Rondo and Ritornello (sixth draft, Nov. 20, 2001)

by William Renwick

This essay provides an overview of issues relating to ritornello and rondo.  I would like to acknowledge the contributions of the students of Music 701 (2001) which have been important in formulating the ideas presented here.


Ritornello has its roots in the Baroque genres of opera and concerto. It is not clear whether the instrumental ritornello forms were adopted from the operatic aria or developed separately.  Cole, p. 172, notes the use of ritonello in both instrumental and vocal forms of the early baroque, specifically in the operatic works of Peri and Monteverdi.

The vocal forms arise through Monteverdi and the Italian and Neapolitan masters, including Alessandro Scarlatti, and appear throughout the baroque opera, oratorio, and cantata repertoire. The ritornello aria forms were broken down by the end of the Baroque era, as less formal and more dramatic forms gained ascendancy. The instrumental forms seem to have developed through Torelli, (perhaps Corelli), Vivaldi, and Bach and Handel, and thence into the classical concerto movements.

A typical da-capo aria form with ritornelli would look like this:

r (I) A (I-V) r (V) A' (V-I) r (I)
B (vi)
r (I) A (I-V) r (V) A' (V-I) r (I)

The ritornello itself usually comprises several, contrasted phrases and concludes with a definitive PAC in I.   Ritornelli are often designed so that they can be repeated later on in abbreviated form, by omitting one or more of their component phrases.

As the diagram indicates, the A section of a ritornello aria is normally a binary form, consisting of A and A'.  Rosen suggests that the "normal" tonal structure of the A' is I-I, not V-I.  This seems like a small point, but in fact it is one of the foundation stones of his theory that sonata form grows out of aria form.  Ratner suggests that the "normal" tonal structure of the A' is x-I, where x is neither I nor V, but a noticeable tonal contrast instead.

The B section is set off as being in a contrasting key, and without ritornelli.

It has been noted that a ritornello form need not necessarily end with a final statment of the ritornello, although this is certainly the norm.  I have to ask what effect the omission of a final ritornello statement may have on the perception of musical form and closure.  One may look to the larger context of an aria within an opera, cantata or oratorio, to find answers to these questions.

Rosen includes a brief description of the basic  concerto form as follows:
r   Solo 1 (I-V)  r  Solo 2 (x-vi)  r  Solo 3  (I)   r

This too may be an interesting platform for comparison with both aria and rondo forms.

It seems logical that the instrumental forms are more likely to make a distinction in thematic content between ritornello and solo, whereas the vocal forms are more likely to use similar material in the ritornello and solo sections.


The rondo is a very basic and ancient musical form, with roots going back to refrain forms and dance forms of the oral tradition. In the common practice period, the rondeau appears throughout the Baroque era as a component of dance suites--principally instrumental compositions. Cole, p. 173, suggests that the evolution of the rondo from the French baroque usage to the classical usage has not been adequately investigated.

In the Classical era, the rondo finds its place as a middle or final movment of sonatas, chamber music, symphonies, and concertos. The early classical rondos were simple in design and of "little true inner value" (Forkel, in Cole, 173)--hence trivial in comparison with the evidently dialectical construction of the sonata form.  Cole makes the important connection of the instrumental rondo of the eighteenth century, with the opera buffa use of the vocal rondo, in a light, simple, pleasing, charming style.

The basic rondo form looks like:

A B A C A D A etc.

where A is the main theme or refrain (or, confusingly, the ritornello), and B, C, D, etc. are the couplets or episodes.

Green stipulates that to count as a rondo the refrain must appear at least three times. Cole, among others, stipulates that the rondo will conclude with the main theme. (p. 172).

