Data in this repository attempts to identify the original language words, in both the Old and New Testaments, translated as quotations (material using "double" and 'single' quotation marks) in various English Bibles. It also attempts to associate speakers with the quotations, where possible, using data from Faith Comes By Hearing (see Sources.md).
This data is an initial effort, and as with many datasets, it can be sharpened. We plan some improvements, but the data as it currently exists is also useful so it is time to release it for those who might be interested in it.
Each Quotation (see Quotation Data below) declares a speaker of the quotation using the field CharacterId or SPEAKER. Information about each of these speakers is available as:
- JSON:
json\character_detail.semantic_data.json. - TSV:
tsv\character_detail.semantic_data.tsv.
This information is a reproduction of FCBH Character Detail information with extra fields added for SDBH and LN, where appropriate. This information includes:
- CharacterId — The identifier of the speaker; this is consistent across OT and NT. This aligns with "SPEAKER" in TSV or JSON that assigns a speaker to a quotation.
- MaxSpeakers — The maximum number of speakers that speak when this speaker is activated. A value of
-1appears to indicate a group of unknown size. - Gender — The gender of the speaker. This primarily indicates whether FCBH intends a male or female voice for the named speaker.
- Age — These are ranges, like "child" or "adult". These primarily indicate whether FCBH intends a child, an adult, or possibly an older adult to voice the speaker.
- Divinity — This field denotes divinity. Items with
Yas value are divine. Otherwise the field is empty. - Comment — Unstructured information regarding the speaker, this field is typically blank.
- FCBHCharacter — An alternate identifier used by FCBH.
- SDBH — A list of possible SDBH (Semantic Dictionary of Biblical Hebrew) identifiers associated with the gloss/speaker. This is a superset and does not presently disambiguate between same-named but different speakers (e.g. "Zechariah").
- LouwNida — A list of possible Louw-Nida (Greek-English Lexicon of the New Testament based on Semantic Domains) identifiers associated with the gloss/speaker. This is a superset and does not presently disambiguate between same-named but different speakers (e.g. "Zechariah").
Please note that we plan future updates to sharpen and expand the data contained in the SDBH and LouwNida fields. These fields will not always have data as there are many instances of unnamed speakers (e.g., "two other disciples" in John 21:3 and 21:5).
Data reflecting the use of quotations in English translations has been processed and created in several steps. Each step may have use, so interim formats have been stored and documented. These include:
alignment-quotations-[translation].tsv— These are tab-delimited files that provide information about quotations (meaning text in double and/or single quotation marks) in a particular translation, but mapped back to the original language (Hebrew, Aramaic, or Greek) based on alignment data available to Clear Bible.SpeakerProjections-[translation].json— Translation-specific quotation material that is mapped to speaker data available from FCBH. This is an initial attempt to identify speakers for word ranges in the original language text. These are the richest form of data with lists of word objects associated with speaker data. Please note: Just because something is “quoted” in a translation does not mean it should have a speaker assigned and be considered a quotation.[translation]-aligned-projections.tsv— Tab-delimited files generated fromSpeakerProjections-[translation].json. Rather than lists of words, these data include start and end with the assumption of contiguous ranges.
Data for the following English translations are available:
- ESV
- NIV84
- NASB95
- NASB77
- NRSV
- CSB
- HCSB
- LEB
New-Testament-Only data representing projections in the OpenText.org analysis are available as well.
Note also that a "consensus" form of the data, representing largest agreement of quoted material, is available with the translation named "Clear". This does not contain or represent an actual translation; instead it simply models a consensus view of quotations in the Bible when comparing information from the available translations.
Files include:
/tsv/alignment-quotations-clear.tsv/tsv/alignment-quotations-csbe.tsv/tsv/alignment-quotations-esv.tsv/tsv/alignment-quotations-hcsb.tsv/tsv/alignment-quotations-leb.tsv/tsv/alignment-quotations-nasb77.tsv/tsv/alignment-quotations-nasb95.tsv/tsv/alignment-quotations-niv84.tsv/tsv/alignment-quotations-nrsv.tsv- There is no OpenText.org data in this format.
These files are generated by a process that evaluates an alignment between an English translation (such as the NIV84) and the original languages which underlie it. The process evaluates the English for single and double quotation marks. It tracks where quotations open and close as well as the level of quotation.
By level (or depth), we mean if the text (level 1) contains, say, direct speech (level 2) which contains another citation (level 3). An example of text with this level of quotation is found in Genesis 3:1 (here in the Lexham English Bible [LEB]):
- LEVEL 1 Now the serpent was more crafty than any other wild animal which Yahweh God had made. He said to the woman,
- “LEVEL 2 Did God indeed say,
- ‘LEVEL 3 You shall not eat from any tree in the garden’
- ?”
- “LEVEL 2 Did God indeed say,
Different translations have different editorial practices which lead to different implementations of quoted material, and thus different levels of quotation.
