Talk:Sandhi

Also in Avestan

Avestan[1] uses this term explicitly as well, though I am pretty sure this is from a Gujarati influence Khirad 11:17, 26 September 2005 (UTC)

Mandarin

However, this is difficult to say, so the tone on nǐ mutates into ní. This is incorrect, the sound does change, but it only changes into a half third tone, not into a full 2nd tone.

Internal Sandhi

I think that the example of internal sandhi is incorrect:

• Internal sandhi features the alteration of sounds within words at morpheme boundaries, as in sympathy (syn- + pathy).

Wouldn't the "syn" change to "sym" simply because of assimilation to the p because it is much easier to say mp than np? Miss Madeline | Talk to Madeline 19:03, 28 March 2006 (UTC)

I think assimilation is a form of Sandhi. --Krsont 18:34, 4 May 2006 (UTC)

Actually the proper name of this article should be "assimilation". That's what it's called in linguistics. The term "sandhi" is probably just used by Sanskrit afficionados, and so belongs solely on sa.wikipedia.org -- bkhl 17:48, 18 September 2006 (UTC)

Assimilation and sandhi are two different things, though assimilation can be a kind of sandhi. If assimilation doesn't happen at a word or morpheme boundary, it isn't sandhi. Synchronically, in that case, you'd be hard pressed to identify it as assimilation if there was no morpheme boundary, but there are plenty of historical assimilations that aren't sandhi (e.g. Latin septem > Italian sette "seven"). And sandhi needn't be assimilation, either; it can be dissimilation, metathesis, lenition, fortition, etc. as well. —Angr 10:38, 26 November 2006 (UTC)

Absolutely right. And furthermore, the details of what happens phonetically at word boundaries can be quite different from what happens elsewhere. As in Sanskrit, final *-as [əs] before a word beginning with a nasal first assimilates to *-az [əz] and then loses the *z with compensatory lengthening of the preceding [ə] (which falls together with the outcomes of smoothened diphthongs), thus aśvo mama "my horse"; but no such thing happens medially, as in asmi [əsmi] "I am". Alsihler 15:58, 15 December 2006 (UTC)

Is sandhi only for phonetics changes?

If not what is the name for the change of the writing of the letters at the beginning and end of the world. Is that a phenomena of easying writing like the given sandhi example whow easing of prononciation. I see a potential parallel here. Spayrard 22:13, 27 November 2006 (UTC)

Sandhi vs. morphophonemics

Future augmentations to this article need to explain this. Hurmata 03:04, 12 January 2007 (UTC)

is there really a sandhi in french ?

French liaison is described as an external sandhi but lack the regularity it should have in such a case... seeing it as a sandhi would also means it is a phonetical phenomenon, but when reading the page "liaison" it is obvious that it is rather an historical process. I believe that the definition of sandhi is the modification of a sound and could not normaly produce the adding of a new sound (or suppressing of it as it is correctly said). excuse my english. stefjourdan@caramail.com (motunono in french wikipedia...)

Yes, that is also what I have read. I need a source to convince me that an apple is an example of sandhi. Were English to have *āpple for a apple, that would be sandhi according to the sources I read. Terminology differs over time and between disciplines, so I'd not be surprised to hear a different POV, though. Alastair Haines (talk) 13:22, 8 May 2008 (UTC)

reflection of sandhi in orthography

The distinction between a and an - if that is even sandhi at all - is not the only place in English where sandhi is reflected in spelling. Further examples include some assimilations, including the very sympathy already mentioned (thus making the article its own counterexample), as well as the common prefix in-, which is spelled to reflect its assimilation in such words as impractical (in-practical), illogical (in-logical) and irregular (in-regular). Other examples that may or may not qualify as sandhi spellings include the variant forms wanna, gonna, and their ilk, and even less standard spellings like doncha (don't you). Tzinacan (talk) 18:16, 1 August 2008 (UTC)

Internal sandhi across morphemes?

