Spend a summer evening in the main square of Kraków and stop for a glass of wine near St Mary’s Basilica and you’ll probably hear the Hejnał mariacki, a trumpet call played every hour from one of the towers. Legend has it that this tune, mentioned in the city’s records as early as 1392, ordered the opening and closing of four city gates. It feels like a slice of medieval life, until it suddenly breaks off mid-melody. One legend suggests this recalls the moment a trumpeter was shot during a 13th-century invasion. However, my tour guide informs me that this story is fanciful; the most likely explanation for the trumpetus interruptus is that another player situated on the city wall would simply respond when the gate was closed.

That vision of late-medieval Kraków has stayed with me: the trumpets, the towers, the old city with its beautiful gates and fabulous wine. In such a place, it is tempting to feel the past is closer than it really is even if it has been filtered through a legend. History is a slippery thing.

Fortunately, there is more to the preserved soundscape of Kraków than this legend: there is also the polyphonic music. Between about 1424 and 1426, a scribe in Kraków compiled a manuscript that contains some theological writings and more than 40 pieces of mostly polyphonic music. Pieces from local sources alongside Italian and French works create a precious snapshot of indigenous Polish music and of the foreign music that passed through this cultural centre in the early 1400s. A lingering glance through the pages of this source reveals different styles of notation that reflect the varied provenances of these works. Sometimes our scribe uses red notes to show changes of mensuration, sometimes void notation (with noteheads unfilled) is used for the same purpose. But what binds these pieces together (if you’ll forgive the codicological pun) is that they all have Latin texts, even when they obviously didn’t originally. Some pieces celebrate the Jagiellonian dynasty, which ruled Poland between 1386 and 1572. This Krasiński Codex, as it is commonly known, is the most important source of Polish music in the late Middle Ages. It’s housed in the Biblioteka Narodowa (National Library) in Warsaw (shelf mark MS III.8054).

Ensemble Dragma, from left: Caroline Ritchie, Grace Newcombe, Marc Lewon, Agnieszka Budzińska-Bennet and Jane Achtman

When I heard that a new edition and recording of this manuscript were being created by musicians and academics associated with the early music powerhouse Schola Cantorum Basiliensis, I journeyed to Switzerland to hear their sessions and find out more. Meeting Agnieszka Budzińska-Bennett, Marc Lewon and Jane Achtman of Ensemble Dragma was a joy. They are the driving force behind this publication, which has been made possible through partnership with the Adam Mickiewicz Institute, supporter and promoter of Polish culture internationally. They are joined by other performers, but the intellectual thrust of the project is theirs.

The story of this manuscript has some twists worthy of a wartime Pathé newsreel. It was housed in the Świdziński library in Warsaw in the 19th century then acquired by the aristocratic Krasiński family, whose library was subsequently torched by the Nazis. The Krasiński Codex was, however, looted just in time by someone clever enough to recognise its value. It was taken to the Bavarian State Library in Munich, where it was kept until about 1948. Budzińska-Bennett’s research suggests that on a visit to the Munich library, the Polish art historian and bibliographer Karol Estreicher saw another reader carry it past his desk. Recognising its importance, he negotiated for its return to Poland. Estreicher also worked to preserve and rescue countless internationally important artworks, such as The Lady with an Ermine by Leonardo da Vinci, items with which the Krasiński Codex can easily compete, being a key witness of musical culture at the eastern boundary of Latin Europe and the transition from the Middle Ages to the Renaissance.

A high-quality scan of the manuscript, catalogued under the general title ‘Miscellanea theologica’, is available to view on the National Library website, polona.pl (it’s a little tricky to find, but the easiest way is to search ‘objects’ for ‘Ciconia’ – one of the composers featured). It’s not an ornate document. It was probably intended as a rough copy from which other manuscripts could later be made. The notation is often unclear, or interrupted by numerous mistakes and scribal errors. But there are also some delightful quirks, such as tenor parts labelled backwards – ‘Ronet’ – which led Willi Apel in his monumental mid-20th-century textbook The Notation of Polyphonic Music, 900–1600 to describe it as ‘an abnormally difficult specimen’. Intrigued? You should be.

