Upamā or rūpaka, and an avagraha that changes the rasa
In a 60-mark oral revision session at a government school, the teacher writes two lines on the board - mukhaṃ te kamalaprakhyaṃ netre nīlotpalopame | vāṇī te 'mṛtavat snigdhā kīrtiś candravad ujjvalā - and asks 54 students to name the alaṃkāra and then recite with correct sandhi. Forty-one students answer rūpaka, because the face and the lotus appear together; nine answer upamā; four answer atiśayokti. During recitation, 33 students drop the avagraha in te 'mṛtavat and chant te mṛtavat, so the line praises the speech as 'like a corpse' instead of 'like nectar'. The class laughs, and the student insists the meaning is the same either way. The written test awards 1 mark for naming the alaṃkāra and 2 marks for a one-sentence justification, and the teacher must also decide whether a recitation error should cost marks in a paper that is otherwise written.
1. Decide between rūpaka, upamā and atiśayokti for the verse, and state the single diagnostic that settles it rather than appealing to how the verse 'feels'.
Read the worked answer
Upamā, four times over, and the diagnostic is grammatical: look for a comparative marker. The verse supplies three different kinds - the suffix -prakhya (kamalaprakhya), the adjective -upama (nīlotpalopame, dual, agreeing with netre), and the suffix -vat (amṛtavat, candravat). Each comparison also names a shared property, which is the fourth element upamā requires: the speech is snigdha like nectar, the fame is ujjvalā like the moon.
Rūpaka would require the identification to be asserted without a comparative word - mukham eva kamalam, 'the face is a lotus'. Atiśayokti would require an impossible hyperbole; nothing here is literally impossible. The 41 students who wrote rūpaka reacted to the juxtaposition of face and lotus, but juxtaposition is not identification.
2. A student chants te mṛtavat and claims the meaning is unchanged. Explain what the avagraha is doing and exactly what has changed.
Read the worked answer
te 'mṛtavat is te + amṛtavat. The final -e of te and the initial a- of amṛtavat coalesce into -e, and the lost vowel is marked by the avagraha. Drop that mark and the reader has mṛtavat, 'like a dead thing', so the line says the speech is corpse-like - the exact opposite of the intended praise, with a different rasa altogether. The class laughter was the correct literary response to the wrong reading.
The avagraha is therefore not a decorative apostrophe: it signals that a vowel has been elided, so it must not be re-pronounced - but it must not be forgotten either.
Compare kīrtiś candravat, where some students restored a full visarga. That changes nothing semantically, since visarga before a consonant becomes ś but a reciter may pause; it is a recitation slip. The avagraha slip changes the meaning and must be corrected on the spot.
3. The paper is written; there is no oral component. Should the mṛtavat recitation error affect the score of a candidate who named the alaṃkāra correctly?
Read the worked answer
No - assess what the rubric names. If the question asks for the alaṃkāra and its marker, a candidate who wrote 'upamā, because of -prakhya, -upama and -vat' has answered it. Importing an oral error into a written mark is unfair and unrepeatable.
The exception is a question that explicitly asks for the sandhi split or the meaning of te 'mṛtavat; then the avagraha is part of the tested content and the mark is legitimately lost. Handle the recitation separately: two minutes of choral reading with the avagraha circled in red on the board fixes the habit faster than a deduction, and the class that laughed has already learned why the mark exists.