 has already shown how late and how thin the Nāth claim to this material is.

Eight compilations later that absence is not an oddity of one paperback. Chapter 10 read the contents pages of all eight and found five different architectures across six houses, organised by named purpose and by named affliction, by ailment under a medical frame, by a ten-act scheme in one case, and by nothing at all beyond the promise in another. Thirty per cent of the entries in the largest of them are titled by a complaint. None of the eight is organised by qualification, by lineage, by grade of practitioner, or by anything that would tell a reader who is entitled to speak. The credential apparatus is not merely unmentioned in the text. It is absent from the table of contents, which is where a book declares what kind of thing it thinks it is.

The answer is in the names.

The book supplies its missing institution by naming one. Chapter 10 prints the roster of who is named in it, and the shape of the roster is the finding: Gorakhnāth far ahead of everybody, Hanumān next, then Īśvar, then the goddess under her several names, then Bhairõ and Mahādev, and after them a long thin tail running through the Yoginīs, the Nau Nāth and Caurāsī Siddh, Kāmākṣā, Rāmcandra, Lonā Camārī, Kālkā, Mansā, Viṣahārī, Mohammadā, Khwāja Khiḍr and the Prophets. Nearly every occurrence of the appeal word in the book stands with one of those names.

Both of those findings carry chapter 10’s caution, and it travels with them here, because a number that leaves its own chapter without its scope is a number a reader cannot weigh. The roster is one reader’s hand-check of a bad scan: sixteen name-families were folded, the folding rule was visual similarity plus contextual identity, it was not applied blind, and it needs another reader and a better scan. The tally of appeal words standing with a name is the same reader’s hand-check of the same scan, and it is approximate on purpose. What is not in doubt is the ordering at the top of the list and the size of the gap between the first name and the sixth.

Those are counts of what the book reaches for. The finding is a distribution, and what it shows is that the speaker very often does not hold the standing he is speaking with. He names somebody who does.

The same shape holds across the corpus, and it holds at higher densities than in this book. Chapter 10 counts the appeal frame mechanically in all eight compilations and finds it in every one of them, against nought in both of the occult books from the same shelf that are not Śābara compilations. It also counts the distinct items standing in the authority slot of that frame, and finds several times as many in the largest compilation as in this one. Those are counts of forms rather than of beings, they are upper bounds by construction, and no form behind any of them is named in this treatise. What they show is that the slot is not a small closed list anywhere, and that it grows with the size of the book.

Ājñā is the obvious word to reach for at this point, in order to ask whose command is being cited in a book with no office behind it. The section above takes it out of the corpus argument. The question survives without it, because the borrowing is visible in the frame count and in the roster, and it does not need a noun meaning command to be visible.

The standing is usually borrowed. Usually, and not always: first-person forms of the appeal word stand ten times in this book, which is a speaker staking something of his own alongside what he has borrowed. That is a distributional fact and it is as far as it goes here. The construction is not printed, in this chapter or any other. An absolute would be an absolute the counts cannot support, and the qualification is what the counts do support.

What stands where the institution would be

If no institution backs this speech, something still has to carry it from one mouth to the next, and one of the eight compilations shows what, in a way no reader would predict from the other seven.

Śābar Mantra-Saṅgrah, published by Kalyāṇ Mandir Prakāśan on Alopī-Devī Mārg at Prayāg, runs to twelve numbered parts. Its founding editor is named on the page as “Kula-bhūṣaṇ” Paṇḍit Ramādatta Śukla, M.A., of Prayāg, and its editor as Ṛtaśīl Śarmā. The fourth edition of parts 1 and 7 is dated to the Gupta Navarātra of Saṃvat 2069, which is 20 June 2012, and they cost ₹30 and ₹50.

Its contents pages do not list mantras. They list people.

Each numbered section of each volume is credited to a named correspondent with a home town, and the towns are the point: Jamnagar and Junagadh in Gujarat, Muzaffarnagar in Uttar Pradesh, Begusarai and Buxar in Bihar, Hamirpur in Himachal Pradesh, Ratangarh in Churu district and Didwana in Rajasthan, Hadapsar in Pune, Sehore in Madhya Pradesh. At least seven states. Each section is titled by what that person sent in: three applications they had tried themselves, material they had obtained from sādhus and ojhās, an item in Punjabi, a set described simply as rare. Only inside a contributor’s section does the material break down by purpose or by deity. Part 1 opens with editorial matter, including an essay by the founding editor on the science of Śābara mantras and another on the role of the guru.