In essence, therefore, the rondo is an open ended form. It presumes literal repetitions of A (the theme) in the original (tonic) key, although the repetitions are often ornamented. The theme itself always ends in the tonic, for it is this ending that will at some point be the ending for the piece as a whole. Caplin, p. 231, suggests that the theme can be ternary, rounded binary or binary, and will close with a PAC. It seems to me, however, that ternary, as defined in McMAC 1999, is less suited for a rondo theme, because ternary is inherently based on contrast. Any contrast set up within the A theme will de-emphasize the contrast that ought to be established at the onset of the B and later themes. Rather, the A section is more typically static, in order that the couplets can develop dramatic tension. In keeping with the static form, Rondo themes usually exhibit regular, symmetrical phrasing.

Couperin was among the first to use abbreviated repetitions of the rondo theme (Cole, p. 172).  Such abbreviations would typically be understood in relation to the degree of repetition inherent in the composition.  Couperin also experimented with transposing the refrain to other keys. Caplin, p. 233, notes that in abridged repetitions of the refrain, it is most usual that the abridgement still ends with a PAC in I.  (How does this compare with ritornello?)

Heinrich Schenker, p. 141, views rondo as the conjoining of two ternary forms thus:

A B A + A C A = A B A C A

In this way, Schenker likens the essential structure of rondo to that of ternary form.

Theorists seem to differ as to whether variety and contrast amongst the couplets is necessary to establish the rondo form. That is, whether A-B-A-B'-A is truly a rondo,or merely an extended ternary form in essence.

The B,C, D, etc. sections--the episodes or couplets or digressions--are based on the idea of contrast to the A theme.   Nevertheless, a continuing issue in rondo is the degree of similarity, complementarity, or difference associated with the sections.  Cole, p. 174, notes that C.P.E. Bach developed couplets of an open design rather than a closed binary or ternary substructure.   Caplin takes this idea further, and suggests that in principle the first couplet will either function as a "subordinate-theme complex", like the transition, second theme and closing theme of a sonata form, or as an "interior theme", similar to the B section of a ternary form (Caplin, p. 231).   Caplin suggests that the former is more typical of a B section and the latter for a C section;  thus, "C" is construed as having more developmental and contrasting quality, as compared to the "B" section.  Caplin, p. 233, strongly suggests that in the case of a "subordinate-theme complex", the various parts are to be considered as all under the umbrella of  "B".  This is a distinction which is perhaps more important for theory than for practice.  It comes into question when we begin to think of classical rondos as variations of sonata form.

Among the important means of contrast is that of key.  Rameau established something of a standard by using the order V and vi for the two couplets of a major key rondeau, and III and v for the two couplets of a minor key rondeau.  Many later rondos bear important relations to these fundamental harmonic plans.

Classical rondos admit of introductions, codas, transitional and re-transitional passages. It may be interesting to see how such passages are used in specific movements. Cole suggests that Leclair, in the later 18th century, was one of the first to adopt the retransition as a linking passage.   The retransition, in fact, often becomes one of the dramatic focal points of the rondo form.

We are finding in the classical repertoire several interesting phenomena. First, the ending of the rondo theme is often not as clear cut as the theory would suggest.  Composers make use of several PACs in quick succession, any of which could be used as the "real" ending of the theme.  Further, we are finding extensive use of transitions and retransitions in the classical rondo, which provide more dramatic and developmental qualities to the couplets. These devices are essentially borrowed from the sonata-allegro form; indeed, in many later rondos, the B section follows the plan of a sonata-allegro transition, second theme and closing group.  Finally, we do find the intermixture of elements from the rondo theme and the couplet themes, particularly near the end of a rondo.

All of these "abberations" may be understood primarily as means of investing the basically static form with various degrees of dramatic and narrative elements.  These trends are increasingly evident in sonata Rondo (see below).   Levy and Levarie, pp. 248-249, make very clear the idea that it is the varied sections of a piece--the solos of a ritornello, the episodes of a rondo--that hold the narrative content of the music.

Cole, p. 174, points to the fantasy as a concept that may be used to understand the great variety of treatments that C.P.E. Bach and later composers employed.  In a similar way, we can see how Beethoven, in his rondo finales, often makes direct or hidden reference to earlier movements, adding another layer of fomal and dramatic complexity.  Accoding to Cole, p. 174, : both Haydn and Mozart "moved from a simple, sectional structure to a complex, integrated form into which he built surprise and variety, and within which he attempted to offest and even exploit the regularity inherent in the traditional layout."

