When Schubert’s songs are presented on the concert or recital stage today, one consistently hears interpretations that adhere strictly to the printed score. This performance convention is clearly related to the concepts of ‘Werktreue’ or ‘Notentreue’, which, as Eric van Tassel argued, only emerged in the last decades of the nineteenth century and is manifested in the theoretical construct of the so-called ‘Urtext’ Editions. Yet the vocal performance practice of Schubert’s time differed substantially from the customs of conventional modern lied singing, as music historiography has made evident. For example, Walther Dürr, lieder editor of the Neue Schubert Gesamtausgabe, repeatedly stated in his publications that in lied performance of the first half of the nineteenth century, it was quite common for singers to make changes to the score and add ornamentation. In particular, he pointed to manuscripts of compositions by the renowned Schubert singer Johann Michael Vogl (1768–1840) in which text-driven, interpretive mannerisms can be found. These written-out examples of successful alterations and ornaments of Schubert’s lieder can also be brought in contact with the so-called declamatory and dramatic school of singing, of which soprano Anna Milder-Hauptmann (1785–1838) and Johann Michael Vogl were regarded as leading proponents.
Apart from changes to the score, even the notated ornamentation in Schubert’s lieder raises questions for the modern performer. Although this has received attention in the Schubert literature, issues remain in regard to the actual execution of specific elements that are usually notated with symbols or abbreviations only. Further discussion of ornamentation in Schubert songs seems worthwhile, and the discussion can be enhanced by information from treatises on the art of singing by August Swoboda (1795–1863) and Anton Rösner (1771–1841), two writers whose works have not received their due attention. Their proximity to Schubert’s interpreters and to the composer himself makes Swoboda and Rösner authoritative witnesses to contemporary performance practice, and their singing manuals offer answers to many relevant questions.
The Vocal Pedagogy of August Swoboda and Anton Rösner
Born in Vienna two years before Schubert, August Swoboda was a member of the Gesellschaft der Musikfreunde (Society of the Friends of Music), which Schubert joined as a member in 1821. Swoboda is also included, along with the well-known Schubert song interpreter Ludwig Tietze (1797–1850), in the register of singers in Anton Ziegler’s address book of musical artists from the year 1823 – a list of musicians who were active in Vienna at the time. Although Swoboda was apparently not a professional soloist he was a composer, music theorist and, from 1828, director of his own music school which existed for at least 20 years. Firmly rooted in Viennese musical life, he seems almost as if predestined to write his singing treatise (Gesanglehre), published in 1827.
While Swoboda was active primarily as a teacher, Anton Rösner was a professional singer, and his career also appears to have been located at the centre of Viennese musical life. In 1801, he was engaged together with his wife at the Viennese Hofoper (Court Opera), after the pair had been working in Dessau. His activity at the Viennese court opera extended until 1814, where he performed among the ensemble of the ‘Deutsche Oper’ alongside Anna Milder-Hauptmann and the later famous Schubert interpreter Johann Michael Vogl; he thus found himself at the centre of the declamatory and dramatic school of singing. As negative reviews indicate, Rösner was not especially successful as a vocal soloist, but, like Vogl, he was also cast as an actor. In 1810, Rösner briefly assumed the post of choir director at the Hofoper, filling in for Philipp Taddäus Korner (1761–1831), who was, at that very time, teacher to the young choirboy Franz Schubert. Rösner was also actively involved in theatre management, filling in 1813 the post of an Economic Adjunct (Ökonomie Adjunctes) at the Hofoper. Finally, in 1822 Rösner attained the post of Vocal Pedagogy Professor at the Conservatorium der Musikfreunde, founded just a few years earlier and without doubt the most important musical education facility in Vienna at this time. Here, too, he succeeded Philipp Taddäus Korner, and he held this professorship until his death. In an evening’s programme of the Gesellschaft für Musikfreunde in 1833 Rösner is listed as a singer of Schubert’s music, and with three colleagues he performed the vocal quartet ‘Die Nachtigall’ (D.724).
A two-volume pedagogical work little known to researchers until now, Rösner’s Leitfaden einer Gesanglehre (Guidelines for Vocal Pedagogy) can be regarded as the publication of his lesson materials at the Conservatorium der Musikfreunde. Especially the second volume (ca. 1835), directed at advanced students, may be consulted as a primary source of information for the ornamentation and performance practice of Schubert lieder, due to the clear proximity of its author to the composer himself. The work had lasting influence on Viennese vocal pedagogy during the first half of the nineteenth century, being republished by Rösner’s colleague Laurenz Weiss (1810–1888) as the Theoretisch-practische Gesang-Schule für das Conservatorium der Musik in Wien (Theoretical-practical Vocal School for the Conservatory of Music in Vienna, ca. 1842).
