I've seen many cases when people who speak different languages make a common mistake spelling words. They add an extra sound (usually, a consonant) while there is no historic or linguistic evidence this sound is necessary nor it was there in the past.

Thai: โรตี [roː tiː] (a kind of puff pastry) spelled as โรตรี [roː triː]
Ukrainian: прецедент [prɛ t͡sɛ dɛnt] precedent spelled as пренцендент [prɛn t͡sɛn dɛnt]
English: drawing spelled as it was "drawring"

The Thai case is especially interesting since an opposite pattern is very common: consonants are often omitted from the consonant clusters:

กลัว [kluːə] to be afraid spelled as กัว [kuːə]
ใคร [kʰrɑi] who spelled as ใค [kʰɑi]

I suppose it might be somehow related to how people produce sounds generally, but I'm not sure what to start with.

Since the mistake is very common, does it have any common origin?
Is there any research on this topic?

  • 7
    I don't know about the others, but 'drawring' mirrors how it is pronounced in many dialects. There's a tendency across languages that disfavours hiatus, and that happens to be how non-rhotic dialects of English deal with it.
    Sep 25, 2012 at 23:00
  • Assuming the others have a similar sort of phonetic basis, the Thai example would be progressive/preservative assimilation, where features from the first syllable bleeds into the second. The Ukrainian example would be the opposite, regressive/anticipatory assimilation. These may be idiosyncratic or perhaps dialectal phonological processes.
    Sep 25, 2012 at 23:08
  • 3
    It seems to me that the question of spelling is incidental: if they are spelt that way, it is because they are pronounced that way, at least by some people. And the incidence of both phenomena in Thai is no surprise at all: it suggests that there are people for whom the difference between, say, [triː] and [tiː] is neutralised, at least in some contexts, so it then becomes a job of memorisation to recall which word is spelt <tr> and which <t>.
    – Colin Fine
    Sep 25, 2012 at 23:16
  • 1
    @ROBOKiTTY Two last Thai examples do not illustrate the question. They illustrate an opposite tendency or cluster assimilation (reduction), which is by itself described very well. On the contrary, [ro-ti] never used to have an extra [r] within; likewise [drawing] does not seem to be a reversal of any prior reduction. Sep 25, 2012 at 23:36
  • 1
    Building off @ROBOKiTTY : I have cited Loanword adaptation as first-language phonological perception by Boersma and Hamann several times on this site. If one accepts their premise of a productive perception phase, influenced by the listener's personal phonology, then they generate URs of words inline with that phonology. Given the phonetic nature of the writing systems in question, it isn't too far of a leap to think that they generating the written form from their own personal UR.
    – acattle
    Sep 26, 2012 at 6:33

2 Answers 2


This is a well known phenomenon called epenthesis.

It is very common in all languages. For example, thunder < thunor, athlete pronounced 'æ thuh leet', intrusive r in non-rhotic dialects, it goes on and on.

The phenomenon of losing a phoneme is just as, if not more, common and is simply called contraction or elision.

For example, probly < probably, not > -n't, lenition, slurring and mumbling of all kinds.

  • ... by the way, how does hapology relate to ellision, is there a difference to be made between conscious omission or fusion of phonemes, one one hand and not learning the phonems of a word correctly in the first place?
    – vectory
    Mar 10, 2019 at 14:16
  • @vectory what do _you _ mean by hapology? urbandictionary.com/define.php?term=hapology ?
    – Mitch
    Mar 10, 2019 at 14:24
  • 1
    TIL, I mean haplology. Oh, the irony! I see that the definition only applies to deletion of syllables. However, "library as /ˈlaɪ.bɹi/ (not /ˈlaɪ.bɹə.ɹi/)" comes pretty close, not to the question, but to your remark. haplology falls under the definition of ellipsis covering words, syllables and letters, so my question would have been for the specific term of elided letter. Turns out that's elision, indeed. I was hoping to improve the answer. Following your given links should suffice, though.
    – vectory
    Mar 10, 2019 at 15:18

Adding to @Mitch's answer, n before dentals, or vice versa, and more (e.g. m before b) is called gemination.

The case of a merger for wr and w, that @ROBOKiTTY pointed out, would be a case of hypercorrection, if the respective speakers learned often enough that words like dowry require an r, and figured that must be the case for drawin. Calling it hypercorrection is not really accurate, if it doesn't stick and due tonawareness of the mistake; Or if it's deemed an accurate transcription, than it would be polemic to denigrate as a mistake. Inhave never seen this one, though. Basically what @acattle said

A t can be realized with or without closing of the vocal tract, same as n. My favorite example would be something, among other slang variants rendered suttin, and thence su'in with just a glottal stop. This can hardly lead to hypercorrection anymore, if alphabetization is high and the standardized spelling seen often enough.

For the rhottic accent, we also have the internet-famous ermahgerd mershed pertaters "Oh my god, mashed potatoes", that's an over-the-top riff on the accent as perceived by outsiders.

  • Vice versa, for lenition there's the way old "I wuv you vewy much" and other such confusions. way old, by the way, seems to fall into the same pattern, as if from very. I can't speak to Thai and the like.
    – vectory
    Mar 10, 2019 at 15:01

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.