The recording sessions are held in the parish church of St Nicholas in Herznach, about an hour west of Zurich by train. It’s an outrageously beautiful setting, rural Switzerland in blazing late June sunshine. The white walls of the church gleam from afar, promising a quiet, cool interior shielded from the aircraft-humming, insect-buzzing summer heat, to say nothing of the pollen. Only the bells in their late medieval wooden frame chiming each quarter-hour pierce the serenity as I pull my suitcase from the bus stop at the bottom of the hill. Time moves slowly in Herznach.

Part of Radomski’s Magnificat in the Krasiński Codex, showing the instruction ‘per bar du num’ in the bottom half of the right-hand page

A guidebook promises me that there has been a church documented on this site since 1185; a catastrophe that befell it around 1341 prompted Pope Eugene IV to grant an indulgence to support its reconstruction. Since then, further renovations have speckled this place with character: the nave and pews were rebuilt in the 1690s, the chancel, altar, pulpit and some paintings and frescoes date from the early 18th century. It’s a joyful cornucopia of styles that serve as a rough guide to ecclesiastical architecture. But it’s not the building that stops me in my tracks, it’s the sound emanating from within: the three-part Credo by Mikołaj Radomski (Nicolaus de Radom). Blazing pure fifths ring out across the courtyard. It’s early music heaven.

Lewon and Budzińska-Bennet are full of impish enthusiasm. Typically for scholars deeply enmeshed in a substantial research project, information spills out in a joyful torrent. Budzińska-Bennet is a Polish-Swiss scholar exploring the historical background of this manuscript, while Lewon is preparing a new edition for publication. They often finish each other’s sentences. We dive in greedily and talk about Nicolaus de Ostroróg’s Gloria, which I also just heard. It’s in two parts, which I followed in the facsimile score. They performed it with three sopranos on top and a wonderful pairing of trombone and vocalising tenor underneath. During the session, Lewon was distracted, perched on a creaky 17th-century pew and fiddling on his laptop until eventually he piped up: ‘We could try F sharps on “tu solus”,’ and showed the performers a quick sketch on his notation software. Budzińska-Bennet couldn’t decide, so they recorded both versions.

Lewon explains his approach to musicology in action: ‘That was one of those cases where it’s a fragment, and adding more and more voices just makes it … well, at some point it becomes arbitrary, it becomes a different piece. But then when I heard it, I thought, “There’s clearly something missing. Maybe we should have a third voice after all.”’ Budzińska-Bennet interjects: ‘I feel the crucial point for us is our proximity to the musical source and to be mindful of how far we can drift from that source and all the information it contains. Plus we bring our own related knowledge, of course: sensitivities and strategies we’ve developed over years through studying these vast repositories. I mean, we can fill in the gaps in the most logical way or in the most probable way, and that’s our aesthetic point – not embellishing when it’s not necessary.’

Lewon adds: ‘There’s actually quite a clear framework within which this kind of art works, but to someone who’s exposed to this for the first time, let’s say, the things we tried – a version that was just two-voice and without musica ficta or a version that has a contratenor and some musica ficta – sound radically different from each other. To us they don’t; the framework is still the same, we’re just colouring it in.’

Lewon and Budzińska-Bennet are later joined by their colleague Achtman. They tell me they are keen to communicate their strategies, and together with the Adam Mickiewicz Institute they are publishing an edition of the manuscript and related pieces which includes academically rigorous transcriptions and reconstructed moments such as the one in Ostroróg’s Gloria. ‘Every single piece is accompanied by a rich essay on concordances,’ Budzińska-Bennet enthuses. ‘We have absolutely updated information, with lots of concordances that we knew about before and some that we’ve discovered, including textual cognates.’ While she speaks, Lewon smiles and angles his laptop screen towards me so that I can see a page or two of the edition. It’s a thing of beauty: reconstructed capital letters in a sumptuous font, incipits to show the original notation, red notation, variants and editorial suggestions clearly marked and beautifully typeset essays introducing each piece. ‘Thank you, Institute,’ says Budzińska-Bennet, grinning from ear to ear with pride. Opportunities like this don’t come around often, and this is special to her as a Polish musician.

