[Journal of the Simplified Spelling Society, J19, 1995/2, p13-17.]
The Dictionnaire Historique de l'Orthographe Française: A Landmark in Historical Spelling Research.
Dr Baddeley has worked for 10 years on orthographies, writing systems and spelling reform for the HESO group at the CNRS (see article below for details), and now also lectures in English at the University of Versailles-Saint Quentin en Yvelines (France). She is an editorial adviser to JSSS, for which she reported previously on French spelling matters in issues J7 1988/1 pp30-31, J10 1989/1 pp10-12, J11 1989/2 pp14-15, J12 1992/1 pp14-15, J15 1993/2 pp3-5. [Also see J23 1998/1.]
Summary.Early in 1995, Larousse, one of France's major publishers of dictionaries, brought out a new and long-awaited work which will be a major contribution to research on the history of French spelling, and to historical linguistics in general. The Dictionnaire Historique de l'Orthographe Française (DHOF), or French Historical Spelling Dictionary, is the fruit of several years of research carried out by Nina Catach and her team at the Centre National de la Recherche Scientifique (CNRS), the French national science foundation in Paris.
Through over 1300 pages, this Dictionary traces the history of French spelling from the 16th century to the present day. It shows how the French spelling system has constantly evolved over the centuries, following (albeit at a distance) changes in pronunciation and in grammar, heeding or ignoring the recommendations of the grammarians, and trying at all periods to strike a balance between the conflicting principles of phonology and morphology, of sound and meaning. Most importantly, it shows that French spelling even today is in a state of flux, with large sectors of variation coexisting alongside what is, by and large, a stable and coherent system.
Introduction.In 1970, Nina Catach founded a research team at the CNRS in Paris, with the name 'Histoire et Structure de l'Orthographe Française' (HESO). The team was given a mission to produce, among other things, a comprehensive survey of the evolution of French spelling from the early modern period to the present day. Over the years, HESO has continually enlarged its research perspectives, producing major works in the field of psycholinguistics (reading and writing acquisition), on the present-day structure of French spelling (with a dictionary of word-families), on the history of writing systems, and in computerized linguistic applications, with a programme that automatically 'translates' any written message into its oral equivalent.
The team and their associates have also been active in the field of spelling reform, drawing up the lists on which the Commission of specialists based their conclusions for the 1990 spelling reform proposals. However, the Historical Spelling Dictionary has always been at the heart of the team's research, and much of the theoretical work it has produced is the result of reflections brought about by the work on the DHOF.
As I had the privilege of working on this unique project during the ten years leading up to its completion, I would like to give SSS Journal's readers not only a conventional review of the dictionary, but also some of the 'inside story' of how it was made.
1. The Corpus: The Dictionary of the Académie Française.When the question first arose of where to find the materials necessary for a comprehensive study of the history of French spelling, the answer soon became obvious. The French language is fortunate in possessing a particularly rich lexicographical tradition, of which the nine editions of the Académie Française dictionary (from 1694 to the present day ) are an especially important feature. External observers in France and elsewhere may look with some condescension on the conservatism of the Académie dictionary, as well as the slowness with which new editions appear; however, for spelling historians and lexicographers, the Académie corpus is a boon. Although new words have consistently been introduced from one edition to the next, old words have seldom been eliminated,  which means that the original corpus of lexical items from 1694 has survived, practically intact, up to the present day. This makes it particularly easy to trace the orthographical development of these items over the centuries.
The DHOF therefore traces the orthographical development of some 20,000 lexical items in the Académie editions over three centuries. Of course, the modern reader may be somewhat bemused to find many words appearing in the dictionary which are no longer in use, and not to find more recent ones.  However, the advantages of having a complete corpus of this importance seemed to us to outweigh the few disadvantages.
This corpus therefore formed the basis for the dictionary, and provided us with the source of all its lexical items. However, as the Académie corpus represents only a part (and the most conservative part) of the French lexicographical tradition, spelling forms from other contemporary dictionaries were also taken to complete the picture. First of all, the items from the Académie corpus were traced back in time, whenever possible, to the 16th- and early 17th-century dictionaries of Robert Estienne and Jean Nicot (1549, 1564, 1606). Also included for reference were Randle Cotgrave's French-English Dictionarie of the French and English Tongues (1611), Pierre Richelet's Dictionnaire François (1680) and Antoine Furetière's Dictionnaire Universel (1690), both almost contemporary with (and competitors of) the Académie's first edition. We also included Féraud's Dictionnaire Grammatical de la langue françoise (1761) and Dictionnaire Critique de la langue française (1787), which are notable for indicating pronunciation systematically. Finally, from the 19th and 20th centuries, we consulted the dictionary of Littré, the Dictionnaire Grammatical of Darmesteter and Hatzfeld, and the contemporary editions of Le Robert and Larousse.
