Classifying a marriage song from Abdasa
During fieldwork in Abdasa taluka, Kutch, you record a marriage song (lagna geet) sung by a 62-year-old woman. Your transcript shows implosive stops (ɗ, ɓ) that standard Gujarati lacks; the first-person pronoun hū̃ where standard Gujarati has hũ; retention of sibilants in several lexical items where standard Gujarati has h; and a quantity of ordinary Gujarati vocabulary mixed in. The singer says she learned the song from her mother, who came from a village twenty kilometres away, and that she 'speaks Kutchi at home but sings in Gujarati'. You are editing a section of 'desi' songs for a Gujarati literature anthology and must label the text and choose an orthography. A reviewer has already objected that calling it Gujarati is 'linguistically wrong'.
1. Which features in the transcript are diagnostic for classification, and which are contact effects? How would you classify the text?
Read the worked answer
Diagnostic features: the implosives and the pronoun hū̃ are phonological/lexical retentions shared with Sindhi and Kutchi and absent from standard Gujarati; the sibilant retention is a conservative feature against Gujarati's h-reflex. These are systemic, recurring features - the kind that define a variety.
Contact effects: the 'ordinary Gujarati vocabulary' is not diagnostic of the singer's base variety; it reflects Gujarati as the prestige and school language of the region, and would appear in any Kutch speaker's repertoire. Code-mixing in a sung genre is even less diagnostic, because song genres travel across dialect boundaries and are often composed in a koine.
Classification: the text is best labelled Kutchi, treated as a variety of the Sindhi group with heavy Gujarati contact, and not as a dialect of standard Gujarati on the strength of the mixed lexicon. The reviewer is right that 'Gujarati' alone is wrong, but 'not Gujarati at all' is also wrong: the anthology's category is literature composed in Gujarat, which is a geographic and literary category, not a linguistic one.
Method: classify on structure (phonology, pronouns, morphology), not on lexicon, and record the singer's own metalinguistic statement as evidence, not as a verdict.
2. What orthography and labelling would you use in the anthology, and how would you answer the reviewer?
Read the worked answer
Orthography: use Gujarati script, since the anthology's readers read Gujarati script, but mark the implosives and the long vowels with a declared diacritic convention (for example a subscript hook for implosives), and add a half-page note on the convention. Do not transcribe phonetically in Latin script; that would make the song unreadable to the anthology's actual users and would visually segregate it from the rest of the volume.
Labelling: 'Kutchi (Sindhi group), collected in Abdasa, Kutch; text in Gujarati script'. Put the linguistic note in the headnote, not the title.
Answer to the reviewer: agree with the linguistic point and correct the label, but distinguish two questions. (1) What language is the text? - Kutchi, by structural criteria. (2) What is the anthology's scope? - literature of Gujarat, which includes Kutchi, Kachhi Gujarati, and other regional varieties. Renaming the section 'Songs of Gujarat' rather than 'Gujarati songs' resolves the objection without excluding the text.
The reviewer's mistake is to treat a literary anthology's categories as linguistic classifications; your mistake would be to let a reviewer's category error push you into Latin-script exoticism.