I had a friend who was very thorough and insisted on pronouncing the g at the end of -ing words. I took college linguistics and remember this table. I tried to explain to her that "ng" may be spelled as two letters but, like sh and ch, is really a single consonant, and linguists give it its own name (the velar nasal) and its own singular symbol (ŋ).
Sometimes we clip the g, and say goin'.
Then there is the standard way to say it, goiŋ.
Then there is the way she was saying it, goiŋ-guh.
I find it fascinating that languages have different sets of sounds. For example, in English p and b are two different letters, and pat and bat are two different words. But other languages don't distinguish them, and their speakers really wouldn't notice a difference if you said pat or bat. On the other hand, I don't hear the difference between p at the beginning of word, like pat, and p at the end of a word, like stop. But one is "aspirated" and one isn't, and in some languages they are different letters, so pat and pʰat could be two different words. The only difference is a puff of air following one and not the other.
They say in the babble of babies you can hear all the consonants of the world, but as we take on our mother tongue we lose the ability to hear or say many of them.
There are English dialects where the "g" sound is present - especially in the Northwest of England (think Mancunian, like Brian Cox, or Liverpudlian, like the Beatles in full-Scouse, lovable mop-tops mode). It's also prevalent among Yiddish-influenced speakers in the US (although that's mostly died out now).
> In German "ng" is always without a distinct g-sound (like German "Finger", or its n and g like "Ungeheuer")
Depends on the dialect. For example, older speakers of Swabian may totally say "Feng-ger" to "Finger", as opposed to the more "modern" Swabian variant of "Fenger".
My chorus director in college was always trying to get us to not use the /g/ when singing German. He often used the finger/singer analogy to explain the difference.
Not in word-final position, no, except (as noted by some commenters) through liaison with an initial vowel in the following word — an "intrusive g" phenomenon similar to the "intrusive r" found in many non-rhotic dialects.
Is there a linguistic name for this returning to the pronunciation or accent of your youth when drinking or around others with the same/similar accent?
I am also from Long Island, and have largely moderated my pronunciation to match a more standard pronunciation, but when I interact with natives of north Jersey, parts of NYC and Long Island, I somehow start adopting the accent a bit, unconsciously.
It’s most commonly found with African Americans and minority/immigrant communities, since in professional or unknown settings people will use more standard registers of English.
I don’t know the name for the transition, but you are talking about what are called registers in linguistics. (I do the same thing when back with old friends)
Also see code-switching. Which is something I have been very amused to find myself doing after moving back home to New Orleans - there's several distinct flavors of English in that city, and watching them pop out of my mouth after twenty five years speaking Newscaster English is fun.
Ha! I grew up on Long Island and was wondering why the Wikipedia say that "finger" and "singer" no longer rhyme in American Engish when I pronounce them exactly the same. It's because I learned most of my English on Lon Gisland!
(My accent is odd because I was born in Brooklyn and we spoke Yiddish at home, grew up on "Lon Gisland" then did college and beyond in Israel, and now spend about 2/3rd of time in US and 1/3 in Israel)
I had an experience like this learning about rhyming words in school when I was little.
"Hog", "log", and "bog" all rhymed, buy we couldn't figure out (not even the teacher) why "dog" (dowg?) sounded so much different. I was telling this story as a teenager when it suddenly occurred to me that I had a very strong accent.
By any chance, was your friend Italian? I distinctly remember some Italian speakers I've met pronouncing it exactly like that, "goiŋ-guh". Probably because they're used to most words ending in a vowel.
No, she was from a small southern town. She wasn't trying to pronounce the final vowel, just the final g. But it's impossible to say a full g without a small uh sound.
It was slight, like a pop. The irony is, she tried so hard to say the g without making too much of it, that she often removed its voice. The unvoiced g is k. So, most of the time she ended up saying goink.
Southern where? So far I've able to deduct not Italy, probably native English speaking, somewhere south. So South Africa? But then why be cryptic about it?
