Bíblia-Belem — Blog - The Blame is on the Sheep

Bíblia Belem AnC 2025 — The Method Behind the Translation

Belem Anderson Costa — Sun, 01 Feb 2026 00:00:00 +0000

Public source text: WLC (Westminster Leningrad Codex) + Nestle 1904. Translation: Bíblia Belem AnC 2025 — literal, rigid, straight from the public códices.

The Translation That Was Missing

There are dozens of Bible translations in Portuguese. Almeida Corrigida. NVI. NVT. NTLH. Almeida Atualizada. Each one made editorial choices — softened here, harmonized there, interpreted elsewhere. All of them deliver to the reader a processed text.

The Bíblia Belem AnC 2025 is different. It delivers the text raw. Morpheme by morpheme. Without softening. Without harmonization. Without implicit interpretation. The reader receives exactly what the códices say — in rough, uncomfortable, and radically faithful Portuguese.

It is the first rigid literal translation in the Portuguese language. The first of its kind.

The Accepted Códices

The translation works exclusively with public domain códices in the original languages. No Latin. No secondary translations. Only the oldest verifiable sources.

Old Testament

Codex	Abbreviation	Description
Westminster Leningrad Codex	WLC	Standard academic Masoretic text — Hebrew + Aramaic

The WLC is based on the Codex Leningradensis (c. 1008 AD), the oldest complete Masoretic manuscript in existence. It is the basis for virtually all academic editions of the Hebrew OT (BHS, BHQ).

New Testament

Codex	Abbreviation	Usage
Nestle 1904	NA1904	Critical text — primary source
Westcott-Hort 1881	WH	Critical text — comparison source
Textus Receptus 1550	TR	Ecclesiastical text — comparison source

The primary source for the NT is Nestle 1904 — a critical edition by Eberhard Nestle based on the collation of Tischendorf, Westcott-Hort, and Weymouth. It is public domain and academically rigorous.

WH 1881 and TR 1550 are used for comparison and recording of textual variants. When there is divergence between texts, the Bíblia Belem AnC records the variant.

REJECTED Source

Source	Status	Reason
Latin Vulgate	REJECTED	Derived translation, not a primary source. Contaminated by ecclesiastical editorial decisions
Any modern translation	REJECTED as source	Translations are derivations — the Belem AnC works only with primary sources
Non-public domain manuscripts	NOT USED	Verifiability requires public access

The Translation Method

Step 1: Source Text Identification

The translator identifies the Greek or Hebrew text in the public domain codex. There are no intermediaries.

Step 2: Morphological Analysis

Each word is morphologically analyzed:

Root/lexeme — dictionary form
Tense/mood/voice (Greek verbs) or binyan (Hebrew verbs)
Case/number/gender (nouns, adjectives, pronouns)
Prefixes and suffixes (especially relevant in Hebrew)

Step 3: Morpheme-by-Morpheme Translation

Each morphological unit receives a Portuguese correspondence. Word order from the original is preserved when possible. When Portuguese grammar requires minimal reordering, it is done — but indicated.

Step 4: Preservation of Designations

Divine designations are kept in their original script with transliteration:

Θεός (Theos), Κύριος (Kyrios), Χριστός (Christos)
יהוה (yhwh), אלהים (Elohim), אדני (Adonai)

Step 5: Zero Interpretation

The translator does not add interpretive notes in the body of the text. Does not soften strange constructions. Does not harmonize apparent contradictions. If the original text is ambiguous, the translation preserves the ambiguity.

What the Reader Finds

The experience of reading the Bíblia Belem AnC is deliberately different from any other translation:

What the reader expects	What the reader finds
Fluid and pleasant text	Rough and literal text
“God,” “Lord,” “Christ”	Θεός, Κύριος, Χριστός
Reorganized sentences	Original order preserved
Embedded interpretation	Zero interpretation
Explanatory footnotes	No interpretive notes

This is intentional. The discomfort is a pedagogical tool. When the reader stumbles on a strange construction, they are forced to investigate. When they encounter a Greek designation, they are forced to research. The text does not deliver answers — it delivers questions.

