*
On May 29, 2025, I tried to edit the English Wikipedia page on the history of machine translation, adding a few lines about Pucci. Here's a screenshot of the edit:
Here is the full text:
In December 1929, the Italian inventor and linguist Federico Pucci presented for the first time in Salerno, Italy, his study on a “mechanical translator”. Two years later, in 1931, he published a 68-pages book, entitled “Il traduttore meccanico ed il metodo per corrispondersi fra Europei conoscendo ciascuno solo la propria lingua: Parte I.”, which is probably the first text in the world about an automatic translation device and the first ever documented rule-based machine translation (RBMT), with two examples of an Italian text from Dante translated according to his method into French and also a French text from Voltaire translated into Italian. From the preface to the reader, written in Salerno on 10 December 1930, the author tells us that he intends to demonstrate that it would be possible to make foreigners correspond with each other knowing only their own language respectively. Pucci never built a working prototype of his mechanical translator, so his contribution to the area of machine translation was only theoretical.
My paragraph was immediately removed. I would really like to understand the logic behind this categorical refusal to talk about Pucci! Who decides this so arbitrarily, and on what grounds?
*
In this 129th anniversary year of the birth of Federico Pucci, I’m publishing here the French translation of the article written for Wikipedia.it in December 2019 by Oriana de Majo (Federico’s granddaughter), together with her husband, which the free encyclopedia unfortunately refused—completely arbitrarily.
Nothing has been changed from the original [except a few added illustrations and minor updates in italics, enclosed in brackets], so readers can make up their own minds about whether this censorship was justified or not!
* * *
Federico Erminio Raniero Carmine Filiberto Pucci (Naples, March 23, 1896 – Salerno, March 6, 1973), more simply known as Federico Pucci, he was a man of genius and culture, a polyglot and a pioneer in machine translation. We owe him the invention of the first documented rule-based machine translation method in the world.
*
Contents
1 Biography
1.1 Family origins and early years
1.2 The mechanical translation system, international exhibitions, and the WWII years
1.3 Post-war period, international press articles, and letters to the CNR
1.4 Final years
2. The Invention
3. Honors
4. Notes
*
1 BIOGRAPHY
1.1 Family Origins and Early Years
Federico Pucci was born in Naples on March 23, 1896, at 10 Via Foria. His father, Arturo Enrico Emmanuele Pucci (Naples, April 13, 1863), was the youngest son of ship captain Emmanuele Pucci, officer of the Royal Navy and, earlier, the Navy of the Kingdom of the Two Sicilies.
Descending from a branch of the Pucci family of Florence, the Pucci family of Naples settled in Sicily at the end of the 16th century for political reasons. Emmanuele’s father, Vice Admiral Ferdinando Pucci, had served in the Navy of the Kingdom of the Two Sicilies from the Napoleonic wars until 1860, ending his career as commander of the 1st Maritime District of the Kingdom of Italy and aide-de-camp to the King.
Emmanuele’s wife, Donna Agata Benzo & Sammartino dei Duchi di Verdura, came from a prominent Sicilian aristocratic family. She was the sister of Giulio Benso della Verdura, praetor and first mayor of Palermo after Italy’s unification, and a senator of the Kingdom.
Enrico Pucci was a widower with two sons, Aurelio and Raffaele, when he married Federico’s mother, Sofia Pisapia Fiore, descended from a family of judges and legal scholars of the Kingdom of the Two Sicilies. Federico was named after his maternal grandfather, Federico Pisapia Fiore, an officer of the Kingdom’s III Chasseurs. Orphaned as a teenager, he was raised by his maternal uncles, Gennaro and Emilia Pisapia Fiore.
[Federico Pisapia Fiore’s wife, Federico Pucci’s maternal grandmother, was Maria Migliorini, daughter of Scipione Migliorini and Cristina Devaux, sometimes recorded as Clarì, likely due to Italianized transcription: Clary.]
