Wiktionary talk:IPA pronunciation key

Definition from Wiktionary, the free dictionary
Jump to: navigation, search

Seeing IPA symbols correctly[edit]

Can anyone give any help with the settings, or whatever, one has to adjust in order to see the IPA symbols correctly on screen? TIA.

The symbols are represented correctly on my browser. You need to make sure that your browser is set to reed Unicode (UTF-8). If that doesn't work you need to download the character set. Eclecticology 05:30, 20 Oct 2003 (UTC)
Which character set? Volvox 19:35, 18 May 2004 (UTC)
Any public domain from where download a true type or free type set for Unicode ??
DejaVu is my favourite. Yngwin 11:36, 19 July 2006 (UTC)

Width of table[edit]

The width of this table needs to be reduced. The primary purpose of this article is to illustrate the use of IPA for the understanding of the English speaker. That makes it sensible to show all English IPA usages that exist. The presence of other languages here should serve that purpose, especially to illustrate sounds that do not exist in English, or are not phonemic in English. The full representation of other languages here should be limited to languages that will significantly contribute to that goal. Most Spanish sounds are easily reproduced by English speakers, so that language does not need full representation to merely illustrate the few problem sounds. Arabic and Hindi, however, present interesting phonetic challenges and should probably be extensively illustrated. In addition to English I would show maybe six columns devoted to specific languages and a wider final column for illustrations in other languages.

Eventually there should be separate pages to illustrate the IPA representations of each language, and how it compares to English sounds. That belongs on some other pages, not here. Eclecticology 05:30, 20 Oct 2003 (UTC)

Strictly speaking, Spanish and English don't share any consonant sounds, and as a native English speaker, I think it's a bit challenging to pronounce the Spanish v (β). However, for the purposes of an English-language dictionary, I would agree with you. – [[User:Mxn|Minh Nguyễn (talk, blog)]] 02:18, 4 Oct 2004 (UTC)
Precisely because Spanish has different consonant sounds from English, it is important to keep Spanish.  One major obstacle English speakers face in learning Spanish is the failure to pay attention to these differences.  The purpose of the I in IPA is to represent correctly these differences.  I recommend deleting other languages containing only sounds already covered.  I recommend adding languages with sounds not in English or Spanish and expanding the table vertically to included all their phonemes, languages like Korean, Arabic, Amharic and Georgian.  Howard McCay (talk) 20:59, 3 October 2012 (UTC)

Need jpegs of symbols[edit]

Can we have jpegs of the symbols so I can make sure my browser and fonts are set correctly? Maybe on a seperate table. --Gbleem 04:40, 25 Mar 2004 (UTC)

Should include interlanguage links[edit]

I suggest include the interlanguage links. This is, instead of [[él]], use [[:es:él|él]]. We could use a floating yellow tag, to indicate "Spanish/Español" ( in native wiktionary version - english wiktionary - and native word language version - spanish wiktionary-). --Anon

Some users writes [[él]] ([[:es:él|-]]) to get both a link to the native wiktionary and to the English. By the way, I suppose you talk about the translations tables? And I don't really get the point of the "floating yellow flag" - what would it be used for? \Mike 10:29, 21 Jul 2004 (UTC)

some sounds I don't know the spelling for[edit]

Please, can someone mention how do you represent the marked sounds in the words below. This is mostly necessary to describe some Romanian words, though I believe others might need it too.

  • constantza
  • cistern
  • horse
  • John

This is how I would spell them in IPA:


 – [[User:Mxn|Minh Nguyễn (talk, blog)]] 02:16, 4 Oct 2004 (UTC)

For constantza I would write [kənˈstæntsə]. [tz] does not occur in English (maybe it does in Romanian). (I won't argue about the 2nd vowel, but it's traditionally transcribed as æ except when transcribing dialect)
The sound of 'h' in horse is in fact [h] but the 'orse' should be [ɔɹs] (at least for en-us). In en-gb it would be [ɔːs]. ɚ is generally used for the vowel at the end of butter. Nohat 19:40, 4 Oct 2004 (UTC)

missing minimal contrastive pair[edit]

There's a missing contrastive phoneme for english amongst the vowels: the shorter æ versus the longer æː. In British and Australian English, at least, these are contrasted in contractions of the names Madolin and Gladys ("Mad", "Glad") against the adjectives for angry and happy ("mad", "glad"). Is this worth adding an extra row for?

