Purcell: A man of the theatre

Music for the theatre takes an important place in Henry Purcell's oeuvre. Today his only opera, Dido and Aeneas, is regularly performed and recorded. In comparison, his semi-operas are lesser known, although they are certainly not neglected. Songs from these works are well-known and part of solo recitals. They were already popular in Purcell's own time, which explains why they were separately published in the collection Orpheus Britannicus.

Complete performances of the semi-operas in our time are extremely rare. According to New Grove, a semi-opera is "[a] play with four or more separate episodes or masques which include singing, dancing, instrumental music and spectacular scenic effects". The inclusion of the spoken text would result in a performance lasting about four hours. Moreover, a complete performance would only make sense if it would be staged, which is rather complicated. It is also questionable how many in an audience, even if they are all native English speakers, would really comprehend the texts of the original play. Today, most performers confine themselves to the musical items from Purcell's pen. Sometimes they include a spoken synopsis, which explains the story to the audience. That may make some sense in a live performance, but would be rather useless in a commercial recording, also because of the language. No wonder that both recordings reviewed here omit any narritive.

The Fairy Queen is based on a libretto by an anonymous author, which is an adaptation of the play A Midsummer Night's Dream by William Shakespeare. It is divided into five acts, preceded by First Music and Second Music, consisting of two instrumental pieces each, and an overture. This piece includes much diversity in forms and scorings, and it was one of the main concerns of Sébastien d'Hérin to create some sort of coherence. "I have wanted to be able to demonstrate that feeling of diversity and of exuberance which I have experienced myself, as well as the work's importance and stature which, I like to think, must have been so eminently evident at the time." It is inevitable that performers take differente decisions, but one would hope that these are all within the boundaries of what is historically tenable. Unfortunately d'Hérin has crossed that line.

The very first piece, the prelude which opens the First Music, is preceded by a solo of the timpani. That is already a bad omen. It is especially with regard to the instrumental scoring that d'Hérin has taken decisions which are debatable at least, and sometimes simply wrong. Purcell's scoring includes two recorders, two oboes, two trumpets, strings, timpani and basso continuo. For unknown reasons, d'Hérin added a cornett and a harp. His ensemble also includes two cellos, a double bass, a regal and an organ. The use of cellos is very questionable, as these instruments were anything but common in Purcell's time. The double bass was never used in Purcell's music. Why a regal was used, is anybody's guess. The entire ensemble is also rather large, including seven violins and two violas.

The decisions regarding the vocal line-up make much more sense. The choral sections are sung by the soloists, and this may well be in line with the performance practice in Purcell's time. "In this score, there is no actual 'role', strictly-speaking [sic], nor any extended musical narrative. One is called upon to highlight a succession of arias and of dissimilar but demanding playlets; little in the way of the psychology of singing characters exists. My choices have thus focused on what I was able to see and understand from among all these loyal and dedicated artists: their personality, their individuality, their strength and their character." No problems here.

What about the actual performance? The decisions regarding the instrumental line-up may be controversial, but that in itself does not prevent a musically satisfying performance. However, that is not the case, I'm afraid. The inclusion of instruments like the cornett and the regal has pretty disastrous effects. The 'Dance of the Followers of the Night' at the end of Act Two is destroyed by the use of the regal. Moreover, here as elsewhere the strings include exaggerated dynamic accents, which are out of place in English music. In The Plaint (O let me weep), the obbligato violin part is played on the cornett, which is highly unsatisfying, for instance with regard to the balance between the soprano and the instrumental part. The cornett just attracts too much attention at the cost of the vocal part. Too often d'Hérin uses percussion, for instance in the chaconne at the end of Act Five. For some reason not discussed in the liner-notes, this piece - the Dance for the Chinese Man and Woman - is placed at the end of the work, after the chorus 'They shall be as happy', which in Purcell's score closes this work.

The singing does not give much reason for joy. Most singers use too much vibrato; Anders Dahlin is the only exception, and he is by far the best of the singers. Samuel Boden does also reasonably well. The Plaint is one of the highlights of this work, but it is destroyed here, not only due to the inclusion of a cornett, but also to Caroline Mutel's wide vibrato. However, the worst part of this recording is her performance of 'Hark! The echoing air', in which she adds some extravagant ornaments, as we are used to hear in a bad performance of a Handel opera. On a positive note: some of the singers are French, but their English pronunciation is surprisingly good. I hardly need to say here that the pronunciation is not historical; that is still the exception in performances of English music of the Renaissance and baroque periods.