In a book on Sanskrit (TY Sanskrit by M. Coulson) I found another category of internal sandhi examples. It occurs for example in the word tiSThati (infinitive is sthaa), where the two consonants s and th become retroflex because of the preceding vowel i. Another one is formed by the group of neutral plural nouns, where the suffix -n- may undergo retroflexion (for instance, in indriyaM -> indriyaaNi, because of the -r- in indriyaM).

Does anyone know of examples like these in other languages? Arjunah (talk) 23:23, 28 October 2008 (UTC)

"Like these" how? Retroflection? kwami (talk) 23:28, 28 October 2008 (UTC)

No, I meant internal sandhi "working" across syllables (sorry, I wasn't too clear I guess) Arjunah (talk) 21:56, 6 November 2008 (UTC)

Couple of questions

1st is how come this article is in philosophy section? Isn't this related to grammar??? 2nd is where are the calssifications like 'savarna deergha' sandhi etc.? Elncid (talk) 07:35, 21 February 2009 (UTC)

external Sandhi

English "a" -> "an" seems a far-fetched case for explanation as sandhi, it would be equally reasonable to call the change in Gaelic from "a" (her) before a consonant to "a h-" (her) before a vowel - perhaps in fact more reasonable since it seems quite likely that "a" originally ended with a consonant (it doesn't cause lenition the initial consonant of a following word) - and I've never heard anyone call it that. Gaelic does of course have external sandhi, sometimes represented in the orthography: what is called urdhubhachadh (caighdean spelling "urú", English "eclipse") in Irish and srònachadh (English "nasalisation") in Scottish is external sandhi, where the final consonant (generally n) of one word interacts with the initial consonant of another; for example something that began as "nan beann" ("of the hills") is now written "na mbeann" in Irish and "nam beann" in Scottish; what started off as "nan craobh" is now written "na gcraobh" in Irish, but this sandhi (present in the pronunciation of most Scottish Gaelic speakers) is not shown in the orthography so we still have "nan craobh" in writing. In Scottish and in Ulster Irish (actually Irish north or east of the cha/ni isogloss) "cha" causes sandhi (although before a consonant it is always written without its final n): for example in Scottish Gaelic "cha tug" is pronounced /xaˈd̪̊ug̊/ (and according to the new orthographic conventions for school examinations "tug" can be written "dug" since grammatically it cannot occur other than after "cha" or "an" so is always pronounced /d̪̊ug̊/). MichealT (talk) 11:46, 29 March 2010 (UTC)

curious observation

Is sandhi the reason we have the English word "Sandhi" and not the original "Samdhi"?

The English orthography is superficially defective here: the 'n' of "sandhi" written with Indic abugidas is not written with the consonant 'ma' (not even with a virama, it does not form a conjunct with the following syllable) but with a diacritical anusvara sign (typically a dot above the base letter in Devanagari, or a small circle above or on the left of the base letter in other Indic scripts); however, its actual phonetic pronunciation may be the nasal consonants /m/, /n/, /ɲ/, or /ŋ/, possibly shortened as /ᶬ/, /ⁿ/, or /ᵑ/, depending on the following consonant, and possibly with the nasalisation of the vowel (noted in IPA with a /◌̃/ diacritic) just before it and normally written with an Indic candrabindu sign). E.g. संधि (romanized as saṃdhí) in Hindi and Sanskrit is pronounced /sɐn.dʱi/ (vedic pronunciation) or /ˈs̪ɐn̪.d̪ʱi/ (classical dental pronunciation). Indic dialects and accents also influence the pronunciation of the orthographic Indic anusvara. So here the original standard romanization is "saṃdhi", but as the diacrited letter 'ṃ' is not part of the English alphabet, it is accurately adapted as "sandhi" to the English Latin orthography with 'n', according to the original phonology /n/ used in Hindi or Sanskrit as well as in English. You'll note as well that English accurately uses its standard Latin orthographic digraph 'dh' to transcribe the single Devanagari consonant 'ध' /d̪ʱ/, and the orthographic vowel 'a' for the phonologic vowel /ɐ/ (for the implicit/unwritten vowel of the Deganagari consonant 'ध'). In other words, no internal sandhi mutation (orthographic or phonologic) ever occurs in the English word "sandhi" translating the original Sanskrit/Hindi word. verdy_p (talk) 10:43, 5 January 2024 (UTC)