Separate languages[edit]

It would be much more helpful to readers to figure out pronunciations if we kept separate pages for every language. Cramming so much information together in table form has major limitations in that there's no room make the transcriptions phonemic, which is the best solution in most cases. There would be plenty of room to explain the summarize the most important allophones and dialectal variations as well as clearing up how prosodic features such as tone and word accents should be transcribed.

Karmosin 11:17, 6 February 2006 (UTC)

The problem is what to call such pages. Other Wiktionaries go with things such as "en.word" or "de:word" but here we don't want to force people to remember a bunch of language codes.
In the future we hope for a way to show only the languages each user is interested in. But so far we're still too small in the scheme of Wikipedia to be able to ask the developers to implement such things for us.
In the meantime, others may tell you to have faith for the much prophecied day that "ultimate wiktionary" appears and solves every problem. — Hippietrail 16:59, 6 February 2006 (UTC)
I'm starting to lose faith in the Ultimate Wiktionary, especially in the concept emerging from nothing. It makes more sense to start from what we have, assessing the varied formats in all Wiktionaries. As to the languages that have en.x and de.x pages (or colon, etc.) what does the page x itself look like? Davilla 12:49, 9 February 2006 (UTC)
I don't really know what you're referring to by the "xx.word"-things, but the problem certainly seem less complicated to me. All we need is a main pronunciation guide page for linking to individual pronunciation guides. Perhaps this one or maybe something along the lines of Wiktionary Appendix:Pronunciation. Then you just attach a link to the correct language page next to the IPA (or other) transcriptions in the articles. Naming seems to me like the least of our problems.
Karmosin 08:29, 8 February 2006 (UTC)
Sorry, I misunderstood. Somehow I thought you were saying we should only have one language per page for all normal entries. Now that I see what you're really talking about I agree with you. — Hippietrail 23:10, 8 February 2006 (UTC)

Mistakes in Dutch examples[edit]

I noticed several mistakes in the Dutch examples, so if you don't mind I'm going ahead and edit them... Yngwin 11:36, 19 July 2006 (UTC)

Excellent resource... and recommended changes[edit]

First, I must say this is an excellent resource. Yes, it is a bit wide, but I’d rather have it be too wide than not at all. I think this would work very well as an Excel spreadsheet where you could freeze the top row while you scroll down so you can still see the language names. Also you could hide columns you don’t need.

Second, recommended changes:

Add Australian English.

Add rhotic vowels (please) – tar, tare, a tear, turn, tour.


  • In Spanish the diphthong in seis is not /e (Alt-618 or little capital i)/ but /ei/. Let me tell you, it makes a big difference when you want to say peine (comb) and you pronounce it like English, and the next thing you know, people think you just asked to use someone’s penis!
  • Spanish: Need to add /au/ as in “auto.” The ʊ (handlebar "u" or Alt-650) sound doesn’t exist in Spanish.
  • Swedish ɛ (Alt-603) and ɜ (Alt-604): These are both given for the pronunciation of häll. If they’re both correct, I recommend including a footnote 4.
  • What’s this: ɤ as in Vietnamese “xương”? It looks like a short upside-down black AIDS ribbon. I thought it was the sound for g in Spanish algo, but I guess that's a tall upside-down black AIDS ribbon. By the way, I vote for more mouse-over names.


  • The ɣ (sound for g) as in Spanish “algo” is only pronounced this way in certain countries.
  • What about adding ɹ (Alt-633 or upside-down r), which is how Rs are pronounced in English? They’re definitely not the same as the r in Spanish.