King Arthur was first performed in 1691 at the Queen's Theatre, Dorset Garden, in London in 1691. The libretto was written by John Dryden and is about the battles between King Arthur - about whom many legends have been woven - and the Saxons. This semi-opera exactly shows why a performance of Purcell's music alone can never be entirely satisfying from a dramatic point of view. The title character is a spoken role, and as a result King Arthur is nowhere to be seen or heard in Purcell's music. It is well summed up in the article on King Arthur in Wikipedia: "King Arthur is a "dramatick opera" or semi-opera: the principal characters do not sing, except if they are supernatural, pastoral or, in the case of Comus and the popular Your hay it is mow'd, drunk. Secondary characters sing to them, usually as diegetic entertainment, but in Act 4 and parts of Act 2, as supernatural beckonings."

There is one similarity between the performances of King Arthur by Vox Luminis and The Fairy Queen by Les Nouveaux Caractères. Like in the latter, the choruses in King Arthur are sung by the soloists. Their number is about the same: twelve and thirteen respectively. But that is where the similarity ends. The instrumental ensemble is much smaller: only two violins and one viola, plus recorders, oboes, bassoon, trumpets and percussion. There are no cellos and no double bass; the string bass is here a bass violin, which is much more in line with the performance practice in Purcell's time. There are no instruments which Purcell did not require.

I have heard Vox Luminis twice with this work. The first time was a concertante performance at the 2015 Festival Early Music Utrecht, the second time a scenic performance during the Purcell Day in Utrecht in 2018. At the latter occasion I did not get the impression that the singers are born actors, but undoubtedly these performances in different settings have helped them to find the right approach for a recording of this work. In the scenic live performance an actor recited a text which informed the audience about the story. That has been rightly omitted here. The synopsis in the booklet can be considered a compensation for it, although it could have been a little more extensive.

I very much enjoyed both live performances, and it was not any different this time. Considering the quality of this performance it is almost impossible to mention some highlights, as this recording is full of them. Let me point out some particularly fine moments. Robert Buckland gives an excellent account of the part of the British Warrior (Act I: Come if you dare). 'Hither, this way' (Act 2) is exquisitely sung by Caroline Weynants. Olivier Berten delivers a refined performance of 'How blest are shepherds' (Act 2). Zsuzsi Tóth and Stefanie True are a perfect match in 'Shepherd, shepherd, leave decoying' (Act 2). The former is at her very best in the air 'Fairest isle' (Act 5). Sebastian Myrus does well as the frozen Genius. The staccato in his air could probably have been a little sharper. The only small disappointment is Sophie Junker, who uses a bit too much vibrato in the part of Cupid; she makes a better impression in the part of Honour in Act 5. One of this recording's assets is also the instrumental playing. Here all the exaggerations of Les Nouveaux Caractères have been avoided. There is no excessive use of percussion, and there are neither exaggerated dynamic accents nor extremely fast tempi. The whole piece has a nice and natural flow.

This is an impressive and highly enjoyable recording of one of Purcell's theatrical masterpieces. It would be nice if Vox Luminis would turn its attention to other theatre music by Purcell. What about The Fairy Queen?

Purcell: The Fairy Queen (Z 629)
Caroline Mutel, Virginie Pochon, Hjördis Thébault (soprano), Caitlin Hulcup (mezzo-soprano), Christophe Baska (alto), Samuel Boden, Anders Dahlin, Julien Picard (tenor), Guillaume Andrieux, Kevin Greenlaw (baritone), Ronan Nédélec (bass-baritone), Frédéric Caton (bass), Les Nouveaux Caractères/Sébastien d'Hérin
Recorded September 2016 at the Théâtre Laurent Terzieff - Ensatt, Lyon, France DDD
Texts included
Cover, track-list & booklet
Glossa - GCD 922702 (2 CDs) [2.03'52"]

Purcell: King Arthur (Z 628)
Vox Luminis/Lionel Meunier
Recorded January 2018 at AMUZ, Antwerp, Belgium DDD
Texts included
Cover, track-list & booklet
Alpha - 430 (2 CDs) [1.37'59"]