pausa

a discussion of how sandhi interacts with pausa would be interesting. (an intended prosodic break will generally prevent sandhi, but it may occur w an unintended break.) — kwami (talk) 15:11, 1 October 2010 (UTC)

Sandhi in English

I think it would be helpful if someone familiar with the topic could put in some more about sandhi in English. For example, the article Morphophonology mentions that the three ways of pronouncing the plural morpheme, [s], [z], and [ɪz], and the three ways of pronouncing the past tense morpheme, [t], [d] and [ɪd], are examples of sandhi. Also, at a word boundary [t] + [j] becomes [tʃ ] ("don't you" → "dontcha") and [d] + [j] becomes [dʒ] ("did you" → "didja"). And I wonder whether the pronunciation difference between UK and US English in words like "education" (UK [d.j] → US [dʒ.]) is an example of internal sandhi.

I don't feel confident putting anything in myself because it's been a long time since I read about it and I don't have access to the references any more. Duoduoduo (talk) 23:38, 28 December 2012 (UTC)

You could also cite the informal American English orthographies "gonna", "gotcha", or "wanna" for "going to", "got you", or "want to", which notes common sandhis occuring in the popular American phonology (now heard frequently as well in UK in popular songs and speech) and as well in various English-based pidgins or dialects. As well as the common orthographies "I'm", "You're", "He's", "She's", "It's", "We're", "They're", "I'll", "You'll", "We'll", "They'll", "isn't", "aren't", "haven't", "don't", "won't", etc. standardized in English and used in all its dialects (for "I am", "You are", ... "I will", ..., "is not", "are not", "do not", "will not", etc. used for emphasis or in very formal speech, but rarifying even in formal texts). verdy_p (talk) 11:26, 5 January 2024 (UTC)

Description of third tone

The article says that the third tone of Mandarin is _often_ pronounced as a falling-rising tone in isolation. Should "often" be "always" (or at least "normally")? --BLebow4500 (talk) 04:41, 18 April 2016 (UTC)

The concept of Sandhi or Punarchi

The concept of Sandhi in Sanskrit and the independent concept of Punarchi in Tamil are well known from two of the classical languages of India. They are the earliest of such descriptions. I added a short phrase to reflect this and it is removed by and an editor. Here is the Diff. The point is not that it occurs in Tamil language, but rather the concept was first recognized in Tamil and Sanskrit and it is of historical importance. I wish to reinstate that phrase at the least. The importance of Punarchi is quite well recognized and elaborated in Tamil. The concept is more general and applicable to a number of languages. Tamil and Sanskrit were two of the languages that elaborated about it some 2000 years ago.--செல்வா (talk) 04:33, 4 December 2016 (UTC)

"North Germanic languages"

The lead claims that sandhi is typical for North Germanic languages. As a native speaker of one myself, I rather doubt the relevance of this unsourced claim. Sure, some sandhi occurs in North Germanic languages, just as it does in most other European languages (other Germanic ones, Romance ones, Slavic ones) so it's not characteristic for North Germanic languages in particular. If we look at European languages, sandhi is nowhere near as important in North Germanic languages as in languages such as Greek, Basque or Breton. Jeppiz (talk) 22:46, 3 January 2021 (UTC)

I agree, there's no reason to single North Germanic out. The article gives examples (morpheme-final [n] → [m] or [r] or [l] in certain prefixes; word-final /r/ in some non-rhotic dialects) in English.

