The Performance Practice of Maria Callas – Interpretation and Instinct

by Dr. Robert E. Seletsky


Maria Callas holds an assured place in the pantheon of great artists, but details concerning the components of her greatness are often mired in platitudes that conflate biography, public persona, and myth with her actual approach to opera.  Moreover, Callas worked in a milieu that itself has always been mercurial and difficult to evaluate.

One of the principal and most frequently repeated achievements attributed to Callas, but one never clearly examined, is her revival of forgotten repertoire and the performance traditions that accompanied it.  The partly accurate, if slightly sycophantic, perspective frequently offered by Callas lovers and apologists is that with her unusual blend of vocal richness and agility, she sought to emulate great singers of the bel canto era like Giuditta Pasta and Maria Malibran, successfully restoring their styles and the operas with which they were associated, many long dormant.  While on the surface, such a statement seems plausible, there are more complex issues involved.  


“Historical performance practice” is a curious catch-all phrase that now encompasses music from the beginnings of notation until the mid 20th century.  As musicologist Richard Taruskin has observed in print on several occasions,[1] the idea that a performer can reproduce the sounds of the past, or even if it were possible, that we would understand or like them, is illusory.  Of course, Taruskin and others refer to the restoration of instruments, techniques, styles, and repertory that were superseded and often largely forgotten.  It is more difficult to discuss performance practice and restoration in terms of opera, as it is not pure music but an amalgam of music with theatre, textual declamation, etc. (hence the generic term “opera”—literally, “work”); moreover, opera was never a stationary art form: it was subject to revision from the first day of rehearsal.  18th-century revivals of operas by such great 17th-century French composers as Lully were reported to be considerably different in texture, tempi, approach, and orchestration despite the fact that opera at the Académie Royale de Musique was the most conservative in Europe, its function propagandistic as well as artistic.  In Italy, from the opening of the first public opera house at Venice in 1637, and anywhere Italian opera was performed, the situation was immeasurably more fluid, making it largely impossible for modern scholars to establish critical editions applicable to more than a few performances.  Italian opera existed at the whim of the principal singers, who generally had to be accommodated by the composers, even if it meant substituting arias from unrelated sources written by others.  The situation stabilized somewhat in the middle of the 19th century but operatic content was by no means fixed until Verdi’s later years.  Not only singers, but producers, patrons, and censors had to be satisfied: thus the numerous versions or alterations of such relatively late works as RigolettoIl TrovatoreDon Carlo, and even Otello with a ballet added for Paris in 1894.

Unlike “historically informed” revivals of early instrumental music, it is difficult to understand the concept of “revival” in opera, an art form where many elements were never entirely abandoned; consciously or not, operatic performers are directly influenced by their antecedents.  Even today, opera has scarcely entered the “academy,” remaining open to conductors and singers who may think nothing of making cuts and rearrangements in works that have long existed in composers’ final approved versions, a methodology that would horrify artists and audiences in any other genre.

It is only against such a backdrop that one can evaluate Maria Callas in the role of revivalist.  There are two principal areas in which her contributions may be examined: the circumstances of the works revived, and their stylistic treatment in her interpretations; the latter category is well documented through extant recordings.


While the citation of Callas as a fountainhead for the re-familiarization of various works in not unfounded, Callas did not herself actually restore long-dormant works.  Indeed, even without her, there was considerable interest in unearthing early operas in Italy during the 1950s and 1960s.  Of the revivals for which she is given credit, [1] nearly none had been absent from the stage for long, [2] they were not operas that Callas herself discovered, and [3] few remained in the regular repertory without her particular genius.  Today a “revival” refers to the rediscovery of a long-neglected composition, a blend of musicology and performance.  Such revivals usually seem to involve pre-19th-century works, typically embracing them under “performance practice” and “historical instrument” rubrics.  Although these criteria certainly have been applied to operas by composers like Monteverdi, Cavalli, Lully, Marais, and Handel, operatic revivals more often involve works formerly in the active repertory that were dropped because of audience saturation or apathy.  Audience taste seems to be cyclical, so few operas, unless they had been entirely forgotten, are absent for terribly long.  A number of operas were revived for Callas because of her special abilities and preferences.  The list of such revivals is fairly extensive, earning Callas the reputation to which she alluded in interviews; e.g., speaking on-air with Norman Ross in Chicago during September 1957, she proudly touted her work as important “for the history of music.”  At the Teatro Eliseo in Rome, Callas was featured in Rossini’s Il Turco in Italia in 1950; at the Maggio Musicale Fiorentino: Verdi’s I Vespri Siciliani and Haydn’s Orfeo ed Euridice in 1951—Vespri also serving as her Scala debut later that year, Rossini’s Armida in 1952, and Cherubini’s Medée[Medea] in 1953; at La Scala: Verdi’s Macbeth in 1952, Gluck’s Alceste and Spontini’s La Vestale in 1954, Bellini’s La Sonnambula in 1955, Donizetti’s Anna Bolena and Gluck’s Iphigénie en Tauride [Ifigenia in Tauride] in 1957, Bellini’s Il Pirata in 1958, and—though not as the principal character—Donizetti’s Poliuto in 1960.

