After conducting the premiere of I masnadieri in London, Verdi traveled to Paris, where he hoped to finally lead the life he wished. He had little time to relax, however: within a week of his arrival on July 28, 1847, the new directors of the Opéra, Nestor Roqueplan and Charles Duponchel, succeeded in convincing the composer to provide an opera for the fall season. Such short notice did not allow Verdi time to compose a new work, and so he negotiated a contract that would let him rework I lombardi, an opera first performed in 1843. Eugène Scribe suggested that Verdi collaborate with librettists Alphonse Royer and Gustave Vaëz, who agreed to fit a modified plot to existing music. We know very little about the collaboration between librettists and composer. With Verdi living in Paris, the parties involved could discuss emerging issues face to face and did not have to rely on the postal system. Unfortunately, the constant presence of two librettists, two impresarios, and two editors (Léon and Marie Escudier) not only deprived posterity of insightful correspondence regarding the project but sufficed to drive Verdi mad, as he himself confessed in a letter to a friend. As to the music, Verdi rearranged some of the numbers, omitted others, composed a few new ones, and replaced transitions where necessary.

Although the librettists had to adapt the Lombardi libretto to a modified plot, they aimed at a free translation, often preserving only key words and the overall sense.

French versification

The French have long taken great pride in the verse forms of their lyric, epic, and dramatic literature. Opera librettos form an important subclass of this corpus, so it is somewhat surprising that comprehensive theoretical works analyzing French verse largely exclude operatic texts. Several reasons may account for this exclusion.

First, librettos are not generally considered to be autonomous works but rather texts subservient to music. Throughout most of the nineteenth century, French librettos were written entirely in verse but at the same time had to provide enough variety of poetic meter, accentual pattern, and stanzaic structure to accommodate the musical style envisioned by the composer. In recitatives, the meters were generally longer and changed more frequently, whereas in arias, they tended to be shorter, more uniform, and more regularly accented to allow for regular rhythms and phrases.

Second, librettos followed neither French drama in maintaining a uniform meter throughout an entire work nor lyric poetry in relying exclusively on stanzas. Thus, the mixture of stanzaic, non-stanzaic, and hybrid forms, as well as the greater freedom librettists took with traditional rules, seem to have caused theorists to regard the French libretto as an unsuitable genre for illustrating the principles of versification, even though many parts of these librettos would have been sufficiently traditional to illustrate particular points. Moreover, there is in fact no single “theory” of French versification. The treatises reflect a wide variety of approaches, some more or less compatible with each other, others contradictory.

Scribe and Verdi discussed the Vêpres project as early as 1853 and may have met in person before rehearsals began on November 16, 1854. Evidence regarding Verdi's prosodic concerns, however, is limited to a few extant letters, which also debunk the long-standing assumption that the composer, with his supposedly inferior knowledge of French, readily accepted Scribe's verses without requesting changes. In fact, the letters portray a composer actively involved in structuring the libretto – from its dramatic pace down to details of poetic meter and accentual rhythm – and a librettist aware of Verdi's talent and graciously complying with his wishes.

On at least three occasions, Verdi requested regularly accented verses. Two of these requests refer to choral sections, and in both cases the composer was looking for meters corresponding with two of the most regularly accented Italian equivalents, the vers de sept syllabes (corresponding to the Italian ottonario) and the vers de cinq syllabes (corresponding to the Italian senario). In a letter of June 7, 1854, Verdi elaborated:

Some of the conventions a composer had to follow when writing for the Paris Opéra have long been well known: a four- or five-act design, grand emotions (i.e., a focus on issues that concern a large segment of society rather than isolated individuals), ballets, extended choral sections, musical forms that complied with French expectations (such as ternary aria forms), and certain dramatic requirements. But when it comes to defining French melody in the broadest sense, including consideration of rhythmic and harmonic characteristics, relationship to prosody and accompaniment, and principles of melodic development, opinions have been vague and often contradictory. One nineteenth-century critic for the Gazzetta musicale di Milano, for example, questioned whether the melodic style even at the Opéra could truly pass as French. In reaction to a review in a Berlin paper that detected French influences in Verdi's Rigoletto, the critic responded:

Aspects of versification and Verdi's musical response

In 1865, Verdi was in frequent contact with Léon Escudier, his publisher in France, about the revisions of Macbeth for the Paris Opéra. Although Émile Perrin, the Opéra's director, preferred to mount this revised opera rather than revive one from the repertoire, he still hoped to entice the maestro to compose an entirely new work. Verdi refrained from committing himself but sent signals, though negative overall, that left a door open for a new commission; it was clear that the final decision would depend on a suitable libretto. Perrin accordingly dispatched Escudier to Busseto, along with a libretto titled Cleopatra and a scenario for Don Carlos. Of these and other possible topics, Verdi considered Don Carlos the best by far, and by the end of August 1865, the Opéra and the composer had come to a basic agreement for a new production. At the beginning of December, Verdi joined Joseph Méry and Camille Du Locle in Paris to work on the libretto for Don Carlos, and by the time the composer returned to Sant'Agata in March 1866, the libretto was nearly complete, except for the final version of act V.

Just as in 1847 with Jérusalem, Verdi's presence on site rendered written correspondence between librettists and composer unnecessary.