4

The great Phoible 2.0 has exclusive list of segment usages based on "Languages", but since some languages such as Lak are only used by handful of people, it won't make any sense if I employ their pronunciations/segments in the language I'm creating , is it?

For example, /k͈ʷ/ sounds are ONLY used in Lak, this sounds would definitely not recommended to be put into my collection of alphabets of my language.

After all, what's the point of making a language if "only handful of people in the whole Earth knows how to pronounce it correctly"!?

And, English and Japanese are both count as "one" language, but we can't never deny the fact that "English speaker" is WAY~ more then "Japanese speaker"!

(~400M native speaker against ~128M according to Wiki)

So, if I choose some sounds that "only" English speakers knows how to pronounce over some "Japanese-only" sounds, you could bet that there are about 272M more people knows how to pronounce that sound correctly.

So, I was wondering if there is some other "outstanding" list out there based their list on "the amount of people using it"?

Much appreciated!

9
  • I’m honestly really confused about what you mean when you refer to ‘employ[ing another] language in mine’ — could you clarify this point please? As far as I’m aware, languages can’t really ‘contain’ other languages in any meaningful way. (Another point: just because Lak has few speakers, that doesn’t mean that non-speakers can’t pronounce it correctly — to give but one example, I (a monolingual English speaker) can pronounce a Lak text more or less correctly, if given in phonetic transcription.)
    – bradrn
    Jan 18, 2021 at 9:54
  • @bradrn I had updated my question, hope it helps. Jan 18, 2021 at 10:24
  • 4
    Thanks @PiggyChu001! I feel that a big part of your problem is that you are running into PHOIBLE’s inadequacies (which are sadly many, so always take it with a huge grain of salt). For instance, your example of /k͈ʷ/ is not actually a sound unique to Lak: it is a perfectly ordinary phoneme found in many other languages as well. However, the person who input the Lak inventory into PHOIBLE decided to transcribe it using a non-standard diacritic for ‘fortis’; no other person has used this exact diacritic, resulting in PHOIBLE reporting it unique. PHOIBLE contains many other such inconsistencies.
    – bradrn
    Jan 18, 2021 at 12:03
  • 1
    … Also, if PHOIBLE lists a segment with strange diacritics or otherwise unusual transcription, or a segment which seems straightforward but is only in very few languages, that can be a sign you need to investigate it further. To give another example, /ⁿd/ is apparently attested in 23 languages — but only because PHOIBLE usually transcribes this phoneme as /nd/, present in 270 languages.
    – bradrn
    Jan 19, 2021 at 6:22
  • 1
    Assuming that it's possible to 'normalize' the segment representations at all, you could generate the list you're interested fairly easily by joining (database theory term) the normalized segment list with a list of languages and the count of their speakers. Mar 6, 2021 at 19:59

2 Answers 2

2

Well, you need to add numbers of speakers from another source (e.g. Wikipedia or Ethnologue, or some other list with speaker estimates). Pro-tip: Use a list where the languages are ordered by the number of speakers. You can use a cut-off (no matter whether you choose 10k or 100k speakers) and set that cut-off for the number of speakers (in statistical lingo, this kind of adjustment is termed smoothing). This frees you from looking up the numbers for languages with fewer speakers.

Be aware of the following: By weighing the languages by the numbers of their speakers you are putting a bias into your language design, heavily favouring Mandarin, English and Hindi/Urdu over the rest of the languages. This kind of bias may be or may not be a good thing in your conlang.

1

All consonants in 50% or more languages:

m, k, j, p, w, n, t, l, s, b, ŋ, g, h

All vowels in 50% or more languages:

i, u, a, e, o

I hope this answers your question

1
  • OP asked about segments based on how many people use them, not how many languages use them, so the question is not answered at all.
    – Cecilia
    Feb 8, 2022 at 12:43

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.