Show Summary Details

Page of
PRINTED FROM Oxford Music Online. © Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single article in Oxford Music Online for personal use (for details see Privacy Policy).


  • David Hiley

A sign placed, in modern notational practice, before a note, which alters its previously understood pitch by one or two semitones. The sharp (♯; Fr. dièse; Ger. Kreuz; It. diesis) raises a note by one semitone; the double sharp (𝄪; Fr. double dièse; Ger. Doppelkreuz; It. doppio diesis) raises it by two semitones. The flat (♭; Fr. bémol; Ger. Be; It. bemolle) lowers a note by one semitone; the double flat (♭♭; Fr. double bémol; Ger. Doppel-Be; It. doppio bemolle) lowers it by two semitones. The natural (♮; Fr. bécarre; Ger. Auflösungszeichen or Quadrat; It. bequadro) cancels a previous sharp or flat. A double sharp is changed to a single sharp by writing ♮♯ or occasionally ♯, a double flat to a single flat by ♮♭ or occasionally ♭.

For a discussion of the addition of accidentals to early music see Musica ficta; see also Editing and Solmization. For the notation of some non-Western music and 20th-century compositions using intonations other than the 12-note system, see Notation, §III, 4, (vi).

1. Early use.

The ♭ sign on the one hand and the ♮ or ♯ signs on the other originate in the forms suggested by Guido of Arezzo for the two possible pitches of the note B: rotundum (‘round b’; also b molle: ‘soft b’) and quadratum (‘square b’; also b durum: ‘hard b’), representing modern B♭ and B♮ respectively (see Table 1). The two shapes appear in the treatises Aliae regulae (GerbertS, ii, 36) and in Micrologus (chap.2: CSM, iv, 1955, p.93), both of which date from about 1030. Guido was not the first to suggest different shapes for the two notes B. The process whereby the notes of plainchant melodies became more or less permanently associated with the letter names that most of them still retain seems to have been completed by the end of the first millennium (see Notation, §III, 1, (iv)). Although it would have been possible to eliminate the need for one of the two forms of B by notating melodies using B♭ a 4th lower or a 5th higher, or conversely by notating melodies using B♮ a 5th lower or a 4th higher, this was never systematically done. Many melodies included both notes; and the alphabetization may originally have assumed two notes E and two notes F as well (see Jacobsthal, on the fate of these and the use of transposition to accommodate them). A notable early witness to the alphabetization of chant is the celebrated tonary of St Bénigne, Dijon, written by 1031 (F-MOf H.159; facs. in PalMus, 1st ser., viii, 1904): this employs an alphabet from ‘a’ to ‘p’; for the two notes B the letters ‘i’ and ‘i’ are used. A copy of the tonary of Odorannus de Sens, made in Sens just before Odorannus’s death in 1046 (I-Rvat Reg.577), uses an inverted letter b, that is, q, for B♮. (For other systems and the decisions of medieval musicians regarding the tonality of chants with accidentals see M. Huglo: Les tonaires, 1971.)

Accidental 1. Early use.: Table 1

View large


Recognition of the two notes B therefore precedes Guido’s work on notation; and although the shapes and nomenclature that he suggested have become standard they are not purely the outcome of his hexachord theory and solmization technique, with which they are usually connected.

Apart from a few exceptional uses of other letters and/or unorthodox use of the coloured line system in early chant manuscripts (see Smits van Waesberghe, 1951, p.43), the ♭ sign and the ♮ or ♯ or quadratum signs were used for lowering or raising a note by a semitone on any degree of the scale. This is not to imply that every note could be both lowered and raised, and in practice it was always one or the other (e.g. C could not be lowered nor E raised) until the chromatic madrigals and fantasias of the late 16th century, when, for example, both G♭ and G♯, D♭ and D♯ might appear in the same piece. This meant that ♮ before E after a series of appearances of E♭ restored E♮, or in a piece with E♭ signature indicated a temporary E♮, but before F meant the modern F♯.

The choice of the quadratum, ♮ or ♯ signs for these purposes depended on the scribe’s training or the printer’s custom. The form ♮ for the quadratum is found as early as the 12th century, and ♯ in the 13th. The B iacente sign (‘recumbent B’: see Table 1 (vii)) gained ascendancy in the second half of the 15th century, the ♯ regaining supremacy in the 18th, by which time the ♮ was restricted to a cancelling function. German printers of the 16th century used the letter ‘h’ for the quadratum, and this passed into common German currency, which has B for B♭ and H for B♮.

The term molle has persisted in the French and Italian names for B♭; and molle and durum survive in the German terms for minor and major keys: moll (with flat or minor 3rd) and dur (with sharp or major 3rd). The application of the term diesis (Gk.) to the ♯ sign is a 14th-century development: the term has quite different origins (see Diesis).

2. Accidentals and solmization.

In Guido’s Micrologus hexachord theory is at an early stage of formation: at first only transposition of the series C–D–E–F–G–A up a 5th was discussed, and not until the 12th century was the distinction between the hexachordum naturale (C–D–E–F–G–A), the hexachordum molle (literally ‘hexachord with the soft-cornered B’: F–G–A–B♭–C–D) and the hexachordum durum (‘hexachord with the sharp-cornered B’: G–A–B♮–C–D–E) regularly made (see Smits van Waesberghe, 1969, p.116).

