Pay attenchun

I've been wondering about this for a long time. I often hear words like "attention" or "mansion" pronounced with a CH sound (like in "cheese"). It seems to me that happens every time there's an N followed by a SH sound.

The question is: does the combination of N and SH really turn into a CH, or is it something different that is happening? No dictionary writes CH for that in its transcriptions, but always N + SH.

So I'd just like to know more about this common feature. Thanks.

I hear this quite often, but only from Spanish speakers who have very poor English, or at least poor pronounciation.

I hear this in some occasions, but only from Spanish speakers who have very very poor English, or at least extremely poor pronounciation.

I hear this in some occasions, but only from Spanish speakers who have very very poor English, or at least extremely poor pronounciation.

The matter is that /ns/ and /nʃ/ are normally realized as [nts] and [ntʃ] respectively in English dialects overall. If anything, realizing /ns/ and /nʃ/ as [ns] and [nʃ] would be a non-native pronunciation, rather than the other way around, even though there may be individual English dialects which have such for /ns/ and or /nʃ/.

Sorry Travis, I am not sure if I understood what you just said. You are saying that I should say /nʃ/ as /ntʃ/, otherwise it would probably sound like a non-native pronunciation, right?
The fact is that all dictionaries write /nʃ/ in their transcriptions, and I still wonder why. Well, Merriam-Webster (which is the one I choose to trust when I don't have a clue how to pronounce a word) always writes something like n(t)-sh. I am now starting to think I should interpret that intrusive t as a sign that two pronunciations are possible, either /nʃ/, or /ntʃ/.
But do native speakers really say /tʃ/ in those cases, or is it something similar, but not quite the same? I think I often hear a pure /tʃ/ sound as in "chip", but not always. So I guess the other times what I am hearing is /nʃ/.

To strictly state such, the matter is that normally there is no contrast between /nʃ/ and /ntʃ/ or /ns/ and /nts/ in English; while one could artificially force /nʃ/ and /ns/ to be realized as [nʃ] and [ns], normally they are realized as [ntʃ] and [nts], like /ntʃ/ and /nts/. If there is any difference at all to begin with in the general case, that would be that /ntʃ/ and /nts/ could have a longer period of actual stopping than /nʃ/ or /ns/.

There is one main exception to this. Many North American English dialects preserve a contrast between /nʃ/ and /ntʃ/ and between /ns/ and /nts/ is in syllable codas, shifting /ntʃ/ and /nts/ to [ʔtʃ] and [ʔts] in coda positions while transferring all nasalization to the preceding vowel while preserving /nʃ/ and /ns/ as [ntʃ] and [nts] in coda positions. Consequently, in such dialects "pants" does not rhyme with "dance", being [ˈphɛ̯̃æ̃ʔts] and [ˈdɛ̯̃æ̃nts] respectively (in my own dialect)

My first assumption would be that this occurs because /n/ is articulated in the same place as /t/.

I've heard that in dementia, eventually and actually, but I prefer plain /S/ (sh) in these words.

Heh - around here such is the normal pronunciation in all registers (except that in careful speech [n] may be preserved in words like "pants"); not having such pronunciation comes off as very stilted to the point of being almost foreign to me. That is unlike some pronunciations, such as pronouncing /ŋs/ and word-final /ŋz/ (when not followed by a word starting with a vowel) as [ŋks] which sound distinctly non-standard and dialect-specific to me (and which commonly shows up in everyday speech here but is largely absent from formal speech here).