And questions are the engine of all investigation.

The Canon: 66 Books

The Bíblia Belem AnC works with the Protestant canon of 66 books — 39 from the Old Testament and 27 from the New Testament. The deuterocanonical/apocryphal books are not included.

Testament	Books	Original Language
Old Testament	39	Hebrew + Aramaic (parts of Daniel and Ezra)
New Testament	27	Koine Greek
Total	66	3 languages

The Author of the Translation

Belem Anderson Costa is not a theologian. He is a police officer, developer, and studied Letters — without completing the degree.

Competence	Application in Translation
Critical textual analysis	Rigorous examination of códices
Morphology	Decomposition of words into morphemes
Syntax	Analysis of Greek and Hebrew sentence structure
Semantics	Mapping of fields of meaning
Pragmatics	Communicational context of passages

The degree in Letters — not seminary — is deliberate. The translator does not carry the weight of a denominational tradition. He was not trained to read the text from a specific perspective. He acquired competencies to analyze the text as text.

Easter Egg #7: The surname “Belem” (Βηθλέεμ — Bethleem) is a transliteration of the Hebrew בֵּית לֶחֶם (Beth Lechem — “House of Bread”). The author carries in his name the same city where the biblical text records the birth of Ἰησοῦς. The suffix “An.C” in the translation refers to “Antes de Cristo” (“Before Christ”) — but inverted: the translation goes from Christ (from the códices) to the present. “Belem AnC” is, therefore, a signature: from the House of Bread, from before Christ, until now.

The Public API

The Bíblia Belem AnC does not exist only as static text. It is available via a public REST API:

URL: https://biblia.aculpaedasovelhas.org

Endpoint	Function
`/api/v1/books`	List of all 66 books
`/api/v1/verses/:book/:chapter`	Verses of a chapter
`/api/v1/verses/:book/:chapter/:verse`	Specific verse
`/api/v1/verses/search?q=term`	Text search
`/api/v1/tokens/:verseId/interlinear`	Interlinear text (Greek/Hebrew + Portuguese)
`/api/v1/tokens/:verseId/morphology`	Token-by-token morphological analysis

The API allows any developer, researcher, or student to programmatically access the Belem AnC text. Integrate with your own systems. Build tools. Verify each translation.

The API is built with TypeScript (Hono framework) and hosted on Cloudflare Workers with a D1 database. The code is open source.

Open Source: CC BY 4.0

The Bíblia Belem AnC 2025 is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). This means:

Anyone can copy and redistribute in any format
Anyone can adapt, remix, and build upon the material
For any purpose, including commercial
As long as proper attribution is given

The reason is simple: if the translation is faithful to the original text, it must be tested by the greatest possible number of people. Access restrictions protect the translator — not the truth. Open source exposes the translator to scrutiny — and that is good.

If there is an error, it will be found. If there is bias, it will be identified. If there is imprecision, it will be corrected. Because public scrutiny is the greatest purifier of Truth.

Integration with exeg.ai

The Bíblia Belem AnC is the textual corpus of the exeg.ai platform. When the user asks a question to the AI, it consults directly the Belem AnC text — not another translation.

The platform offers:

Semantic search — finds similar passages by meaning (FAISS)
Interlinear analysis — Greek/Hebrew text + literal translation side by side
Easter Egg Engine — detection of lexical patterns between passages
Intertextual mapping — traceable OT/NT connections

All based on the rigid literal translation. The AI does not soften, does not harmonize, does not interpret. Just like the translation.

The Invitation

The Bíblia Belem AnC 2025 is not for everyone. It is for those who accept the discomfort of literalness. For those who prefer a rough but faithful text to a fluid but interpreted one. For those who want to investigate rather than consume.

Each reader becomes an investigator. Each verse becomes a piece of evidence. Each reading becomes a forensic act.