Callas is often, and rightly, credited with the revitalization of bel canto operas.  Her work sparked a generation of unusually agile or inventive singers like Sutherland, Sills, Caballé, and Gencer to take up the cause after she left the operatic arena, often with greater fidelity to the scores; but the enthusiasm for a bel canto revival was not especially long-lived, and most of these operas departed from the repertory again in the 1980s.  Sadly, performances of similar worthy but arcane works are now the exception rather than the rule.  It must also be asserted that Callas and her musical cohorts were seldom pioneers in the choice of works to be revived.  Few operas in the above list were entirely unknown and others had received smaller-scale performances shortly before they were picked up for Callas: Il Turco in Italia had been performed at Tanglewood in a production mounted by Boris Goldowsky in 1948, Haydn’s Orfeo ed Euridice was recorded in 1950 by the Haydn Society (though Callas was in the first staged performance, as the planned 1791 premiere never occurred), Macbeth was staged at Glyndebourne in 1938 and 1939, Anna Bolena was revived at Bergamo in 1956, and even the very obscure—however wonderful—Il Pirata had been performed in 1935 with Beniamino Gigli in the principal tenor part.  La Sonnambula never left the repertory, nor did the Gluck operas or I Vespri Siciliani in various textual permutations, and La Vestale was closely linked with Rosa Ponselle.  Medée, in the Italian-language pasticcio called Medea containing Franz Lachner’s 1855 recitatives that Callas adopted, had been a 1909 vehicle for Esther Mazzoleni.  Even Donizetti’s Poliuto, in its French incarnation Les Martyrs had been performed in the 20th century.  In essence, that leaves Armida as the only true Callas “revival.”

Ultimately, the Callas revivals apparently occurred because her riveting overall persona turned them into vehicles for the design and directorial skills of artists like Luchino Visconti or Franco Zeffirelli, overshadowing the care given to the far more significant and enduring musical achievements of the composers.

Performance Practices

The other important question applies to the revivals as well as the more standard earlier operas upon which Callas left her indelible mark, e.g., NormaPuritaniLuciaTraviata: what were the parameters of Callas’s performances with regard to stylistic matters?  Curiously, in the cold light of scholarly reality, Callas’s external attitude toward opera was often frustratingly unadventurous and ill-informed.  Not only was she content to observe so-called “traditional cuts” in standard operas—even in studio recordings, mechanically defending their necessity in order to “keep the action moving,” but her 1950s mentors, especially Tullio Serafin to whom she was most devoted, introduced further cuts—to which Callas never objected—than had been taken in many 4-minute acoustical and electrical recordings; e.g., second verses of “Vien diletto” from Puritani and “Addio del passato” from Traviatawere recorded by Luisa Tetrazzini in 1912 and 1913 respectively, but are never heard in Callas versions.  The frequently stated motivation for cuts in 1950s performances and recordings was fear that listeners would find the music unfamiliar and become impatient; such a proposition requires one to believe that in under forty years, opera lovers had forgotten an entire repertory and had become bored with the performance traditions it shared with standard works.  Unusual, unnecessary internal cuts even appear in Callas’s performances and recordings of veristic works like ToscaButterfly, and Pagliacci.  As for the revivals of dormant (or, as indicated, not so dormant) operas, most editions made for Callas were eviscerated, rearranged, and even re-composed to a point that the hand of the composer was sometimes scarcely perceptible.  In addition to the removal of entire arias and scenes, the editing consisted of numerous smaller splices that ruined the phrase structure and obscured the original character of the music; the cuts never really move the action forward as purported, but radically compromise the composers’ styles and forms.  Yet inexplicably, long and dreary sections of music, like the ballets in Armida and Macbeth, were retained in Callas performances largely untouched.  That Callas could proclaim herself a champion of bel canto composers (“great geniuses who often died in poverty”), and yet defend the unrecognizably mutilated editions prepared by her conductors, is an enormous contradiction.  Indeed, in the few cases where “traditionally” deleted passages were restored by Callas’s conductors, notably Karajan in the 1954 Scala and 1955 Berlin Lucias, and the 1956 EMI Trovatore, Callas never insisted that the restored music remain for future use: in later performances and her second studio recording of Lucia, the cuts are back; and in retrospect, one must wonder whether Karajan would have expanded Trovatore for an August 1956 EMI recording had not Tebaldi and her ensemble opened the same cuts, in addition to twelve more measures at the coda of Act I, for a Decca Trovatore one month earlier.

Even more surprisingly, Callas’s understanding of performance practice, as we now think of it, was quite threadbare and unstudied.  One can find many examples in Callas’s studio and live broadcast recordings where obvious and expected unwritten cadential trills and appoggiature are omitted; perhaps these errors would be understandable if the tradition had been long dead but it was not.  The cadential trill was basic for all performers from the 17thcentury through at least 1930; that professional, and some very famous, singers seem to have forgotten this extraordinarily basic device and its application—and in some cases, even the technique to execute it—makes one incredulous.  Not only can we hear these trills in recordings by early 20th-century bel canto exponents like Luisa Tetrazzini, but even a later non-specialist like Giannina Arangi-Lombardi, most noted as Gioconda, Aida, and Santuzza supplies the obvious, though unwritten, cadential trills on the final leading-tones of “Casta diva” and “Com’è bello” (from Donizetti’s Lucrezia Borgia) in electrical recordings made in 1929 and 1933 respectively; both arias were sung by Callas—the former practically her signature—without trills.  Callas didn’t know or attempt to learn that the double notes in the recitatives or arias of operas composed through the mid 19th century—often appearing after the downward interval of a third or a second, generally constituted a notational convention, the first of the two identical written pitches to be performed as an upper appoggiatura. Curiously, even some of Callas’s contemporaries seemed more aware of this practice, including, of all people, Renata Tebaldi, who interprets a noticeable double-note figure during “Sempre libera” in her 1954 Decca Traviata as the appoggiatura-resolution intended, while Callas never does.  It is true that in the wake of omnipresent verismo, a sort of collective stylistic amnesia settled upon many singers in the 1940s and 1950s, but one could have expected Callas, as the putative champion of dormant traditions, to inform herself better about relevant performance practice issues, especially as it might only have meant listening to a few recordings that weren’t even very old.