Wishful thinking: I wish someone would acknowledge the widespread pronunciation of “dr” (as in "dry") which is not d+r but like j in "jump" + an English r or /ʤr/(Alt 676 + r/). And also “tr” as in “try” in English and “tres” in Costa Rica and Chile. They're not pronounced /tr/ but as a ch + r /ʧr/ (Alt 679 + r). Food for thought. DBlomgren 01:18, 24 October 2006 (UTC)

I acknowledge that widespread pronunciation. However, it is not /ʧr/ with phonemic slashes, it is [ʧɹ] obsolete or nonstandard characters (ʧ), invalid IPA characters (ʧ), replace ʧ with t͡ʃ with phonetic brackets.
I think that some of these can be taken out, Arabic should go from the vowels section because it only has three vowels, some of which aren't even on this list. I'm not sure if Catalan is different enough from Italian or notable enough to include (why not Lingala or Swahili?). We've also got to face the fact that nobody's going to touch Ukrainian.
The list does a poor job of indicating which sounds are phonemes and which are contextual allophones. I think that it ought to be limited to just phonemes unless the status as only an allophone is somewhat dubious (like with german ch or Russian ы)
I'm not so hot on the ordering, which seems pretty nonsensical to me. It doesn't have to be alphabetical, heck Spanish and Japanese could go together. Ƶ§œš¹ IPA: [aɪm ˈfɻɛ̃ⁿdˡi] 19:49, 20 December 2006 (UTC)
To both of you: be bold! henne 13:20, 25 December 2006 (UTC)

Mistakes in greek examples[edit]

I noticed some mistakes in the greek examples, so if you don't mind I'm going ahead and edit them... --Lou 22:20, 5 March 2007 (UTC)

gray cells[edit]

I thought a nice idea would be to gray out cells that a language doesn't have so that it's more obvious what needs to be filled. I started it out but I don't know all the details for many of these languages. Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:18, 12 November 2007 (UTC)


Knowing all I know about phonologies of multiple languages, I only see the tables in this page continuing to expand until they are not only ridiculously wide and tall, but even more sparse than they are now, and that much more difficult to browse due to sheer size of dimensions. I think this page should just cover general IPA descriptions (like with the IPA standard reference charts), and provide a list of links to other pages of pronunciation keys for individual languages. It is conceivable, however, that that because this is the English multilingual Wiktionary, the English IPA guide (including British, Irish, American, Canadian, Australian, etc. articulations) can be provided here, or at least linked to near the top. - Gilgamesh 11:43, 14 November 2007 (UTC)

Yes, there are too many languages right now and there's no criterion for exclusion. How should we change this? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 21:17, 24 November 2007 (UTC)
I think Omniglot provides a good model of inspiration. - Gilgamesh 23:39, 24 November 2007 (UTC)
You need to be more specific. What part of Omniglot has a table of sounds? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 02:00, 26 November 2007 (UTC)
Most of the pages linked here. Omniglot actually helped me memorize the finer nuances of IPA through providing examples for language phonologies I already knew. I am of course not saying that we copy Omniglot's data. It just provides a useful model of inspiration for demonstrating pronunciations of spelling in different languages. A giant monolithic table (without clear criteria for inclusion or exclusion) will inevitably grow too big and without getting any less sparse. Instead of making each language a table column, it would be better to dedicate separate tables for each language. See Wiktionary:About Greek/Pronunciation for an example in development. - Gilgamesh 04:09, 26 November 2007 (UTC)
So then we convert this page to a link hub for all the different language pronunciation pages? Ƶ§œš¹ [aɪm ˈfɻɛ̃ⁿdˡi] 07:29, 29 November 2007 (UTC)
That would certainly seem like a more efficient use of Wiktionary's inherent resources. - Gilgamesh 19:13, 29 November 2007 (UTC)

RFD discussion[edit]

Green check.svg

The following information passed a request for deletion.