Approaching Viennese ornamental practice of lieder at the beginning of the nineteenth century
While in the eighteenth century ornaments were freely added to published music by most singers, the general trend gradually over the next century would appear to be one of composer-determined ornaments. Nevertheless, even in 1827 August Swoboda regarded ornaments as elements to be added to an interpretation by singers, independent of the composer: ‘Manieren [manners] are what one names certain ornaments of singing, which the singer usually applies at his own discretion’. In his vocal textbook Anton Rösner also refers explicitly to the autonomy of singers vis-à-vis the notes on the page. However, he also demands that this practice fit the character of the composition:
The singer must [know] how to ornament on his own, and to apply it to the proper places, but is also warned not to sing too many ornaments, so that the melody is not made unrecognizable. The ornaments must harmonize with the character of the piece, since every vocal piece has its own character, according to the sentiment that rules within it.
One might expect that different genres of vocal music were handled differently in relation to the praxis of ornamentation, so that, for instance, a grand form such as an aria would be more ornamented than a German lied. Yet in his Beytrag zur Lehre von den Verzierungen (Report on the Teaching of Ornamentation, 1814) Friedrich Rochlitz (1769–1842) specifically includes the small vocal genres as being within the bounds of ornamentation and formulates almost identical thoughts to that of Rösner on this matter, connecting ornamentation to the character of a piece of music:
the small, simpler, polyphonic piece, the Romanze, the Canzonette with its relatives, the German lied in its broad diversity, and more of the same – these songs are, and indeed every piece, very special, closely and tightly bound to its text, and so one can do nothing more than providing a piece of general advice: consider what is called the character of the piece of music and do it consequently, and according to this, ornament or do not ornament, add ornaments abundantly or sparingly, choose the ornaments in the same way, and perform them also in this way.
Ornamentation in lieder was common practice: this we know from concert reviews during Schubert’s time. When they did not fulfil the premises of Rochlitz and Rösner, however, they were criticized, as is made clear by this review from the year 1817:
‘Adelaide’ by L. v. Beethoven, sung by a dilettante, whose voice is bright and clear; only some of the ornaments added to this piece were not really in the spirit of this expressive composition.
In such concert reviews of the time, successful ornamentation was rarely singled out separately; in a stylistically appropriate interpretation, ornamentation was a matter of course. If during a lied recital an excess of elements such as coloratura, trills and other mannerisms were heard, only then would one take the opportunity to admonish:
‘Sehnsucht’ by Göthe, composed by Mr. Häusler, royal Bavarian Music Director, sung by Mr. Siebert. Who will deny that Mr. Siebert sang quite artfully, when this artful singing consists of endless coloraturas, trills, ornaments and leaps?
The bass mentioned in this review, Franz Siebert (1788–1858) is furthermore proven to have been a Schubert singer; he sang the bass part in the premiere of Schubert’s duet inserted into Das Zauberglöckchen, which took place in 1821 at the Kärntnertortheater. And an anecdote of Schubert’s friend Franz Lachner (1803–1890) assures us that the composer and singer knew each other privately as well.
Grace notes, turns and trills
We shall now consider specific ornaments that interpreters meet in Schubert’s lieder. These ornaments are usually not written out, but are indicated by symbols or abbreviations. Fortunately, uncertainties regarding the execution can be almost completely resolved in consultation with the treatises from the Schubertian environment.
One finds in Schubert’s lieder various types of turns, as well as the mordent and its inverted form; however, the trill plays no role in his lieder, in distinction to his instrumental music and his arias. This observation along with the information from the aforementioned sources lead to the conclusion that in the contemporary performance practice of the Schubert lied, all ornaments except for the trill could also be added by the interpreter independently, if they match the expression of the work.
In both the autograph manuscripts of Schubert’s lieder and Rösner’s treatise grace notes appear exclusively without crossed-out note stems. Beyond the self-evident exceptions (for example, a quaver before a minim or semibreve, which indicates a short grace note) the rule applies that crotchets and quavers receive a long grace note, while semiquavers receive a short one, as Rösner points out (Figure 1). Therefore, contrary to the usual custom in modern performance, quavers or crotchets with crossed-out note stems should most likely to be understood as long grace notes.
Figure 1 Grace notes in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 9.
In a dotted note with a grace note, the grace note receives two parts of the duration, the main note one part (Figure 2).
Figure 2 Dotted note with a grace note in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 9.
Although some scholars have advocated the performance of short grace before the beat, in music of this period, contemporary vocal tutors are clear: all grace notes, whether long or short, should fall directly on the beat.
As to the question of how grace notes are to be resolved at the end of a phrase, a so-called ‘syllabic’ solution has been most often proposed, and this is the solution offered by the new Schubert Gesamtausgabe. The ‘syllabic’ solution entails an exchange of the main note for the grace note. In contrast, in the so-called ‘melismatic’ solution the grace note as well as the main note are sounded in the duration of the main note, thus creating a small melisma on the syllable that underlies the main note. In a lengthy discussion about ‘the long appoggiatura’, David Montgomery argues in favour of the so-called ‘melismatic’ solution and thus opposes examples of the ‘syllabic’ grace notes at the end of a phrase, which are convincingly cited by Clive Brown. Rösner seems to answer this interesting question in favour of the melismatic solution, as can be seen in bar 2 of Figure 3. He indirectly indicates his reasoning for this execution, in explaining that the essence of the existence of grace notes lies in their accentuation (note also, in bar 5 of Figure 3, that Rösner treats the semiquaver grace notes before a quaver as long):
The long grace notes always receive more emphasis than the note to which they belong, otherwise the grace note would cease to be an ornament.