Sorry, I meant the southern United States. What is the term for when you're talking on the Internet and forget that you might be talking to someone halfway around the world?
Interestingly, although we think of [iŋ] as the standard pronunciation and [in] or [In] as low-prestige nonstandard variants, careful transcription reveals the alveolar variants frequently in even high-prestige dialects.
As I mentioned, it's part of the folklore of sociolinguistics that the [in] pronunciation exists, and that it's (apparently) not stigmatized in the way that the [ɪn] pronunciation is, and that in fact it may be heard as [ɪŋ] or [iŋ] and even transcribed that way in some studies.
As for singing diction, singing diction is prejudiced against high vowels (like [i]) to a sometimes absurd degree; you often hear the “-ing” suffix pronounced [ɛn] in popular music, because it lowers the vowel and thus increases the volume. The surprising thing is that this lowering is obvious to me, even though my dialect has the pin–pen merger, which would normally neutralize that contrast. A similar lowering transformation lowers things like "baby" [beɪbi] to the outlandish [beɪbɛ]. When I was a kid, this bothered me a lot. Now I no longer think there's anything wrong with it, but I think it's strong evidence that singing pronunciation has important differences from spoken pronunciation.
There was a Language Log post on a student project analyzing differences in the same speaker's language between musical performances and interviews, focusing in particular on the alternation between [ɪn] and [ɪŋ], at https://languagelog.ldc.upenn.edu/nll/?p=3687. There's some good discussion in the comment thread there about how sung pronunciations often reflect not only the exigencies of singing to project (perhaps more important before microphones, but still practiced today) but also musical-genre-specific diction traditions. Some discussion also ensues about the [i]/[ɪ] alternation.
> I find it fascinating that languages have different sets of sounds. For example, in English p and b are two different letters, and pat and bat are two different words. But other languages don't distinguish them, and their speakers really wouldn't notice a difference if you said pat or bat.
This is related to a question I've been wondering about for a long time.
One of the first stages in language processing is converting a sound's raw physical waveform into a sequence of phoneme tokens. Different languages divide different sounds differently -- as you note, [p] and [b] are distinct in many languages, but not all of them.
What I'd like to know is how nonexistent sounds get classified. Take [θ], the thorn sound. It is rare, but physically similar to several other common sounds -- the stop at the same place of articulation would be [t̪], the closest nearby tongue-based fricative is [s], and [f], while not tongue-based, is even more similar.
All of those are attested substitutions for the difficult [θ] sound. And they appear to be predictable -- English-speaking babies reliably substitute /f/, which they can pronounce, for /θ/, which they can't. Mandarin speakers hear /θ/ as /s/. Cantonese speakers hear it as /f/. (As, I believe, do Russian speakers.) In Ireland, the sound is replaced by some kind of /t/. But all of the languages I just mentioned offer the full set of /f/, /s/, and /t/ as available options.
To a certain degree this is conventionalized. I'd believe that modern Irish pronunciation is based more on what you hear from other Irish people than on a difficulty with thorn. And in English, the Mandarin sound /tɕʰ/ has sometimes been recorded as "ts" and sometimes as "ch", and if you pronounce it to untrained listeners and ask them what they heard, you'll get a mix of both answers. (Today it is less helpfully spelled "q".)
But nobody trains babies that if they can't produce a correct /θ/, the proper thing to do is to fall back on /f/, because /s/ would be a faux pas. The babies just do their best, and they all independently agree that their best approximation to /θ/ is /f/.
I think it should be possible, if you know the phonology of a language, to determine which phoneme (or possibly phonemes) a particular foreign sound would be understood as by a speaker of that language. As things stand today, it isn't -- we know that different languages will cause the same sound to be heard differently, and we know experimentally what the likely options are for a given sound, but we can't predict an individual mapping. And there's enough consistency in what people hear that we _should_ be able to do that.
