One label, two questions — and one of them has never been measured
The argument is unresolvable as stated because two questions are being answered at once. The first is structural distance — what Kloss called Abstand, "a language variety or cluster of varieties with significant linguistic distance from all others". The second is social elaboration — Ausbau, "a standard variety, possibly with related dependent varieties" 28. Kloss's 1967 article was not opened for this page; the definitions here come from a Tier-C encyclopaedia entry, and nothing below turns on any finer reading of his framework than those two sentences will bear 2928.
An earlier version of this page answered the Abstand question by counting authorities, and that was the same category shift it convicts the claim of. Abstand is a measurement — feature distance, lexicostatistics, mutual intelligibility — and Kloss devised it precisely to be capable of overriding inherited taxonomic labels. What the authorities supply is inherited taxonomy. Grierson's nesting of Dakhinī under Urdū 2, Glottolog's level field 22, the encyclopaedia infobox 30 and Casadei's opening classification 19 agree with one another and descend from one another. No source opened for this page reports an intelligibility test between Dakhni and Standard Urdu, a feature-distance metric, or a lexicostatistical count. On the Abstand question the honest report is that it has not been measured here, and that the unanimity is an inheritance rather than a result.
On the Ausbau question the answer is historical, and the tense does the work. Dakhni was an elaborated court and literary language in the Deccan before Delhi's Rekhta poetry existed 101219. It is not one now, and the record names the mechanism: under the Nizams Urdu replaced Persian in the courts and administration, but "the Urdu chosen as the language of the state was the northern variant spoken in Delhi", and in 1919 the foundation of Osmania University with Northern-Urdu as its medium "led to the further segregation of Dakhni to the non-official sphere, limiting or even eradicating its use in the official public domain" 19. Under Kloss's criterion elaboration is what makes a variety a language; de-standardisation by conquest and administrative substitution does not retroactively unmake the elaboration that occurred. The Ausbau answer is therefore yes-then-no. Stating only the present tense — "it is not an elaborated standard now" — reports the demotion as though it settled the question, and it does not.
There is a third question underneath both, and it is the one the popular claim answers correctly: is the Deccan's literature older than standardised northern Urdu? It is, by centuries 1011. The claim takes that chronological answer and reports it as a taxonomic one — but this page's own taxonomic answer has to be stated as an inference, not a finding.
What the reference works actually record
The classification most often cited against Dakhni is Grierson's. Volume IX Part I of the Linguistic Survey of India — title page dated 1916, in the volume titled Specimens of Western Hindī and Pañjābī 1 — states on its first page: "Urdū, itself, has two varieties, the standard literary form of Delhi and Lucknow, and the Dakhinī, spoken, and used as a literary medium, by Musalmāns of Southern India" 2. The nesting is explicit on the same page, where the returns are broken out as "Hindōstānī— Local Vernacular 5,282,733 / Literary Hindōstānī (including Urdū and Hindī) 7,696,264 / Dakhinī 3,654,172" 2. The section devoted to it is headed "DAKHINĪ HINDŌSTĀNĪ OR MUSALMĀNĪ" 3, and the specimen text is headed "INDO-ARYAN FAMILY. CENTRAL GROUP. WESTERN HINDĪ. HINDŌSTĀNĪ (DAKHINĪ OF BOMBAY DECCAN)" 6.
Anyone re-checking these pages should note that the title page reads 1916 while some library metadata attached to the scan gives 1907 1; the earlier date belongs to other volumes and to the Imperial Gazetteer.
The nineteenth-century surveys say the same thing in weaker terms. Cust in 1878: "Taking the Hindustáni of Dehli as the standard of purity, we have in Southern India another variation in the so-called Dekhani, and it is admitted, that there are points of difference in it from the Northern Lingua-franca" 8. The Imperial Gazetteer of 1907 adds the archaism observation and, notably, the Deccan's priority: "and here Urdū literature took its rise. 'Dakhinī Hindostānī,' as it is called, differs somewhat from the modern standard of Delhi and Lucknow, and retains several archaic features which have disappeared in the north" 7.