Rondo and Ritornello Compared

Occasionally one finds the term ritornello used to denote the theme of a Rondo. This terminiological confusion is indicative of the complex relationship between the two forms.  The fact that the term ritornello means literally a "small return" helps us to understand that the ritornello by no means should be equated with the rondo return in terms of its structural primacy.   Wallace Berry refers to "the rondo principle in the da capo forms", and points particularly to the use of multiple trio sections in instrumental forms such as the Bach orchestral suites.  It is not clear at this point why these are considered da capo forms, however, except that the composer typically uses the designation "Menuet I da Capo" for example, instead of writing the movment again.  The typical form here would be A B A C A, where each section is fully closed.  This is another important way in which the Rondo idea also connects with the ternary form, as Schenker suggests (vide supra), as A B A conjonied to A C A.

Interestingly, William Caplin notes the similarity of thematic structures between rondo and ritornello. [source]   Cole, p. 172, notes that "parallels between the later rondo and the ritornello principle and the rondo cantata need to be more thoroughly investigated."  Both forms have in common the idea that something--Green, p. 153, calls it a refrain--is repeated from time to time.

The essential difference--at least in theory--is that in rondo it is the main idea that is repeated, whereas in ritornello, it is subordinate material (interludes) that is repeated.  How true is this in each of the cases we are studying?   In either case, the material is normally a self-contained, harmonically "closed" passage (Green, p. 153).

[Levy and Levarie, p. 249, make the distinction between whether the material in question is "an appendix" or the "main section of the piece".  Unfortunately they give opposite meanings to the terms ritornello--as the main theme--and refrain (read rondo) as an appendix.]

According to Cole, p. 174, the ritornello idea is to an extent transferred to C.P.E. Bach's rondos as he uses transposed repetitions of the main theme.

Joel Galand portrays the complexity of the issue with sensitivity:
"The late eighteenth-century rondo cannot be distinguished from expanded-binary or sonata form on the bases of harmonic plan or patterns of thematic recurrence, development, and contrast.  If one points to the regular return of the refrain in the tonic, another might note the existence of the modulating rondo and its recognition by theorists like Turk.  But the return of opening material in other keys articulates ritornello forms generally, and ritornello construction in turn characterizes not only concertos but also symphonies and chamber music, which often bear concertante traits.  We might categorize rondos as song forms, but in the eighteenth century they are often ritornello forms; we hear a return as a new beginning, and impetus for further expansion, rather than as a discrete frame for a contrasting middle section." (p. 30).   "The rondo, in short, was not so much a form as a loosely defined genre that could be adapted to any number of formal procedures. . . .  Genre characteristics cut across formal boundaries, and formal procedures cut across generic categories."    Despite this view, composersfrequently labelled movements as rondos, implying a formal procedure.   Perhaps here we need better definitions of form and genre before the argument can proceed.  Galand, p. 37, goes further and suggests that "Two principles need to be distinguished:  that of variation or modification, and that of contrast and return.  Both, of course, depend on the recurrence of a readily identifiable theme, and this appears to be the one necessary requirement for a rondo."

Sonata Rondo.

Classical rondo and da-capo form: It is interesting to note that a closer rapprochment of rondo and ritornello appears when we consider what Green calls the "classical rondo". Its form is A B A C A B A, where the first B is in V. This is more usually called the sonata-rondo; as its name suggests, it is a hybrid of ""rondo design" and "sonata-allegro tonal plan".  (Cole, p. 175.)

