Two Judges, One Verse: Utprekṣa versus Atiśayokti and a Scan That Will Not Settle
During a revision class for 38 Class 12 students, a four-mark question on a prescribed Telugu verse asked candidates to name and justify the figure of speech in the third line, where the poet says the moon, seeing the heroine's face, wrapped itself in a cloud out of shame. Twelve students wrote utprekṣa, fourteen wrote atiśayokti, and the rest wrote upamā. A second part of the same question asked for the metre. One student scanned fourteen syllables per pada and another scanned thirteen, because the second had split a sandhi in the printed line before counting; a third argued that a short vowel before a doubled consonant should still count as laghu. The teacher must publish one defensible key, award marks that do not punish an arguable minority reading, and show the class how a scan is actually performed, all in the 20 minutes left before the bell.
1. Which figure of speech should the key prefer, and how should the fourteen atiśayokti answers be marked?
Read the worked answer
Distinguish by the speech act, not by the presence of a comparison.
- Upamā compares openly: 'her face is like the moon' - a marker such as vale, la, pola, bhāti is present.
- Rūpaka identifies: 'her face is the moon'.
- Atiśayokti asserts an impossible superiority: her face defeats the moon, with no imagined scene.
- Utprekṣa imagines the impossible as if it happened: the moon, ashamed, wraps itself in a cloud.
The line gives the moon a motive and an action, so utprekṣa is the primary answer - provided the student quotes 'wrapped itself in a cloud out of shame' as the trigger word or phrase. Atiśayokti stays defensible, because the purpose of the image is to exalt the face, and this is a well-known overlap. Marking rule: full marks for label + quoted trigger + one-line reason, with either label accepted where the reason matches the label used. Penalise only a label whose stated reason actually describes a different alaṃkāra. On this evidence the fourteen are not wrong, and the upamā answers are wrong only because no comparison marker is present.
2. One student counts thirteen syllables, another fourteen. State the rule that decides the count and what the published key must show.
Read the worked answer
Two rules settle it.
1. Count the akṣaras of the printed line. Sandhi is a writing convention laid over sounds that are already there; a printed coalescence neither adds nor removes a syllable. Splitting a sandhi in order to count is exactly how a student reaches thirteen where the line has fourteen. The scan is performed on the printed pada.
2. Guru/laghu is a property of the vowel and what follows it: long vowel → guru; short vowel followed by a conjunct (saṃyukta) → guru; short vowel followed by anusvāra/sunna → guru; otherwise laghu. The third student's claim - that a short vowel before a doubled consonant stays laghu - is the wrong rule, and it is the likeliest source of the missing syllable.
The key must print the evidence, not just the name: the pada, its akṣara count, the guru/laghu string, the resulting gaṇa sequence, the yati position, and the prasa check (second letter identical across padas). A metre named with no scan behind it should earn nothing in future tests.
3. You have 20 minutes with the class. What goes on the board, in what order, and why those three items?
Read the worked answer
Three items, chosen because they transfer to any verse in the paper:
- One table, four rows - upamā / rūpaka / utprekṣa / atiśayokti, each with a one-line example and the trigger word that identifies it. Students copy it; it answers half of any appreciation question.
- One live scan - the disputed pada goes on the board, the class calls out the rule as each akṣara is marked guru or laghu, then the gaṇa pattern is read off and prasa checked. Slow, once. The point is the procedure, not the answer.
- Two-sentence exam rule - 'Quote the words that trigger the alaṃkāra, then name it and explain in one line. For chandassu, always show the gaṇa division.'
Close by showing why both labels were accepted on the disputed line, so the fourteen see that justification earns marks and guessing does not. Skip the remaining verses; a half-taught method is worse than a taught one.