The modern catalogue agrees. Glottolog files dakh1244, "Dakhini (Urdu)", with the explicit level label Dialect, under Urdu, with no ISO 639-3 code shown 22. What none of these four records contains is an argument, a measurement, or a bibliography addressed to the question. They are consistent with each other; consistency among descendants of a single nineteenth-century classification is not independent corroboration.
Grierson's argument runs against the "corrupt Urdu" framing — and against his own chronology
The page that classifies Dakhni as a variety of Urdu is also the page that refuses the hierarchy usually smuggled in with that classification. "It is commonly said that Dakhinī is a corrupt Hindōstānī. Historically speaking it would almost be truer to say that literary Hindōstānī is a corrupt Dakhinī, for we have seen that Hindōstānī literature took its origin in the Deccan. Without, however, going so far as that, it is not correct to look upon Dakhinī as a corrupt form of speech" 3.
The reason given is chronological, not sentimental: "The Musalmān armies carried vernacular Hindōstānī to the south before it had been standardised, and it then contained many idioms which are now excluded from literary prose. Some of these have survived in the Deccan. Such are, for instance, the employment of the oblique genitive as a base for declension in phrases like mērē-kū, to me, where the standard would have mujh-kō" 3.
That passage is the single most useful thing to put in front of a reader who believes the colonial survey settled the hierarchy. But Grierson's chronology, on the same page, is wrong by roughly 270 years, and an earlier version of this page suppressed the fact. Immediately before the anti-"corrupt" passage he writes: "It was in the Deccan that Hindōstānī, under the form of Urdū, first received cultivation, and it was at the hands of Walī of Aurangabad (cir. 1700), the 'Father of Rēkhta,' that a standard of literary form was given to it" 3. That date post-dates the entire Qutb Shahi and Adil Shahi corpus — Fakhr-e Dīn Nizāmī at fl.1434, Muhammad Qulī Quṭb Shāh, Wajhī 12. Grierson refuses the derivation hierarchy and then dates Deccan literature from the one Deccan poet whom Delhi's tazkirah tradition had canonised, which is the hierarchy operating in another register.
So Grierson is a good witness against "corrupt Urdu" and a bad witness on when Deccan literature begins. The accurate formulation is not his 1916 sentence but Faruqi's, which puts Gujrī, Dakanī and Rekhtah on a level as regional names of one language 17.
The differences, stated in full
Conceding a shared core does not require minimising the divergences, which are substantial. One scoping note first, because it changes how far the inventory generalises. Casadei's descriptive section states its object explicitly: "some salient syntax features of Dakhni as it is spoken in Rayalaseema, in Andhra Pradesh" 19. Following Mustafa (2008), she records that Andhra Dakhni splits between the Rayalaseema variety, "which has features imported from Telugu", and the Telangana variety, "called Hyderabadi Urdu, which shows more resemblance to the Standard Urdu" 19. Everything below is drawn from the most Telugu-converged sub-variety described in the article, and should be read as such.
On phonology: "Although it generally shares with Standard Urdu (SU) the same sound system, it presents the aspirates /mh/, /lh/, /rh/, /vh/ and /yh/ and the nasal and lateral retroflex phonemes, /ṇ/ and /ḷ/" 19.
On pronouns, the system itself is restructured: "Rather than the three-level pronominal system present in Hindi and SU (tū, tum, āp), Dakhni instead presents a two-level hierarchy: tu... and tūme... The honorific form āp is not used in Dakhni" 19. There is a clusivity contrast — exclusive ham logāñ against inclusive apan/apal logāñ — which Hindi and Standard Urdu do not have 19.