Contributor names are not printed in this treatise. The towns are, because the geography is the finding.

That is not an anthology. It is a periodical bound into book covers: a reader-submission magazine running for twelve numbered issues, edited in Allahabad, filled by practitioners across north and west India who wrote in and were credited by name and by place.

Which answers, for one house at least, the question this chapter has been asking sideways throughout.

The material does not come down through a lineage that certifies its holders. It comes in through the post. The editor’s function is not to qualify anybody. He collects, he prints, he credits, and the authority attaching to any given item is the authority of a named man in a named town who says he tried it. That is a reputational economy rather than an institutional one, and it is exactly the economy a book of borrowed authority would be expected to have: nobody is ordained, everybody is cited.

It also cuts against a habit of description. To speak of “the compiler” and “the compiler’s choices,” as though every one of these books were assembled by one person exercising judgement over a stock, does not fit three of the eight witnesses counted here. There is no such person there. There is an editor with a postbag.

This is one publishing house, and generalising from it would be the error the whole eight-witness comparison exists to prevent. What can be said is that at least one substantial part of the printed corpus reproduces by reader submission, that this is visible on the contents pages rather than inferred, and that any account of the corpus as a closed bazaar-press product manufactured in a shop has to reckon with it.

The poet uses the spell-word

Now go back to the beginning of this treatise, to the four lines everybody quotes, and read the couplet that closes the block they sit in.

Chapter 1 established what that passage is. The Śābara verse stands in the maṅgalācaraṇ of the Bālkāṇḍ, where Tulsīdās speaks in his own voice about his own undertaking, and it is there as a precedent. The gods made potent speech out of ill-matched vernacular syllables. Therefore a poem in Avadhi can be potent too. The most-cited verse in the popular literature on this subject is a poet’s argument for his own book.

Here is how he seals the argument.

सपनेहुँ साचेहुँ मोहि पर जौं हर गौरि पसाउ। तौ फुर होउ जो कहेउँ सब भाषा भनिति प्रभाउ॥१५॥

sapanehũ sācehũ mohi para jauṁ hara gauri pasāu, tau phura hou jo kaheũ saba bhāṣā bhaniti prabhāu If in dream and in waking Hara and Gaurī are gracious to me, then let all I have said of the power of a vernacular composition come true.

Rāmcaritmānas, Bālkāṇḍ, dohā 15 #

“If Hara and Gaurī are gracious to me, in dream and in waking, then may everything I have said come true, the power of a bhāṣā bhaniti, a vernacular composition.”

Editions differ at this couplet and the differences should be on the page rather than in a note. The line above is the Gita Press line of text, carried from chapter 1. The proofread Hindī Wikisource transcription of a printed edition with ṭīkā reads साँचेहु for साचेहुँ, placing the candrabindu on the first syllable rather than the last, and कहउँ for कहेउँ. Neither variant touches फुर होउ, which is the word this section rests on.

The couplet is a prayer for grace and a wish about a poem. Tulsīdās asks the favour of Hara and Gaurī, Śiva and Pārvatī together, the same pair he had named four lines earlier as the makers of the mantras, and on that favour he stakes the truth of what he has written in the vernacular. He is not asking for readers or for fame. He is asking that his book turn out to be true.

And of all the verbs available to him for that wish, the one he takes is phur.

तौ फुर होउ. Tau phura hou. Then let it come true.

That identification is the load-bearing philological claim of this chapter, and the standard Hindī dictionary settles it rather than this treatise. Dāsa’s Śabdasāgara lemmatises Tulsī’s फुर under हिं० फुरना, the same head as the modern verb, and derives the family from Sanskrit स्फुरण. The sixteenth-century Avadhi form and the form printed in a 2016 Delhi paperback are one lexeme by the judgement of the dictionary that Śyāmasundara Dāsa built.

There is a third witness, and it is the same book that raised the doubt about the fourth item. Mānas-pīyūṣa, in the same place, explains why these mantras bear fruit when four ordinary conditions for a mantra’s efficacy are unmet, and says that a Śābara mantra takes effect instantly. The verb the commentator reaches for is फुरता.