He enrolled at the Institute of Accountancy, the only high school at the time offering advanced studies in foreign languages and mathematics. He began self-studying several foreign languages. A deep love of languages and culture was a family trait, along with a habit of speaking French even in everyday life, a habit he maintained and passed on to his daughters.
Given his extraordinary gift for languages, he entered the Royal Oriental Institute of Naples, Europe’s oldest sinology and Oriental studies institution. There he studied several Eastern languages, including Chinese and Korean, but never graduated, unjustly accused during the final exam of allowing another student to copy his work. Deeply offended, he refused to retake the exam.
Seeking financial independence, he entered a competitive exam for translator-interpreter at the Italian State Railways, placing first for 15 languages and was hired in 1915 in an executive role.
He later moved to Salerno and married Gilda De Filippis on January 13, 1924. She bore him four daughters, including Anna, the mother of Oriana and Marina de Majo, keepers of their grandfather’s memory. [Top]
1.2 The Mechanical Translation System, International Exhibitions, and the World War II Years
In December 1929, Federico Pucci presented in Salerno his first study on mechanical translation, which also appeared in the Italian press in 1930.
That same year, he exhibited for six months (March–November) his text on the "French-Italian Mechanical Translator" at the Bolzano Fair (literary section), earning a silver medal.
Also in 1930, he presented the same text at the Coni Trade Fair and again received a silver medal.
[In 1931, he published the world's first modern text on a "mechanical translator," as well as the first of about ten books he wrote over 30 years on his invention, still entirely unknown today.]
In 1934, the National Research Council (CNR) accepted his application to the Levant Fair in Bari, the first international exhibition of inventions.
In 1935, he won a silver medal at the Paris Fair for his invention: "The Method for Translating Languages Without Knowing Them." In 1936, he participated in the International Exhibition of Inventions at the Leipzig Fair. On recommendation of the Minister of Communications, he was knighted in October 1936 in the Order of the Crown of Italy.
This recognition likely stemmed from his outstanding participation in these fairs, especially Paris.
In February 1936, during the Italo-Ethiopian War, Pucci published a short geopolitical essay, “L’Europa non vuol morire” [1]. Its purpose was to dispel suspicions of his anglophilia and possible ties to Masonic circles (rumors that were not entirely unfounded), given his cosmopolitan reputation. At the time, he frequently traveled abroad, especially to France, and was known to be friends with French jurist Henri Demont, author of “Pour supprimer ce crime: la guerre,” and an early advocate (in 1908) of a “League of Nations” [2].
During this period, Pucci developed a mechanical translator for military use. In 1940, he proposed a translation system to the Ministry of War involving a transmitter (C) and a receiver (D), and intended to present it at the 1940 Exhibition of Technology.
Although the Ministry expressed interest and considered funding a prototype, Pucci declined, fearing that reliance on a technician would compromise confidentiality.
Following Italy’s entry into the war in June 1940, Pucci was appointed by the Salerno prefecture to censor civilian and military correspondence in 30 foreign languages. This is confirmed by a certificate issued after the war by Prefect Cenami.
In performing this task, he used what he called "Mechanical Translator E," which, through a Cartesian coordinate system, could identify the source language by evaluating the value of each letter—provided the text was written using a system of literal constant substitution. [3]
He reviewed correspondence from central and southern Italy, and even foreign mail, especially from Germany. Although others in Salerno spoke German, he alone could read Gothic script fluently.
In 1942, he achieved what he called "automatic grammatical analysis," approved by the CNR in opinion no. 11095 on October 30, 1942. However, like in Marconi's case, the CNR never recognized Pucci’s theoretical studies as an "invention" due to the lack of a physical machine.
In late June 1943, anticipating an Allied landing in the Gulf of Salerno (which occurred on September 9), Pucci moved inland to Ajello di Baronissi and stopped working for the prefecture. There, he served as translator-interpreter for a German unit (16th Panzerdivision), where he reportedly saw the Enigma machine—finding similarities with his own mechanical translator.