Until the period of Verdi’s mature works, operas were crafted in forms that not only permitted, but required, the creative collaboration of the performer, exemplified by the ornamentation/paraphrase of repeats in strophic arias and cabalettas.  Yet, such repeats are usually among the extensive cuts observed by Callas.  Callas was overly cautious in her approach to ornamentation, variation, and even in notated fioriture.  Indeed, her omissions in bel cantoworks, whether by her or her conductors’ design, often extend beyond cautiousness into musical distortion; prime disappointments, all in the music of Bellini, include the omission the repeats in the cabalettas “Vien diletto” inPuritani and “Ah bello a me ritorna” in Norma—the former recorded and embellished by Tetrazzini in 1912, and the latter by Rosa Raisa in 1929, and the disfigured finale to Pirata.  Viewed as indispensable, “traditional” interpolations like the standard, overly long cadenza for the Mad Scene in Lucia—never heard by Donizetti—were included by Callas from hackneyed textbook examples with little original material.  Likewise, when embellished repetitions were unavoidable, as in “Ah! non giunge” that concludes Sonnambula, Callas relied on unoriginal formulas largely identical to those in acoustical recordings by singers like Tetrazzini.  Improvisation or even prepared additions did not come naturally to Callas despite her perceived affinity for bel canto.

At the same time, Callas would not part with 20th-century performance mannerisms unknown to the composers of the older music she claimed to adore: e.g., until her upper register shredded entirely after 1959, Callas interpolated every high dominant or tonic note that the music would bear, frequently substituting them for more expressive or harmonically interesting written notes.  Conversely, again to cite Tetrazzini’s 1912 recording of “Vien diletto,” although that singer had the easiest and most plentiful high notes, she refrains from capping her inventively ornamented second verse with an interpolated dominant high E-flat.  In every performance of “Vien diletto,” Callas cuts the second verse and coda altogether, jumping to an interpolated E-flat, which then becomes the focal point, actually distorting the structure of the cabaletta and extinguishing any sense of Bellini’s real compositional style in favor of self-indulgent, anachronistic vocal-pyrotechnic fashion.

It is difficult to avoid the specter of Arturo Toscanini who espoused two dicta that, on the surface, seem reasonable, but in reality, spell the death of the varied interpretations expected by composers: [1] if one performs every written note and indication, the artist will have fulfilled the composer’s true intentions, and [2] “tradition” is the last bad performance.  The irony is that neither Toscanini nor Callas, who admired him, worked strictly in accordance with these principles.  Both introduced instantly recognizable and distinctly personal interpretations, and both adhered to traditions of the recent past, some genuinely of the “last bad performance” variety, including cuts—a practice that certainly contradicts any notion of performing everything notated by the composer.  The problem is that although her overall results are brilliantly inimitable—unlike those of many who lean on Toscanini’s admonitions, Callas could support the less desirable elements of her preferences—cuts, the omission of repeats and ornamentation—with one or the other of Toscanini’s artificial, contradictory axioms.

The Callas Paradox

Many observers are fond of reiterating the paradoxes that inform Maria Callas’s biography, most of which involve superficial issues of personal mismanagement (sometimes melodramatically seen as “operatic”); however, one paradox, never really addressed, concerns her perspective as a performing musician.  Callas was quite adamant about her distaste for strict specialization: she said on many occasions that despite a singer’s preferences, he or she must be able to sing everything.  Yet her career shows a great deal of specializing, and in fact, can be divided into two discrete periods: the dramatic soprano years and the bel canto years.  When Callas emigrated to Italy in 1947, she sang heavy dramatic parts like Gioconda, Aida, Tosca, Turandot, Isolde, Kundry, Brünnhilde virtually without exception; even Norma and the Forza Leonora were somehow included in the “dramatic” category at the time.  After 1949, when she famously replaced an indisposed Margherita Carosio in Puritani on a week’s notice while still singing several Brünnhildes, she began to drop “dramatic” repertoire in favor of the more flexible bel canto roles to which she remained most dedicated at the apex of her fame.  Beginning in 1953, however, EMI needed her to record standard roles that she had discarded or had never sung; perhaps because of the schism between her performing and recording careers, she was able to give the impression of “singing everything.”

Despite the divergent repertoire that she sang during different segments of her career, Callas was not a specialist in the contemporary sense.  A period performer today will not only eschew inappropriate repertoire but will learn specific stylistic vocabularies of musical expression and rhetoric.  Callas treated all repertoire the same, creating the illusion of specificity with her unassailably polished approach to music as a whole.  Callas was convinced that everything she needed was in the score and in the relationship of the music to the text.  While this view did allow her to fall into the “Toscanini trap,” in its essence, her contention was largely sound.  Rudolf Bing, as is well-known, indicated that having once heard Callas, it was difficult to listen to anyone else sing the same music.[2]  Callas was extraordinary in being utterly convincing while providing the listener with incomplete information.  It is absolutely true of music written during and after Verdi’s later years that scrupulous attention to detail may yield profound interpretive insights: no one has ever equaled Callas on this account.  The difficulty is that in earlier works, all the information is not on the page—a deliberate move on the part of the composers, who anticipated the aforementioned creative collaboration with the performer.  Callas’s efforts to sculpt every phrase and syllable, while producing results artistically in a class of their own, were unfinished because she did not educate herself about pertinent stylistic matters.