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

I'm not quite sure how this page would be useful. We already have pronunciation keys for individual languages, so this page seems to be used mostly to compare pronunciations in several different languages. I see several problems with this approach as of right now... The list of languages is rather small and seems biased towards European languages (why Finnish but not Arabic, Mandarin or Hindi?), and there are too many languages and too many specific details in each language to fit all of them into one table. The table also doesn't consider the issue of phonemicity. Long vowels are not phonemic in French yet are listed as if they are, while the reverse is done for Dutch. And consider what would happen if Arabic were added: would words containing /a/ be added to that column even when the allophonic range of that phoneme can go anywhere from [ɛ] to [ɑ]? I would consider this 'comparing apples and oranges' at best. —CodeCat 20:20, 8 March 2012 (UTC)

I've always used the page as a one-stop pronunciation lookup, to get an idea of how the symbols are pronounced. Of course, Wikipedia has more extensive converage of all the symbols, and a one-stop-shop is only useful to those who speak several of the listed languages (monolingual people can use monolingual pages). I'm on the fence about (perhaps weakly in favor of) keeping the page. I'm strongly in favor of keeping the shortcut WT:IPA, though, to point to information on the IPA symbols (even if that means making it a soft redirect to WP). - -sche (discuss) 21:20, 8 March 2012 (UTC)
w:de:Liste der IPA-Zeichen is a pretty nice list of IPA symbols. If someone could translate it and expand it with missing letters, we would have the perfect replacement right there. -- Liliana 23:27, 8 March 2012 (UTC)
Isn't that redundant to WT:IPA already? —CodeCat 00:12, 9 March 2012 (UTC)
Delete/merge with Wiktionary:International Phonetic Alphabet. There's clearly a lot of hard work gone into this page, sadly not all of it accurate (it claims that in French on and son don't rhyme, d'oh) but how can we justify including some languages and not others. It looks to me like something that should be on someone's user page, where they can be as POV as they like. Mglovesfun (talk) 19:50, 9 March 2012 (UTC)
I want to keep the link to this on WT:IPA so that someone searching for IPA has a comparative one-stop shop that has collected a lot of work. I think it should stay on a user page or user sub-directory until it is slightly more universal. Below is a list of 20 languages i use for starting point Swadesh lists.
ar-Arabic, bn-Bengali, de-German, el-Greek, en-English, es-Spanish, fa-Persian, fr-French, he-Hebrew, hi-Hindi, id-Indonesian, ja-Japanese, kr-Korean, nl-Dutch, pt-Portuguese, ru-Russian, tr-Turkish, sw-Swahili, ta-Tamil, zh-Chinese
I started with the six (6) w:Official languages of the United Nations, ar, en, es, fr, ru, & zh. The mere six (6) de facto official w:Languages of the African Union, ar, en, es, fr, pt, & sw only require adding in the last two. The four (4) official languages of the w:Union of South American Nations, en, es, nl, & pt only require adding nl. To cover the rest of the Americas, the four (4) official languages of the w:Organization of American States, en, es, fr, & pt are all already included. In detail, the w:Central American Integration System official language of es; the w:Caribbean Community official languages of en, es, fr, & nl; and the w:North American Free Trade Agreement official languages of en, es, & fr are all already included. Australia's one 1 official language, en, is already included and also the working language of the w:Pacific Islands Forum. Antarctica is desolate. Eurasia is less integrated. Because of the unwieldy number of 23 w:Languages of the European Union (EU), the over-representation of European languages already and to come, and the fact that the most widely spoken language, en, is already included, the only two (2) added to the list are the most natively spoken language, de, and the only one (1) with a script not yet represented; also the backbone of Christianity; el. The only (1) official language of the w:Commonwealth of Independent States (CIS), the w:Eurasian Economic Community, and the w:Customs Union of Belarus, Kazakhstan and Russia is the already included ru, which is also co-official and mutually intelligible