The hexachord system and its solmization (fitting the syllables ut–re–mi–fa–sol–la to the hexachord, always understanding a semitone between mi and fa) facilitated the memorization of the position of semitones in a plainchant melody, and in the case of B♭ and B♮ helped to avoid the need to think of a note which, written on the staff, had two possible meanings. Most music in the Middle Ages and Renaissance was performed from memory, and its intervallic structure was learnt not as a series of visual signs on a page but in terms of hexachord constructions and solmization syllables. This affected the attitude to written music of even the most accomplished solo performers who could sight-read their parts (but whose training would have been dominated by solmization from their earliest years). A proper skill in solmization obviated the need for a use of accidentals as explicit as in modern practice: once the performer had decided that a passage belonged to certain hexachords, its intervallic structure was determined; only the accidentals necessary to ensure the correct choice of hexachord, and thus the correct solmization, needed to be entered in a copy of the music. They were often written several notes in advance of the one they affected most directly: in terms of solmization they affected all the notes of that passage. Reconstruction of the decisions in solmization likely to have been made by performers of medieval and Renaissance polyphony is a prerequisite of editing and performing the music.

The use of ♭ as a prefacing signature at the beginning of a staff (or even, in the absence of any other letter, as a clef in early English sources: see Clef) is found in the earliest manuscripts using the staff (11th–12th centuries). A two-flat signature appears in the conductus Hac in die rege nato (I-Fl 29.1, f.332r) in both voices for six whole systems (f.333r); the piece begins and ends without signature. Such signatures affected the interval structure of the whole piece (or a long section of a piece), but the intervals were as liable to modification in detail as a piece without signature. Much early polyphony relied on the interval of the 5th for most of its structurally important harmonies; this meant that the voices of, for example, a two-part composition might very frequently be in a relationship parallel to that of the hexachordum naturale (with next highest note B♭) to the hexachordum durum (with B♮). Consequently pieces with one more flat in the signature of the lower part(s) than that of the upper part(s) are not rare in the 13th century and became very common in the 14th.

3. Use from the 17th century.

The introduction of bar-lines to mark off regular metric periods in music (as opposed to bar-lines that merely help to coordinate voices of polyphony notated in score, which are of no metrical significance) is found in German organ tablature of the 15th century, and in printed music of the late 1520s. But not until the end of the 17th century were bar-lines generally understood to terminate the effect of accidentals. By this time too the influence of modal theory on the use of signatures had waned. Discrepancies to be found in early 18th-century music between modal and the modern tonal practice chiefly occur in pieces written in the modern ‘minor keys’, but notated in Dorian (d–d′ without signature) or transposed Dorian modes (one flat less than modern practice). Such pieces are Bach’s ‘Dorian’ Toccata and Fugue (bwv538) and, from the Clavierübung, iii, bwv680, 681–2 (one extra sharp), 683–5 and 689. The modern forms of double flat and double sharp were also accepted generally by the 18th century.

The restriction of the efficacy of an accidental to the note immediately succeeding it was first suggested and practised by composers of chromatic madrigals at the end of the 16th century (e.g. Lassus, Vincenzo Ruffo etc.: see Kroyer, p.81). The difficulties caused by strongly chromatic and atonal music became critical at the end of the 19th century. While Schoenberg, in many works from the last movement of the Second String Quartet op.10 (1907–8), prefaced every note with an accidental (♯, ♭ or ♮) which was to apply to one note only, Dieren (Six Sketches, 1911) and Busoni (Sonatina seconda, 1912) used ♯ and ♭ but not ♮, again to affect only the note they immediately preceded.


  • G. Jacobsthal: Die chromatische Alteration im liturgischen Gesang der abendländischen Kirche (Berlin, 1897/R)
  • T. Kroyer: Die Anfänge der Chromatik im italienischen Madrigal des XVI. Jahrhunderts (Leipzig, 1902/R)
  • J. Wolf: Handbuch der Notationskunde (Leipzig, 1913–19)
  • W. Apel: The Notation of Polyphonic Music, 900–1600 (Cambridge, MA, 1942, 5/1953)
  • J. Smits van Waesberghe: ‘The Musical Notation of Guido of Arezzo’, MD, 5 (1951), 15–53 [Eng. trans. of pp.47–85 of De musico paedagogico et theoretico Guidone Aretino, 1953]
  • J. Smits van Waesberghe: Musikerziehung, Musikgeschichte in Bildern, 3/3 (Leipzig, 1969)
  • H. Besseler and P. Gülke: Schriftbild der mehrstimmigen Musik, Musikgeschichte in Bildern, 3/5 (Leipzig, 1972)
  • L. East: ‘Busoni and van Dieren’, Soundings, 5 (1975), 44–54
  • B. Stäblein: Schriftbild der einstimmigen Musik, Musikgeschichte in Bildern, 3/4 (Leipzig, 1975)
Musical Antiquary
Florence, Biblioteca Medicea Laurenziana
Montpellier, Bibliothèque Inter-Universitaire, Section Médecine
Châlons-en-Champagne, Bibliothèque Municipale
Musica disciplina
Rome, Biblioteca Apostolica Vaticana