Callas’s ultimate paradox is that her preparation allowed her to inhabit roles in ways no one ever has, each a complete living individual-in-music, but her identical approach to all repertoire, however brilliant, caused a lacuna in stylistic specificity: complete distinctness ironically emerging from complete similarity.  When she switched from dramatic soprano to bel canto operas, she invented a new category for herself: the “dramatic coloratura,” bringing the thrilling boldness of her Aida and Isolde to Elvira and Lucia, with an unusual vocal flexibility, brilliance, subtlety, and precision of voice, technique, and word inflection.  So, from her perspective, she could very well say that a professional must sing “everything.”  But she was not alone: singers did perform whatever they were hired to sing, and we find countless examples like Lehmann, Tetrazzini, and Raisa singing bel canto as well as Aida, Amelia, or Tosca without a second thought, although they were better versed in stylistic distinctions than Callas.  It is impossible to claim that Callas was emulating Pasta or Malibran, because while they did “sing everything” as Callas claimed inaccurately had not been the case since their time, compositional styles as remote as those of Puccini or Wagner were not issues for them, so comparisons with Callas are specious.


Having aired Callas’s stylistic and intellectual anomalies and the logical flaws in typical evaluations of her contributions, it is interesting to determine exactly what, in Callas’s performances, continues to make her the operatic artist whose recordings outsell all other singers’ after fifty years, whose name is recognized by persons for whom opera is alien and often anathema, and whose results set the standard by which other singers are measured, whether or not they admit it.  The single word that sums it up is authenticity.  In historical performance circles, the word has a bad reputation, often describing performances with correct external and musicological trappings that lack musical or artistic vitality: a sense of the academy or the museum rather than the stage.  As such, “authenticity” is the second subheading of this paper, placed in quotes.  It is now invoked without irony.  Despite lapses in details of performance practice—the inclusion of anachronistic elements and the exclusion of appropriate ones, the use of corrupt scores, and inaccurate historical assumptions, Callas’s approach to opera raised to an unparalleled peak its most significant, universal element: the complete—and in her case, uncanny—fusion of musical complexity and textual significance.  With Callas, one often feels that the original essence of opera (and its forerunner, monody) has been achieved: the quality described by the poets, musicians, and philosophers of the late 16th-century Florentine Camerata as “heightened speech.”  Through the centuries, opera has ebbed and flowed between this aesthetic and that of pure singing unconcerned with text (the latter satirized as early as Benedetto Marcello’s Il teatro alla moda, ca. 1720 ); every “reform” in opera, whether successful or not (Gluck, Wagner, etc.), was an effort to reinstate its first intent.  Callas is the only musician, certainly in the 20th century, who demonstrated that neither music nor text had to be compromised; she created a synthesis that was more potent than either.

Detractors often label Callas a “singing actress.”  Such a characterization misses the point of her art.  On stage, she acted very minimally, moving far less than most other singers.  Her apologists say that her acting was in her singing.  They too are mistaken.  There is no “acting” in Callas’s work at its best.  Rather, she used her infinitely layered approach to music, with an inimitable subtlety of note-length and stress, flexible rubato controlled over an unerringly resilient tactus, expressively realized dynamics, and complex phrase shapes to create living musico-textual entities.  In the rare films of her performances, either staged or in concert, even when she is not singing, one is struck by her active listening, the silences enveloped into her overarching concept.  Her results are greater than the sum of their parts: not simply musical or dramatic, they are stylized musical poetry.  Whatever she sang feels inevitable even when the listener is fully aware of problems with performance practice and editorial mishandling.  Callas does not need to be seen; her insights were entirely based on elements in the music so she is equally riveting in recordings.  Callas’s musical insights are authentic in the most profound sense, her art a transcendent probing of the music and an evocation of its inherent humanity.  We are moved by her authenticity because, solely through music, she taps into the most genuine and significant of our shared experiences.  Opera, the most stylized and artificial of forms, ironically becomes the medium for the revelation of artistic truth, mirrored in Callas’s performances as the most immediate expression of our own identification with the Self.

Eight Epiphanies

An enormous, perhaps disproportionate, amount has been written about Callas’s special treatment of the operas with which she is identified or for which her performances have become legendary: e.g., NormaSonnambula,TraviataMedeaPuritaniArmidaMacbethGiocondaTosca (the latter somewhat anomalous given her low opinion of the opera).  I prefer to note several less discussed moments of Callas’s transcendent interpretation.  It quickly becomes apparent in all repertoire that Callas never took a note or syllable for granted, elevating the music and text with an exhaustive, multi-leveled realization of their inherent potential.

1. Mimì in La Bohème is a role that Callas only learned for EMI, yet her development of its musico-rhetorical content is beyond the accomplishments of other recorded Mimìs.  Callas’s moments of musical and human revelation in her complete 1956 EMI Bohème are so numerous that they would constitute a separate article, but it is almost her first utterance that sets her apart.  Near the beginning of her introduction aria, “Sì, mi chiamano Mimì,” Callas “pronounces” the following simple line unlike any other artist: “La storia mia è breve.”  Singers generally don’t bother much about the line; it is intoned on a single pitch and is less interesting to them than what precedes or follows it.  Callas, on the other hand, treats it as a necessary focus, with a shape that carries everything else forward.  The first four words are lifted with a slight crescendo to “breve” on which Callas stresses the first syllable, relaxing, shortening, and covering the second, singing it more softly—using vocal weight, time, dynamics, and vocal coloring to bring it to life.  It is all apparent in the instrumental scoring, yet no other singers take their cue from the music which, in turn, reflects natural spoken rhythm.  Callas narrates the line as “heightened speech”: “breve” consists of a longer, stressed syllable followed by a shorter unstressed one.  Callas was not even a native Italian speaker, yet not one other recorded Mimì, Italian or otherwise, finds the correct interaction of music and text here.  Callas shaped the line similarly when she recorded the aria for her 1954 EMI Puccini recital, but in the complete 1956 recording, she is so much more attuned to every nuance of the music that her manner of intoning the line has far-reaching interpretive ramifications.