The poet, the commentator and the compilation all use the same word.

The identification stands on Dāsa, and it is not a decorative parallel. A man with no institutional standing to write scripture in Avadhi wished, in the vocabulary of charm, that his own poem come true, and named the divine pair on whose grace the wish depended. Then he wrote the book, and the book did come true, in the only sense that matters to a book.

And a commentator got there first.

Mānas-pīyūṣa, in the ṭippaṇī on this same dohā 15, says it outright. The word phur stands in the śābara mantras, and that, the commentator holds, is why Tulsīdās gave the word phur here: he has already likened his own composition to their ill-matched letters, and he carries the same point through to the seal. The commentator then draws the parallel this section draws, setting pragaṭa prabhāu maheśa pratāpa, the power is manifest, the glory of Maheśa, against jo kaheũ saba bhāṣā bhaniti prabhāu, the power of a vernacular composition. The same word prabhāu stands in both, and the second takes its case from the first.

So the identification is not this treatise’s. The observation belongs to the Hindī commentarial tradition, and it has been in print since 1968, in the notes on the very couplet this chapter is working from, in a volume this chapter cites three times. What this chapter can say is that the poet’s seal is in the vocabulary of a charm, that the commentators saw it first, and that a reader who wants the argument in Hindī can go to that ṭippaṇī and find it.

The typology nobody has run

Charm scholarship has done this work for other traditions, and it has a vocabulary this treatise can legitimately borrow.

Jonathan Roper’s “Typologising English Charms,” in Charms and Charming in Europe (Palgrave Macmillan, London, 2004), surveys how the English material has been catalogued, and one of the schemes it sets out sorts by kind rather than by content: commands, comparisons, requests, three different speech-acts doing different work. Roper’s own caution should travel with the scheme: narrative charms resist it, and he says they call for a less logical system.

What the axis does not do is license anyone to place this material on it by counting words, and this chapter does not attempt it.

The methodological point is what carries forward. European charm scholarship built a typology and ran its corpus through it, on properly edited texts. The Śābara material has never been run through any typology, and cannot be until such texts exist. It has been described by what it supposedly does, by whom it supposedly comes from, and by how supposedly ancient it is, and almost never by what kind of words it is made of. Doing that much for one printed book is a contribution this treatise can actually make, on evidence it can actually reach.

The register is the sociology

Which brings the chapter to the thing the counting was really for, and to the point where it has to be most careful.

Set the vocabulary of this book beside the apparatus chapter 8 laid out. There, a mantra is made effective by procedure: the ten saṃskāra, the initiation, the qualified teacher, the sealed transmission, the whole graded architecture by which a syllable is made fit to be used and a person is made fit to use it. That architecture is a credential system. It answers the question “why do these words work?” with an answer about the speaker’s standing and the syllable’s processing.

Here, the answer is different in kind. The words work because the word is true, because it is wished into effect, and because somebody with power has been named.

Consider who talks that way.

A brahmin with initiation does not need to insist that his word is true. His position asserts it before he opens his mouth. He has no use for a formula that certifies his speech, because his speech is already certified by the fact that it is his.

Insistence is what you produce when you have no such backing. Truth-assertion, in a mouth without credentials, is a substitute for credentials. And borrowed authority is what a person invokes when the authority is not his to hold.

That is a sociological reading of a word list, and it is inference rather than evidence. Its width is what matters. The reading does not rest on one book. Stated at the width the material supports: across eight independently compiled books from six houses, the vocabulary of qualification is absent from all of them, the vocabulary of assertion and appeal is dense in all of them at rates one to two orders of magnitude above two occult books from the same shelf that are not Śābara compilations, and one house’s contents pages show the material arriving by post from named practitioners in seven states. The population these eight are drawn from is still unbounded, and a ninth book could still be unlike all of them. The inference is an inference. It is simply no longer an inference from a single instance.

Two theories of why words work, running side by side for centuries in the same country. One credentialist. One assertive.

The credentialist theory wrote the treatises. What the assertive one wrote, we now have eight printed samples of, and they agree.

The next chapter asks about the names, including one set that the seven new witnesses show to be larger than a single compilation suggests.

The word claims to be true.

Someone else has to be named before it can be.