Arabic has separate consonants for /ð/ (ذ) and /z/ (ز). But it also has a consonant ظ that can be pronounced as either /ðˤ/ or /zˤ/. For me as an English speaker trying to learn Arabic it was confusing because /ðˤ/ and /zˤ/ don't sound the same at all, but evidently to Arabic speakers they do and the pharyngealization makes all the difference.
Malay/Indonesian also turns fricatives into sibilants. So /θ/ becomes /s/ e.g. ath-thulathaa /aθ-θulaθa:ʔ/ (الثلاثاء) becomes Selasa. /ð/ would become a /z/ but that sound isn't native to most languages in that part of the world either, so it can become all kinds of sounds including /dʒ/ and even /l/!
Be careful to distinguish borrowings and substitutions. They are not the same.
Contemporary Indonesians substitute /ts/ for Arabic /θ/ if they have only so-so Arabic pronunciation. But this is on the basis of the contemporary transliteration of Arabic. Indonesians do not have the English speakers' concept of spelling - only transcription. In English, they're more likely to use /t d/ but sufficiently many people who speak English have learnt proper Arabic in school that they can pronounce the English sounds.
The use of /s/ is not a contemporary substitution, it's a historical borrowing. It's like saying "English substitutes french /s/ with /ʃ/" on the basis of words like "fashion". Yeah, at one point, French had an /s/ sound that sounds like English /ʃ/, but the pronunciations of both has changed and that doesn't happen any more.
The replacement of /ðˤ/ with /l/ in e.g. lohor is likewise a borrowing - but it seems to be ancient. ض is reconstructed as a velarised voiced lateral fricative in Wikipedia. And that borrowing is widespread; Indian Muslims going full Indian on me are surprised that I recognise "Lohor" from Indonesian. For my part, I pronounce it as thoohoo (book), but most native Australian English speaking Muslims seem to use /s z z/ for Arabic /θ ð ðˤ/ - they learnt the Arabic words from their Pakistani/Indian parents - and therefore say "zohaw" (hot). Even in recitation - Arabs seem to ask them to stop leading salah~sholat, but that hasn't happened to me.
Also, it's not actually /dʒ/ but more of a palatal stop. My wife - and even I - can certainly hear my accent in Indonesian. I also can't hear the difference between c and j unless I already recognise the word.
> Be careful to distinguish borrowings and substitutions. They are not the same.
They are the same. You're comparing substitutions that happened hundreds of years ago to substitutions that happen today. They follow different rules for obvious reasons, but they're not different things.
Sure, every language collapses multiple sounds into unified phonemes. In modern Mandarin, [v] is a kind of /w/. In Old English, [v] was a kind of /f/.
But that's a little different from what I asked about. These correspondences can be learned directly -- a child growing up to speak Old English would have known to treat [v] as a form of /f/ because he would have heard [v] all the time, as a form of /f/. The language had an explicit opinion, and [v] was a common sound.
But that isn't the case for foreign sounds. A child growing up in a Mandarin-speaking environment will never hear [θ] at all. Instead, the correspondence of [θ] to /s/ is an implicit rule, driven by the rules that actually do exist. I want to use the explicit rules, like [v] = /w/, to predict the implicit ones, like [θ] = /s/.
It's not aspiration, but rather the delay in onset of voicing. /b/ begins voicing before release, /p/ begins very close to release, and /pʰ/ begins after release.
You might also be thinking about the fact that, where English speakers would use /pʰ/, they tend to hear /p/ as /b/. So while an English speaker would say <pat> as /pʰat/, they will hear /pat/ as <bat>.
Right, /p/ in positions where it is contrastive with /b/ is distinguished by both voicing and aspiration. So if someone says [pæt] instead of [pʰæt], those two features are out of sync; the listener can either error-correct it to [bæt] or to [pʰæt]. (And this is why "b" is used in hanyu pinyin to represent unaspirated [p], and analogously "d" for unaspirated [t].)
You can do an experiment to find out. Try saying "spa" and "sba". In English plosives are not aspirated after an S. So if the only difference is aspiration then it should sound exactly the same.
