The proper chants of the mass and office were not communally sung: in large enough institutions, a choir and cantor would be responsible for these. How this element of the liturgy was handled in smaller and poorer situations is quite unknown. The justification for the incorporation of such a sensuous medium as music into the liturgy was its power: it should promote understanding of the word of God, and move worshippers to greater devotion. In such ideals, Carolingian ambitions for music were no different from those of the Roman church in earlier periods: but the Carolingians construed those aims in accordance with their more general concern with the quality and correctness of Latin texts. Therefore concern with the primacy of the words of chants directed their work on the texts, while concern with audibility and projection of meaning governed new levels of musical control. With new intellectual means (in the form of music theory) and materials (in the form of well-made books, some with musical notation) to manage the delivery of chant texts, Carolingian singers could be expected to perform eloquently, so that, in the words of Hrabanus Maurus, their singing might "turn the mind of the people upwards to heavenly things and direct it to divine contemplation." For Carolingian liturgical writers, such expectations rested entirely on the view that good performance depended on correctness: that correctness further depended on following instructions - at the minimum the chant text - which could be verified against a written record when necessary. It now became unacceptable for individual singers to express themselves freely, except through the delivery of set texts, singing these using melodic shapes considered appropriate for that delivery.

There is no implication in the surviving manuscript sources that musical notation was introduced as a means of regulation of musical delivery: nevertheless, once available, it could function in that way. The very detailed notations written into a small number of late ninth-century music books (above all, Laon 239 and SG 359) show how far good musical scribes could go in offering information about how to sing texts distinctly and with eloquence. The primary objective in terms of the practice of liturgical chant circa 800 had been the availability of an approved repertory, and its appropriate, corrected, expression. In terms of repertory and text versions, that objective had been fulfilled by the middle of the century: it is likely that musical delivery had also been brought under control by that time, using standardized melodic formulations and a shared melodic grammar, based on newly-rationalised music theory. By circa 900, the singing of "Roman" or "Gregorian" liturgical chant had been stabilised, its forms of expression widely transmitted in written texts with musical notations. An early Carolingian aspiration had reached ample fulfilment. Cutting across this narrative of the history of chant, however, patterns of its codification are less stable: the changeover from books containing texts to books containing notated texts is but one element in a more complex picture. Where, from the late eighth to the late ninth century, priests were expected to have books of chant, it is unclear whether this expectation was maintained into the tenth century, and how it was then met.

Books containing chants integrated with prayers, rather than copied in a separate section of a codex, were becoming available: the many examples from Italy are mostly fragments but there is one extant full "integrated" plenary missal (Milan D 84 inf.), with chants, prayers and readings set out in the order in which they would be required in the mass. From northern Europe, the examples are more limited, and there is none in which this integrated layout has been adopted.