Early Spanish cookbooks (1894–1920)
Four Spanish cookbooks from 1894 to about 1918 trace how home cooking in Spain moved from thrift and leftovers, through regional and national pride, to the modern household. Each book has a Fundamentals page, with its body of knowledge and concept map, and a Recipes page with twelve recipes transcribed from the scanned original.
| Book | Author | Year | Scan |
|---|---|---|---|
| El Practicón | Ángel Muro | 1894 | BNE Digital |
| La cocina práctica | Picadillo (Manuel M.ª Puga y Parga) | 1908 ed. (1st ed. 1905) | BNE Digital |
| La cocina española antigua | Emilia Pardo Bazán | [1913?] | BNE Digital |
| La cocina moderna | Carmen de Burgos (prologue and arrangement) | [ca. 1918] | BNE Digital |
All four are public-domain scans from the Biblioteca Nacional de España. Recipes keep their original Spanish names and quantities, with the source page for each.
Across the four books, and how the recipes were transcribed
The four books share an argument about Spanish identity but split on France: Muro and Pardo Bazán write against French dominance, Picadillo against French fashion, while Carmen de Burgos mixes French and foreign dishes in without comment.
- Thrift. Muro's leftovers appendix and Picadillo's "de a pie" cooking treat economy as a virtue; Burgos frames it as household management.
- Region and nation. Picadillo writes from Galicia; Pardo Bazán assembles regions into a national cuisine and borrows from both Muro and Picadillo.
- The same dishes recur. Bacalao á la vizcaína appears in three of the books (Muro, Picadillo, Pardo Bazán); Muro's and Pardo Bazán's versions share almost word for word the sauce of fried onion, toasted bread, choricero peppers and flour. Gazpacho and the cocido or puchero also appear more than once.
- Women and the kitchen. Pardo Bazán turns to cookery after her feminist series stalls; Burgos presents cooking as a learned, professional skill.
- Measures. All four mix old units (libra, onza, cuartillo, cuarterón) with grams and litres; Burgos's plum-cake is given wholly in grams.
Scan quality and OCR model per book
| Book | Source used | Clean words in BNE OCR | Model | Unreadable words |
|---|---|---|---|---|
| Picadillo | BNE OCR text | 87% | Claude Haiku | 2 |
| Pardo Bazán | BNE OCR text | 85% | Claude Haiku | 0 |
| Muro | BNE OCR text | 78% | Claude Haiku | 2 |
| Carmen de Burgos | PDF page images | 39% (unusable) | Claude Sonnet | 1 letter |
"Clean words" is the share of OCR tokens that are plain words, measured on each full text. Where the library's OCR was good, the cheapest model only had to clean it up; for Burgos the scan was sharp but the library's OCR had failed, so a stronger model read the page images directly. One recipe per book was checked against the source text or page image and matched.