On case, the genitive has been eaten by the dative: "Arora and Subbarao (1993) observed that the Dakhni language went through a process of 'dativisation', losing the genitive structures that are common in both Hindi and SU" 19, as in kā ku jāre? 'Where are you going?' 19. (The reference is to Arora and Subbarao; Casadei's own bibliography carries only their 1988 paper, not a 1993 one 20.) On the verb there are categories with no Hindi-Urdu equivalent: a "habitual present-cum-future (maiñ jātũ)", a permissive-imperative in -ñde, and a relative adjectival participle in -so (maiñ bolya so bāt, suno!) 19.
A second description points the same way, with a heavy caveat about its standing. Reporting Vashini Sharma's chapter, Ferro lists features that make Dakhinī "resemble Telugu more than Urdū/Hindī": "the use of the ergative particle ne does not follow the rules of Standard Hindī"; "Dakhinī is 'degenitivising' for a dativisation"; "the use of short versions of emphatic particles, such as hī, bhī, to, which are -ī, -īṁ in Dakhinī"; agglutination of the auxiliary (rahā + hai > rā / ray / raiṁ); honā for cāhie; the Marathi forms हौ (which Ferro transliterates hai) and नको nako; and the quotative bolke / karke "calqued from the Telugu form -ani" 21. This is second-hand, and worse than second-hand: Ferro states that Sharma's chapter "is unfortunately just a draft, since Sharma passed away before the publication of the volume" and that "the conclusions of this paper are absent" 21. An unfinished posthumous draft, reported in a review, does not converge with anything; it points the same way, which is a weaker thing.
One widely repeated feature was wrongly dismissed in an earlier version of this page and is restored here. The emphatic -ič is not attested in either academic article opened for this page, but it is not unattributed: an encyclopaedia entry records "suffix -ich [it͡ʃʰ] (mai idharich hũ)" against standard hī, as an "Emphasis marker", with the note that "Matthews comments that this is 'probably from Marathi'" — referring to Matthews's 1976 London dissertation on Dakhani language and literature 3020. Ferro/Sharma's short emphatics -ī/-īṁ are a different item, about the shortening of hī, bhī and to; they neither exclude nor compete with a separate borrowed clitic, and the earlier version's "instead" was a non-sequitur. A borrowed emphatic clitic displacing inherited hī is core-grammar borrowing of exactly the kind that bears on structural distance. It is recorded here at Tier C, and confirming it means opening Matthews 1976, which was not done.
That correction generalises into a warning about this page's own method. Verifying only what could be opened produces false negatives on every feature whose primary literature was not reached, and the verified-only method is doing heavy work in the verdict below. "Not in the sources opened here" is not "not in the literature", and the earlier version reported one as the other.
Grierson's 1916 specimen headnote catches part of the same picture, with his own hedge attached. The specimen was "prepared in the office of the Collector of Customs of Bombay" and "is said to illustrate the dialect of the 'Musalmāns of the Deccan'" — an office translation exercise, not field data 5. The headnote reads: "Typical of Dakhinī are the use of jidhar to mean 'when' and 'where' instead of 'whither;' wŭ for woh, he; kōĭ-nē for kisī-nē; and athā for thā, he was" 5.
Convergence, and what it does to the shared-architecture claim
An earlier version of this page said the differences "sit inside a shared postpositional, participial and lexical architecture". That sentence is falsified by the sources cited two paragraphs above it, and it is withdrawn. The genitive postpositions are being lost to dativisation; the pronominal system is reduced from three honorific levels to two and gains a clusivity contrast; the finite verb acquires categories with no Hindi-Urdu equivalent; the reflexive possessive apnā is reported replaced by tumhārā; the conjunctive participle is joined by a Telugu-calqued quotative 1921. Those are the postpositional, pronominal and participial systems — precisely the domains named as the shared core.
What survives as genuinely shared is narrower and should be stated at its real size: the Perso-Arabic script, the Indo-Aryan core lexicon, the postpositional rather than case-inflecting typology, and enough shared morphology that Delhi could read a Deccan dīvān in 1720 without a translator 14. That is real. It is less than "one architecture".