Figure 3 Melismatic grace notes in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 11
In fact, the four known versions of the song ‘Erlkönig’ (op. 1, D.328) demonstrate that Schubert himself may have treated the question of the suspension at the end of a phrase differently, and thus Clive Brown’s argument can be supported with further examples. Of the four versions of ‘Erlkönig’ the first is preserved only in a transcription by Schubert’s friend Albert Stadler (1794–1884), while the fourth is based on the first edition from 1821 by Cappi and Diabelli. However, autographs of versions 2 and 3 have also survived and prove to be particularly relevant to this topic. In bar 26 of ‘Erlkönig’, a syllabic version is written out in Stadler’s first version, while Schubert himself provides a grace note with a quaver in both manuscripts (see Figure 4). And yet the first edition depicts a crochet grace note. For Schubert, there must have been a difference in execution between a crochet and a quaver, otherwise he would have given the grace notes in bars 26 and 28 the same duration, just as the publication ends up levelling them out.
Figure 4 Franz Schubert, ‘Erlkönig’ D.328, bars 25–32.
At the end of the phrase, Schubert actually means to suggest an exchange of the main note with the grace note in bar 30; this becomes clear by comparing the different versions of bar 134 of the same song. Here (as in the first edition) a quaver grace note appears in the second version, while in the third version Schubert clearly decides in favour of the syllabic version (see Figure 5, bar 134).
Figure 5 Franz Schubert, ‘Erlkönig’ D.328, bars 133–140.
In bars 134 and 139 Stadler omits the grace note (Figure 5). Yet Rösner confirms that at the end of a phrase and before rests, instead of two identical notes without a grace note, ‘the upper grace note of a whole tone above’ should be used (see Figure 6). When a phrase ends with two notes on the same pitch, the necessity of a raised first note is due to ‘incontrovertible rules of declamation’, a practice also argued at length by Schubert’s friend Leopold von Sonnleithner (1797–1873) as late as 1861.
Figure 6 ‘Recitative’ in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 9
Thus, an upper neighbour of a quaver may be used both in response to a quaver grace note and also (as in the last bar of Figure 6) where a repeated quaver ends the phrase with no grace note. This is not a universal rule, however. Appropriate text expression can overrule habitual ornamentation. For example, in the well-known lied ‘Der Doppelgänger’, the ‘calm’ (ruhen) in the phrase ‘es ruhen die Gassen’ is painted by the repetition of the F# on ‘Gassen’ in bar 8 (Figure 7). Here, an ornament on Gassen would destroy the effect. In this context, one can recall the above-mentioned passages by Rösner and Rochlitz that the ornaments must, above all, harmonize with the character and sentiment of a vocal piece.
Figure 7 Franz Schubert, Der Doppelgänger, from: Schwanengesang D.957, bars 5–8.
According to Rösner and all contemporary Viennese sources, turns indicated before the main note are to be executed directly on the beat, just like grace notes (see Figure 8).
Figure 8 Turns in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 9.
The turn with more than two notes (also known as gruppetto and Mordant (mordent in English)) deserves closer attention. In his Gesang-Lehre (ca. 1815) Joseph Preindl (1756–1823) knows only the sign (𝆗) for the ornament, and he assigns to this an execution beginning from the note above the main note; the inverted form of the mordent is written out in his work from the note below the main note at the beginning. Friedrich Starke (1819) assigns the same sign to an execution from the note above. For the ascending turn Starke explicitly adds the note below. In his Allgemeinen Theorie der Tonkunst (1826) August Swoboda also offers, along with the already-known symbol (𝆗) and its execution from the note above, the descending turn and the symbol (𝆘). Yet in his Gesanglehre published a year later (1827), Swoboda curiously exchanges the two symbols: ‘If the beginning cusp of the figure appears upwards [𝆘], then the same [turns] are made from above to below, and if it appears downwards [𝆗], from below to above’ (see Figure 9). This is not a mistake by the author, but an attempt to differentiate between the symbols for the turns. Johann Nepomuk Hummel (1778–1837) explains the turn in his piano treatise (Vienna 1828) in a way similar to Swoboda a year before: ‘It is presented and executed in three ways, namely: a) from the main note itself with an added little note (𝆘); b) from the note above (𝆘); and c) from the note below (𝆗)’.