The text is open. The códices are public. The translation is verifiable. The method is documented.

All that is missing is the investigator.

“You read. And the interpretation is yours.”

Rigid Literality — Translating Without Interpreting

Belem Anderson Costa — Sun, 01 Feb 2026 00:00:00 +0000

Public source text: WLC (Westminster Leningrad Codex) + Nestle 1904. Translation: Bíblia Belem AnC 2025 — literal, rigid, straight from public códices.

The translator’s dilemma

Every translation faces a choice: fidelity to the original text or fluency in the target language. The more faithful to the original, the less fluent. The more fluent, the more interpretive.

Most biblical translations in Portuguese chose fluency. The Bíblia Belem AnC 2025 chose fidelity. Total. Non-negotiable.

This produces a text that sounds strange in Portuguese. Sentences that do not flow naturally. Constructions that demand effort from the reader. And this is intentional. Because the goal is not linguistic comfort — it is access to the original text.

What rigid literality is

Rigid literality means translating morpheme by morpheme, not meaning by meaning. Each minimal unit of meaning in the Greek or Hebrew text receives a direct correspondence in Portuguese.

Approach	What it does	Result
Dynamic equivalence	Translates the general meaning of the sentence	Fluid, interpretive text
Formal equivalence	Translates word by word	Reasonably literal text
Rigid literality	Translates morpheme by morpheme	Raw text, without editorial treatment

Rigid literality is the most extreme degree of fidelity to the original. The translator does not soften. Does not harmonize. Does not “improve” the text for the modern reader. He delivers the text as it is — raw, rough, unpolished.

Concrete examples

The differences are visible and measurable:

θηρίον (therion)

Translation	Choice	Problem
Almeida Corrigida	“besta” (beast)	Adds pejorative charge that the Greek does not possess
NVI	“besta” (beast)	Same problem
Bíblia Belem AnC	“fera” (wild animal)	Literal translation of θηρίον — wild animal, without value judgment

The Greek θηρίον simply means “wild animal, beast.” The Portuguese translation “besta” carries a semantic charge of stupidity, moral brutality, repugnance — none of these connotations exist in the Greek. The translator who chooses “besta” has already interpreted before translating.

ἄγγελος (angelos)

Translation	Choice	Problem
Most translations	“anjo” (angel)	Implies winged celestial being — concept not present in Greek
Bíblia Belem AnC	“mensageiro” (messenger)	Literal translation — someone who carries a message

The word ἄγγελος means “messenger.” It can be human, it can be celestial — the Greek does not specify. When the translator writes “angel,” he has already decided that the messenger is a celestial being. That decision is interpretation, not translation.

ἐκκλησία (ekklesia)

Translation	Choice	Problem
Most translations	“igreja” (church)	Implies organized religious institution
Bíblia Belem AnC	“assembleia” (assembly)	Literal translation — group summoned for gathering

In classical and koine Greek, ἐκκλησία is simply an assembly of summoned citizens. There is no temple. There is no hierarchy. There is no denomination. The translation “church” projects 2000 years of institutionalization onto a text that describes gatherings of people.

σταυρός (stauros)

Translation	Choice	Problem
Most translations	“cruz” (cross)	Assumes specific shape without textual basis
Bíblia Belem AnC	“estaca” or “madeiro” (stake/timber)	Literal translation — σταυρός = vertical post

The Greek σταυρός designates a vertical post, a stake. The cross shape (with horizontal crossbar) is a later artistic tradition. The Greek text does not specify the shape. The literal translation preserves this ambiguity that the original possesses.

The principle of non-softening

Conventional translations apply three editorial processes that the Bíblia Belem AnC rejects:

1. Softening

Making the text more pleasant for the reader. Example: transforming rough Hebrew constructions into fluid sentences in Portuguese.

Rejected. If the Hebrew is rough, the Portuguese of the translation will be rough.

2. Harmonization

Making apparently contradictory passages seem consistent. Example: adjusting divergent genealogies.

