I've read that some people attribute it to influence from the Slavic languages. But it doesn't just appear in Slavic loans — it also shows up in obviously Latin-derived words like câine 'dog' (from CANEM). On the other hand, not all words with a in Latin have â in Romanian — e.g. FRATER went to frate rather than *frâte. So was this a sporadic sound change, or a regular change conditioned by some environment, or what?

  • Turkish also has a similar sound in ı (dotless i) and Romania was part of the Ottoman Empire, so I wonder if this is also an influence. An important thing not to forget about Romanian though is that â/î contrasts with ă which is like a schwa in English and French. Its's particularly strange for an Indoeuropean language to have two contrasting central vowels. Nov 9, 2011 at 12:35
  • 4
    And while we're counting, Albanian and Armenian also have phonemic schwa -- even in stressed syllables, not just as a result of vowel reduction. Seems like it cropped up in IE languages all through Eastern Europe and Central Asia. Nov 9, 2011 at 18:21
  • 1
    Those sounds are not abnormal in a Romance language. Similar ones can be found in Portuguese and many Italian dialects/languages. That doesn't exclude a possible Balkan connection, especially through Albanian.
    – cipricus
    Nov 19, 2019 at 13:12
  • 1
    Stressed schwa is arguably a feature of the Balkan Sprachbund, and one more reason that Armenian is arguably a member. If you can find a loan with it to or from Armenian, you can establish if the pronunciation of ը shifted or not - given that Armenian has been written since about 400 AD, before any Turco-Mongol invasions, like Greek, and South Slavic - and split from them far earlier. Sep 26, 2022 at 16:34
  • 1
    That said, Turkish influence seems unlikely, because: 1) Armenian and Georgian still have a pretty complex sound system 2) Persian and Greek don’t have phonemic schwa 3) None of these languages have vowels that are actually unique to Turkish, namely ö and ü 4) Actual loans from Turkish don’t result in this sound, e.g. en.wiktionary.org/wiki/zaiflâc#Romanian Sep 26, 2022 at 16:35

3 Answers 3


wrt. the other answer:

  1. Romanian sfânt does NOT come from Latin sanctum, but from Slavic svętъ (compare with Polish święty); most Romanian (and Hungarian) Slavic borrowings turned the nasal vowels (which were de-nasalized in most Slavic languages) into vowel + n (vz. luncă, rând, mândru, muncă, smântână, etc).

  2. Romanian /ɨ/ may also come from any vowel + /n/ [1], not just /a/; off the top of my head: /e/ (pavimentum -> pământ, tener -> tânăr, fenum -> fân), /u/ (aduncus -> adânc). Neither of those are later loans.

  3. The same process from 2) also happened extensively with borrowings from other languages (e.g. gond -> gând).

Taken all together, it may be probable that the /ɨ/ (which btw, sounds more like /ɯ/ with most speakers) evolved by merging a series of nasal vowels that existed at some point in time; a hint in that direction is that /ɨ/ was written with Cyrillic ѫ in the old Romanian alphabet (the letter used for ǫ /õ/ in Church Slavonic).

Once established, the /ɨ/ became an euphonic alternative for almost any vowel, depending on the consonants surrounding it.

[1] I'm pretty sure that it also happened with vowel + /r/, though I cannot think of any example right now ;-)

  • Romania "sfânt" does come from slavic, but it was superimposed upon an older word that's actually still used, "sânt"/"sân" (as in many village names "Sânpetru", "Sânandrei", etc.) that does come from Latin. In fact, most primitive Christian/religious words come from Latin, in Romanian.
    – Dan
    Apr 3 at 12:11

In words of Latin origin, Romanian /ɨ/ arises from a few different places:

  • Latin /a/ followed by /n/: cânt < canto, sfânt < sanctum. The diphthong in pâine, câine is due to assimilation to the following vowel.
  • The Latin word /in/. This is not a regular sound change in that not all Latin instances of /i/ followed by /n/ are centralized, but all forms of the preposition "in" have /ɨ/.
  • The vowel /i/ preceded by /rr/ or initial /r/: râu < rivum, urî < horrire.