Moreover, Hummel demands a distinction between these signs be made in music publications. Swoboda’s later version and Hummel are corroborated by both Rösner (ca. 1835) and Weiss (ca. 1842): ‘if the turn is to be sung from above, then the sign must begin from above, 𝆘; however, should it be made from below moving upwards, the symbol begins from below 𝆗’. In Schubert’s lieder the symbol 𝆗 appears most often, and it is to be assumed in the majority of cases that according to custom, an execution from above is meant. However, since we have seen that Swoboda and Rösner operated in close proximity to Schubert, and that Hummel was also active for many years in Vienna around this time, it can be deduced that an execution of this ornament can be realized from the note below as well.
Figure 9 Groppo, in August Swoboda, Gesanglehre (1827), p. 17.
In the Singschule (1794) of Ferdinand Kauer (1751–1831), the inverted mordent (𝆘) as well as the mordent sign (𝆗) are named ‘Zwiker’, (pince-nez) whereas in Preindl’s Gesanglehre (ca. 1815) only the inverted mordent sign (𝆘) appears. For Preindl, a mordent is the same as a turn. The execution remains clear, however: the inverted mordent is formed from the note above the main note, the mordent from the note below. Friedrich Starke’s Wiener Pianoforte-Schule (Viennese Pianoforte-School, 1819) and August Swoboda’s Allgemeine Theorie der Tonkunst (General Theory of Musical Art, 1826) differentiate between short trills and short as well as long mordents or ‘Beisser’ (Biters) more precisely and almost identically (see Figure 10). Both ornaments appear to have lost their meaning by this time, however: in Swoboda’s Gesanglehre published in the following year (1827), the short trill from above is called a ‘mordent’ while the actual mordent is left entirely unmentioned. Finally Rösner documents the decline of both ornaments (and the reprint from Weiss attests to this), in that he no longer differentiates between the sign 𝆗 and 𝆘, labelling both as ‘Mordant’ and characterizing them as being executed using the note below the main note.
Short trills after a semitone step down, as found in Schubert’s lieder such as ‘Der König in Thule’, D.367, are very likely to be executed in the manner described in Starke as well as Swoboda’s Theory of Musical Art (compare ‘Der Pralltriller’ in Figure 10 to Figure 11), and not as a short grace note or turn before the beat, as is most usually heard today.
Figure 10 Friedrich Starke, Wiener Pianoforte-Schule in III Abtheilungen (1819), p. 18; identical to ‘Mordant’ in August Swoboda, Allgemeine Theorie der Tonkunst (1826), pp. 48–49
Figure 11 Franz Schubert, ‘Der König in Thule’, D.367, bars 1–8
It should be noted that in both Schubert’s autograph and Diabelli’s first edition of ‘Der König in Thule’, the slur in bar 7 applies only to the first two notes c and b, and thus implies the execution handed down in the treatises (Figure 11). However, all modern editions perceive the slur in question as a sign of a melisma, and incorrectly lead it up to the third note (a).
In the vocal pedagogy works of Schubert’s time, portamento is described as a primary element of performance practice and is not considered a type of ornamentation. Portamento has a bad reputation in singing today, but it was central to vocal artistry of the nineteenth century and is worth discussing here.
As is made clear in the vocal pedagogy sources, portamento in Schubert’s time was regarded as a kind of performative execution in which during the succession of two tones, the first is bound to the second tone in lightly anticipated articulation of it. This involves neither a legato nor a glissando in the modern sense. After August Swoboda, in a section entitled ‘To Carry the Voice’ (‘Vom Tragen der Stimme’), describes the connection of two notes without portamento (i.e. legato), he explains the achievement of portamento with the following words:
§. 79. The second way to carry the voice occurs between two tones, which form a smaller or larger interval, and follow one another only in distant steps. It consists in letting the voice glide by means of an imperceptible bond between the two tones, so that the second tone is already approached a little beforehand. … §. 80. If such a portamento is applied from the depths to the heights, the voice passes from the weak to the strong by means of an imperceptible help of the throat, and vice-versa.
In other words, Swoboda observes that in portamento there also lies an inherent dynamic component. When the melody is led from a lower note to a higher one, a crescendo must be applied, and a decrescendo in the opposite case. For a better understanding of what has been said, he adds a musical example to his description (see Figure 12).
Figure 12 August Swoboda, Gesanglehre (Vienna 1827), p. 11: Portamento.
Anton Rösner confirms that portamento must be accomplished without glissando and expresses himself quite similarly to Swoboda in regard to its correct execution:
one needs to take special care that in a slow or heavy tempo, the sounds in between are unnoticeable: when transmitting in long notes, the first is broken off somewhat, as can be seen in the following example … In order to be able to fully effect the portamento, the lower sound must begin with moderate voice, the connection with a crescendo and the higher [tone] must be struck strongly; the same applies from above, that the higher sound is sounded strongly, yet the lower sound weakly.
Rösner supports his textual explanation of portamento with a musical example in which the anticipation of the second note is clearly notated (see Figure 13).
Figure 13 Portamento, in Anton Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 4; quasi identical to Laurenz Weiss, Theoretisch-practische Gesang-Schule vol. 3 (ca. 1842), p. 4.
