Each Elan file have to contain the following tiers:
| NB! | Tier in ELAN | Comments |
|---|---|---|
| ! | Speaker_Transcription-txt-andi | text |
| Speaker_Lemma-txt-andi | lemmas | |
| Speaker_POS-txt-en | Info associated with the lexical item e.g. Part of speech | |
| ! | Speaker_Morph-txt-en | division into morphemes |
| ! | Speaker_Gloss-txt-en | English glosses |
| ! | Speaker_Translation-gls-... | free translation into ... language |
| Speaker_Gloss-txt-... | glosses in ... language | |
| Speaker_Translation-gls-... | free translation into ... language | |
| Speaker_Participant-note-... | comments on ... language |
The name of each layer has a clear structure:
The name of the speaker and the name of the layer are separated by a lower underscore, all other fields by a hyphen. You can only change the speaker's name and ISO language code.
[speaker_name of the layer]-[layer type]-[code of the writing system]
What are the [layer type]:
- txt - the main line of transcription (Baseline)
- gls - for the sentence, this type means free translation
- lit - literal translation
- note - notes
[code of the writing system] --- ISO code (can be found here: https://www.ethnologue.com/browse)
NB! There can be several layers in the markup with the same name, but different languages. For example, if you have a translation into two languages (for example, Russian and English), the names of the layers will look like this:
- Speaker_Translation-gls-en
- Speaker_Translation-gls-en
The Transcription layer does not depend on any layers. The layers Morph, Translation and Participant depend on Transcription. The Gloss layer depends on the Morph layer. If you have a speaker A and the source language is Andi (andi), the markup will look like this:
picture
As a separator, only spaces are used, but they should not be used anywhere in the layer. Separator for the gloss is a hyphen, everything else needs to be separated by something else (for example, Baba_Yaga).
In the template you can use word-by-word division, as, for example, here:
picture
Types of all layers:
| title | layer type |
|---|---|
| A_Transcription-txt-andi | - |
| A_Morph-txt-andi | Symbolic Subdivision |
| A_Lemma-txt-andi | Symbolic Association |
| A_Gloss-txt-en | Symbolic Association |
| A_POS-txt-en | Symbolic Association |
| A_Translation-gls-ru | Symbolic Association |
| A_Participant-note-en | Symbolic Association |
The list of possible fields in the metadata table lies in meta_template.csv. Bold in those columns that should be required. The metadata file must be in the .csv format and have the following appearance (each line is one speaker in the file):
| file_title | speaker_id | ... | ... | ... |
|---|---|---|---|---|
| andi_pirozki_1.eaf | A | |||
| andi_pirozki_1.eaf | B | |||
| andi_pirozki_2.eaf | A | |||
| andi_pirozki_2.eaf | B | |||
| andi_pirozki_2.eaf | C |
A file with glosses is a table (.csv format), in which first there is what is in the markup, and in the next column it is written which large category this gloss refers to. For example:
| voc | case |
|---|---|
| adnum | case |
| m | gender |
| f | gender |
| n | gender |
| mf | gender |
| tran | vType |
| intr | vType |
| pf | aspect |