Under French liaison, word-final consonants that have been lost in isolation are restored, in certain cases, in front of word-initial vowels (and joined to them across the syllable boundary); in at least a couple of cases (the numbers six = "six", neuf = "nine", and dix = "ten), the consonant is pronounced in isolation, dropped before a consonant, and voiced before a vowel. Examples in isolation, [sis]; six jours ("six days"), [si ʒuʁ]; six ans ("six years"), [si zɑ̃]]. Also, French has the same internal sandhi as English for the same prefixes—as do all the other Western Romance languages I'm aware of.

So even in Europe there's no reason to single out the North Germanic languages.Largoplazo (talk) 23:17, 3 January 2021 (UTC)

I just removed it. In addition to what I wrote above, French, Spanish and Portuguese bear witness to the same sandhi as English when it comes to prefixes "con-", "in-", "en-", and "syn-"/"sin-" in front of labials, velars, and, for some of those, liquids. (In French the effect on pronunciation is gone because the terminal consonant in those has been reduced to nasalization of the previous vowel when the prefix precedes any consonant, but the historical sandhi is attested, in front of labials, by the same change in spelling as in English.) This internal sandhi actually took place in Latin (as attested by Latin spelling), but it seems to me that it remains productive in all those languages: if someone were to coin a verb consisting of "en-" + "plasticate" to connote the wrapping of an item in plastic, I have no doubt it would come out "emplasticate", because that's how our phonology for those prefixes works. Largoplazo (talk) 12:23, 8 February 2021 (UTC)

I'm supposing the word pairs "a"/"an" in English and "с"/"со" and "в"/"во" in Russian could be described as consequences of sandhi. Also, conjunctions "y/e" and "o/u" in Spanish; articles "un"/"uno", "una"/"un'", "il"/"lo", "il"/"l'", "la"/"l'" and "i"/"gli" in Italian; and "le"/"l'", "la"/"l'" in French. Largoplazo (talk) 12:32, 8 February 2021 (UTC)

Standard contextual elision of articles and pronouns "ce"/"c'", "le"/"l'", "la"/"l'", "ne"/"n'", "je"/"j'", "me"/"m'", "te/t'", "se"/"s'", "que"/"qu'" in French; and informal elisions "tu"/t'", "elles"/"l'", not always necessarily before a vowel-starting word (plus the non-eliding external, sandhi "ce"/"cet" before vowel-starting words, and the informal eliding internal sandhis of "ce"/"c'te" and "cette"/"c'te", and other informal internal and external sandhis "cet"/"c't'", "c'est"/"c't'" and "je te", "j't'", or the informal external sandhis "Il y a"/"Y'a"/"Ya", "Il n'y a que"/"Y'a que", "Il n'y a qu'à"/"Y'a qu'à"/"Yaka", etc.). Plus false liaisons, that were standardized in the orthography in some cases, using an epenthetic "-z-" or "-z'" particle, such as "Vas-y !" (the normal imperative conjugated verb here should be "Va !" but the 's-' here is an epenthetic external sandhi required before the locative pronoun "y", which was formed in old French by elision of the Latin word hic meaning 'here' and not as the newer informal external "y" for "il", "il y") sometimes used informally as "Z'y va", "Z'y-va", "Z'yva", "Z'y vas", "Z'y-vas", or "Z'yva". Sometimes as well "Z'ont" for "Ils ont" or "Elles ont". The negative or restrictive particle "ne" may be also not just elided contextually as "n'", but dropped completely in informal speech if it is followed by the restrictive adjective/adverb or preposition "que", 'only' or 'that', or by the negative adverb/adjective "pas", or the negative pronoun "rien", or the negative pronouns/adjectives "aucun", "aucune", etc. ~~

Renaming

Into euphonic combination? These are native Greek and Latin words, suitable for English. --170.64.211.105 (talk) 00:33, 9 May 2024 (UTC)

You seem to have left a few words off the beginning of your first sentence, and what does "these" refer to? In short: What did you intend to say/request? Largoplazo (talk) 00:53, 9 May 2024 (UTC)