The frame the descriptive literature itself uses is convergence, and this page never used the word. Casadei: "The special linguistic situation in South India favored the bidirectional convergence of Dakhni towards Dravidian languages, e.g., Telugu" 19. Her bibliography is a convergence bibliography: Arora and Subbarao, "On extreme convergence: the case of Dakkhni Hindi-Urdu" (1988); Pray, "Evidence of grammatical convergence in Dakhni Urdu and Telugu" (1980); Mustafa's 1985 Aligarh dissertation and his 2008 encyclopaedia article on Dakkhni 20. None of these was opened for this page; the bibliographic details are transcribed from Casadei's reference list, which was read 20. The earlier version cited Arora and Subbarao once, for dativisation, and left the title of their paper out.
Stated at full strength, the case against the classification is this. Dakhni is one of the standard South Asian examples of contact-induced structural convergence; the affected domains include core grammar rather than lexicon and colour; and no classifying source opened for this page has weighed that convergence against its own classification. Glottolog's Dialect label carries no supporting comment 22. Grierson wrote before the convergence literature existed. Casadei describes the convergence and files Dakhni as a variety of Urdu in the same article without reconciling the two 19. A classification that has never been tested against the strongest evidence available against it is a convention in good standing, not a finding — and that is the reason this page's verdict is contested rather than settled.
The contact literature also disagrees with itself about which neighbour did the work, and nothing opened here resolves it. Casadei says the borrowed words and grammatical features came "from Marathi and, to a smaller extent, local Dravidian languages, e.g., Kannada and Telugu" 19. Ferro, reporting Sharma, gives features that "make Dakhinī resemble Telugu more than Urdū/Hindī" 21. Matthews, at second hand, "states that Dravidian influence was much less" than Marathi 30. Sharma's is the single most damaging formulation on this page for the position it argues, and it deserves an answer rather than a listing. The answer available from here is partial and should not be mistaken for a refutation: the claim is about a set of syntactic features, not about the script, the noun morphology or the core lexicon; it is asserted in a review of an unfinished posthumous draft whose conclusions are missing 21; and the varieties in question are the Telugu-converged ones, which Casadei and Mustafa distinguish from the Telangana variety that "shows more resemblance to the Standard Urdu" 19. Those qualifications reduce the claim's reach. They do not dispose of it, and no source opened here does.
The chronology, where the claim is strongest
The load-bearing evidence for the claim's historical half is Faruqi's account of the names: "Early names for the language now called Urdu were Hindvī, Hindī, Dihlavī, Gujrī, Dakanī, and Rekhtah, more or less in that order, though until about the middle of the nineteenth century Dakanī continued to be the name for the form of the language used in the Deccan" 10. On Faruqi's account Dakanī is not the name of a different language; it is an older name of the same one. That account is itself a position in a live dispute, which the transmission section below states rather than hides.
"'Urdu' as a name for the language seems to occur for the first time around 1780" 11. Before that the term meant something else: "The name 'Urdu' seems to have begun its life as zabān-e urdū-e mu'allā-e Shāhjahānābād (the language of the exalted City / Court of Shāhjahānābād, that is, Delhi). It originally seems to have signified Persian and not Urdu" 11.
Against that date, the Deccan corpus is very old. Fakhr-e Dīn Nizāmī is dated "fl.1434", writing "in the Deccan" 12. His surviving work is the masnavī Kadam Rāo Padam Rāo, edited by Jamīl Jālibī; neither the text nor Jālibī's edition was opened for this page, and no date for it is asserted here beyond Faruqi's "fl.1434" 42. Wajhī's Sab Ras was edited by Maulvi Abdul Haq and published by the Anjuman-i Taraqqī-yi Urdū in 1932, and is standardly described as the earliest work of Urdu literary prose 34. Zore's edition of Muhammad Qulī Quṭb Shāh's Kulliyāt appeared in 1940 35. Two figures asserted in an earlier version of this page — a page count for the Kulliyāt, and edition details for Sab Ras including a note dated 2 January 1953 — could not be confirmed against any scan and are withdrawn.