A B (V) A C (vi) A B' (I) A

This form bears close affinity to the da capo ritornello form described above,and begins to look rather like ternary as well.   Caplin, p. 235, calls sonata rondo "perhaps the most complex of the classical forms."   Caplin seems to consider the final A statement of the Sonata-rondo to be a coda; as he says: "the coda is a required element of sonata-rondo, because that section includes the final return of the main theme." (Caplin, p. 235).   I must admit puzzlement here, because, in the usual notion of Coda, the music is inessential to the completeness of the work, as an appendix might be for a book.  But if a coda is necessary, how then is it a coda?  Caplin, p. 239 admits that this is problematic.  He clarifies, however:  ""there is not consistent relationship between the beginning of the coda and the beginning of the final refrain." P. 235.  --"The former embraces the latter. . . .the rondo refrain always appears somewhere in the coda."  The interesting thing, then, is why this final statement should be so functionally different from the previous ones.    Is it still a rondo if the final refrain is omitted??

This appears to resemble the da-capo form if A is considered the ritornello (see Caplin, p. 233). However, the dynamic qualities tend to be much different. In particular, the da capo form A-section would typically be a binary form in itself. Further, the first thematic statement in da capo would be I, not V.   A question arises out of this:  what factors in the music seem to demand a continuation beyond the A B A into C  A  and possibly beyond?  Why doesn't the music simply stop at the end of the repeat of the refrain?  Also, in the case of ABACABA, what are the factors that demand a return of B?

Caplin, p. 235, also seems to liken later returns of A to a recapitulatory function.

Apparently, Mozart's K 157, string quartet, contains the earliest known example of a sonata-rondo.
Mozart later tended to abbreviate the form by omitting the third statement of A.

However, Sonata Rondo form begs the question of whether the form is really sonata or really rondo:  Rosen implies the former when he calls it "sonata-finale", and goes on to say that it is sonata form with extra thematic statements interposed;  Ratner implies the latter when he calls it a "couplet-form".

Postscript:  Prelude Form:    Because Jane is having an existential problem over her analysis of Telemann's apparent ritornello instrumental music, I thought I would add here a short summary of what one might call "prelude" form.   In Bach's Precepts and Principles, we find several figured-bass exercises that are simply founded upon the transposed repeition of materials according to the following pattern:

A (I)  A' (V)  A'' (vi) A (I).

Interestingly the final A in not written out, but simply indicated as da capo.    These forms demonstrate a very simple basis for creating musical form out of any material at hand;  they exhibit some features that we associate with rondo, and some of ritornello.  Similar structures are also found, often with figurated link passages, in The Langloz Manuscript which stems from the Bach circle.  The figurated passages illustrate how the basic form grows into something more like a regular ritornello form:

A (I)  bridge A' (V)  bridge A'' (vi) bridge A (I).

where the bridges may be developmental or episodic.

Joel Galand, p.39, suggests several  formal plans that combine ritornello with simple binary and with exposition-recapitulation binary.  His models point up the idea of an A following B as perhaps a new begining rather than a repetition of old material.  This connects well with the techniques often seen in sonata forms, where the music after the double bar commences with a repeated or varied form of A.  This connects also with the question of whether intermediate rondo or ritornello statements are seen as returns to home base or as bridges between developmental sections.

Bach, J.S., Precepts and Principles  ed. Pamela Poulin, OUP
Berry, Wallace, Form in Music  (1976?)
Cole, Malcolm, "Rondo", New Grove XVI-172.
Caplin, William, Classical Form, New York, Oxford, 1998.
Galand, Joel, "From, Genre, and Style in the Eighteenth-Century Rondo", Music Theory Spectrum XVII-5 (Spring, 1995), 27-52.
Green, Douglass, From in Tonal Music.
Hutchings, Arthur, "Ritornello", New Grove XVI-57
Levy, Ernst, and Siegmund Levarie. Musical Morphology: A Discourse and a dictionary.  Kent State, 1983.
Ratner, Leonard, Classic Music  ML 195 R38
Renwick, William, The Langloz Manuscript, OUP, 2001
Rosen, Charles, Sonata Forms  ML 1156 R67 1988 (on reserve)
Westrup, Jack, "Aria", New Grove I-573-579.