Purcell: Dido and Aeneas

It is an intriguing coincidence that some of the great masterpieces of music history leave many questions which have not been answered yet, and probably never will be Examples are Monteverdi's Vespers, Bach's B minor Mass and Mozart's Requiem. To that list one can add Henry Purcell's opera Dido and Aeneas. For a long time it was thought to have received its first performance in 1689 at Josias Priest's Boarding School for Young Gentlewomen, but first it was discovered that an earlier performance may have taken place at that school two years before. Right now it is assumed that it may have received its premiere at the court of Charles II, as early as 1684. From that perspective the statement of Robert Matthew-Walker, in the liner-notes to Christopher Monks's recording, that "[there] is also a subtle reference to contemporary events: the first chorus: 'When monarchs unite, how happy their state,' was - in 1689 - surely a direct reference to the accession to the throne of the joint British monarchs William and Mary the previous year", is highly questionable.

He also does not mention the problems with regard to the sources, which are discussed by Ellen Hargis in the booklet to Fabio Bonizzoni's recording. We know next to nothing about the first performances. The earliest musical source of the entire opera dates from about a century later. More interesting is a performance in London in 1704, the second at the public stage; the first public performance was in 1700. In 1704 Purcell's opera was preceded by the masque Mars and Venus by John Eccles and Godfrey Finger. "The conjunction of these two masques also provides information on the vocal ranges in Dido and Aeneas, as the singers in the masque of Mars and Venus are known. In all, Mars and Venus calls for one boy, two men, and five women (...)", Ellen Hargis writes. They can be correlated to Purcell's score, and that results in the Sorceress being scored for a bass rather than a mezzo-soprano or alto, as is the case in many recordings. That is not to say that the latter option is wrong; for the performance at the boarding school that part may have been transposed up an octave. Another difference is the role of the Sailor, mostly performed by a tenor, but apparently intended for a soprano. Bonizzoni follows these indications in his recording. In some recordings attempts are made to reconstruct dances which may have been part of the original performances. The only specimen in Monks's recording is a Guitars' Chaconne in the first act. Bonizzoni plays here a kind of improvisation as well, and if I am not mistaken, it is based on a piece by the Italian Bernardo Storace, which is a little odd.

The liner-notes to Monks' recording are not really up-to-date, and in a way that goes for the performance as well. I haven't heard anything I had not heard before. In no respect this recording offers a new perspective, which is disappointing, considering the large number of recordings in the catalogue. That would be less of a problem if the performance would have been really good, but it is not. Most of the singers use quite a lot of vibrato, and that includes Rachael Lloyd, who takes care of the role of Dido. It severely damages the famous lament, which makes too little emotional impact anyway. The most stylish singer is Roderick Morris as the Sorceress, but he seems vocally overstretched, and does not make much of his role, which is anything but fearsome. As is so often the case in performances of Dido and Aeneas, the Witches make a caricature of their part, producing the conventional nasal sound and singing deliberately out of tune. However, this opera is a tragedy, not a comedy. Bonizzoni understands that, and in his performance the Witches have to be taken seriously. In that respect there is more consistency with the role of the Sorceress, who is really threatening, thanks to Iason Marmaras' excellent interpretation. On the other hand, here the bickering between Dido and Aeneas at the end - "I'll stay / Away, away" - is unsatisfying: the tempo is too slow and as a result the agitated character of this dialogue doesn't come off. In this recording the performance of the part of Dido by Raffaella Milanesi does certainly not lack emotional depth, but stylistically it is disappointing because of her incessant vibrato.

Her diction also leaves a bit to be desired; the text is not always clearly intelligible, which is especially regrettable as we come here to one of the ground-breaking aspects of this recording: the use of historical pronunciation. The starting point was the fact that words which are supposed to rhyme, don't in present-day English. In order to correct that, the performers have adopted a kind of historical pronunciation. As a result "destiny" rhymes to "defie" (Aeneas, act 1) and "wounds" to "hounds" (Belinda, act 2). This is of great importance, but I have the feeling that the performers are a bit half-hearted. I remember to have heard that at that time, for instance, the pronucnation of the "r" was closer to present-day American English than to 'Oxbridge' English. The differences could be much more far-reaching than this recording suggests. It seems there is still some work to do in this department.