One correction the claim usually needs. Dakhni is not the oldest stratum of this literature. Faruqi puts the first substantial surviving corpus in Gujarat: "The first person whose Hindvī survives in substantial quantity, and with whom Urdu literature can seriously be said to begin, is Shaikh Bahā ud-Dīn Bājan (1388-1506)" 12. Gujri and Dakani are siblings, not parent and child — which is exactly Faruqi's point when he writes that Valī "showed conclusively, and for all time, that Rekhtah / Hindī was capable of great poetry, just as Gujrī / Hindī and Dakanī / Hindī were, at their best" 17. All three names carry the same second element. This correction bites against a Deccan-primacy stronger than the exemplar quoted on this page actually states 32; it is offered as a limit on the claim's strong form, not as a refutation of its stated one.
The 1720 test, and its limits
The sharpest available intelligibility test is what happened when a Deccan poet's collected works reached Delhi. Faruqi writes that "we know Valī to have already been a substantial poet when he visited Delhi in 1700" 15. "When his dīvān arrived in Delhi in the second regnal year of Emperor Muhammad Shāh [the second regnal year began in October 1720], it took Delhi by storm, and everybody, young or old, adopted Valī's style of poetry" 14. Mushafī, writing in 1794-95 and quoting an eyewitness, says its "verses became current on the tongues of young and old" 14. Language boundaries do not behave like that. One caveat on the source: points 1-3 are Faruqi's summary of the received Gulshan story, and he accepts only the third, writing that "There is fairly firm ground for accepting the third assumption" 14.
Faruqi also reports Maulvī 'Abd ul-Haq's judgement "that in the first half of the eighteenth century, the language registers of Delhi and Aurangabad were practically indistinguishable" 1816.
The counter-evidence is on the same pages and is not suppressed here. Valī's language "had a greater tilt toward the Persian-mixed Rekhtah of Delhi", and "Most of the 'Dakanī' component of Valī's language is tadbhav, and a good bit of it is to be found in Delhi's register as well" 16. What Delhi embraced may have been the most Delhi-like Dakhni available. The harder case is Nusratī, "perhaps the greatest Dakanī poet" — but the judgement usually quoted against him is not a northern one. Shafīq Aurangābādī, "a major historian, poet and tazkirah writer of the South" 15, writes that Nusratī's poems come "heavy on the tongue because of their being in the mode of the Dakanīs", and Faruqi glosses it by adding that "Hasan Shauqī's poetry is comparatively gentler on the Aurangabadi ear" 16. That is Aurangabad on Bijapur. It is still a real limit on the intelligibility argument, and arguably a sharper one, because it shows the distance inside the Deccan rather than across the Vindhyas.
A twentieth-century remark often quoted alongside it needs its context restored. In a private letter of 1967, Habīb ur-Rahmān al-Siddīqī Merathī wrote "'My problem is that I've forgotten Dihlavī, and I never did master Dakanī'" 13. Faruqi introduces the footnote as "Interesting evidence of Delhi's continuing bias against Valī"; two weeks earlier the same correspondent had written that Valī "needs to be debunked", and Faruqi notes that al-Siddīqī "regarded his own speech as that of Delhi" 13. It is a partisan aside in private correspondence, not neutral testimony about intelligibility, and the earlier version of this page built an inference on the verb "master" that the passage will not support. It is retained as colour and used for nothing.