Other instances of /ɨ/ are generally loans. Also, there are numerous words of Latin origin which appear to violate the rules above, but these are usually later reborrowings from Latin or sometimes French.

  • 5
    sfânt is not from Latin sanctum, it's from Slavic. But sânt is. Jun 19, 2021 at 20:18
  • @AdamBittlingmayer - Isn't it probable that Slavic svętъ gave sfânt by contamination with the Neolatin sânt?
    – cipricus
    Sep 26, 2022 at 10:51
  • 2
    @cipricus Possible, but the answer from Guest convinced me that svętъ to sfânt is expected - cf smântână. Sep 26, 2022 at 15:54

The problem with explaining away the Romanian /ɨ/ as coming straight from either Latin or Slavic is that it ignores the local substratum which just happened to have been in the area between Latin and Slavic and could not but accept/borrow this feature. I suggest that in fact /ɨ/ or a type of /ɪ/ might have been a shared vowel in a larger Indo-European area. That is, if on the Slavic end of this area, there had been an /ɨ/ or an /ɪ/¹, it is very likely that /ɨ/ or /ɪ/ (his possible ancestor) could also have been a linguistic trait of the local variants of the Indo-European language, aka Thracian idioms. On the Latin side of the continuum, I suppose that /ɪ/ or even an /ɨ/ was what Latins described as being their short 'i'. Once we have established that both sides of the continuum, the Latin and the Slavic ones, have the /ɪ/ or even an /ɨ/ the hypothesis that /ɨ/ was a Latin or Slavic influence is not necessary. My take is that proto-Romanian, similar to Latin or the Romanian Transilvanian dialects, inherited from its Thracian substratum some pairs of tense and relaxed (long-short) vowels that differed by quality. I suppose that some of those relaxed vowels, especially the /ɪ/ vowel changed into /ɨ/ when the vowel system got simplified, at some point down the line, if it had not been there already as I mentioned. Consequently, the borrowings from other languages just followed the pattern. This might be the case of the Hungarian ‘domb’ or the Germanic ‘damm’, which is ‘dɨmb’ in Romanian. However, I wouldn’t rule out an indigenous /dɨmb/ or /dɪmb/.

¹ - https://www.veche.net/novegradian/history

  • 1
    Your answer could be improved with additional supporting information. Please edit to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers in the help center.
    – Community Bot
    Oct 23, 2023 at 17:14
  • Scholarship, my... As long there is no written text of Proto-Albaniani or Thracian, for that matter, we are trading conjectures, nothing more. At least I wrote MY speculation, whereas you just produced two names. The ancestor of Albanian... you speak of it as if you have known at least a few speakers of that idiom. Oct 27, 2023 at 21:09
  • 1
    You can find here a significant contribution against the trend of looking for non-Latin substrate of the main features of Romanian. Also note that /ɨ/ is a normal evolution in Romance, very common in southern Italy, parts of northern Italy and Portuguese. You are a new user but I still have down voted you for your bad reaction to criticism and your dismissal of scholarship as such. What is the element you say is so familiar that it doesn't need citation?
    – cipricus
    Nov 1, 2023 at 18:15
  • 1
    Nobody said it is coming from Latin or Slavic, it is a Romanian evolution, of Latin or non-Latin forms (mostly from Latin for the simple reason that the core vocabulary is Latin). The sound is in fact mostly absent in the southern Slavic languages that influenced Romanian. (Romanian has î where Bulgarian has ă: Târgoviște/Tărgovishte). That it was present in Dacian or Thracian is possible but there is no proof of that, as we don't have those languages. It is absent in Albanian. The only language that has the sound and is a source of some words with the sound is Ottoman Turkish (calabalâc).
    – cipricus
    Nov 1, 2023 at 19:02
  • 1
    We don't have to get outside the Romance areal in order to explain such vowel shifts, French and Portuguese are even more creative. In this sense Romanian is closer to Italian (especially Italian local languages), as suggested by the work linked in previous comment. Like in all Romance, this must be an internal/specific Romanian (post-Latin) development. Imagining it as ante-Latin is otiose.
    – cipricus
    Nov 1, 2023 at 19:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.