Silesian Folk Belief

Editorial Method

The texts collected here were taken down between roughly 1850 and 1930 by folklorists working directly from spoken accounts. What they recorded is not literary prose. It is speech: unpunctuated, repetitive, digressive, held together by "and" and "then" rather than by sentence structure. That is not a defect of the record. It is the record.

Sources

Every text is transcribed from the original printing or scan from official online libraries. No modern edition, anthology or existing translation has been used, and each entry carries its full reference so that it can be verified against the source.

Prejudice in the sources

Some of these texts carry the prejudices of the society that produced them — against Jews above all, and also against Travellers, neighbouring people and the poor. Those passages are not removed, softened or quietly edited. They are part of the record: they show what the people who told these stories believed, and what the persons who collected them thought worth writing down. Reproducing a text is not endorsing it. The material is presented as a historical document, to be read as evidence of its own time rather than as a statement addressed to ours — and where a prejudiced attribution appears, it is marked in a note, because in most cases it is a detail of one telling rather than a part of the story itself.

Orthography

The original spelling is kept, including the long ſ in the German texts and the phonetic notation of the Polish and Silesian ones, where collectors used diacritics to record pronunciation rather than standard spelling. These marks carry information about how a word sounded in a particular village. Normalising them would quietly discard it.

Punctuation

The originals are frequently printed without sentence divisions. In the source text this is left untouched. In the translation, sentence breaks and punctuation are added. This is the one substantial editorial intervention, and it is made deliberately: without it the reader meets a wall of words and stops reading, which serves neither the story nor the person who told it. Where a break could fall in more than one place, it is placed to preserve ambiguity rather than resolve it.

Translation

The aim is fidelity to the speaking voice, not to literary English. Repetition is kept. Loose connectives are kept. Where a speaker uses a German loanword inside a Polish sentence, or the reverse, the mixture is part of the evidence and is noted rather than translated away. Obvious printing errors are corrected only where the reading is certain, and marked where it is not.

A note on certainty

These texts are the transcription of a transcription. A speaker told a story aloud, in a dialect that shifts between Polish, Czech and German from one village to the next and was never standardised. A collector wrote down what they heard, in a phonetic spelling of their own devising. That printed page was later scanned, and the scan read again, word by word, sometimes against nothing more than a single occurrence anywhere in the source. Each of those steps adds a layer of interpretation of its own — this is not a flaw introduced at the end of the chain, it is the condition of the material itself.

The editorial process therefore distinguishes between three different operations: transcription, which aims to reproduce what is printed; interpretation, which attempts to establish what an obscure or dialectal form means; and translation, which renders that established or provisional meaning into English. These operations are not interchangeable. A word may be transcribed with confidence while its meaning remains uncertain, or its meaning may become clear from context even when the printed form is difficult to read.

Context, parallel passages and comparison with other occurrences in the same collector's corpus may be used to resolve obscure forms. Reader input may also contribute to a reading. Such evidence is treated as evidence, not as automatic correction: a plausible interpretation is not silently promoted to certainty. Where a word's sense rests on a single attestation, a plausible etymology, contextual reconstruction, reader input, or an educated guess rather than a settled reading, that is stated in the entry's translator's note rather than folded silently into confident English.

A reading is revised when better evidence turns up, and the note says so. The translator's notes are not an appendix to the real work; for a dialect this fragmentary, they are part of it.

Process

Translations are drafted with AI assistance and then reviewed, corrected and adapted by hand against the source text. Illustrations are drawn entirely by hand in charcoal, graphite or pencil on paper, without AI-generated imagery or third-party visual reference material. Original 3D sculpts made in Nomad Sculpt by myself serve only as pose, shape and shading references for the drawing process. Finished drawings are digitised by camera and post-processed in Krita. No text or image on this site is published without editorial review by myself.

What is not done

Nothing is abridged, combined, modernised or improved. Variants are not merged into a single "best" version. Where a story is strange, incomplete or ends without resolution, it is left that way.

More