From this it can be assumed that portamenti were left mainly without notation by the composers, and that this manner of performance was left entirely to the interpreter. In the prints of songs composed by the famous singer and friend of Schubert Johann Michael Vogl, one can even discover ornaments that can be clearly identified as portamenti. For example, in bar 7 of his ‘Song of Desdemona’ (see Figure 14), there is the anticipation of the F# of the following bar, added as the last quaver of the third crotchet, which is undoubtedly to be performed as a portamento. In bar 3 of this lied, it is also likely that Vogl indicates a kind of portamento – here, however, the anticipatory note is notated as a main note, while the lower note of the connection seems to appear as a grace note. In contrast to this, one could argue that what is notated in bar 5 (third crotchet) is an anticipation with two written quavers and a legato slur. These three similar tone combinations are undoubtedly fine gradations – the figure in bar 5 may denote a normal legato, while the other two examples appear to represent different portamento variants.
Figure 14 Johann Michael Vogl, Lied der Desdemona, from: Fünfzehn Lieder (ca. 1815).
The concepts of ‘Werktreue’ and ‘Notentreue’ appear to have led modern singers and pedagogues into a trap that limits personal interpretive expression in a way Schubert and his contemporaries would have found curiously bland. As sources presented here as well as others verify, ornamentation was also an integral part of vocal music during Schubert’s time. The genre of lied was expressly included in this art of ornamentation.
For questions pertaining to the exact execution of the ornaments in Schubert’s songs, usually only indicated by abbreviations and symbols, both Swoboda’s and Rösner’s tutorials can be used profitably, since the authors worked in the immediate vicinity of the composer and his main interpreters. Yet, as the illustrated case of the appoggiaturas at the end of a phrase indicates, in difficult cases the composer’s autographs ultimately remain the most meaningful witnesses of a lost performance practice.
With regard to a performative element that has been frowned upon today but was central at the time, portamento represents a special case. Both Swoboda and Rösner attest to the relevance of this contemporary stylistic device, which is easy to apply to the vocal performance of the lieder of Schubert and his time. It is to be hoped that future readers will accept the challenges to add their own ornamental inflection sensitive to each lied’s character.
 Eric van Tassel, ‘“Something Utterly New”: Listening to Schubert Lieder. 1: Vogl and the Declamatory Style’, Early Music 25/4 (1997), 702–14, at 708ff; Wolf-Dieter Seiffert, ‘“Urtext”. Ein historischer Überblick samt Begriffsdefinition’ G. Henle Verlag online, www.henle.de/de/feuilleton/was-ist-urtext/.
 Walther Dürr, ‘“Manier” und “Veränderung”, in Kompositionen Franz Schuberts’, in Zur Aufführungspraxis der Werke Franz Schubert, ed. Roswitha V. Karpf and Vera Schwarz (Munich: Katzbichler, 1981), 124–39, Dürr, ‘Virtuosität und Interpretation – Schubertlieder ausgeziert?’, in Internationales Symposium Musikerautographe. Bericht (= Publikationen des Instituts für Österreichische Musikdokumentation Vol. 16), ed. Ernst Hilmar (Tutzing: Katzbichler, 1990), 145–62, Dürr, ‘Geschriebenes und Ungeschriebenes: Vortrags-“Manieren” und Verzierungen’, in Schubert Handbuch, ed. Walther Dürr and Andreas Krause (Kassel: Bärenreiter, 1997), 104–11. See also Thomas Seedorf, ‘Aspekte des Liedgesangs’, in Neues Handbuch der Musikwissenschaft, Vol. 11, ed. Hermann Danuser (Laaber: Laaber, 1997), 339–41. On Vogl see: Andreas Liess, Johann Michael Vogl – Hofoperist und Schubert-Sänger (Graz: Böhlau, 1954).
 Martin Günther, Kunstlied als Liedkunst. Die Lieder Franz Schuberts in der musikalischen Aufführungskultur des 19. Jahrhunderts (Stuttgart: Steiner, 2016), 146–72. See also Van Tassel, ‘Something Utterly New’.
 Hartmut Krones and Robert Schollum, Vokale und allgemeine Aufführungspraxis (Vienna: Böhlau, 1983). At the end of the 1990s, a dispute arose about the practice of ornamentation concerning Schubert in general, in which the ornamentation of Schubert’s songs played a central role. See Malcolm Bilson, ‘The Future of Schubert Interpretation: What Is Really Needed?’, in Early Music 25/4 (1997), 715–22, David Montgomery, Robert Levin and Walter Dürr, ‘Exchanging Schubert for Schillings’, in Early Music 26/3 (1998), 533–35, David Montgomery, Franz Schubert’s Music in Performance (Hillsdale: Pendragon Press, 2003), 173–89. A recent PhD thesis by David Greco attempts to elaborate integrating aspects of the nineteenth-century vocal sound world. See David Greco, Waking the Dead Diva: Recovering the Expressive Sound World of Forgotten Nineteenth-Century Singers (PhD diss., University of Melbourne, 2020). While Greco admittedly addresses relevant issues, his thesis remains methodological questionable, as he completely omits a considerable body of literature that points to significant changes in vocal technique during the nineteenth century (e.g. several articles on the topic by Mauro Uberti, Marco Beghelli and Thomas Seedorf/Bernhard Richter). Particularly problematic is the omission of two seminal and easily accessible artistic dissertations in the field of nineteenth-century voice research by Alexander Mayr and Sarah Potter. See Alexander Mayr, Die voce faringea – Rekonstruktion einer vergessenen Kunst (PhD diss., Universität für Musik und darstellende Kunst Graz, 2014) and Sarah Potter, Changing Vocal Style and Technique in Britain During the Long Nineteenth Century (PhD diss., The University of Leeds, 2014).