While Enigma was electromechanical, its logic was akin to Pucci’s mechanical invention. After the German retreat and the Allied landing, he returned to work for the Italian railways. His two eldest daughters worked for the Allied military administration. [Top]
1.3 Post-War Period, International Press, and Letters to the CNR
In late 1949, denied promotion at the railway despite years of service, Pucci took early retirement to focus on his research. He renamed his invention the "dynamo-mechanical translator," possibly inspired by Enigma.
On August 25, 1949, United Press sent out a news release from Salerno about his invention. This sparked international interest, with the story picked up by major outlets, including The New York Times and the News Chronicle (UK) on August 26.
He presented his project at the Paris Invention Fair from September 16–29, 1949, receiving a certificate of guarantee from the French authorities to protect his intellectual property—despite having no prototype.
On July 10, 1949, Pucci sent a letter to the CNR to seek formal recognition of his invention, followed by another on October 17. The CNR replied with bureaucratic detachment, referring him to the Ministry of Industry’s Inventions Office.
In his July letter, Pucci cited progress in the U.S. by Dr. Harry Huskey, who was building an "electronic brain" for literal translation of foreign languages—though Pucci believed his own system to be more effective and cost-efficient.
Huskey, a computing pioneer who had studied with Alan Turing, worked on the Standards Western Automatic Computer (SWAC), in Los Angeles. Pucci noted that while Huskey’s project was funded by the U.S. Navy, his own required far fewer resources.
On August 25, 1949, United Press sent out a news release from Salerno about his invention. This sparked international interest, with the story picked up by major outlets, including The New York Times and the News Chronicle (UK) on August 26.
He presented his project at the Paris Invention Fair from September 16–29, 1949, receiving a certificate of guarantee from the French authorities to protect his intellectual property—despite having no prototype.
On July 10, 1949, Pucci sent a letter to the CNR to seek formal recognition of his invention, followed by another on October 17. The CNR replied with bureaucratic detachment, referring him to the Ministry of Industry’s Inventions Office.
In his July letter, Pucci cited progress in the U.S. by Dr. Harry Huskey, who was building an "electronic brain" for literal translation of foreign languages—though Pucci believed his own system to be more effective and cost-efficient.
Huskey, a computing pioneer who had studied with Alan Turing, worked on the Standards Western Automatic Computer (SWAC), in Los Angeles. Pucci noted that while Huskey’s project was funded by the U.S. Navy, his own required far fewer resources.
He even wrote to President Truman about his invention and, in 1953, to U.S. ambassador to Italy, Clare Boothe Luce.
In 1949, Pucci published several works:
- “Serie delle grammatiche dinamiche, pratiche, ragionate, storico-comparate: Parte I. Per coloro che in pochi giorni desiderano acquistare una conoscenza elementare della lingua straniera. [fasc. I.] Inglese” ;
- « Le traducteur dynamo-mécanique: L'invention pour traduire les langues de l'occident sans les connaitre presque sans dictionnaire. Op. I: anglais-français. Sous-titré : « Perfectionnement de l'invention primée (traduction mécanique) avec diplôme de médaille d'argent à l'Exposition Concours International des Inventions, Foire de Paris 1935 »;
- “Il traduttore dinamo-meccanico: Serie A. L'invenzione per la traduzione immediata e rapida nelle lingue dell'Occidente senza conoscerle e quasi senza vocabolario... [fasc. ] 1. francese – italiano” ;
- “Il traduttore dinamo-meccanico: Serie A. L'invenzione per la traduzione immediata e rapida nelle lingue dell'Occidente senza conoscerle e quasi senza vocabolario... [fasc. ] 2. Inglese – italiano”.
1.4 Final years
During the 1950s, Pucci continued his research and studies, publishing several works despite the lack of funding:
- 1950: Grammatica dinamica della Lingua tedesca: (linee fondamentali); Il traduttore dinamo-meccanico: Tipo libro macchina. Serie A. L'invenzione per la traduzione immediata e rapida nelle lingue dell'Occidente senza conoscerle e quasi senza vocabolario. [fasc. ] 1. Italiano-Inglese ;
- 1952: Il traduttore dinamo-meccanico: Serie B. L'invenzione per la traduzione immediata e rapida nelle lingue dell'Occidente senza conoscerle e quasi senza vocabolario... [fasc. ] 1. Italiano – Francese.
