पाठभेद

By ear

Most of these texts were learnt by listening long before anyone printed them, and each family keeps the version its elders sang. Śubhaṃ Karoti is the clearest case: the printings we compared all differ, one asking that the lamp burn till midnight, दिवा जळो मध्यान्हरात, another all night, सारी रात. Neither is a mistake. Both are how families say it.

• शुभं करोतिShubham Karoti

By copying

Websites copy one another, and an error copied once is copied everywhere. In the Mahālakṣmī āratī, a 2007 Webdunia page reads कुश्चित, which is not a Marathi word, and drops तद्रूप from the fourth stanza, leaving the line a beat short. Drik Panchang, Brahmabhakti and several other pages print the same text, errors included. Printings that do not descend from it read कुत्सित, ‘wretched’, and keep तद्रूप. Five pages that agree because they copied one another are one witness, not five.

• महालक्ष्मीची आरतीMahalakshmi Aarti

By the press

Spellings drift too. सहस्त्र for सहस्र, ‘a thousand’, appears in many printings, including several we compared. Even a careful source can be uneven: the version of the Śrī Sūkta at the head of its Sanskrit Wikisource page has many typing errors, while a second version on the same page is clean.

Different traditions

Some differences are not errors at all but separate traditions. The Śrī Sūkta is recited one way in Maharashtra and another in the south: ईश्वरीं here, ईश्वरीग्ं there, which writes the nasal as southern reciters sound it. The text edited by Isidor Scheftelowitz in 1906, from a single birch-bark manuscript from Kashmir, differs again, and arranges verses 13 and 14 the other way round.

• श्रीसूक्तShri Sukta

Different lengths

Texts differ, too, in how much of them is said. Homes chant twenty-one verses of the Mahiṣāsuramardinī Stotra; sanskritdocuments.org gives thirty and numbers the familiar ones 7 to 27. Printed Śrī Sūktas run on after the sixteenth verse with verses that differ from book to book.

• महिषासुरमर्दिनीMahishasura Mardini Stotra

An old problem

None of this is new. In the late sixteenth century, about three hundred years after it was written, Eknāth said he had restored the Dnyāneshwarī to its original form, and warned that ‘neither change nor addition should be permitted’.

• पसायदानPasaydan

How we choose

For every text we compare several printings and editions, and we count independent witnesses, not copies. Where they disagree we prefer the reading that is a real word, makes sense in its line and fits the metre; where traditions differ, we follow the one recited in Maharashtrian homes. We do not hide the alternatives: the note beside each verse records them, and every page lists what it was checked against.

If a reading here is not what your family says, it may not be wrong. We would like to know.