The third recording to be reviewed here is a staged performance, released on DVD by Alpha. Vincent Dumestre, director of the renowned ensemble Le Poème Harmonique, is responsible for this performance which took place in 2014 at the Opéra of Rouen. Those who have enjoyed the recording of, for instance, Lully's Le Bourgeois Gentilhomme, may expect a comparable approach here: historical staging and costumes, baroque acting gestures and a use of historical pronunciation. They will be severely disappointed. As far as I know the stage directors have no experience with baroque opera, and although they deliberately avoided a "consciously modernising transposition", they did not opt for "a pure historical reconstruction". Incorrectly they call both options "extremes" - in my view a basic misunderstanding. If one aims to do as much justice to the intentions of the composer as possible, the adherence to what we know about the performance habits of his time is the most logical option. That said, in comparison to so many stagings of 18th-century operas, there is little here that is outright annoying. That is at least something. As you will understand, historical acting and pronunciation are also ignored here.

I already mentioned the scoring of the various roles. Like in Bonizzoni's recording, the role of the Sorceress is sung here by a baritone (Marc Mauillon). However, he also sings the role of the Sailor, which was originally intended for a soprano. Unfortunately that is not the only unhistorical aspect of this recording. The orchestra is relatively large, including eight violins and three violas. Even more problematic is the participation of instruments Purcell did not include in his scoring, such as recorders, oboes and bassoons. In the ensemble we also find a double bass, although it is known that in Purcell's time this instrument was not used (it is also part of the Armonico Consort; the instruments in Bonizzoni's performance are not listed). There are also guitars, which mostly play in the reconstruction of some dances. It is assumed that there were several dances in the score, but the music has mostly not survived.

The performance does not give much reason for enjoyment. Vivica Genaux is pretty horrible. I find her voice unpleasantly harsh, and she has quite some intonation problems. Moreover, she uses a pretty wide vibrato on virtually every note, and as a result the famous Lament is severely damaged. Henk Neven is alright as Aeneas, but is rather bland in the account of his role, and there is little 'baroque' in his style of singing. Vocally speaking Marc Mauillon is by far the best of the ensemble. Ana Quintans as Belinda does rather well, but I can't really warm to her singing either. The minor roles are not more than mediocre.

All in all, neither of these three recordings does really satisfy me, as neither consistently applies what we know about performing habits at the time. If I had to choose between these three, I would go for Bonizzoni, as he is most consistent in the scoring of the various roles, and makes use of historical pronunciation, Moreover, he also offers Mars and Venus by John Eccles and Godfrey Finger, which preceded Dido and Aeneas in the London performance of 1704. It is here performed after Purcell's opera, but if you want to 'reconstruct' that performance, you can easily do so by programming your CD player accordingly.


Rachael Lloyd (Dido), Elin Manahan Thomas (Belinda), Eloise Irving (Second Woman, First Witch, Spirit), Jenni Harper (Second Witch), soprano; Roderick Morris (Sorceress), alto; Robert Davies (Aeneas), Miles Golding (Sailor), baritone
Armonico Consort/Christopher Monks
Recorded October 2014 at the Church of St Augustine, Kilburn, UK DDD
Texts included
Cover, track-list & booklet
Signum Classics - SIGCD417 [50'45"]

Raffaella Milanesi (Dido), Stefanie True (Belinda), Michela Antenucci (First Witch, Sailor), soprano; Anna Bessi (Second Witch, Spirit), mezzo-soprano; Richard Helm (Aeneas), Iason Marmaras (Sorceress), baritone
Coro Costanzo Porta; La Risonanza/Fabio Bonizzoni
(+ John Eccles (1668-1735) & Gottfried Finger (1685-1717), The Love of Mars and Venus, 1680)
Recorded live 25 February 2016 at Cité de la Musique et de la Danse, Soissons, France DDD
Texts included
Cover, track-list & booklet
Challenge Classics - CC72737 [76'11"]

Ana Quintans (Belinda), Caroline Meng (First Witch), Jenny Daviet (Second Woman), soprano; Vivica Genaux (Dido), Lucile Richardot (Second Witch), mezzo-soprano; Nicholas Tamagna (Spirit), alto; Henk Neven (Aeneas), Marc Mauillon (Sorceress, Sailor), baritone
Choeur Accentus; Le Poème Harmonique/Vincent Dumestre
Recorded May 2014 at the Opéra, Rouen (Haute-Normandie), France DDD
Subtitles in E/F/D/ES
The entire performance is available on YouTube
Alpha - 706 [1.20']