The traffic also ran south — but the earlier version overstated how far. Faruqi endorses as "accurate" 'Abd us-Sattār Siddīqī's account, which is explicitly geographical: by 1590/1 there were "two forms of the Hindustani language in the Deccan. One, which was current in Dravidian[-dominated] areas of the Deccan, outside the territory of Daulatabad, and found few opportunities to renew its connections with the language of Delhi....The other form of the language was that which was prevalent in Daulatabad and its surrounds" 16. The Delhi re-infusion is about the second: "People from Delhi came to Aurangabad in very large numbers in the times of Shāhjahān and Aurangzeb, and brought Delhi's high Urdu with them. It renewed and refurbished the language of the territory of Daulatabad, and the Aurangabadis happily adopted the new language of Delhi. And that is the language that we find in Valī" 16. Golconda and Bijapur — Quli Qutb Shah, Wajhī, Nusratī, the corpus the chronology rests on — are in the first zone, which Siddīqī explicitly exempts. And Faruqi continues: "Once the Deccan became more or less independent of Delhi in the 1750's, the language of the Daulatabad-Aurangabad area lost touch with Delhi, and gradually tilted back to the main Dakanī mode" 16. So the documented Delhi re-infusion is confined to the Aurangabad-Daulatabad marches and was reversed within a century. It is not fatal to the independence claim outside Aurangabad, and the earlier version's "fatal" is withdrawn.
A weaker statement of derivation does stand, with a weaker warrant: "starting with Sayed (1968), scholars share the opinion that Dakhni developed from Urdu, which arose in and around Delhi" 19. That is one summary sentence in a ten-page overview, crediting a 1968 Mysore publication that was not opened here. It is reported, not relied on.
The forms in the poems
Grierson's diagnostic example of an archaic retention is "the employment of the oblique genitive as a base for declension in phrases like mērē-kū, to me, where the standard would have mujh-kō" 3. The Deccan poetic corpus carries the same stratum. Muhammad Qulī Quṭb Shāh: "'qutub' shah na de muj divāne ko pand / divāne kuuñ kuch pand diyā jaa.e nā", and "dekhtā huuñ har kahāñ dasta hai tuj mukh kā safā" 37. Valī: "kiyā mujh ishq ne zālim kuuñ aab āhista āhista", "yaad karnā har ghaḌī tujh yaar kā", "tujh lab kī sifat la.al-e-badaḳhshāñ suuñ kahūñgā" 38.
Three honesty notes on that convergence, one of them new. First, Grierson's example (mērē-kū) and the poets' usage (muj/tuj before a noun, kuuñ for ko, suuñ for se) are related features of the same stratum, not the identical form; the poets do not illustrate Grierson's example. Second, these transliterations come from a modern digital archive that states no source edition, no manuscript basis and no editorial policy, so they are Tier C and cannot serve as a text-critical witness. Third, and the reason their status here is now "unverified" rather than "fetched": rekhta.org returns HTTP 403 to automated retrieval, so the transliterations could not be re-checked, and internally they mix two incompatible romanisation registers across the three records 373839.
That the older Deccan register needed glossing for later Urdu readers is a point of intelligibility evidence rather than a note on poetics. Abdul Haq's twentieth-century volume on Nusratī is reported to carry a farhang, a glossary of Dakhni — but that report could not be re-confirmed against the scan and is recorded here as unverified; nothing rests on it 36.
Codes, schedules and speaker counts
No ISO 639-3 code has ever been assigned to Dakhni: the reference entry shows "ISO 639-3: – (none assigned)" with only a Glottolog code 30, and Glottolog's Dakhini page shows none either 22. The Urdu code page lists no alternate or dialect names at all 25.
The code dcc, sometimes produced as Dakhni's, is a different entry. It is registered as "Language Name: Deccan. Status: Active. Scope: Individual. Language Type: Living" 24, and Glottolog maps it to Deccan Marathi with an explicit comment that it conflates two things: the entry "closely matches the information from George A. Grierson 1905 ... on the 'broken dialects of the Dekhan'" while "also contains information matching the similarly named Dakkhini Urdu variety of the Urdu [urd] ... such as the primary usage of Arabic script", concluding "Hence this entry is spurious either as the Marathi variety or as Dakkhini Urdu" 23.