2. Callas had only limited experience with Gilda in Rigoletto—two performances in Mexico City on 17 and 21 June 1952 that were undermined by insufficient rehearsal as well as mediocre colleagues (except Giuseppe di Stefano) and instrumentalists.  Callas is, of course, thrilling and moving, her skill with the material typical of her approach, and the impeccable ease with which she sings emblematic of her pre-diet voice.  Yet it is the slim Callas, in her EMI recording of 1955, who achieves deeper insights into the music than any other interpreter—including herself in 1952.  Writers commonly discuss the rapturous approach needed to illuminate the set of variations that comprise “Caro nome”; only Callas infuses them with mounting human passion as the figuration increases in speed and complexity.  The moment that interests me is the end of the cadenza.  Verdi writes b/b-sharp/c-sharp/c-double-sharp at the end of an ornamental sixteenth-note melisma followed by a half-note trill on d-sharp with a termination consisting of a nachschlag e/d-sharp slurred to a dotted eighth-/sixteenth-note c-sharp/d-sharp—all on “sa-,” resolving to e—“rà.”  Callas combines a stylish musico-dramatic reading with a modern interpolation.  She phrases during the melisma before the chromatic four-note ascending b/b-sharp/c-sharp/c-double-sharp—“ca-ro no-me,” begins the d-sharp trill with a stressed appoggiatura on e like an 18th-century trill, and instead of a direct resolution to e after the trill termination, she elides the d-sharp of the termination to a long appoggiatura on g-sharp—all on “tu-”—slurred to a short, unstressed f-sharp resolution—“-o,” and sings a high b—“sa”—as a separate interpolated climactic note to the e resolution—“rà.”  Many other singers have opted for the notes that Callas sings in their extensions to Verdi’s original cadenza, but only Callas understood how they could be used to enhance the sense of rapture. The added appoggiatura e to the trill on d-sharp is musically and emotionally right. Like other singers, Callas elides the termination of the trill to the upper g-sharp, but unlike any other singer, she continues through to the f-sharp, making the g-sharp/f-sharp an analogous appoggiatura-resolution to the e-appoggiatura/d-sharp trill.  By stark contrast, when all other singers elide the termination of an appoggiatura-less trill to the g-sharp and insert a break, then make of the f-sharp a short anacrusis to the interpolated high b, one is left wondering if they ever had a day’s training in harmony or have any sense of musical, rhetorical, and emotional language.  Callas certainly must have realized how right it was because it is not original with her, but rather, she derived it, note-for-note (in the “Caro nome” key of E major), including the appoggiature, from Verdi’s own flute/clarinet introduction to “Me, pellegrina ed orfana” from La forza del destino—a work which she had recorded the previous year.

3. I now choose a moment in a role that is one of Callas’s signatures: Violetta in Traviata.  But I will discuss a detail that, while obvious in the music, is somehow missed by every other Violetta, however great (though after Callas, other versions only seem valuable if they contain passages cut in Callas’s edition).  At the end of the opera, thinking she is on the verge of recovery, the dying Violetta sings “O gioia!”: the stressed syllable “gioi-” is set on a double whole-note high b-flat, the “-a” on a brief quarter-note a-flat.  Clearly this is another appoggiatura-resolution moment.  It is the ultimate word-painting in opera, the b-flat representing false hope and the a-flat, despair; yet only Callas seems to be aware of it: in every performance throughout her career, and in her 1953 Cetra recording, the b-flat is full and triumphant, soon tapering to a tiny, nearly inaudible a-flat resolution as Violetta collapses.  Callas makes it incomparably moving despite the quaint period melodramatics of the scene.  She does it through purely musical means, with nothing beyond a scrupulous reading of the score and an accurate use of the most common expressive written note—the appoggiatura—raised to a summit of intensity.  Other singers seem not to notice that the b-flat is an appoggiatura at all, blithely singing both notes with equal weight and volume, unaware of Verdi’s intended melding of music, text, and human experience, thereby destroying a moment when the listener might otherwise have been transported beyond the artificiality of opera.

4. After such a lofty and gripping moment, I turn to something lighter.  As noted, Rossini’s Il Turco in Italia was one of the works that had a small revival before it was tagged for Callas in 1950.  Sadly, no complete live recordings have surfaced of what were, by all accounts, glorious, if cut, performances by Callas in Rome on 19, 22, 25, 29 October 1950,[3] but the slender Callas recorded the work for EMI in 1954.  She sings the role of Fiorilla with incandescent and subtle humor, illuminating the music and text scrupulously from beginning to end.  The moment that stands out for me is a single line of recitativo semplice.  In Act I, scene 2, as Fiorilla serves coffee to a smitten Selim with a very relaxed, dignified, if playful, manner, Callas introduces a slyly seductive tone: she could even wring complexity out of the simplest musical form, the secco recitative.  She inquires “Il zucchero è bastante?”—“Is there enough sugar?”  The second syllable of “bastante” is elongated and tapered in a subtly suggestive manner: she is clearly not merely asking about the sugar in the coffee, but is teasing Selim, saying in essence, “Have I excited you enough?”  The simple line becomes a sexy double entendre, and yet no one but Callas ever seems to grasp its delicious potential, at least not enough to find a musical means of expressing it; yet it’s right in the score.  Callas was said to lack a sense of humor, but evidently music animated and transformed her at every level, just as her use of music transformed the listener.  Indeed, in the 1955 Scala production, there was a line that Callas reportedly wanted to speak but could not deliver convincingly, finally performing it as recitative: no “singing actress.”[4]

