It’s often said of J S Bach’s works in particular that they’re ‘too academic’, ‘too boring’, and yet nothing could be further from the truth. Tallis shows us in his strangely-scored if beautiful Miserere Nostri (à 7) that it’s possible to marry broad brush strokes of harmony with intricately conceived counterpoint. I often think that Tallis is showing off a little in this piece – he dazzles too in other large-scale works, Spem in Alium being an obvious example. Here though, canon, or rather canons is/are the order of the day. The canon at the unison between the two superius parts is easy enough to spot – it’s further down that things get quite a bit more interesting. We have, according to the volume’s index, 6 partes in duabus, cum, uni parte ad placitum, in other words a six in two canon with one freely-composer part. We have what is called a mensuration canon of four voices – that is, all start simultaneously, but at four different speeds. The Discantus part starts us off, with the Contra tenor in canon in double augmentation (the notes are four times longer). The two Bassus parts are in canon per Arsin et Thesin, that is, they are inverted so that upward intervals in the starting part are downward in the answer. Bassus 2 is augmented (the note values are doubled) and Bassus 1 is triple augmented (the note values are eight times longer). That just leaves us with the Tenor, which is freely composed. The motet comes from the famous collection Cantiones Sacrae of 1575, Britain’s first music publication. Interestingly, the Discantus part book attributes this piece to William Byrd, although this is believed to be an error (one of many in the publication, witness the missing time signatures evident in my prefatory staves!) since Byrd’s Miserere Mihi appears as No.29 in the collection (Miserere Nostri, the last item, is No.34). The highly melismatic setting of the text, (Lord have mercy, have mercy upon us) lets the interweaving of the parts shine through.
I’ve done my edition in G, but as Alistair Dixon points out, another semitone would enable the altos to tackle the Discantus completely on their own (I often end up on this part as a tenor) and shove all the tenors onto the Contra Tenor and Tenor parts. Tallis’s scoring is a little perplexing to modern choristers expecting a more-or-less equal scoring, but it’s in large part as a result of the canonic writing. The sonic effect is never thin, however – he exploits fully the range of each individual voice. Ian Harwood’s two pitch hypothesis about performing pitches in the Elizabethan period is also worth considering, although it largely concerns the mixed consort. His argument is that there were basically two main pitches for instrumental music, drawing this conclusion largely on the measurements of surviving instruments of various kinds. The high pitch is a’=465Hz, and a’=392Hz is the lower pitch. Nowadays, Tudor music appears commonly transposed up about a minor third from the original (notable exceptions in modern practice including Tallis’s Lamentations of Jeremiah, both parts of which are commonly performed by men’s voices, that is ATTBB, and at or close to the written pitch). I’ve seen editions of Miserere Nostri in F (for AATTBarBB), G, and Ab, and I think to be honest that it just depends on your available forces – there’s no strong historical case for any particular one being ‘correct’. G seems like a sensible compromise to me, given the undesirability of sticking it into Gb or F# from a reading point of view.
It’s an intense wall-of-sound type piece, the major mode not really detracting too much from the penitential nature of the text (conventions of major and minor viz a viz happy and sad weren’t yet really established).
My edition: PDF