The consonants are neatly divided into five categories, depending upon where the tongue touches the roof of the mouth/teeth/lips (back to front) - guttural (k as in kite, g as in go), palatal (ch as in chair, j as in jug), retroflex (t as in tap, d as in dog), dental (th as in thin, dh as in them), and labial (p as in pet, b as in bear).
Each of these also have an aspirated version, and there's a nasal at the end - so 5 in each category times 5 categories = 25 main consonants.
Then there are the vowels, semi-vowels, and spirants as different categories by themselves, each of which can be systematically tacked on to each of the consonants.
This systematic categorization of sounds and writing forms is amazing for a language that was developed a few thousand years ago.
There are of course some sounds that are not part of the system, like x, z and f. Some vowels too - like the a in bad, but all in all, the amount of thought put into the system is amazing.
The sounds "not a part of the system" are not native to Sanskrit, so they were added in with diacritics. And they're mostly done pretty sensibly: फ़ fa is फ pha with a dot, etc.
my unproven pet Theory is that Sanskrit still had the F sound but it was lost and re-interpreted as “i”:
पि
This theory is motivated by the observation that words with पि have an ( inclination towards) F in other branches of Indo-European (Also just the left-hand side of पि without the 'P' which cannot be pasted due to Unicode restrictions)
Accidentally learning IPA because I needed social studies credits in university is probably one of the best things that has happened in my life. It’s so useful to know how to correctly pronounce a word, instead of reading poor English approximations.
Not necessarily because pronunciation varies substantially across dialects and often also according to prosodic stress and phonetic context. For instance, the word "what" can be pronounced as /wɒt/ /wɒʔ/ /hwɒt/ /wʌt/ /hwʌt/ /wɑːt/ /hwɑːt/ and a few more...
Drifts also occur in time as well as space. Chaucer's early Middle English or Shakespeare's later Middle English are still mostly readable to a modern English reader, but a lot of the pronunciations have shifted almost entirely. There's a sort of archival benefit to the divorce between writing and pronunciation, the slower pace in which written forms drift versus spoken forms. (Which would also seem to hint why it's somewhat of a natural language outcome of speech being somewhat more ephemeral and prone to drifting than written text that outlasts it in libraries and schools.)
Which version of English shall we use in this simplification?
While English could probably withstand some changes to orthography, there are also advantages to the spelling being somewhat disjoint from the exact pronunciation.
When I was a teenager I remade the English alphabet taking freely from the ghost of English past and other Western European languages.
Ended up with, if I recall, about 47-53 distinct “letters”, although it was a flat hierarchy so a, æ, e, and è were all distinct letters from one another for sorting and reference purposes. Part of my motivation was to reduce the character count of individual words by combining digraphs and trigraphs into a single letter… so I could fit more on the Twitter. Secondary was to have one letter to one sound, although I don’t think I covered every sound in English.
When I soured on Twitter, I also lost interest in my modified alphabet, but my main takeaways: remaking the English alphabet from scratch can be done in about an afternoon, including a reference word list to act as a guide for new spellings. Your main problem is the only one that will be literate in your new alphabet is you, not an insurmountable problem in itself, but it’s probably not worth re-educating society for something you put together in an afternoon.
If you want to see a more successful manifestation of an alphabet put together in a way you suggest, try Hangul. It’s flexible enough to adapt to other languages as well.
For example, in Year 1 that useless letter "c" would be dropped to be replased either by "k" or "s", and likewise "x" would no longer be part of the alphabet.
The only kase in which "c" would be retained would be the "ch" formation, which will be dealt with later.
Year 2 might reform "w" spelling, so that "which" and "one" would take the same konsonant, wile Year 3 might well abolish "y" replasing it with "i" and iear 4 might fiks the "g/j" anomali wonse and for all.
Jenerally, then, the improvement would kontinue iear bai iear with iear 5 doing awai with useless double konsonants, and iears 6-12 or so modifaiing vowlz and the rimeining voist and unvoist konsonants.