Rejected. If the códices diverge, the translation reflects the divergence. The reader decides what to do with it.

3. Implicit interpretation

Choosing a translation that already implies an interpretation. Example: translating πνεῦμα (pneuma) as “Spirit” with a capital letter in certain contexts and “spirit” with a lowercase letter in others — a decision that the Greek does not make (Greek does not have upper/lowercase in the same way).

Rejected. The translation does not make interpretive decisions for the reader.

The sovereignty of the reader

This is the central principle: the reader has absolute sovereignty over interpretation.

The translation delivers the raw text. Any processing — softening, harmonization, interpretation — is done by the reader, not the translator.

This inverts the power dynamic. In conventional translations, the translator makes hundreds of micro-interpretive decisions that the reader never sees. The reader receives an already-processed product and believes they are reading “what the text says.” In reality, they are reading what the translator decided the text says.

Easter Egg #5: The expression λίθον λευκὸν (lithon leukon — “white stone”) in DES 2:17 receives in the literal translation exactly those two words: “white stone.” Conventional translations sometimes add explanatory notes about Roman voting customs. The literal translation does not add them. The reader finds “white stone” and researches on their own. The explanation does not come embedded — because explanation is interpretation.

The text as an untouched crime scene

In criminal forensics, the crime scene must be preserved intact. Each piece of evidence must remain where it was found. No one reorganizes the scene so it “makes more sense” for the photographer.

Rigid literal translation operates by the same principle. The original text is the crime scene. The translation is the photograph of the scene. If the photograph is “improved” — objects reorganized, stains cleaned, artificial lighting — it loses probative value.

Conventional translation	Rigid literal translation
Reorganizes the scene for the reader	Photographs the scene as it is
Cleans the textual “stains”	Preserves every stain
Adds interpretive lighting	Shows with natural light
Produces a beautiful image	Produces a faithful image

A beautiful translation can be a lying translation. An ugly translation can be the true translation.

The exeg.ai platform and literality

The artificial intelligence platform exeg.ai follows the same principle as the translation: it does not apply normalization automatically.

When the user asks a question about a passage, the AI searches the data of the Bíblia Belem AnC and returns the literal text. It does not soften. Does not harmonize. Does not interpret. It offers tools — semantic search, lexical analysis, intertextual mapping — and the user decides what to do with the results.

The AI is a microscope. It is not a pathologist. The report belongs to the investigator.

Practical comparison: DES 13:1

Let us see how different translations handle the same verse:

Greek (Nestle 1904):

Καὶ εἶδον ἐκ τῆς θαλάσσης θηρίον ἀναβαῖνον, ἔχον κέρατα δέκα καὶ κεφαλὰς ἑπτά

Almeida Corrigida:

“And I saw rising from the sea a beast that had ten horns and seven heads”

NVI:

“I saw a beast that came out of the sea. It had ten horns and seven heads”

Bíblia Belem AnC 2025:

“And I saw out of the sea wild-animal ascending, having horns ten and heads seven”

The Belem AnC translation preserves:

θηρίον = “fera” / “wild-animal” (not “beast”)
The Greek word order: “horns ten and heads seven” (not inverted for fluency)
The preposition: “out of the sea” (ἐκ τῆς θαλάσσης — the exit is emphasized)

Is the text less fluid? Yes. Is it more faithful to the original? Also yes. And that is the choice rigid literality makes — in every verse, in every word, without exception.

The first of its kind in Portuguese

The Bíblia Belem AnC 2025 is the first rigid literal translation in the Portuguese language. Literal translations exist in English (such as Young’s Literal Translation of 1862), but in Portuguese this approach had not been attempted with this degree of rigor.

The project is open source, licensed under CC BY 4.0, and the API is publicly available at Bíblia.aculpaedasovelhas.org. Anyone can verify the translation choices. Anyone can contest. Anyone can propose corrections.

Because rigid literality is not a dogma — it is a method. And methods are refined by public scrutiny.

“You read. And the interpretation is yours.”