How do you solve a problem like Maurice-Quentin de La Tour?

Warning: unless you are a lexicographer, copy editor or bibliographer, don’t read this post. It has nothing to do with art history (nor for that matter with musical theatre). It may be the most boring post I have ever written. And the answer is given away in the title.

In fact there are two problems – although both have the same solution. They are, in inverse order, how to print the artist’s family name; and whether to hyphenate his forenames.

Just to confuse you further, it is the hyphenation problem that is of more general interest, so let me take it first after all. For general purposes, French publishers employ any of the three possible hyphenation conventions: (a) don’t hyphenate any forenames (e.g. Jean Baptiste or Maurice Quentin); (b) hyphenate only compound names (Jean-Baptiste, Maurice Quentin); or (c) hyphenate all forenames (Jean-Baptiste, Maurice-Quentin). But which is best?

I’m assuming you all realise (although many older writers seem not to have known) that Quentin is a forename, not the family name (but see below): i.e. his siblings were not called Quentin. It was in fact quite a common forename in Saint-Quentin. Maurice was the name of his parrain at baptism. Thus the names came from different sources; there is no saint Maurice Quentin (well, there is for some of us, but not in the established church). And (although this is the tricky bit) if you knew him intimately, you probably wouldn’t have called him Maurice-Quentin at every turn (although we don’t in fact have any idea whether his friends called him Maurice or Quentin, as such oral uses were not recorded; and people in those days didn’t use forenames the way we do now).

So put simply, he had two forenames, Maurice and Quentin, rather than a compound name (or “prénom composé”) of the kind borne by his rival, Jean-Baptiste Perronneau (saint Jean Baptiste was one saint). But what of the third pastellist, Jean-Étienne Liotard? Authorities are split down the middle. When he signs he does so in full, or abbreviates to “J. Etienne” or “J. E.”, suggesting that he did not see his names as severable. His twin was “Jean-Michel”, and an older brother was simply Jean. Another case is that of the pastellist Jean Pillement, who signs thus, but was actually baptised Jean-Baptiste. While one would never now separate Marie-Antoinette, she was of course baptised Maria Antonia Josefa Johanna.

The fact is that there is no way of determining now which apparently compound eighteenth-century forenames are “composés” or just a series of simple forenames. You can’t appeal (as many people assume) to documents in which people signed their name, as this simply reveals how aleatory eighteenth century orthography was: people often signed with just their family name, they almost never used accents; spacing and capitalisation were random, and hyphens never appear even in names like Jean Baptiste or Marie Antoinette. Notaries sometimes did separate names, often with ambiguous marks which look like commas or strokes: e.g. Jacques/Antoine/Marie. But in contemporary printed material, the broad (but not universal) consensus among genealogical tomes (the only area where forenames habitually appear) was to hyphenate all forenames.

What confuses the matter is that France is a country with legislation that we in England would regard as bizarre. In 1803 Napoléon brought in a law restricting parents’ choices of names to those of calendar days and those from ancient history (that of course is why prénoms composés became popular), relaxed finally in 1993 (although you still can’t have silly names or ones with foreign diacriticals such as ñ). But for prénoms composés laws remain in place that govern their punctuation in legal documents: the parents can choose whether to separate them with a hyphen or a space – but all other forenames are separated by a comma. So for example “Jean-Baptiste, Marie” or “Jean Baptiste, Marie”. (But outside these legal documents, the commas are never used, thus undermining the case for following these styles more widely.) Further, prénoms composés can only have two components. So to a modern French person a name like “Jean-Claude-Gaspard Sireul” needs to lose a hyphen (the second invariably chosen because “Jean-Claude” “sounds right”, even though as we have seen Jean and Claude came from different sources). And Marie Antoinette needs one, even though she never had one when she was alive.

These modern rules simply didn’t apply in the ancien régime, although there is a growing trend to try to impose them retrospectively. This is in effect the result of one or another category confusion; neither autograph signatures nor modern legislation are of any help here, in what properly is simply a question of a publisher’s choice of printing convention.

One would have thought that the matter was settled by bibliographers and lexicographers in the nineteenth century when convention (c) above was almost universally adopted. The Lexique des règles typographiques en usage à l’Imprimerie nationale is categoric (3e éd., p. 151):

Les prénoms français ou francisés se lient par des traits d’union.

That, for example, is what you will find in the BnF Catalogue general. The benefits of the rule are obvious: you only have to know which is the family name (which you need to know anyhow – as in Jean-Claude Richard de Saint-Non, where, without the policy, you might think that Richard was a forename; it is not) and there is no further ambiguity.

I can see that convention (a) can at least claim to match holographic evidence, and is just as easy to use as (c). But the continued encroachment of convention (b) is to be deplored, not only because its intellectual justification is based on error, but because it is almost impossible to apply consistently. Any book (not only multi-author exhibition catalogues) that tries to follow it seems invariably to end up with dozens of errors. It requires an iron discipline for copy editors to achieve consistency, and I have rarely seen the task succeed. It cannot in short be recommended for books relating to the eighteenth century.

* * *

Let me now turn to the similarly pointless debate over the proper spelling of La Tour’s family name (de La Tour, de Latour, Delatour etc.). You’ll all be aware of the basic rules for French proper names: if the particle is separated, you capitalise and alphabetise from the bit containing the definite article (La or Du, but not de). The “de” normally introduces a territory or estate which at some stage has been acquired by a member of the family, and (if noble) is usually assumed as the name, with the family name dropped when their ascent is sufficiently clear (something only peers do in England, although Scotland makes more use of estates, without driving out the family name); this gives rise to some flexibility and a great deal of aspiration which I’ve discussed before. Thus for example a M. Legendre becomes Legendre de Villemorien, and then signs (as a witness to Perronneau’s marriage in this example) simply “De Villemorien”. But that doesn’t mean that anyone would index him under D. He remains under L until he (or his descendants) might move to V (but never D) if they became really grand. (This all reverses after 1789, with many former aristocrats rapidly closing up the spaces to conceal their status: but that’s another story.)