1.1. From 20,000 words to 2,000 articles.
It soon became apparent that the 20,000 words of the 1694 corpus were more than enough for a history of spelling. Indeed, of the 20,000 items, almost half (in fact, 48.7%) had not changed in their spelling at any time between 1694 and the present day. This figure includes many monosyllables and 'grammatical' words, as well as many of the most common words, such as homme, femme, enfant, pain, vin, etc. Having eliminated these, there still remained a substantial corpus. However, devoting an article to the spelling changes of all these 10,000 words would have entailed a huge amount of repetition, as the spelling changes involved were very often the same: introduction of an accent (27.69%), replacement of I by J, loss of etymological consonants, etc. Eventually, we reduced the corpus to the 2,000 words which seemed to us to be the most interesting, or the most representative. These words cover 148 different types of spelling changes.
This explains why, in the dictionary, there are two types of entries: some simply give the spelling forms in the various dictionary editions, followed by a reference number indicating the types of spelling change represented, while the others are followed by a more substantial article. Originally, the two sets of words were meant to be presented separately; however, the publishers insisted that a dictionary was, by definition, in alphabetical order, and the two lists were therefore merged together.
2. Structure of the Dictionary.The Dictionary is made up of several parts: a 15-page Introduction, a page of statistics taken from the data used, a bibliography, the Dictionary itself, 89 pages of 'Paragraphes de synthèse', a list of the dictionary entries classed according to their type of spelling modification, and finally a 100-page Index listing all the spelling forms given in the dictionary.
2.1. The 'Paragraphes de Synthèse'.
One of the most original features of the dictionary are these paragraphs (148 in all), which describe in full detail the complete history of each of the types of spelling change identified in the dictionary, and almost make up a history of spelling in their own right. The paragraphs are organized as follows: vowels, semi-vowels, accents and auxiliary signs, consonants, double consonants, internal mute consonants, final consonants, Greek and Latin notations, homophones, compound words, lexical pairs and changes in grammatical category. This sequence ranges more or less from the phonetic to the purely graphic and semantic.
These paragraphs make up the backbone of the dictionary, and give both an overview of the features presented in the individual articles, and a more detailed and documented analysis than was possible in the space of a short article.
2.2. The System of Cross-References.
Another characteristic feature of the dictionary is its unique system of cross-references. In the 'bare' articles, which just give the spelling forms in each edition, we felt that it was sufficient simply to indicate the paragraph number for the type of spelling change involved. However, in the case of the 2,000 'full' articles, another solution was needed. On the one hand, we could not repeat the same information every time the same type of spelling change was involved; on the other hand, just giving a paragraph number and expecting the reader to look it up every time seemed insufficient. Finally, an economical and original solution was found. We chose, for each of the 148 types of modification, a word that exemplified this change: for example, for all words featuring an I replaced by a J, we chose the word jaser, and for the introduction of the circumflex accent, bât. In this way, the reader would come to associate an orthographical change with a particular word, which is both easier to remember and more evocative than just a paragraph number.
Therefore, in the course of reading an article, readers will be referred to one or more of these 'base words' (indicated by a double asterisk), where they will find more information about the type of spelling change involved, and then, if they look up the base word, another reference will send them to the appropriate paragraph. In the published version, the typography helps to distinguish the different parts of the dictionary and the different types of articles.
The work thus succeeds in meeting the demands both of specialists and of the general public. The specialist who is in search of particular information can find it easily either by looking up entries in the alphabetical order of the dictionary, or in the comprehensive index of spelling forms at the back. Although a dictionary is not the sort of thing one tends to read from cover to cover, casual readers can also 'browse', and, guided by a set of graded cross-references which lead them step by step to more and more detailed information, will be able to make up their own picture of this particular aspect of the history of the French language.
[See typical page from DHOF 96kb on another page.]
2.3. The Structure of the Articles.
The 2,000 or so 'full' articles are all organized along the same pattern. First, we present the spelling forms taken from the dictionaries that make up the main corpus (1549, 1564, 1606, then the Académie editions). A short paragraph next gives the etymology of the word, and, where appropriate, spelling forms from Old and Middle French. We then present, in separate paragraphs, the spelling changes undergone in the word, generally in chronological order. When appropriate, we give information about the pronunciation of the word, changes in meaning, and, finally, give the homophones or homographs of the word.
However, the transformation of our manuscripts into a published edition took a considerable time, and did not always go smoothly. On the one hand, we authors wanted to keep our particular terminology, and give as much information as possible; on the other hand, the publishers wanted to make the work as widely available as possible, and not swamp the potential reader with too many erudite details. It was often difficult to compromise, and also to adapt our 'home-made' production to the latest publishing technology.
3. Publishing the Dictionary.Today, the dictionary is available in a very handsome edition, in the 'Trésors du Français' collection by Larousse. At the CNRS, we also have a computerized version on a database, which has enabled us to carry out a certain number of statistical analyses, which are given in the Introduction to the dictionary.