- 1958: Vocabolario mobile italiano - francese: (parte Traduttore Meccanico).
In 1960, he published his last work:
- Il traduttore dinamo-meccanico: Serie A. L'invenzione per la traduzione immediata e rapida nelle lingue dell'Occidente senza conoscerle e quasi senza vocabolario Tedesco – Italiano, namely the translator applied to German.
2. The Invention
To understand the significance of Federico Pucci’s invention, a brief introduction is needed on what “machine translation” meant before computers.
Modern consensus defines machine translation (MT) as translation by a computer without human intervention.
Before the advent of computers, people only spoke of “mechanical translation” or even “translation machines,” a distinction that persists in the English name for the discipline: “Machine Translation,” or MT for short.
After the computer, sometimes referred to as an "electric brain" or "electronic brain," the first "software" for machine translation consisted of so-called "rule-based" systems. These systems were almost exclusively used for about fifty years before the emergence of hybrid systems or more advanced ones.
And the first demonstration in history of a Rule-Based Machine Translation (RBMT) system is known in its smallest details: date, place, team, languages, procedure, etc.
In fact, it was more of an anecdote than a true scientific demonstration: on January 7, 1954, in New York, at IBM’s headquarters, the team was a collaboration between Georgetown University (Paul Garvin for the linguistic part) and IBM (Peter Sheridan for the programming part). The language pair was Russian and English, with a carefully selected lexicon of 250 words, a few dozen sentences, and 6 rules!
The next day, IBM announced in a press release :
In fact, it was more of an anecdote than a true scientific demonstration: on January 7, 1954, in New York, at IBM’s headquarters, the team was a collaboration between Georgetown University (Paul Garvin for the linguistic part) and IBM (Peter Sheridan for the programming part). The language pair was Russian and English, with a carefully selected lexicon of 250 words, a few dozen sentences, and 6 rules!
The next day, IBM announced in a press release :
And the giant computer, within a few seconds, turned the sentences into easily readable English.This same press release quoted Professor Leon Dostert of Georgetown University, who predicted that within a few years, machine translation could become a reality:
Doctor Dostert predicted that “five, perhaps three years hence, interlingual meaning conversion by electronic process in important functional areas of several languages may well be an accomplished fact.”The optimistic tone of this statement primarily encouraged the U.S. government to allocate significant funding for research. From this perspective, the goal was achieved! However, in reality, the “Georgetown University – IBM” experiment was followed by a decade that all experts in the history of machine translation agree to call “the great disillusionment.”
Remarkably, five years before this event, Pucci predicted the same in a letter to the CNR dated July 10, 1949:
The considerable sums allocated by the American government for building the electronic brain... concern a device of no commercial use and unsuitable for its intended goals. Which proved entirely accurate.In summary, this is the experiment that the scientific community, as of 2025, considers the starting point of machine translation—a judgment that should be reconsidered in light of Pucci’s invention.
*
Just 25 years before the Georgetown-IBM experiment, in December 1929 (90 years ago), Mr. Federico Pucci presented his method in Salerno, invented entirely from scratch, which he conceived well before the advent of computers, likely in the mid-1920s, and which can be perfectly defined as the first “rule-based mechanical translation” system.
This system was documented in detail two years later in a 68-page booklet published in 1931, titled: “Il traduttore meccanico ed il metodo per corrispondersi fra europei conoscendo Ciascuno solo la propria Lingua : Parte I.” (“The mechanical translator and the method for Europeans to correspond, each knowing only their own language, Part I”). Here is the cover:
An extremely sophisticated method for its time, it had nothing to envy in the American experiment that would take place a quarter of a century later. It combined, in a single framework, the three typical approaches of a rule-based machine translation (RBMT) system (later distinguished only thanks to the increasing computational power of computers). To simplify:
- Direct transfer: Morphological analysis, word-for-word translation based on bilingual dictionaries, vocabulary supplements, and tables of vowel and consonant correspondences between languages. Example: Italian to French.