There is a referent problem on the other entry too, and it removes Glottolog from the authority column rather than qualifying it. Glottolog's Dakhini entry links, as its own article, to "Hyderabadi Urdu" 22 — and Hyderabadi Urdu is, on Casadei's account following Mustafa (2008), the Telangana sub-variety that "shows more resemblance to the Standard Urdu", as against the Rayalaseema variety "which has features imported from Telugu" 19. If dakh1244 is anchored on the least divergent sub-variety, its bare Dialect label is no better targeted than the dcc entry the catalogue itself calls spurious. This page therefore does not count Glottolog as evidence for the classification, and it should not be counted against it either.
India's Eighth Schedule lists twenty-two languages, including Urdu; Deccani is not among them 26. No body has granted Dakhni separate official status.
On numbers: the only speaker figure this page can vouch for is Grierson's total of 3,654,172, which appears both in the summary on p. 1 and as the total of the table on p. 59 24. Its basis is stated on p. 58 — "Taking the figures of the Census of 1891 as our basis" — so it is an 1891 return published in 1916, not a 1916 count 3. Grierson footnotes the Bombay Feudatories sub-figure of 254,282, not the total: "Many of these speak the standard form of Hindōstānī, but it is impossible to separate them out" 4. No modern figure is asserted here.
Who is answering which question
The popular statement of the claim is explicit that both alternatives are errors: "Typically, 'Hyderabadi Hindi' is what north Indians wrongly call Dakhni. Please, DON'T call it Hindi... Many Urdu speakers also mistake it for a form of localized Urdu, which, again, is incorrect" 32. Read as a claim about structural distance, it is not supported by anything opened for this page — and, on the argument above, it is not refuted either, because nothing opened here measures the distance. Read as a claim about elaboration and chronology — that the Deccan had a court and literary standard before Delhi's Rekhta, and that the name Dakanī is older than the name Urdu — it is well supported 101112.
One source used four times in an earlier version of this page was misfiled, and the correction matters to the shape of the argument. Dakhni.org was described there as "the most careful source stating the separate-language claim", and its statements were presented as concessions wrung from an advocate. It states no such claim. Its summary field reads "Family: Southern register of Urdu", and its opening sentence calls Dakhni "the southern register of Urdu that took shape in the Deccan between the fourteenth and sixteenth centuries" 33. It holds this page's own position; it can concede nothing to it; and every "even its own advocates concede" construction built on it has been deleted. Its provenance also went undisclosed: the site carries a standing banner stating that its content "is compiled with AI assistance from historical sources" and recommending independent verification, and it is a single-maintainer heritage project with no named editorial authority 33. It is retained below only as an example of how the variety classification is stated in popular Deccan heritage writing.
That leaves the position being adjudicated with a single exemplar: a 2020 web essay 32. The thinness is this page's fault and not the position's. A monograph literature on Dakhni exists, and none of it was opened: Sayed, The Value of Dakhni Language and Literature (Mysore, 1968); Hashmi, Deccan men Urdu (1963); Khan, Qadim Urdu (Osmania, 1965); Schmidt, Dakhini Urdu: History and Structure (New Delhi, 1981); Matthews, Dakhani Language and Literature (London PhD, 1976); Mustafa's 1985 Aligarh dissertation 20. Nor was Jamīl Jālibī's Tārīkh-e Adab-e Urdū, the standard Urdu-language treatment 42. Whether any of these argues separate-language status is not known from here; the word "language" in a title settles nothing, since Casadei uses it throughout while classifying Dakhni as a variety of Urdu 19. What can be said plainly is that the strongest scholarly form of the opposing position was not read for this page, that Schmidt's title promises exactly the structural treatment this page says is missing, and that Matthews is the source behind the one feature the earlier version dismissed. The verdict is set accordingly.
The formulation that survives all of this is narrower than the earlier version's. Dakhni is the older southern variety of a language later standardised in the north; it was an elaborated literary standard and was de-standardised, after 1687 and again through the administrative and educational substitution of northern Urdu in the Nizam's state 19. Whether the accumulated convergence has carried it past whatever threshold "separate language" would name is a question no source opened here has asked, let alone answered. What can be refused without hesitation is the hierarchy — the "broken Urdu" framing Grierson was already rejecting in 1916 3.