Regarding accuracy of data, the data is only as accurate as the quotation marks encoded in the alignment data that this process analyzes. For example, there are known issues with the quotation marks in the Lexham English Bible (LEB). These shortcomings affect the derived quotation data.
The tab delimited data is essentially a list of each original language word unit with information about the reference, word, and quotation level. The word stream is interrupted with an indicator of quotation start and stop. The fields are as follows:
- depth — The depth of the quotation, as described above. Text defaults at level 1; any level above 1 indicates quoted material of some sort. A level of 0 (zero) indicates that the original language word was not directly aligned with a word in the translation, so its presence or absence in the quotation cannot be confirmed.
- id — A numeric identifier taken from the alignment data the extracted quotation material is based on. The numeric identifier includes information on the book number, chapter number, verse number, word number, and word part:
BBCCCVVVWWWP. - clearId — The identifier to resolve in Clear Bible's MACULA data. This is almost exclusively the same numeric identifier as the
id, but witho("Old Testament") orn("New Testament") prepended. - USFM — A modified form of a USFM (Universal Standard Format Marker) Bible reference. It includes a Bible book abbreviation (e.g. "GEN") with book, chapter, verse and word numbers. This data is modified with a dash after the word to also include word part.
- textContent — The Hebrew, Aramaic, or Greek word at word position in the original language data.
The word stream is broken to delimit start and end of quotations (at level 2) using +++ and ---, in manner like the following example from Genesis 3:1–2:
1 010030010101 o010030010101 GEN 3:1!10-1 אֱלֹהִ֑ים
1 010030010111 o010030010111 GEN 3:1!11-1 וַ
1 010030010112 o010030010112 GEN 3:1!11-2 יֹּ֙אמֶר֙
1 010030010121 o010030010121 GEN 3:1!12-1 אֶל־
1 010030010131 o010030010131 GEN 3:1!13-1 הָ֣
1 010030010132 o010030010132 GEN 3:1!13-2 אִשָּׁ֔ה
+++ ==> START QUOTE
2 010030010141 o010030010141 GEN 3:1!14-1 אַ֚ף
2 010030010151 o010030010151 GEN 3:1!15-1 כִּֽי־
2 010030010161 o010030010161 GEN 3:1!16-1 אָמַ֣ר
2 010030010171 o010030010171 GEN 3:1!17-1 אֱלֹהִ֔ים
3 010030010181 o010030010181 GEN 3:1!18-1 לֹ֣א
3 010030010191 o010030010191 GEN 3:1!19-1 תֹֽאכְל֔וּ
3 010030010201 o010030010201 GEN 3:1!20-1 מִ
3 010030010202 o010030010202 GEN 3:1!20-2 כֹּ֖ל
3 010030010211 o010030010211 GEN 3:1!21-1 עֵ֥ץ
3 010030010221 o010030010221 GEN 3:1!22-1 הַ
3 010030010222 o010030010222 GEN 3:1!22-2 גָּֽן
--- ==> END QUOTE
1 010030020011 o010030020011 GEN 3:2!1-1 וַ
1 010030020012 o010030020012 GEN 3:2!1-2 תֹּ֥אמֶר
1 010030020021 o010030020021 GEN 3:2!2-1 הָֽ
Files include:
/json/SpeakerProjections-clear.json/json/SpeakerProjections-csbe.json/json/SpeakerProjections-esv.json/json/SpeakerProjections-leb.json/json/SpeakerProjections-nasb77.json/json/SpeakerProjections-nasb95.json/json/SpeakerProjections-niv84.json/json/SpeakerProjections-nrsv.json/json/SpeakerProjections-OpenText.json(though the schema differs slightly; usesGreekWordinstead ofOriginalWord)
This JSON is the most descriptive form of data representing word-level quotation data mapped to verse-level speaker data available from Faith Comes By Hearing (FCBH). There are two files from FCBH that are relevant.
- CharacterDetail.txt — Detail information on each unique named character or entity assigned quotation material in the Biblical text.
- CharacterVerse.txt — Verse level assignment of named character or entity as speaker.
This JSON provides character level data from both of these files merged together for a reference range assigned to a particular speaker/character/entity. It also provides word-level quotation data (also known as "projections") for quotation ranges within the reference range. This data uses a "speaker key", which is a concatentation of the first verse, last verse, and speaker (e.g. GEN 1:3|GEN 1:3|God) which then has a SpeakerInstance and Projections (which is a list of ProjectionData) associated with it.
