Ke nneth Levy is well known to early medievalists for his work on early medieval music and neumatic tradition. For more than three decades Levy has written with sustained intelligence, opening up new subjects and contributing some lucid and comprehensive analysis to our understanding of early medieval music, and more precisely Byzantine and Carolingian plainchant. In Gregorian Chant and the Carolingians Levy seeks "to consider Gregorian chant's passage from oral to written transmission" (p. 3), and he does so by reproducing several of his most relevant papers and by adding to them some new studies. The result is an important and most illuminating book, which traces the evolution and dissemination of Latin plainchant in the early medieval West.
The book itself contains eleven chapters. The first chapter, entitled "From Gregory to the Ottonians", is a general introduction to the entire volume, and in it Levy gives a short survey of the various problems he discusses in the following chapters. As he explains, "between the sixth and the tenth centuries, European plainchant went through three more or less simultaneous processes of change" (p. 6). First, local repertories were suppressed and substituted by a single, "authoritative" Gregorian repertory; second, a change from variable to fixed melos took place; and third, the solely aural transmission was gradually replaced by notational transmission, in which neumes served as memory aids. These three intertwined changes were all completed by the end of the ninth century. Given the fact that the fixing of melodic substances did not begin before the reign of Pippin III and Charlemagne, the inevitable question that needs asking is what happened to the music during the ninth century, before the initial neumatic transmission?
According to Levy (pp. 10-13), three scenarios were suggested by scholars in recent decades. The first scenario, entitled "the late independent" scenario, argues that the Gregorian musical substances reached fixed, definitive melodic states in the late eighth century, and since then were maintained without change by aural means until neuming started independently at different places ca. 900. The second scenario, "the reimprovisation" scenario, agrees with "the late independent" scenario as to the late, independent start of neuming, but it also adds that no musical uniformity was maintained by the Carolingians. According to this scenario, improvisational strategies and great freedom continued to characterise Carolingian music, even after neuming was introduced. The third scenario, "the early archetype" scenario, perceives the neumes as essential agents in reaching a fixed melodic state, and accordingly it maintains that an archetype Carolingian model of neuming was developed ca. 800, and thereafter served as a basis for nearly all Carolingian neumations. This is the scenario favoured by Levy, and this is the theory he develops in Chapters 4 and 5, "Charlemagne's Archetype of Gregorian Chant" [originally published in Journal of the American Musicological Society 40 (1987), pp. 1-30] and "The Origins of the Neumes" [originally published in Early Music History 7 (1987), pp. 59-90].
Chapter 2, "A Gregorian Processional Antiphon" [originally published in Schweizer Jahrbuch fuer Musikwissenschaft 2 (1982), pp. 91-102] discusses the antiphon Deprecamur te, which, according to Bede, was sung in 597 by Augustine of Canterbury on approaching the city. After examining the various neumatic witnesses of this antiphon, Levy argues that "the Carolingian melody is likely to have been received already stylized from Italy, with little added in the way of 'northern' or 'Frankish' retouching" (p. 30), and that the ultimate Italian source was probably Roman and not Beneventan.
Chapter 3, "Toledo, Rome and the Legacy of Gaul" [originally published in Early Music History 4 (1984), pp. 49-99] explores the antecedents for certain Gregorian offertories, suggesting that Gallican rather than Roman tradition is reflected in many of them. In this chapter Levy focuses on "close multiples", that is, neumed versions of a certain plainchant which, although written in different times and different places, have some melodic agreements. These "close multiples" point to a shared aural version behind the various neumed versions, and through these "close multiples" one can get a rare glimpse of the archaic melodic states, before neuming was introduced.
"Close multiples" is also the subject of Chapter 6, "On Gregorian Orality" [originally published in Journal of the American Musicological Society 43 (1990), pp. 185-227]. Here Levy compares the various neumed recensions of the old-Gallican offertory for Saint Stephen, Elegerunt apostoli. The melodic agreements which these recensions represent point to a stable melos, even before its music was neumed. However, the lack of a full agreement indicates that the melos was not yet firm during the period of neumless transmission.
Chapter 7, "Abbot Helisacher's Antiphoner" [parts of which were published in Journal of the American Musicological Society 48 (1995), pp. 171-184], and Chapter 8, "Aurelian's Use of Neumes", are dedicated to particular texts. The former examines Helisacher's letter to Archbishop Nidibrius of Narbonne (dated to ca. 820), while the latter discusses Aurelian of Rome's Musica disciplina (dated to ca. 850). Both texts, according to Levy, discuss a melos that was rather fixed, down to the fine details, and quite possibly profiled in neumes. Thus, submits Levy, these texts "should lay to rest any notion that improvisation was a continuing option in the ninth-century Gregorian melodic transmission" (p. 194).
In Chapter 9, "Plainchant before Neumes", Levy turns to study differences of musical behaviour in the notational states in an attempt to understand the aural states, before there was any neuming. Consequently, Levy distinguishes between four classes of Gregorian chant, that is, "remembered melodies" (idiomela); "accommodated melodies"; "Psalmic matrices", and " centonate compilations", each of which "having something of its own behaviour, and each reflecting perhaps a different process of aural generation, or, a different path from aural to notational delivery" (p. 195).
Chapter 10, "The Carolingian Visual Model", is, in more than one respect, the culmination of the entire book. It "ties together proposals made in earlier chapters concerning neumes origins and the start of neumed Gregorian states" (p. 15), and it persuasively presents "the early archetype" theory. By ca. 800, argues Levy in this paper, a full, authoritative Gregorian repertory was established (probably in Metz), in which melodic refinements were attained by visual-notational control. This repertory -- Charlemagne's Archetype -- provided choirmasters throughout the Frankish realm with the memory support they needed to maintain the Carolingian-Gregorian melodies. Thus, according to Levy, neumes were already in regular use during the late eighth century.
Chapter 11, "Memory, Neumes, and Square Notations", concludes the book, and in it Levy traces in broad lines the nature and development of melodic variance from the tenth to the eighteenth century.
Although Levy's hypotheses and discussions may generate some criticism and controversy, there is little place to doubt that Gregorian Chant and the Carolingians is an important contribution to the study of early medieval music. It provides a strong and quite convincing case for the "early archetypal" scenario, and it contains much that will be of interest to academics and research students of early medieval music and liturgy.