5. Another light moment that sets Callas apart occurs in her 1957 EMI recording of Il Barbiere di Siviglia.  Although her performances of the work at La Scala in 1956—16, 21 February and 3, 6, 15 March—were poorly received, one hardly has to defend the EMI recording, thought even by many non-Callas enthusiasts to be definitive, despite extensive cuts totaling thirty minutes of music.  In an incomparable “Una voce poco fa,” Callas sees an opportunity for a subtle use of word-painting in the line “E contenta io resterò”: she broadens “resterò”—“I will remain” with a hint of rubato, quite literally “remaining” on each syllable very slightly longer than the orchestra.  Indeed, her repose makes complete sense of the text “and happy I will remain,” subtly setting the tone for a multi-dimensional, musically elastic Rosina.  Her 1954 EMI recording of the aria on her Lyric and Coloratura disc is sung more deliberately and with more effusive embellishment, but in the 1957 version she achieves a nexus of music, text, and sentiment.  Both this and her irreplaceable musical treatment of Fiorilla in Turco serve to refute the assertion made by her friend and most prolific chronicler, the late John Ardoin, that her Rossini singing “lacked charm.”[5]

6. Callas uses a fascinating permutation of word-painting in an aria of a very different nature.  In “Tu che le vanità” that opens the final act of—and in a sense, summarizes—Verdi’s Don Carlo, she imparts an unsettling quality to the line “Il cor ha un sol desir: la pace dell’avel”—“My heart has one desire: the peace of the grave.”  In keeping with her complete involvement and identification through music, Callas allows a singular, almost unreal quality in her voice to emerge on the word “avel,” a smothered, “bottled” sound.  This quality that sometimes crept into Callas’s voice is often described as a flaw, but as she often did with her vocal anomalies, Callas used it in the service of high art.  In her 1958 EMI Verdi Heroines disc as well as live performances on 15 May 1959 in Hamburg, 19 May 1959 in Stuttgart, and 30 May 1961 in London (the latter with piano only, played by Sir Malcolm Sargent), Callas’s unearthly, smothered sound permeates the stressed second syllable of “avel.”  In effect, she not only expresses the world-weariness and resignation felt by Elisabetta, but gives the listener a chilling, visceral musical depiction of the grave itself.  In the performances of 11 July 1959 in Amsterdam and 4 November 1962 in London, Callas performs “avel” instead with a crescendo and a diminuendo respectively; although both are less effective to these ears, Callas, shown as a probing, instinctive artist, is never more present.

7. Turandot was a role that Callas last undertook on stage in 1949.  She recorded “In questa reggia” compellingly in her 1954 EMI Puccini recital, and returned to the role only for a belated complete 1957 EMI recording, by which time she was unable to draw upon the kind of rich vocal resources she had at her disposal in 1949.  Yet she brings subtleties to the 1957 recording that she may not have considered as a younger musician.[6]  Two instances, both during “In questa reggia,” come to mind for their complex interplay of music, text, and human expression.  Callas’s Turandot begins the aria with the necessary stentorian quality as she addresses the Unknown Prince and the crowd.  In the seventh line of text, as she begins to narrate the tragic story of her ancestor, she uses the name “Lo-u-Ling” as a pivot point, with liquid portamento and dynamic-coloristic tapering, to arrive at an entirely different vocal tone filled with wistfulness and nostalgia.  Callas’s Turandot seems suddenly hypnotized as she slips not only into narrative mode but into personal identification with the subject of the narrative; Callas thus projects a not inappropriate schizophrenia in the character.  Yet, as usual, she distorts nothing in the music, taking Puccini’s shift to a dream-like orchestration as the basis for her own approach.  She and Puccini thus indicate that Turandot is possessed, and using only musical means, Callas draws us into the half-lit world of imagined memory.  Again we ask: how could everyone else miss something obvious in the score that has the ability to deliver the listener into an unsuspected sublime realm?  The second event is more word-painting: at the end of the second verse, Turandot sings “là nella notte atroce, dove si spense la sua fresca voce”—“there on that horrid night, her young voice was silenced.”  Callas covers the sound of her voice throughout the line, and “voce” is sung with a muted, still sound, the second syllable very short and tapered—music, text, narrative simultaneously encompassed to invoke the pain and regret of innocence lost through violence.  As with the Bohème aria discussed above, Callas was aware of the interpretive possibilities even in her 1954 EMI Puccini recital disc, but her interpretation is considerably more complex in the complete Turandot recording.