 Carl Friedrich Pohl, Die Gesellschaft der Musikfreunde des österreichischen Kaiserstaats und ihr Conservatorium (Vienna: Graumüller, 1871), 16, and Otto Erich Deutsch, Schubert – Die Dokumente seines Lebens (Kassel: Bärenreiter, 1964), 141.
 Anton Ziegler, Adressen-Buch von Tonkünstlern … in Wien (Vienna: Strauß, 1823), 131, Schubert can be found as a pianist ibid., 133. For example, Tietze first performed Schubert’s song ‘Auf dem Strom’. See Allgemeine Theaterzeitung und Unterhaltungsblatt 37 (25 March 1828), 148.
 Christian Fastl, ‘Swoboda, August’, in Oesterreichisches Musiklexikon online, www.musiklexikon.ac.at/ml/musik_S/Swoboda_August.xml.
 August Swoboda, Gesanglehre (Vienna: Anton von Heykul, ). This pedagogical work was published without a date, but an advertisement in the Wiener Zeitung confirms that it was released in November 1827. See ‘Musikalische Anzeige’, in Oesterreichisch-Kaiserliche privilegierte Wiener Zeitung, 5 November 1827, 1139.
 The following sources were consulted for Rösner’s biographical information: Zeitung für die elegante Welt 3 (6 January 1801), 19; Wiener Theater Almanach (Vienna: Riedl, 1804–1814); Oesterreichisch-Kaiserliche privilegierte Wiener Zeitung 6 October 1810, 1431; Pohl, Die Gesellschaft der Musikfreunde, 127; Moritz von Prosky, Das herzogliche Hoftheater zu Dessau (Dessau: Bauman, 1885), 32; Beate Henneberg, Das Konservatorium der Gesellschaft der Musikfreunde in Wien (Vienna: Praesens, 2013), 269.
 In this context it should be remembered that Schubert, like Vogl, considered Anna Milder-Hauptmann to sing compositions that he is said to have written to suit her as perfectly as possible. See Christine Schumann, “The Role of Singing Schools for Schubert’s Vocal Works’, in Schubert Yearbook 1996, ed. Klaus-Gotthard Fischer and Christine Schumann (Duisburg: Bärenreiter, 1996), 114–15.
 Zeitung für Theater, Musik und Poesie 20 (31 May 1807), 121, and 15 (10 October 1807), 39.
 Christine Schumann, ‘Der Sängerknabe Franz Schubert – und die zeitgenössische Gesangschulung’, in Schubert Jahrbuch 2006–2009, ed. Volkmar Hansen and Silke Hoffmann (Kassel: Bärenreiter, 2009), 13–26.
 Otto Biba, ‘Franz Schubert in den musikalischen Abendunterhaltungen der Gesellschaft der Musikfreunde’, in Schubert-Studien, ed. Franz Grasberger and Othmar Wessely (Vienna: VÖAW, 1978), 7–31, here 23.
 Anton Rösner, Leitfaden einer Gesanglehre für Schüler, vol. 1 (Vienna: Anton Strauß Witwe 1832), vol. 2 (Vienna: Anton Diabelli, ca. 1835).
 On Weiss see Pohl, Die Gesellschaft der Musikfreunde, 127–28.
 ‘Manieren nennt man gewisse Verzierungen des Gesanges, die der Sänger meistens willkührlich anbringt’; Swoboda, Gesanglehre, 15.
 ‘Der Sänger muss selbst Verzierungen zu machen, und am gehörigen Orte anzubringen wissen, sich aber vor zu vielen Verzierungen in Acht nehmen, damit durch dieselben die Melodie nicht unkenntlich gemacht werde. Die Verzierungen müssen dem Charackter des Stückes harmoniren; denn jedes Gesangstück hat seinen eigenen Charackter, nach der Empfindung, die darin herrscht’; Rösner, Leitfaden einer Gesanglehre vol. 2/1, 19, and Laurenz Weiss, Theoretisch-practische Gesang-Schule, vol. 3: Theorie von der höhern Bildung des Gesanges mit Beispielen von der Aussprache des Textes (Vienna: A. Diabelli u. Comp., ca. 1842), 13. With this statement and the one prior to it, as well as the new sources investigated here, David Montgomery’s opinion that the singers of Schubert’s time did not use improvisational ornamentation is conclusively refuted. In this context it is also astonishing that Montgomery cites Swoboda’s singing theory as well as Weiss’s reprint of Rösner’s guide to singing in his bibliography, but obviously either does not know its content or knowingly withholds it (indeed the 2nd volume of Rösner’s guide is missing from Montgomery’s list). See Montgomery, Franz Schubert’s Music in Performance, 200, 296, 302.
