Marking Meena's Hindi Composition: Dialect Transfer or Error?
Meena, Class 7, is one of 34 students in a Hindi-medium school in Khunti; 26 of them speak Santhali, Mundari or Kurukh at home. For the composition 'मेरा गाँव' she wrote about 120 words describing the mahua trees, her mother carrying gobar to the field, and the bus that comes only twice a day. Three of her sentences read: 'हमरा गाँव बहुत सुन्दर बा।', 'चारो ओर जंगल बा।', 'सुबह में हमर माई खेत जाले।' The teacher, who does not speak any local language, circled eleven forms, wrote 'अशुद्ध' in the margin, deducted 8 of 20 marks under 'भाषा', and told the class that children who speak dialect at home cannot write Hindi. Meena's next composition was 40 words long, with no dialogue and no local detail; she now says 'मुझे हिंदी नहीं आती' and asks to copy from a friend. The half-yearly examination is in six weeks and the paper will carry a 10-mark free-writing question.
1. Was the 8-mark deduction under 'भाषा' defensible? Analyse Meena's forms and the rubric.
Read the worked answer
The instinct to teach standard forms is right; the 8-mark deduction and the diagnosis behind it are not.
- The forms are rule-governed, not careless. 'बा' is the copula in Sadri and related Jharkhand varieties; 'हमर' marks possession; 'जाले' carries person and tense. These are systematic transfers from a language she speaks fluently - exactly what second-language writing looks like while the learner maps L1 grammar onto L2 words. Random error would look different.
- On content the piece is strong. 120 words, specific, sensory, time-ordered, with a point of view and no padding. Any writing rubric credits content, organisation and voice.
- The rubric did not separate the criteria. A single undivided 20 with 8 removed for language tells the child the whole text failed when only one dimension was weak. That is an information failure, not just a harshness failure.
- The statement to the class was a claim about capacity, not about this text, and it is not supported by eleven circled forms. Children read such statements as judgements about identity and family.
Defensible version: award content and organisation marks in full, record language accuracy as a separate line, choose two or three high-frequency items to teach ('है' for 'बा'; possessive 'मेरा/मेरे'; verb agreement). Standard Hindi is an additional school register, not a replacement for the home language.
2. Rewrite the marking and the written feedback you would give on this composition.
Read the worked answer
A workable split for a 20-mark composition: 10 for content and organisation, 6 for language accuracy, 4 for presentation and length. Meena scores close to full on the first and part on the second, which is a truthful and motivating profile.
Feedback rules that follow from the case:
- Respond to meaning first. One sentence naming what worked: 'The detail about the bus coming twice a day makes the village real - keep that.'
- Select two focus items, not eleven. 'In your next piece, use है in place of बा in at least five sentences. Keep everything else exactly as you wrote it.'
- Keep a personal error log with three columns: what I wrote / standard form / rule. Only high-frequency items enter it.
- Make the contrast visible without shaming. Put one sentence on the board and ask the class to say it in Sadri, in Hindi and in English. The home language is treated as a resource and the standard form becomes an object of study rather than a correction.
- Assess the standard forms again in the next piece only; do not re-correct old items.
Correction overload is precisely what produced the collapse to 40 words; the feedback design, not the child's attitude, controls whether she keeps taking risks on paper.
3. Meena's next composition is 40 words and she asks to copy. Interpret this, and state what it implies for assessment.
Read the worked answer
This is avoidance after punishment, not a loss of ability.
Reading the signs: she has removed exactly the features that were marked wrong - dialogue, local detail, expressive vocabulary - and kept only short neutral sentences. That is a rational response to feedback: if expression is penalised, stop expressing. The self-description 'मुझे हिंदी नहीं आती' and the request to copy are evidence of reduced self-efficacy and increased risk aversion, both predictable consequences of a public capacity judgement.
Assessment implications:
- A 40-word sample cannot measure composition ability. Any conclusion drawn from it ('she cannot write') is a measurement artefact and would set off the very loop the head teacher is worried about.
- The teacher should restore a low-stakes sample: a topic Meena chooses, oral rehearsal before writing, permission to keep one local word in brackets, and response to content before language.
- Keep one language focus per piece, so success is achievable and visible.
- Watch for the same withdrawal in other children who were present when the class was told that dialect speakers cannot write Hindi.
The diagnostic question to hold on to: is she less able, or has the classroom made writing feel unsafe? The evidence supports the second.