| Property | Type | Description |
|---|---|---|
| StartVerse | string | USFM start verse of range |
| EndVerse | string | USFM end verse of range |
| Verses | List of strings | List of verses in range, in order |
| CharacterIds | List of strings | List of FCBH CharacterIds; first is default |
| Delivery | string | delivery from FCBH data |
| Alias | string | alias from FCBH data |
| QuoteType | string | quote type from FCBH data |
| DefaultCharacterId | string | Default / preferred FCBH character Id |
| ParallelPassage | string | not exhaustive or reliable but sometimes useful. From FCBH data |
| Alternates | List of SpeakerInstance | Some areas are ambiguous and the speaker is not certain so alternates are included |
| Property | Type | Description |
|---|---|---|
| StartWord | string | identifier for word in Hebrew or Greek MACULA |
| EndWord | string | identifier for word in Hebrew or Greek MACULA |
| ClearSubjectReferents | List of strings | identifiers for word(s) in Hebrew or Greek MACULA that resolve either implicitly or explicitly stated potential speakers |
| ClearSubjectReferentLabels | List of strings | glosses or labels in English for word(s) in Hebrew or Greek MACULA that resolve either implicitly or explicitly stated potential speakers |
| Words | List of ProjectionWord |
Specific lists of words in order |
| Depth | integer | Depth of quotation. Initial text level is 1; incremented for subsequent levels of quotation. |
| StartVerse | string | USFM start verse of projection |
| EndVerse | string | USFM end verse of projection |
| Verses | List of strings | List of verses in range, in order |
| Property | Type | Description |
|---|---|---|
| Id | string | identifier for word in Hebrew or Greek MACULA |
| Text | string | Hebrew, Aramaic, or Greek word (or word part) |
| Gloss | string | English equivalent gloss of Text, not included in data presently |
| Depth | integer | Depth of quotation. Initial text level is 1; incremented for subsequent levels of quotation. |
Files include:
/tsv/Clear-aligned-projections.tsv/tsv/CSBE-aligned-projections.tsv/tsv/ESV-aligned-projections.tsv/tsv/LEB-aligned-projections.tsv/tsv/NASB77-aligned-projections.tsv/tsv/NASB95-aligned-projections.tsv/tsv/NIV84-aligned-projections.tsv/tsv/NRSV-aligned-projections.tsv/tsv/OpenText-aligned-projections.tsv(no alternate speaker data)
Tab-delimited files generated from SpeakerProjections-[translation].json. Rather than lists of words, these data include start and end with the assumption of contiguous ranges. The data consists of the following fields:
- KEY
- START VS
- END VS
- SPEAKER (FCBH)
- ALT SPEAKER (FCBH)
- SPEAKER REFERENT (CLEAR)
- SPEAKER REFERENT LABEL (CLEAR)
- QUOTE TYPE
- QUOTE DELIVERY
- PROJECTION START
- PROJECTION END
- CLEAR START
- CLEAR END
Note that the KEY may be duplicated if a range assignment consists of multiple projections but one speaker. Below is an example of this (GEN 18:13) as well as inclusion of alternate speaker information (GEN 18:9).
GEN 18:3|GEN 18:5|Abraham (Abram) GEN 18:3 GEN 18:5 Abraham (Abram) o010170260051 Abraham Normal GEN 18:3 GEN 18:3 o010180030021 o010180050132
GEN 18:3|GEN 18:5|Abraham (Abram) GEN 18:3 GEN 18:5 Abraham (Abram) o010180020061 men Normal GEN 18:5 GEN 18:5 o010180050151 o010180050181
GEN 18:6|GEN 18:6|Abraham (Abram) GEN 18:6 GEN 18:6 Abraham (Abram) o010180060021 Abraham Normal GEN 18:6 GEN 18:6 o010180060071 o010180060141
GEN 18:9|GEN 18:9|Abraham (Abram) GEN 18:9 GEN 18:9 Abraham (Abram) o010180020061 men Normal GEN 18:9 GEN 18:9 o010180090031 o010180090052
GEN 18:9|GEN 18:10|angels at Sodom, two GEN 18:9 GEN 18:10 angels at Sodom, two God o010180070041 Abraham Normal GEN 18:9 GEN 18:9 o010180090071 o010180090083
GEN 18:9|GEN 18:10|angels at Sodom, two GEN 18:9 GEN 18:10 angels at Sodom, two God o010180010031 Yahweh Normal GEN 18:10 GEN 18:10 o010180100021 o010180100102
GEN 18:12|GEN 18:12|Sarah (Sarai) (old) GEN 18:12 GEN 18:12 Sarah (Sarai) (old) o010180120021 Sarah Normal GEN 18:12 GEN 18:12 o010180120051 o010180120111
GEN 18:13|GEN 18:15|God GEN 18:13 GEN 18:15 God o010180130021 Yahweh Normal GEN 18:13 GEN 18:13 o010180130051 o010180140101
GEN 18:13|GEN 18:15|God GEN 18:13 GEN 18:15 God o010180150021 Sarah Normal GEN 18:15 GEN 18:15 o010180150041 o010180150051
GEN 18:13|GEN 18:15|God GEN 18:13 GEN 18:15 God o010180130021 Yahweh Normal GEN 18:15 GEN 18:15 o010180150091 o010180150111