with be-Belarusian in the two-language (2) w:Union State. Because much of the geographical extent of the CIS contains people familiar only with trk-Turkic languages, the largest member of the language family, tr, is added and also the official language of the w:Turkic Council. The one (1) official language of the w:Arab League as well as the w:Cooperation Council for the Arab States of the Gulf, ar, is already included as well as being the backbone of Islam but the language, he, of the country that disconnects the Arab League geographically, Israel, has a script that's not yet represented and is the backbone of Judaism and of import to Christianity and of interest to Islam; as such a language added is he. The one (1) official language of the budding w:Economic Cooperation Organization, en, is already included but the predominant language, fa, is inter-intelligible with Tajik & Dari and the backbone of Zoroastrianism, Yazdanism, Ahl-e Haqq, Sufism, Babism, & the Bahai Faith; as such a language added is fa. The only one (1) official language of the w:South Asian Association for Regional Cooperation (SAARC), en, is already included but the most widely spoken language, hi, is inter-intelligible with ur-Urdu; the largest member of the predominant language sub-family of the region, Indo-Iranian; besides its script, Devanagari, is not yet represented and the backbone via Sanskrit of Hinduism, Buddhism, Sikhism, & Jainism; as such a language added is hi. The one (1) working language of the w:Association of Southeast Asian Nations, en, is already included, but the largest member, id, of the predominant language family map-Austronesian is inter-intelligible with Malay; as such an included language is id. The last three nations of Asia, China, Japan, and Korea cover the already represented language, zh, the backbone of Taoism & Confucianism as well as scripts not yet represented which are therefore added in ja & kr, which also round out the w:ASEAN Free Trade Area#ASEAN_Plus_Three of China, Japan, & South Korea. Covering linguistic diversity through these w:Supranational unions covers most families, geography, religion, and script through only gives eighteen (18) languages; the other two (2) came from looking at w:Global Internet usage#Internet_users_by_language (en, zh, es, ja, pt, de, ar, fr, ru, ko), which re-justified de, ja, & ko; w:Linguistic demography#Most_spoken_languages (zh, hi, es, en, ar, bn, ru, pt, ja, de), which highlighted the next added language from the SAARC with its not yet represented script, bn; w:List of languages by total number of speakers (zh, en, es, hi, pt, ar, bn, ru, fa, pa; fr, en, ru, pt, ar, es, fa, zh, de, ja; zh, en, es, ru, fr, hi, pt, ar, bn, fa), where each top ten entry was included except pa-Punjabi, which is not national and does not have a distinct script or a great deal of difference from hi/Urdu; w:List of languages by number of native speakers (zh es en hi ar bn pt ru ja pa), as before; w:World language#Living_world_languages (en es fr zh hi ar pt ru de id fa sw ta it nl ja bn), which highlighted the next added language, ta, from the not yet represented dra-Dravidian family and script as well as it-Italian, which does not have a distinct script or a great deal of difference from es. Four (4) national scripts notably absent are the Ge'ez alphabet of am-Amharic as well as ti-Tigrinya, the hy-Armenian alphabet, the ka-Georgian alphabet, and the Thaana abugida of dv-Dhivehi. Also notably absent are the indigenous languages of Australia with its largest family, the ~codeless Pama-Nyungan (e.g., nys-Nyoongar); those of South America, (e.g., qu-Quechua, official in Bolivia); and those of North America the (e.g., azc-Uto-Aztecan nah-Nahuatl, ~codeless Na-Dene nv-Navajo, aql-Algic abe-Western Abenaki, iro-Iroquoian moh-Mohawk, & esx-Eskimo-Aleut iu-Inuktitut official in Canada's Nunavut Territory. For more information, see w:Lists of languages. Warmest Regards, :)—thecurran Speak your mind my past 05:58, 30 March 2012 (UTC)


I still maintain that this page can never do what it intends to do, because it does not consider phonemicity. From experience with talking to others I know that subphonemic differences are often not even apparent to speakers, which makes it hard for them to interpret this table. And since it only has a few languages in it, it is very selective and POV-ish. —CodeCat 17:48, 31 October 2012 (UTC)