Bai iear 15 or sou, it wud fainali bi posibl tu meik ius ov thi ridandant letez "c", "y" and "x" -- bai now jast a memori in the maindz ov ould doderez -- tu riplais "ch", "sh", and "th" rispektivli.
Fainali, xen, aafte sam 20 iers ov orxogrefkl riform, wi wud hev a lojikl, kohirnt speling in ius xrewawt xe Ingliy-spiking werld.
For those who don’t recognize it: That text is rather famous, and was written in 1971 by M. J. Shields¹. It has since been frequently mis-attributed to Mark Twain, who only wrote about a similar topic in 1899².
A common alphabet for Africa has been proposed for sounds not represented by the basic Latin alphabet of colonial languages such as French and English.
I am wondering that neurons in artificial neural networks might be learning it.. those trained model of voice AI solutions using artificial neural networks.... Is it possible to reverse engineer ANN model to check when it's firing neurons in the middle layers of the network.. not a Ann expert....
When I first started my career a senior dev told me a funny story about 2 competing standards for some obscure technology. There was no clear winner and the community couldn't decide what did so what did they do? They created a new standard no one used.
I don't understand how that relates. The international phonetic alphabet is pretty widely used in its niche. Read any decent article about phonetics, phonology, historical linguistics, and you'll see it.
English-language dictionaries and the reading textbooks I used in elementary school in the US in the 1980s use a different standard. A much worse one. And in ASCII-only media, there is X-SAMPA. eSpeak uses a variant of X-SAMPA; recent versions support IPA too, but only for output, and the conversion is lossy. Mandarin phonetics are often written with bopomofo, especially in elementary school, and Japanese people often use kana for writing down pronunciations—universally for elementary-school kids learning kanji, but sometimes even in contexts so inappropriate it's self-sabotage, like learning English pronunciation. And of course many languages have sufficiently systematic orthography that you don't need a separate convention for indicating pronunciation, except for linguists investigating dialect variation or foreign-language students.
But these are really marginal cases. The IPA is where it's at.
> English-language dictionaries and the reading textbooks I used in elementary school in the US in the 1980s use a different standard.
Perhaps I'm being too nit-picky, but that's actually more just like an alternate English alphabet than it is a phonetic alphabet. The goal of a phonetic alhpabet is to map symbols any sound that is distinguishable in any language, not simply the ones that are distinguishable in English.
X-SAMPA is a lot more like the IPA. It's basically an incomplete encoding of IPA in ASCII. It's a little hack to get around the insufficiencies of inputing characters into computer devices.
For Uralic linguistics, there is the Uralic phonetic alphabet[0]. While relatively few new texts are ever created in it (partly because all the dialectal fieldwork was already done in the late 19th/early 20th century), it is the transcription used in all the standard references that linguists in this subfield still use and cite today.
There's the Americanist phonetic alphabet (https://en.wikipedia.org/wiki/Americanist_phonetic_notation), which saw widespread use in North and Central America in the early parts of the 20th century. I think it's mostly been replaced by the IPA.
For the most part, those funny little phonemic notations that dictionaries come up with are just alternate alphabets, and not intendend to be a standard phonetic alphabet.
Sometimes we clip the g, and say goin'. Then there is the standard way to say it, goiŋ. Then there is the way she was saying it, goiŋ-guh.
I find it fascinating that languages have different sets of sounds. For example, in English p and b are two different letters, and pat and bat are two different words. But other languages don't distinguish them, and their speakers really wouldn't notice a difference if you said pat or bat. On the other hand, I don't hear the difference between p at the beginning of word, like pat, and p at the end of a word, like stop. But one is "aspirated" and one isn't, and in some languages they are different letters, so pat and pʰat could be two different words. The only difference is a puff of air following one and not the other.
They say in the babble of babies you can hear all the consonants of the world, but as we take on our mother tongue we lose the ability to hear or say many of them.