Two Syllables, Three Rasas: Finalising a Poetry Key Overnight
A half-yearly unit test for Class XII contains a 4-mark item on a prescribed Odia poem, marked 1 for scansion, 1 for rasa and 2 for alaṃkāra with justification. Forty scripts show three problems. First, the line set for scansion has 15 akṣaras; at akṣara 5 the vowel is a short 'a' and the next akṣara begins with the conjunct 'mbh'. Twenty-two students mark akṣara 5 as laghu, eighteen mark it guru. Second, asked for the prevailing rasa, nine students write vīra, citing the poet's 'lion-like' image of the speaker's resolve; the rest write karuṇa, citing the lament of the final line. Third, on the expression 'his face was a winter lotus', thirty students write rūpaka, six write upamā because a lotus implies a comparison, and four write 'both'. The head of department wants the key finalised tonight and refuses a re-test.
1. Which group is right about akṣara 5, and how should the 22 wrong answers be marked given that they are the majority?
Read the worked answer
The eighteen are right. The rule is not 'count the vowel length' but 'a short vowel followed by a conjunct in the following akṣara becomes guru'. Akṣara 5 is therefore guru, and the metre must be scanned on that basis.
Marking: the 22 take 0 on that sub-item. Being in the majority does not make a scansion error a legitimate alternative - there is one rule and it is checkable, unlike a rasa judgement. Awarding half marks to protect the majority would devalue the eighteen correct answers and would misrepresent a rule-governed skill as a matter of taste.
The majority error is diagnostic of teaching, not of marking. Sixty-five per cent of the class holds the same misconception, which means the rule was taught as 'long vowel = guru' without the conjunct clause. Record it as a class error, re-teach with three minimal pairs on the board (short vowel + single consonant; short vowel + conjunct; long vowel), and put a compensatory scansion item in the next cycle. That is fairer than retroactive leniency, and it leaves the key defensible if a script is shown to a parent.
2. How do you decide between vīra and karuṇa for the rasa mark, and what would you do differently if you were rewriting the item?
Read the worked answer
Rasa is carried by the dominant sthāyī bhāva of the passage, not by its most striking single image. The test is coverage: which reading explains more of the passage without contradiction? If the lament of the closing line is where the poet settles the emotional key, and the 'lion-like' phrase is one clause of self-assertion inside an otherwise elegiac movement, karuṇa is the key and the nine students have mistaken a subordinate image for the governing mood.
But note what has happened: the item as written asks only for a name. That makes a genuinely arguable question into a one-mark guess, and a one-mark item with two defensible answers is a defective item. The correct professional response, before results are published and not after, is either to double-key it or to void the sub-item, and to say so on the key.
If rewriting: make it 2 marks - 'Name the prevailing rasa and justify it with two expressions from the passage.' Vīra with two supporting expressions would then earn 1, karuṇa with the closing lament would earn 2, and a bare name with no support would earn 0. The item then tests judgement rather than luck.
3. Settle the alaṃkāra marking for the three groups (rūpaka, upamā, 'both').
Read the worked answer
The expression asserts identity - the face is a winter lotus - and contains no comparison word (like, as, matu, pari). That is rūpaka: the upameya (face) and upamāna (lotus) are fused, and the common property (pale, cold beauty) is left to be inferred.
- rūpaka with a correct justification (30 students) - 2/2. The justification must identify face as upameya, lotus as upamāna, and note the absence of a comparison word.
- upamā (6 students) - 1/2. They have correctly identified the two terms being compared, which is half the analysis, but the name is wrong: upamā requires four elements - upameya, upamāna, a shared property and a word of comparison. Credit the identification, refuse the label, and write the four-element definition on the script.
- 'both' (4 students) - 1/2 as well, unless the line genuinely contains two clauses, one predicating identity and one drawing an explicit comparison; in that case 2/2. A single clause cannot be both.
Consistency point: apply the same 'half for correct analysis, nothing for a wrong name with no analysis' rule elsewhere in the paper, so that the deduction looks like a standard rather than a mood.