8. My last example comes from a period after the zenith of Callas’s career.  Callas never performed the role of Desdemona in Otello, yet the long opening scene of the fourth act, recorded for EMI on 16 and 17 December 1963, comprises nearly half of her second Verdi recital.  In common with the Turandot example above, the scene combines a layering of time and narrative.  Desdemona confides in her maid Emilia about her troubled state of mind as she prepares to retire for the night.  Her prosaic actions are interspersed with long reveries about the unhappy romance experienced by her mother’s young servant Barbara and the “Canzon del Salice”—“Willow Song”—that she sang, and which Desdemona repeats.  Callas creates an even more compelling reality of Verdi’s already brilliant multi-textured scena.  Using several vocal tones, she emphasizes Desdemona’s growing paranoia and presentiment of death each time she returns to reality; the nostalgic reveries, intended to distract herself, eventually connect her distant thoughts with her present predicament.  For the foreground, Callas sings in a weary, almost parlando style; the mid-ground reminiscences about Barbara are lyrical and melancholy; and the “Willow Song,” beginning as a background event, is performed in the desolate, almost hypnotic manner suggested by Verdi’s scoring, a sad lullaby that Callas sometimes almost hums rather than sings, its unaccompanied refrains of “Salce” vibrato-less, distant, and hollow.  As the scene and the form—reality/narrative/”Willow Song”—proceed, they are eventually fused, Callas’s Desdemona passing through vague depression, terror, and finally, resignation where her voice in the present is as doleful as it was in the narrative, the three levels resolved with her tender, careworn “Emilia, addio.  Come m’ardon le ciglia!  È presagio di pianto.  Buona notte.”—“Goodbye Emilia.  How my eyes burn!  It means that tears are coming.  Good night.”  As the realization of her fate strikes her at last, Desdemona’s wrenching “Ah, Emilia, Emilia, addio, Emilia, addio!,” in stark contrast with everything that preceded it, is made achingly poignant by Callas’s impassioned use of dynamics and phrase-shaping.  The scena is emotionally shattering and transfiguring for the listener.  At this stage in her life, Callas’s voice was in tatters, but even lacking her earlier vocal riches, she extracts every grain of almost unbearable expressivity contained in the score.


The foregoing examples are the proverbial “tip of the iceberg,” not even moments that are generally discussed by Callas lovers.  That the scrutiny of every musical utterance produced by this artist seems to reveal a similar wealth of detail is breath-taking and nearly unimaginable.  It is frustrating that many grumble about the longevity of Callas’s renown, thinking it is generated only by exploitative tabloid journalists and by the endless crass marketing to which her recorded legacy continues to be subjected especially by EMI nearly three decades after her death.  Sensationalistic, often unfounded, gossip posing as biography, a focus on the early decay of her great instrument, and her posthumous treatment as a commodity all diffuse the reality of Callas’s irreplaceable artistry.  As stated earlier, she was not equipped with the necessary stylistic vocabulary for entirely accurate representations of all the music she chose to sing, but what she did bring to music was self-evidently far more important.  I have often mused that had Callas been transported to a rehearsal of an opera led by a composer like Rossini, Donizetti, or Bellini, any of them would have praised her interpretation as the pinnacle of their intentions, while adding “but you might like to think about adding cadential trills in measure xappoggiature in recitative y, and embellishments for cabaletta z, as we are going to observe the repeat.”  The priorities are clear: considerations of performance practice are important but do not outweigh a sublime elevation of music to the level of the human spirit achieved by Maria Callas.

Recordings cited (with Callas unless noted)

CD catalogue numbers indicate most easily available versions; if several are equally available, the best is given.  Second catalogue numbers are listed for less accessible but better editions when the first listing is much inferior.  Citations do not denote approval, especially of EMI.  Indeed, as I wrote in “Callas at EMI: Remastering and Perception,” The Opera Quarterly, Vol. 16, No. 2 (Spring 2000), pp. 240-55, EMI CD remasterings of 1997 and later often have compromised sound and disfiguring edits; therefore, Rigoletto, for example, though cited in its presently available version, EMI 5 56327 2 (appallingly missing the “s” on the high B “sarà,” relevant to this paper) is also listed in the superior previous remastering, EMI 7 47469 2.

Cited in Performance Practices section:

Luisa Tetrazzini: Bellini, I Puritani, “Vien diletto,” rec. 10 June1912; Verdi, La Traviata, “Addio del passato,” rec. 26 September 1913 (Tetrazzini arias in Phonographe 5099)
Donizetti, Lucia di Lammermoor, 18 January 1954, La Scala, Milan (SRO 831-2)
Lucia di Lammermoor, 29 September 1955, Berlin (EMI 5 66441 2; Melodram 26004)
Verdi, Il
 Trovatore, rec. 3-9 August 1956 (EMI 5 56333 2)
Renata Tebaldi, etc.: Il Trovatore, rec. July 1956 (Decca 448 743-2) 
Giannina Arangi-Lombardi: Bellini, Norma, “Casta diva” rec. July 1929
Arangi-Lombardi: Donizetti, Lucrezia Borgia, “Com’è bello,” rec. 7 December 1933
(Arangi-Lombardi arias in appendix of Naxos 8.110112-14—Ponchielli, La Gioconda)
Tebaldi, etc.: La Traviata, rec. August 1954 (Decca 430-250-2): CD-1, Track 7, 1:27
I Puritani, rec. 24 March-3 April 1953 (EMI 5 56275 2 [budget packaging: 5 85647 2])
Norma, rec. 23 April-3 May, 1954 (EMI 5 56271 2)
Rosa Raisa: Norma, “Ah bello a me ritorna,” rec. 1929, unpub., (in Marston MR 53001)
Bellini, Il
 Pirata, 27 January 1959, New York (EMI 5 66432 2; Melodram 26013)  

Cited in Eight Epiphanies section:

Puccini, La Bohème, rec. 20-25 August, 3-4 September 1956 (EMI 5 56295 2; EMI 7 47475 2)
Puccini Arias, rec. 15-21 September 1954 (EMI 5 66463 2; EMI 7 47966 2)
Verdi, Rigoletto, 17 June 1952, Mexico City (Opera d’Oro 1253)
Rigoletto, rec. 3-16 September 1955 (EMI 5 56327 2; EMI 7 47469 2)
La Traviata, rec. by Cetra 15-19 September 1953 (Naxos 8.110300-01)
Rossini, Il
 Turco in Italia, rec. 31 August-8 September 1954 (EMI 5 56313 2)
Rossini, Il Barbiere di Siviglia, 16 February 1956, La Scala, Milan (Opera d’Oro 1264)
Il Barbiere di Siviglia, rec. 7-14 February 1957 (EMI 5 56310 2; EMI 7 47634 2)
Lyric and Coloratura Arias, rec. 15-21 September 1954 (EMI 5 66458 2; EMI 7 47282 2)
Verdi, Don Carlo, “Tu che le vanità”:
Verdi Heroines [Verdi Arias Vol. I], rec. 19-25 September 1958 (EMI 5 66460 2)
15 May 1959, Hamburg (Gala GL-325)
19 May 1959, Stuttgart (EMI 5 62682 2; Palladio PD-4188)
30 May 1961, London (Gala GL-322)
11 July 1959, Amsterdam (EMI 5 72030 2)
4 November 1962, London (Gala GL-322)
Puccini, Turandot, rec. 9-15 July 1957 (EMI 5 56307 2; EMI 7 47971 2)
Maria Callas Sings Verdi Arias [Verdi Arias Vol. II], rec. 16 December 1963-24 April 1964 (EMI 5 66461 2)

Robert E. Seletsky is an independent scholar and professional musician with published work appearing in New Grove II (2000), Early MusicThe Opera QuarterlyRecent Researches in the Music of the Classical EraThe New Harvard Dictionary of Music, and elsewhere. This article is revised and lengthened from a version initially appearing in The Opera Quarterly, Vol. 20, No. 4 (Autumn 2004), p. 587-602.

[1] e.g., Richard Taruskin, “The Pastness of the Present and the Presence of the Past,” Nicholas Kenyon, ed., Authenticity and Early Music (London: Oxford University Press, 1997), pp. 137-210.

[2] Interview with former Metropolitan opera general manager Rudolf Bing in the 1978 PBS (WNET) documentary Callas written by John Ardoin and narrated by Franco Zeffirelli.  Bing repeated the sentiment in several subsequent documentaries about Callas.

[3] For notes about Turco’s 1950 reception, see: Michael Scott, Maria Meneghini Callas (London: Simon and Schuster, 1991; 2/Boston: Northeastern University Press, 1992), pp. 65-6.  In October 2004, an exceptionally important, newly discovered recording of “Non si dà follia maggiore” from the 19 October 1950 performance was issued on Tima Club Clama CD 48 as a bonus track to a CD otherwise devoted to Greek tenor Ulisse Lappas.  Callas’s 1950 approach is far freer, with typically cleaner, easier articulation and richer voice than the brilliant but more conservative 1954 EMI recording; she also interpolates a marvelous, unexpected cadenza terminating on a magnificent high E-flat.  The performances of 19 and 25 October 1950 were broadcast and RAI recorded them for later re-airing.  The sound on the CD is excellent, its background scratch indicating a customary RAI transcription disc as its source.  It is hoped that the entire set of discs containing the performance will surface.

[4] Zeffirelli, who staged the 1955 Scala production of Turco, recalled that “Rossini left a line, ‘Che Turca impertinente!  Osa a Fiorilla l’amante disputar! [That impertinent Turkish girl!  Daring to vie for Fiorilla’s lover!], as a kind of ad lib without music.  Maria said, ‘I want to speak it to prove I can act.’  But she was embarrassingly bad and finally our conductor, Gianandrea Gavazzeni, added some music under the line.  Then Maria did it perfectly, with marvelous charm and humor”; in: John Ardoin and Gerald Fitzgerald, Callas: the Art and the Life, the Great Years (New York: Holt, Rinehart and Winston, 1974), p. 111.  Zeffirelli seems to be citing the wrong lines as these, in Act II, scene 1, are already written as recitativo semplice by Rossini and are heard as such in the 1954 EMI recording that predates the Scala production by more than seven months.

[5] John Ardoin, The Callas Legacy (New York: Scribner, 1977, 2/1982), p. 186.  In discussing the Rossini and Donizetti Arias LP, recorded 4-23 December 1963 and 13-24 April 1964 (first issued as Columbia SAX 2564 and Angel 36239, now on CD as EMI 5 66464 2) he writes, “Oddly, given her abundance of rich expressive gifts, Callas almost consistently lacked charm, if one measures her Rossini singing against a Supervia.  Callas could use her voice trippingly, even beguilingly, but the feminine charm with which Supervia continually disarms one on record is missing.  Fortunately, the vast majority of Callas’s repertory was such that charm was not an essential ingredient.”  It is strange that Ardoin commented on the matter in a recording made so late in Callas’s career—one considered perhaps her least successful; yet he ignores it in when discussing the two brilliant complete Rossini comic operas she recorded for EMI during her great years.

[6] No audio document of those 1949 performances survive except for a two-minute fragment from the end of Act III.  The “Turandot Act II” issued by Rodolphe Records in 1984, purporting to be from a live 1949 Buenos Aires performance, has been proven fraudulent; see my report in The Opera Quarterly, Vol. 16, No. 3 (Summer 2000), pp. 538-9, as well as a complementary article on this site by Milan Petkovic.

©2004/8 by Robert E. Seletsky