Malwai Punjabi: regional speech in words and recordings

The Malwai corpus

The Central Institute of Indian Languages’ Punjabi speech-corpus documentation places Malwai in a regional framework that includes Ludhiana, Moga, Sangrur, Barnala and other areas. It records collecting speech separately from Malwai, Doabi and Puadhi regions. This documentation project does not claim that every resident speaks one variety or that speech stops at district boundaries.[1]

What a recording can measure

Arora and Singh’s comparative study built a corpus of 125 isolated agricultural words spoken by one hundred participants. Sixty were classified as Malwi speakers, compared with twenty each for Majhi and Doabi. The Malwi group included thirty participants in each of the study’s male and female categories. This uneven regional sample matters when interpreting the comparison.[2]

The researchers recorded with a laptop microphone and a mobile phone and analysed pitch, intensity and vowel-related acoustic measurements using PRAAT. Their materials included tone-related words, gemination and nasalisation. These measurements describe selected utterances under stated recording conditions. They do not give one pitch or voice quality that identifies every Malwai speaker, and an isolated-word corpus cannot reproduce the full range of conversation.[2]

Written variation

A 2015 study by Arvinder Singh and Parminder Singh instead investigated written conversion between standard Punjabi and Malwai or Doabi forms. Its system combined dictionaries with rules for word structure. The authors gathered material from linguistic resources and informal written settings, illustrating that regional variation enters writing as well as speech.[3]

Their task was to identify and replace particular words or word parts, not to translate all possible Malwai conversation. The distinction helps explain why a written conversion table and a recording project answer different questions. Both document variation; neither makes a household’s ordinary way of speaking an error merely because it differs from a chosen written reference form.[3]