Dialect in the Tamil classroom: error or register?
Selvi teaches Class 5 in a Tamil-medium school in a village near Madurai, with 36 children, most of whom speak the local dialect at home. In an oral retelling task, 29 of 36 retold a story fluently, using forms such as 'போறேன்', 'வாங்க', and 'அவளுக்குத் தெரியல'. Writing the same answers, the children used the same forms and lost marks. The unit test carried 8 of 10 marks for spelling and grammar and only 2 for comprehension. After the test, a girl who had given the best oral retelling in class wrote only two lines and stopped. Her mother, a daily-wage labourer, says the child 'cannot write good Tamil' and has asked about private tuition. Selvi's own training notes mark dialect forms as 'பிழை' beside 'வாங்க'. The head teacher wants the pass percentage up before the term exam, and Selvi is considering two weeks of drilling the written forms.
1. Are the children's forms errors or dialect variation? How would you decide whether the girl's two-line answer shows comprehension failure or a register problem?
Read the worked answer
Use three tests before deciding.
- Is the form shared and rule-governed? 'போறேன்', 'வாங்க', 'தெரியல' are systematic spoken-Tamil forms used across most of the class, not one child's invention. Shared, consistent forms belong to a variety, not to a mistake.
- Does Tamil have a spoken–written split? Yes, and it is well documented. The written register uses 'போகிறேன்', 'வாருங்கள்', 'தெரியவில்லை'. The children's speech is not a failed attempt at these; it is a complete system doing a different job.
- Collect three pieces of evidence for the girl: (a) an oral retelling in her own speech - she already did this best in the class; (b) the same content retold in writing, marked on content only; (c) one oral 'why' question answered in dialect.
If the content survives all three, comprehension is intact and only the register is missing. Her two lines after a strong retelling point to affect and task demand - her best performance was marked wrong - not to poor understanding. Conclusion: teach the written register as an additional code; stop marking dialect as 'பிழை'.
2. Redesign the 10-mark unit test so that comprehension and language accuracy are separately reported. State what each part measures.
Read the worked answer
The present test under-samples the aim: 8 of 10 marks go to form and label dialect speakers as weak in Tamil when their comprehension is fine.
Proposed split, same 10 marks, two separate scores.
- Comprehension (5): two short answers on events, one 'why' question needing inference, one 'what would you have done' response. Mark content only; dialect forms accepted; spelling not penalised. Measures whether meaning was built.
- Language accuracy (5): three marks converting five taught spoken forms into written register; two marks on correct register in the child's own three sentences. Measures control of the code being taught.
- Oral component, recorded on a checklist, not folded into the written total: retelling and one inference answer. Measures oral comprehension.
Why: content and form are different constructs, so a single mixed score cannot be interpreted. The girl can now score well on comprehension and still get honest, specific feedback on register. Report both scores to parents; do not average them into one 'Tamil mark'.
3. Describe one classroom routine that teaches the written register without stigmatising home speech, and explain why it works.
Read the worked answer
Routine: 'two-voice' telling, five minutes, twice a week.
- A child tells the day's incident in home speech. The class listens for content; the teacher asks two 'why' questions.
- The teacher re-tells the same content on the board in the written register, without comment, keeping the child's content intact.
- In pairs, children 'translate' one sentence each of their own telling into the written form.
- A class chart collects pairs - 'வாங்க / வாருங்கள்' - as an interesting list of two codes, not a list of mistakes.
Why it works: the register is presented as an addition, not a correction of the child's family speech, so there is no identity threat. The contrast is explicit, which makes transfer deliberate rather than accidental. Pair work removes the exposure of one child being corrected before 35 others.
It also avoids the two-week drill plan, which would raise scores on memorised forms while leaving comprehension - the class's actual strength - untouched and unmeasured.