One day, the people from Larousse told us that the dictionary was too long compared to the other works in the same collection, and we had to find a way of shortening it by several pages. Not wanting to cut out any of the contents, we did this by introducing a large number of abbreviations. However, at a later stage we were told that there were too many abbreviations, and they made the articles unreadable. So, we had to go back over our list, and put the abbreviated words back in their full form in all the 1300 plus pages of the dictionary. Fortunately, we were able to do some of this work automatically, but not without a few 'hiccups': for example, while replacing the abbreviation ds by dans, one of the keyboard operators forget to tell the machine to replace it only in whole-word contexts, with the result that words like poids, lourds, ronds etc came out as poidans, lourdans, rondans, and all these words then had to be corrected 'by hand'.
Similar problems occurred with a few special characters: the notation of Latin long and short vowels, and certain phonetic symbols. In some cases, for example, with the special characters invented by spelling reformers, we simply had to find a simpler notation, which we regretted. However, that was the price to be paid for publication, and we felt that it was more important to reach a wider audience than to have the dictionary exactly as we wanted it, and for it only to be available to a happy few. So far, reactions to it have been overwhelmingly favourable, from the scientific community as well as from the public at large.
Conclusions.We hope that our Spelling Dictionary will put an end to the widespread myth that French spelling is something fixed and immovable. We have charted the variations and hesitations that accompanied every change in spelling, and have shown that, although part of the evolution can be said to be due to a unified purpose and desire to control, spelling changes are more often than not the result of the system 'adapting itself' to the changing role of written language in society. From this point of view, each article contains to a certain extent, in miniature, the whole story of the evolution of French spelling, and, within it, the whole history of society.
The 16th-century dictionaries, based on the written usage of the cultivated Parisian liberal professions, tended to stress etymology (when it was known), and to give spelling forms which were as close as possible to Latin and Greek origins. This was useful at a time when pronunciation was extremely variable and when most people who knew how to read and write also knew Latin.
As reading and writing became more widespread during the 17th century, the Académie editions reflect this extension of literacy: the first editions are more interested in recording varieties of usage (both spoken and written) and linking them to different social groups: hence the numerous comments on how words are pronounced, and the attempts to 'grade' spelling variants according to their social acceptability. The 18th century took this normative tendency even further, and attempted a large-scale 'rationalization' of the spelling system, eliminating variants and standardizing spelling forms according to established principles (prosody, word-families, etc). However, even a man with such linguistic insight and force of character as the Abbé d'Olivet  was, at times, obliged to back down when confronted with the irregularities produced by the conflict between pronunciation and morphological principles, new 'rational' uses and the age-old force of habit, the needs of a modern spelling system and the limitations of the Latin alphabet.
The numerous inconsistencies that the spelling system had spawned over the centuries were carried over into the 19th century, when, with the centralization of the state, the appearance of a huge number of civil servants and of compulsory state education, the need arose to fix the spelling system, and the Académie dictionary, despite all its faults, became the Bible for printers' correctors and for all those engaged in written production of any sort.
The Dictionary shows us, at last, that the process of change is ongoing, and that there are whole sectors of the spelling system (compound nouns, loan words, doubled consonants) where hesitations still exist and a certain amount of 'clearing up' is still needed.
We hope also to have restored some credibility to the Académie dictionary, which has been superseded this century by the more popular and more readily available Robert and Larousse editions. Contrary to popular belief, the Académie has never had any mandate to take decisions or to impose its own recommendations concerning spelling: it has always described itself as the 'custodian of usage', and the DHOF shows that it has consistently fulfilled this role over the centuries: never being in too much of a hurry to record fashionable new spellings that were unlikely to last very long, but rarely failing to do so when majority use prevailed. However, with its latest edition (1986), the Académie has taken the unprecedented step of recommending new, 'rectified' spellings as variants, even before they were adopted by the 'mainstream' dictionaries, Robert and Larousse.
Finally, we hope that our dictionary will not only provide convincing evidence of what spelling was like in the past, but may also give some indication as to where it is going. While we were preparing the lists for the 1990 'Rectifications', we frequently used data taken from the DHOF to show that the reforms proposed (for example, generalization of the grave accent on syllables containing an 'open' E) accorded with the general tendencies of spelling evolution over the centuries. Without this historical perspective (which has, until now, not been readily available), there can be no real, informed debate about the present structure of French spelling (or that of other languages) or about its possible reform.
 Dates of the editions used: 1694, 1718,
1740, 1762, 1798, 1835, 1878, 1935, and 1986 (unfinished). We disregarded simple
reprints, and the numerous counterfeit editions.
 According to Académie regulations,
any modification in the existing dictionary articles required the votes of at least
two-thirds of the Academicians present; however, the addition of an article only
required a majority vote (ie, over half).
 In the Académie's first edition, many
common words were left out by mistake, and were included in the following
 Joseph Thoullier D'Olivet was a grammarian and Academician, who brought about the full-scale modernisation of the spelling of the Académie dictionary (mainly by the systematic introduction of accents) in 1740.
Back to the top.