 ‘das kleine, einfachere, mehrstimmige Stück, die Romanze, die Canzonette mit ihren Nebenzweigen, das deutsche Lied in seiner großen Mannigfaltigkeit, u. dergl. mehr – dieser Gesang ist, und zwar jedes Stück ganz speciell, nahe und eng an seinen Text gebunden; und so lässt sich denn im Allgemeinen nur rathen: Erwäge diesen, und mithin das, was man den Charakter des Musikstücks nennet, und diesem gemäß verziere oder verziere nicht, verziere reich oder spärlich, wähle deine Verzierungen eben so, und führe sie eben in dieser Weise aus’. Friedrich Rochlitz, ‘Beytrag zur Lehre von den Verzierungen’, in Allgemeine Musikalische Zeitung 8 (23 February 1814), cols 130–31.
 ‘Adelaide von L. v. Beethoven, gesungen von einem Dilettanten, dessen Stimme hell und rein ist; allein einige in diesem Musikstücke angebrachte Manieren waren nicht ganz im Geiste dieser ausdrucksvollen Composition’. Anon., ‘Concerte’, in Allgemeine Musikalische Zeitung mit besonderer Rücksicht auf den österreichischen Kaiserstaat nr. 38 (18 September 1817), col. 328.
 ‘Sehnsucht von Göthe, componirt von Herrn Häusler, königl. Bayr. Musik-Director, gesungen von Herrn Siebert. Wer wird läugnen dass Herr Siebert recht kunstreich gesungen hat, wenn der kunstreiche Gesang in endlosen Coloraturen, Trillern, Verzierungen und Sprüngen besteht?’; Anon., in Allgemeine Musikalische Zeitung mit besonderer Rücksicht auf den österreichischen Kaiserstaat nr. 28 (6 April 1822), cols 221–22.
 Montgomery’s view that ‘of the nearly 100 other singers (many of them professionals) who are known to have performed Schubert’s music during his lifetime … only Vogl is singled out for supposedly having embellished his music in performance’ is again placed in perspective; Montgomery, Franz Schubert’s Music in Performance, 200. It can be ruled out that Siebert’s ornamental praxis differentiated between music by Schubert and other German composers in the repertoire he performed.
 Neue Musik-Zeitung 10, nr. 12 (1889), 149.
 Crossed-out note stems are found in only a few prints published during Schubert’s lifetime. See Dürr, ‘‘Manier’ und ‘Veränderung’, 125–26.
 Rösner, Leitfaden einer Gesanglehre vol. 2/1, 8. The reproductions from the original tutors by Rösner and Swoboda are printed with permission of the Archiv der Gesellschaft der Musikfreunde in Wien. The license is limited to the online version of this text.
 Barry Cooper has convincingly pointed out, crossed-out quavers were used synonymously with semiquavers with two parallel flags in Vienna at this time and were not widely adopted as a means of indicating a short appoggiatura until about the 1830s. See Barry Cooper, ‘Beethoven’s Appoggiaturas: Long or Short?’, Early Music 31/2 (2003), 165–78, at 165f.
 Cooper, ‘Beethoven’s Appoggiaturas’.
 Oddly, after having provided several examples of short grace notes on the beat, David Montgomery, who proceeds in an otherwise positivistic manner, suddenly argues as follows without any source evidence: ‘In reality, however, there was and is a conflict between taking an appoggiatura short and on the beat at the same time’; Montgomery, Franz Schubert’s Music in Performance, 185. Also, Hartmut Krones and Robert Schollum are of the opinion that sometimes one ‘has to’ execute short grace notes before the beat, although they cannot offer a source for this practice. Krones and Schollum, Vokale und allgemeine Aufführungspraxis, 192.
 Dürr is consistent in all of his publications pertaining to the syllabic solution, and he is joined by Krones and Schollum, Vokale und allgemeine Aufführungspraxis, 193–206.
 Montgomery, Franz Schubert’s Music in Performance, 170–83.
 Clive Brown, Classical and Romantic Performance Practice 1750–1900 (Oxford: Oxford University Press, 1999), 465–67.
 ‘Die langen Vorschläge bekommen immer mehr Nachdruck als die Note, zu welcher sie gehören, widrigenfalls der Vorschlag aufhören würde, eine Verzierung zu seyn’; Rösner, Leitfaden einer Gesanglehre vol. 2/1, 8.