Achtman is both a fiddle player and the ensemble’s expert in early music logistics. She makes a good point: a project like this requires foresight and strategic funding, from photographing and preserving manuscripts to making them available in facsimile so that scholars can explore their contents and relationships more easily and then share their findings with colleagues. She reminds me that medieval musicology benefits from a range of technological advances, from high-resolution scanning to the many databases that record relationships between manuscripts.

This whole recording project benefits from the high-quality images available for the Krasiński Codex. I think about this as I zoom in on Radomski’s Magnificat primi toni. One of the larger ensemble numbers on this recording, it’s intriguing in terms of notation and is a musical highlight. In fact, it is oft-cited in textbooks as a relatively rare example of a fauxbourdon in a continental manuscript. Our scribe – of whom I’m beginning to feel rather fond – has helpfully labelled it ‘per bar du num’, which means that at least one other voice should be added as long as certain rules are followed, rules that performers were expected to know. I ask Budzińska-Bennet about the instrumentation in their performance, and we talk about the piece’s structure, which is similar to that of Dufay’s setting. ‘We perform this Magnificat in a very rich way on purpose. Of course, it’s a three-part piece, basically, with the cantus texted and two other voices not texted. There is even one more part not written out in fauxbourdon sections. The texture differs between sections, and this led us to emphasise these differences with our instruments.’

I ask if in an all-vocal performance she would advocate the use of text in each part, and she agrees, adding: ‘The fauxbourdon parts, obviously, are the easiest since they are syllabic and you just follow the cantus. This is not a problem at all. But there are some moments in the two untexted lower voices that are quite crazy rhythmically, where it would be really difficult, and not very natural, to add text. You can do it, but it would be a little bit extreme. So the value of the text would get lost in such moments.’

We talk briefly about an old argument that some complex medieval musical lines are too difficult for singers so therefore they must have been instrumental parts. ‘This is a long discussion that leads to many wrong assumptions,’ Budzińska-Bennet warns me. Lewon adds: ‘We have famous contratenorista singers, composers of contratenor parts, who prided themselves on being able to sing such wide leaps. I mean, it’s a sign of virtuosity, vocal virtuosity, and I’m sure they were absolutely able to sing like that. You can’t say something is instrumental just because it’s difficult.’ (I think to myself, ‘He’s got a point.After all, people do sing Donizetti operas!’)

The resulting performance has three upper voices, a tenor and a baritone, plus two fiddles and a trombone. ‘For me it was a nice idea to have really quickly moving instruments and to look for places where the cantus firmus could be elevated above them. I think it emphasises the function of this piece as well as offering variety.’ Budzińska-Bennet is particularly passionate about this piece and refers to it at one point as a ‘manifest of Polish culture, the most known piece of the 15th century in Poland or maybe even of early music in general from Poland’. I love the final result, particularly Grace Newcombe’s silvery voice coupled to the busy, agile textures of the medieval fiddles. It’s all much more poignant and moving than I expected from such a simple fauxbourdon.

Achtman sums it all up by referring to Radomski’s specifically Polish way of composing song and Mass movements. ‘It’s somehow very simplified and very direct and it just goes right to your heart.’

I ask Achtman for an elevator pitch for this project, and like Radomski, she’s direct and touching all at the same time. ‘It’s basically about intercultural exchange in medieval Europe,’ she says, and looks into the distance. ‘And I think for me the astonishing side of it is how emotionally accessible it is to a modern-day audience. It speaks to you across the centuries.’

The next time I take an evening stroll through Kraków’s main square and hear the Hejnał mariacki, I’ll put this recording on my headphones and steep myself in the late medieval city soundscape which still resonates gently across the centuries. It’s survived theft, fire and goodness knows what else besides, and now it’s found a sympathetic group of international experts in a church in rural Switzerland.

This recording is available digitally on the Raumklang label