Signs for People Who Already Knew the Tune

Open a ninth-century chant manuscript and you will not be able to sing from it.

There are the words, in Latin, and above them small marks. A little stroke slanting up. A dot with a tail. A shape that doubles back on itself. These are neumes, and they tell you that the voice rises here, falls there, turns and comes back. They do not tell you how far.

To a modern eye this looks like notation that has not been finished yet, and that is the wrong way to read it. The neumes were not a failed attempt at what we do. They were a complete solution to a different problem.

The problem was uniformity.

A cantor in a large church had to know an enormous repertory: chants for every day of the liturgical year, proper to the feast, the season and the office, running into thousands of items. He knew them by heart, because that was the only way anyone knew anything of that size. Teaching them to the next generation meant singing them until the boys had them too.

That works in one building. It does not work across an empire. When the Carolingian court wanted the chant sung in Metz to be the chant sung in Rome, the question stopped being how do we remember this and became how do we check that they are doing it the same way we are. Notation appears exactly there.

Which tells you what the marks were for. They were a reminder for a singer who had the melody, and an audit trail for whoever wanted the melody to stay put. Not a recording.

And because that was the job, the neumes are extraordinarily precise about something modern notation barely addresses: how notes are grouped.

A neume is a single sign that can cover one note or five. The five are written as one shape because they are one gesture, sung in one movement, and the next sign starts the next gesture. Staff notation gives you five separate noteheads of equal visual weight, and the grouping has to be reconstructed by the performer or supplied by an editor with a slur. The old signs carry it inside the notation itself.

They are also precise about ornament. Some neumes indicate a note that is pressed, or shaken, or approached from below with a particular quality of attack, and these distinctions were lost when the system was rationalised into noteheads on lines. Chant scholarship spent much of the twentieth century going back to the oldest manuscripts for exactly this reason.

So the practical position for anyone editing chant is to use two kinds of source at once. The early adiastematic manuscripts, with no pitch height, for shape, grouping and nuance. The later ones, with lines, for which note is which. Neither is sufficient and together they are quite good.

There is a general lesson sitting in here that is worth taking away from the specific one.

Writing systems do not arrive because people want to preserve things. They arrive because somebody wants to check on somebody else at a distance. Accounting tablets, legal codes, chant notation: the sequence is the same each time, and the preservation is a side effect that turns out to matter more than the original purpose.

Nobody in the ninth century was writing neumes so that we could hear what they sang. They were writing them so that the next monastery would sing it their way. The fact that we can now partly recover the music is an accident of an administrative tool.