 For the sources of the lied, see Otto Erich Deutsch, Franz Schubert: Thematisches Verzeichnis seiner Werke in chronologischer Folge (Kassel: Bärenreiter, 1978), 198–99. The first print of ‘Erlkönig’ is available at: https://onb.digital/result/10036EB1. The autographs of the first and fourth version (‘Stichvorlage’) are lost; see Deutsch, Franz Schubert.
 The vocal parts of the four versions are very similar and represent more or less variants of the same composition. The piano part of the second version has quavers instead of triplets in the right hand. The second version is held at the Staatsbibliothek zu Berlin, https://digital.staatsbibliothek-berlin.de/werkansicht/?PPN=PPN655180990; the third version is at the Morgan Library, in New York, www.themorgan.org/music/manuscript/115640.
 Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 10: ‘der obere Vorschlag eines ganzen Tones’.
 Leopold von Sonnleithner, ‘Bemerkungen zur Gesangskunst V’, in Recensionen und Mittheilungen über Theater und Musik nr. 9 (3 March 1861), 129–38.
 Joseph Preindl, Gesang-Lehre (Vienna: Steiner, ca. 1815), 21–22.
 Friedrich Starke, Wiener Pianoforte-Schule in III Abtheilungen (Vienna: self-publ., 1819), 18.
 August Swoboda, Allgemeine Theorie der Tonkunst (Vienna: self-publ., 1826), 48.
 ‘Steht die Anfangsspitze der Figur nach oben [𝆘], so werden selbe [Doppelschläge] von oben nach unten, und steht sie nach unten [𝆗], von unten nach oben gemacht’; Swoboda, Gesanglehre, 17.
 ‘Er wird auf dreierlei Art vorgestellt und vorgetragen, nämlich: a) von der Hauptnote selbst mit einem Zusatznötchen (𝆘) b), vom obern Hülfston herab (𝆘), und c) vom untern Zusatzton hinauf angefangen (𝆗)’. Johann Nepomuk Hummel, Ausführliche theoretisch-practische Anweisung zum Piano-Forte-Spiel (Vienna: Haslinger, 1828), 398.
 Hummel, Ausführliche theoretisch-practische Anweisung zum Piano-Forte-Spiel: ‘Da die Notenstecher fehlerhafterweise dem Doppelschlagzeichen einerlei Stellung geben, der Doppelschlag mag von oben herab, oder von unten hinauf gemacht werden sollen, so werden die Herren Musikverleger darauf aufmerksam gemacht, ihren Stechern den Unterschied der zweierlei Stellungen dieses Zeichens einzuschärfen und sie zur Befolgung anzuhalten.’
 ‘soll der Doppelschlag von oben herabgesungen werden, so muss das Zeichen von oben anfangen, 𝆘; soll er aber von unten hinauf gemacht werden, so fängt das Zeichen von unten an 𝆗’; Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 10, and Weiss, Theoretisch-practische Gesang-Schule, vol. 3, 7.
 Ferdinand Kauer, Singschule nach dem neuesten System der Tonkunst (Vienna: Artaria, 1794), 11–12; Preindl, Gesang-Lehre, 21.
 Starke, Wiener Pianoforte-Schule, 18; Swoboda, Allgemeine Theorie der Tonkunst, 48–49.
 Swoboda, Gesanglehre, 19.
 Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 11–12, and Weiss, Theoretisch-practische Gesang-Schule, vol. 3, 8–9.
 Schubert’s autograph is in the Staatsbibliothek zu Berlin, https://digital.staatsbibliothek-berlin.de/werkansicht?PPN=PPN655180834&PHYSID=PHYS_0006&DMDID=DMDLOG_0002. Diabelli’s first edition is available online from Harvard University Library, https://iiif.lib.harvard.edu/manifests/view/drs:12131866$11i.
 ‘§. 79. Die zweyte Art, die Stimme zu tragen, findet zwischen zwey Tönen statt, die ein kleineres oder größeres Intervall bilden, und nur in entfernten Stufen nach einander folgen. Es besteht darin, daß man die Stimme, mittelst einer unmerklichen Bindung beyder Töne von einem Ton zum andern übergleiten lasse, so daß man den zweyten Ton schon etwas vorher nimmt. … §. 80. Wird ein solches Portamento von der Tiefe zur Höhe angebracht, so geht die Stimme vom Schwachen, mittelst einer unmerklichen Nachhülfe der Kehle, zum Starken über, und umgekehrt’; Swoboda, Gesanglehre, 11.
 ‘besonders muss man sich in Acht nehmen, dass im langsamen oder schweren Zeitmasse die dazwischen liegenden Klänge nicht bemerkbar werden. Beym Übertragen in langen Noten, wird der ersten etwas abgebrochen, wie in folgendem Beyspiele zu ersehen … Um das Portamento vollkommen bewirken zu können, muss der untere oder tiefere Klang mit gemässigter Stimme angefangen, die Verbindung crescendo und der höhere stark angeschlagen werden, dasselbe gilt auch von oben herab, dass der höhere Klang stark, der untere aber schwach angeschlagen wird’. Rösner, Leitfaden einer Gesanglehre, vol. 2/1, 4–5.