But some names beginning De aren’t noble and should be spelled solid (and filed under D) – although the owners may like to pretend. And when roturiers have names that come from places (that they didn’t own, but by which they have always been known), there’s no right or wrong answer: just convention and usage.

Working out just what contemporary usage was is of course tricky, and raises the same ambiguities as discussed above re hyphenation of forenames. People didn’t often write their names in full, nor did their family names normally start sentences: you would write to “M. de Villemorien” without dreaming of a capital D, which only appears when he signs. For the same reason, dozens of examples of La Tour’s signature (almost all of which take the form shown above, Dela_Tour, although there are also a few cursive De_la_tour examples) tell you nothing about whether the D should be capitalised or be the point of alphabetisation when the name is given in full or set in twenty-first century type. The flexibilities of handwriting allowed subtleties such as the linked but discernable gaps between the components, as well as internal capitals which no modern copy editor would tolerate.

And in print (e.g. almanachs or annuaires of the Académie), La Tour appears as “De la Tour” (alongside “De Lagrenée”, “De la Joue” and “De Larmessin” although no one is threatening to file them under D), or, as in the salons livrets, as “de la Tour”:

Among contemporary critics, the overwhelming preponderance was for “de La Tour” or “de la Tour”. Even that ultimate snob, La Font de Saint-Yenne, who had a nose for imposture and pretension, consistently uses that form, as in this famous passage:

Indeed he refers in places to “l’ingénieux la Tour”, which defeats the idea that the “de” was considered integral at the time. I personally find that “les pastels de La Tour” “sounds right”, while “les pastels de Delatour” does not.

The hunt for the family name in previous generations also fails to justify putting La Tour in his place as peasant. He was of course the son of a writing master, and a progression may be seen in his father’s increasingly elaborate penmanship: whether on La Tour’s own baptismal entry

or by the time (1726) of the baptism of the pastellist’s half-brother Jean-François:

his father was clearly separating the particle from “La Tour”, as did his own father Jean de La Tour, a maître maçon. Jean’s signature is found in numerous parish registers, usually accompanied by his monogram (which may also be his mason’s mark), JLT in a circle:

The invariant in all of this was some separation (by space or capitalisation) of the Tour or La Tour element; never does the form Delatour appear. This clearly was the view not only of the artist himself, but of his elder brother Charles, whose signature appears above Maurice-Quentin’s in this 1761 document:

But whatever the arguments, they were conducted in full in the nineteenth century when a clear consensus was established in favour of “de La Tour”, indexed under L. That is where you will find him almost everywhere: in the last great catalogue raisonné, B&W (1928), in more recent monographs such as Debrie 1991 and Debrie & Salmon 2000, as well as the major retrospective La Tour 2004, and in all standard art historical dictionaries, the BnF and Getty ULAN. There would have to be a very good reason to try to overturn such weight.

But I wouldn’t have written this post if there were not some who disagree. A former head of the drawings department at the Louvre felt strongly about the issue, and insisted on labelling his work “Delatour”. At Saint-Quentin the formulation “De La Tour” is in use, the form adopted by Élie Fleury for his 1904 catalogue, but changed to Delatour in the later version co-authored with Gaston Brière. In a footnote they explain that this is to follow the spelling in official documents, while recognising that the artist used a capital T. Hervé Cabezas followed this too, explaining in a Revue du Louvre article (2006) that “De La Tour” allowed him to keep the capital D found in every example of the artist’s signature (but this is the “De Villemorien” argument above).

The idea that the commonly accepted form is “wrong” and should be “corrected” springs I think from a similar category error as the hyphenation confusion. It frankly doesn’t matter whether La Tour was or was not entitled to something that might be confused with a noble particle, even if we could work out any basis on which such a debate could be decided. It does matter that when you go to a library you get out on the right floor to consult all the books on this artist together, and that you need only take out one volume of the reference works in which he appears. We should challenge authorities when they are wrong factually – to root out error and confusion – not when they have adopted conventions which are now well established when we might have preferred the other choice: that merely sows confusion.

Postscript (February 2018)

This debate was revived by the new introductory note to Christian Michel’s Académie royale (2018): see the note at the end of my blog post.

Postscript (November 2018)

Perhaps we should seek Enlightenment from the source. In a 1753 letter to Mme Du Deffand, d’Alembert began a sentence “Latour a voulu absolument faire mon portrait”, which excludes “Delatour”; while Voltaire sent the artist a letter in 1735 with this envelope (now in the New York Public Library, who kindly provided this scan):

Postscript (January 2020)

One further piece of contemporary evidence: in 1762 La Tour’s friend Jacques-Charles Roettiers, graveur général des Monnaies de France, made a medallion portrait of the pastellist with his name inscribed around the circumference: MAURICE QUENTIN DE LA TOUR.

Postscript (May 2021)

There is even an article by Elie Fleury in the Journal de Saint-Quentin, 27.vii.1897 complaining about the rue Latour on a street sign in Amiens and rehearsing the usual discussion. I should perhaps have added La Tour’s earliest signature, on his 1719 apprenticeship contract, in the form Q. DeLaTour.