- Indirect transfer: Analysis of syntax and grammar, source and target representations, and transfer rules to generate the target text from the source text. Among other things, the book provides 9 pages to locate a dictionary entry by determining the value of modifications due to inflections (declension, conjugation, etc.). Example:
Each variation (see the third columns from the left) expresses an inflection. Pucci provides the example of the Italian sentence: l’uomo viene, where uomo is the base word, a masculine singular noun, and viene is a conjugation of the verb venire, representing the modification of the concept expressed by the infinitive to that expressed by the third person singular of the present indicative.
Thus, using the international key X to express the variation from the indefinite concept to the present indicative concept, we obtain:
- Italian: L’uomo viene = l’uomo venire + X
- French: L’homme vient = l’homme venir + X
- German: Der Mann kommt = der Mann kommen + X
- English: The man comes = the man come + X
- Pivot Transfer: The pivot can be either a natural interlanguage (English, Esperanto, etc.) or a symbolic representation, as in Pucci’s case, made up of international keys and ideograms enabling the transmission of concepts through graphic elements. Example:
As Pucci himself states, ideograms offer several advantages:
- Allowing the recipient to grasp concepts more quickly (if a Frenchman wants to convey the idea of a horse and draws one, we understand faster than if he simply says or writes cheval).
- Representing grammatical and syntactic concepts through graphics, or even explaining a grammatical concept absent in one language but present in another.
Moreover, in the "book-machine" that Federico Pucci published after the war, he presents a model of his invention incorporating the following four modules (handwritten words):
- Mobile vocabulary (A)
- Mobile vocabulary supplement (B)
- Syntactic corrector (C)
- Morphological corrector (D)
Incredibly, with each "book-machine" published, he included a hand-crafted, annotated model in his own handwriting to explain how to build his "translation machine" by following his instructions!
*
Let’s imagine, in conclusion, that Federico Pucci had been able to build a prototype of his invention: it would likely sit today on the same dusty shelves of history as the machines of Georges Artsrouni and Petr Smirnov-Trojanskij, regarded as quaint, outdated curiosities.
However, he left us far more than an obsolete device, as he documented, in black and white for posterity, as early as 1931, the world’s first documented rule-based machine translation system, enriched with two treasures: the first two texts translated "mechanically" —an excerpt from Dante’s Vita Nuova translated from Italian to French, and an excerpt from Voltaire’s Zadig translated from French to Italian. In the foreword, he advises the reader on learning his method:
Read the translation example on page 66 and verify, using the vocabulary, how I translated it. Then, after translating a few lines, read the tables on page 36 to avoid having to consult the dictionary too often.
The scholar will read everything in the order of presentation, and anyone who struggles to understand the book’s structure will find an explanation in the second part.
The discovery of Federico Pucci and his invention only occurred in 2017, thanks to Jean Marie Le Ray, a French blogger and translator-interpreter [5].
Let us hope that, sooner or later, a university or one of the major players in machine translation worldwide will take up Pucci’s work and insights, finally recognizing his role as a pioneer in the history of machine translation, or even building a functional prototype of his "translation machine"… [Top]
3. Honors
Knight of the Order of the Crown of Italy, October 27, 1936
[Many honors received in France]
4. Notes
1) L’Europa non vuol morire, Federico Pucci, Salerno, 1936
2) Pour supprimer ce crime: la guerre, Henri Demont Plan of 1908, developed and proposed to the Allies in 1918, Henri Demont, Paris, 1938
3) Letter to the CNR, July 10, 1949
4) Letter to the CNR, July 10, 1949
5) https://adscriptum.blogspot.com/p/federico-pucci.html
[Top]
This work is licensed under a
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
