There are few possible syllables in Japanese, (96, I think) and not many more in Italian.
I was pondering to myself, how many standalone monosyllables could exist in English? That is, how many monosyllables would meet the normal constraints of English pronunciation: componsed of English phonemes in orders that would not seem out of place in English? I realise there would be some fur on this concept, and any analysis would be subject to dispute but I will have a go at a ballpark estimate.
In my own dialect (general Australian), as in RP, there are usually 44 phonemes counted. 24 consonants, 7 short vowels, 6 pure long vowels, 7 diphthongs.
A syllable could be represented as [initial consonant sequence][vowel][final consonant sequence], where consonant sequences could be null. We could just work out all the possible ICSs, vowels, and FCSs, and multiply the three counts but this would give an overestimate, as many of the combinations would not seem authentically English.
I gave some thought on how to deal with the schwa: it’s a real part of English but in any real monosyllabic _word_, the schwa is standing for something else. I probably say “to” as /tə/ half the time, but still, the formal pronunciation is /tu:/. (I would contrast this, say, to the word “television” in which the schwas make up part of the proper pronunciation.) I decided to leave the schwas out of this analysis. Also, I will neglect syllabic nasals and liquids.
The following restrictions would apply, based again solely on my dialect:
1/ Only a long vowel (pure or diphthong) can end a standalone monosyllable with no FCS. There are 13 of these.
2/ Only a short vowel could be followed by /ŋ/. There are 6 of these, as I’ve dropped the schwa.
The set of ICSs is as follows:
We have our tsetse, vraic, pshaw, and our Dvorak and so on ... I have more or less arbitrarily ruled that the initial consonant sequences don’t _sound_ like real English.
Also, the set of ICSs starting with ʃ would be shorter were it not for the Yiddish influence on English but there you have it. I’ve listed 72, 73 with the null.
So, what about the arse-end?
Again, I left out some arguable ones like /nf/ from inf. There are also some like /mzd/ that I could have included ... can’t think of any words that end in /mzd/ but I can _imagine_ someone saying something like that in English without batting an eyelid. Same with /mdz/. Or /bzd/. One day Hobbes might be a verb, and we’ll say “You’ve been Hobbesed!” and it won’t sound odd. But it’s not one day yet, so I left them out anyway.
I also left out /zm/, /zmd/, /zmz/ ... Spasm has two syllables in my dialect.
There are 133 in my list, 134 with the null.
Allowing for the restriction 1/ above, 133 of these FCSs could go with the 6 short vowels (remembering I’ve kicked the schwa). 126 could go with the long vowels.
73 * (126 * 13 + 133 * 6) = 177828
By making different assumptions you could get significantly more or fewer than this, but no matter how you slice it, it’s a lot, compared to Japanese.