Keep. - -sche (discuss) 21:13, 31 October 2012 (UTC)
Keep. The nominator makes a good case, but I'm afraid the table is just too useful. ~ Nelg (talk) 00:52, 20 February 2013 (UTC)

kept as no consensus. --ElisaVan (talk) 17:31, 5 October 2013 (UTC)

Glottal stop is missing[edit]

I couldn't find the glottal stop symbol in the chart. Could somebody please add it? Floppy1974 (talk) 00:24, 30 December 2013 (UTC)

None of the languages in the list have a glottal stop phoneme. This chart is really bad anyway, it's very incomplete and it can never be completed because there are too many languages. —CodeCat 00:35, 30 December 2013 (UTC)
That's not exactly true. English speakers use a glottal stop instead of t in certain positions depending on the dialect. There's even a Wikipedia article about it (en.wikipedia.org/wiki/T-glottalization). In American English t-glottalization is fairly specific, as mentioned in the article, but it's also fairly standard; if a true t is used in words like "mitten" or "Manhattan" it sounds awkward and stilted. T-glottalization is less specific in British English but no less common. That said, I agree with you about the chart itself. Ed Oppty (talk) 16:21, 27 April 2014 (UTC)
She said none of the languages in the list have a glottal stop phoneme. The glottal stop in English is an allophone of /t/, not a phoneme. —Aɴɢʀ (talk) 18:21, 27 April 2014 (UTC)
I think that really shows what is fundamentally wrong with this list as a concept, though. There is no way we could ever arrange the phonemes of the world's languages in a table like this. The glottal stop is common worldwide, but out of the languages listed, only Arabic has it. Other phonemes which are pretty common, like [q], are only represented by Arabic as well, and [χ] is curiously missing altogether. Breathy voicing is not represented either, despite being widespread in India (hundreds of millions of speakers). The retroflex consonants from this area are also missing. What is perhaps even more confusing is that there is widespread dialectal variation in many of these languages, and the sounds may be nowhere near their "canonical" IPA representations in many of them. But because the words are listed on the same line, the impression is given that the sounds are somehow equivalent. English soon and German Kuh don't have the same vowel at all, nor do Dutch deur and German Ökonom. And someone who pronounced Catalan neu with the same diphthong as Dutch eeuw would get some odd stares, I'm sure. Languages are far too complicated to simplify them into a single coherent phonemic system even within one language. Trying to do it across all languages in the world is madness. —CodeCat 18:49, 27 April 2014 (UTC)
The English Wikipedia entries on the individual phonemes do a much better job at explaining their use in various languages, at least in my opinion. -- Liliana 19:22, 27 April 2014 (UTC)

Background color of "-" in tables[edit]

Is there any difference between white background of "-" cells in tables and gray background? Should they be the same color?



Vitalik (talk) 08:57, 8 May 2014 (UTC)

[χ] vs. [x] distinction[edit]

Currently, the table lists the German word Buch under [x]. However, strictly speaking the correct pronunciation is [χ], which is currently not listed in the table at all. Is it worth adding a new row for? — Timwi (talk) 06:10, 22 June 2016 (UTC)

Buch normally has [x] rather than [χ]. See the last paragraph of w:Standard German phonology#Ich-Laut and ach-Laut for the distribution of the two allophones of the ach-Laut. That said, there probably should be a row for [χ], which could use Bach as an example. Other languages listed in the table that could be used to adduce examples are Dutch, Spanish, French, Swedish, Arabic, and Portuguese. —Aɴɢʀ (talk) 10:42, 22 June 2016 (UTC)
Dutch is a difficult case because the backness of the velar varies quite widely by dialect. —CodeCat 13:39, 22 June 2016 (UTC)
The footnote section shows that we already have some example words that apply only to specific Dutch dialects. Just pick one whose χ is most unambiguously uvular. —Aɴɢʀ (talk) 15:16, 22 June 2016 (UTC)