-
Notifications
You must be signed in to change notification settings - Fork 3
03 Transcription and Transliteration Policy
In IE-CoR, each lexeme in each language should be represented on up to four levels :
- In the normal, original orthography for that language, e.g. Cyrillic or Devanāgarī.
- If that orthography is not a Roman script, then also in an established transliteration.
- In IPA phoneMic transcription, following an established phonological analysis of that language.
- In IPA phoneTic transcription, to the level of distinguishing major allophones.
As an example of the four levels for a single language, see: http://www.cobl.info/language/AncientGreek/wordlist/Jena200/
In orthography, transliteration and transcription, IE-CoR faces some contrasting imperatives. We need to respect already long-established conventions specific to a given language and how it is represented. But we also need to make the database public, so that it can be interpreted and analysed by all, in consistent ways.
The best balance between local conventions and universal standards is different, then, on each of our four representation levels:
- In the orthography, it is fine and expected that you will stick closely to language-specific (or region-specific) conventions in spelling.
- In the transliteration into Roman script, again some language-specific or local usage will often be necessary, although when in doubt, please tend to usage that is more modern and more compatible with other languages.
- Since the phoneMic level is still abstract, although it may be possible to tolerate some language-specific conventions, that is to be avoided wherever possible. Please aim for as much standardisation as possible around the precise IPA symbol for the major, default or ‘elsewhere’ allophone of that phoneme.
- On the phoneTic transcriptions should use only, and strictly, the symbols of the IPA, in Unicode, not any idiosyncratic language-or region-specific conventions.
- Please use the citation form most consistent with that used in most other languages. So in Latin, for example, verbs are given not in the first person singular, even though that is traditional in Latin dictionaries, but in the infinitive, for consistency with most other languages.
- Please do not add extra free (unbound) morphemes that are extraneous to the root, just because they are often used in citation forms. For example, in English verbs are not to be cited with preceding 'to', i.e. 'see', not 'to see'. IE-CoR is fundamentally about root cognacy, and about lexicon, not the grammar of whether the infinitive form in most contexts requires a supporting other free morpheme like 'to'. Note that no normal dictionary of English cites all verbs under 'to …'. IE-CoR should not do so either.
All orthographic entries should be given in lower case. If there are already some entries for your language given in upper case, then they must be corrected to lower case. The only cases where words should be allowed with an initial upper case letter are those where the standard orthography for that language requires it, i.e. on nouns in German. So German Haus but English house.
Entries in upper case should also ring alarm bells: these are very old, legacy and often very unreliable entries. Pease pay special attention to these entries, to consider whether they are really the appropriate lexemes. You should have no hesitation in excluding or correcting them as necessary. Do not consider that as existing entries they need to be kept!
IE-CoR aims to represent and be in line with the leading scholarship on each of the languages covered. Often, specialists in particular languages have their own conventions for spelling, transliterating and transcribing individual languages. These differ from script to script, and language to language. Ancient languages, of course, often had their own dedicated but also idiosyncratic scripts.
The main, orthographic entry for a lexeme is thus always to be given in whichever is the most established and widely used standard for that language.
Where there are competing alternatives, the general preference is for the more modern one, and the one more compatible with orthographies for other languages, especially if closely related.
IE-CoR is a very broad comparative database, spanning languages that use a multitude of scripts, including ancient inscriptions. Very few viewers of the website are likely to be highly familiar with all of them. So for any language whose orthography is given in a non-Roman script, we also need a transliteration into Roman script.
There can be competing approaches for a single language, e.g. for transliterating Russian into Roman script. In such cases, the expert for that language should choose the most widely used and respected. This will often be the most recent alternative, and an ISO system. If in doubt, contact the IE-CoR administrators.
The IE-CoR database will also be used for qualitative and quantitative searches and analyses in both phonetics and phonology, e.g. to find all tokens in all languages of a particularly sound, or sound sequence. That is why we also ultimately need both phonemic and (major allophone) phonetic transcriptions of all lexeme entries. It is also why transcriptions on both levels need to be as consistent and standardised as possible.
If absolutely necessary, and since the phoneMic level is still abstract, some language-specific conventions may sometimes be tolerated here -- but please check first with the IE-CoR administrators. Nonetheless, the preference is definitely for as much standardisation as possible around the precise IPA symbol for the major, default or ‘elsewhere’ allophone of that phoneme – not just the simplest symbol to type on a keyboard!
This is because users cannot be expected to familiar with language-specific conventions in every language.
So for example, in Standard British English, the phoneme often represented /r/ is of course not a trill, and should ideally be ɾ represented /ɹ/ instead, since [ɹ] is its default allophone.
On IE-CoR’s precise and strict usage of IPA, see the next section below.
Obviously, phoneTic transcriptions need to be in symbols of the International Phonetic Alphabet, and IPA only, nothing else.
Please do not use other regional approaches such as the traditional ‘Americanist’ transcriptions, even if they are common for researchers in your language. So for example, please:
- Use [y] (not ü) for the rounded front high vowel.
- Use [ʃ] (not š) for the voiceless palato-alveolar fricative.
- Use the dedicated raised characters such as labialised [ʷ] and palatalised [ʲ], not superscript of a normal w and j (not possible anyway in our web entry form).
For more examples, see under the conventions for Indo-European in the next section below.
Please also take care to use the correct, precise IPA symbols where there are multiple similar ones, e.g.
- Use [β] (not 'Eszett' ß) for a voiced bilabial fricative, and its lowered equivalent for the corresponding approximant
- For accent marks and diacritics, use the correct combining characters, not the standalone characters with space.
- Some diacritics have multiple variants in Unicode, for thinner target characters like [i], or wider ones like [m]).
- Please mark vowel length, with the correct IPA symbols [ː] (not colon :), [ˑ], [ ̆ ]
- Please do mark primary stressed syllables, with the correct IPA symbol [ˈ].
- Please be accurate with exact Unicode symbols like apostrophes, rough breathings, etc..
- Please distinguish correctly between palatal stops (e.g. [ɲ]) and (secondary articulation) palatalised stops (e.g. [nʲ]).
There are of course occasional cases where many linguists do not follow IPA usage, e.g. in marking tone. If you are minded to use any notation different to IPA, however, please check with the IE-CoR administrators first.
Wikipedia has many good, detailed pages on IPA phonetic symbols and their corresponding Unicode character(s). Just type en.wikipedia.org/wiki/ followed by the symbol character: <e.g. en.wikipedia.org/wiki/ʃ>.
You can also search for characters on the Wikipedia page on the IPA, and just copy individual symbols from there. That should give you the Unicode symbol to paste directly into the IE-CoR website.
Another useful 'trick' is to paste a single symbol into a Microsoft Word document, select it, then do Insert —- Symbol. That will give you a description and the Unicode reference number for that character.
If you need further advice on setting up dedicated keyboard shortcuts for special characters, ask the IE-CoR administrators.
Each branch of Indo-European may have its own conventions for transcription and transliteration, so please see the following pages for each.