Tunings, Scales and Tonal Systems in Antiquity: Possible Origins of Pitch-Related Concepts
Tunings, Scales and Tonal Systems in Antiquity: Possible Origins of Pitch-Related Concepts
Tunings, Scales and Tonal Systems in Antiquity: Possible Origins of Pitch-Related Concepts
Music need not employ pitches.1 If it does, these need not stand in fixed relations
(tunings). If they do, they need not be perceived as scales – and on the other hand,
if there are scales, these need not imply a fixed tuning. Finally, if a particular music
culture does use scales that implement more or less standardised intervals, these
still need not be conceptualised as a tonal system. And yet, we find such
conceptualised systems both in cuneiform and in Greek written sources.2 In this
contribution I am going to revisit these, with a focus on the extent to which these
written expressions of musical matters are still tied to their original background in
music-making, or have emancipated themselves as objects of reasoning for its own
sake, in a development towards music theory. As the primary criterion for the latter
case I regard the evolution of conceptions that are no longer of immediate use for a
member of the music-producing classes (which include not only performers but
also instrument makers). The distinction is not straightforward, since abstract ideas
may well, secondarily, re-enter the domain of music-making; but I hope that with
some caution the secondary element in such instances can be detected. In any case,
I think past scholarship on ancient musics has rarely erred in the direction of
attributing too much to the practitioner, especially the instrumental practitioner:
rather, one was prepared to assume elaborate intellectual enterprises of the
speculative sort on the part of the ancients, based perhaps on vocal melodies and
suitable experiments.3 It is against such a background that I will try to show how
much of what we know about the ‘tonal systems’ of ancient Mesopotamia and
Greece can (and should) be understood on the basis of direct experience with
instruments that were common at the time, and how theory of a more abstract kind
would have risen from there.
1 I express my gratitude to Dahlia Shehata for sharing her thoughts on important points, and to Tosca Lynch for a
lot of suggestions in the course of editing my style.
2 The study of ancient Near Eastern music necessarily involves frequent reference to the Greek (and hence
Hellenistic) world, of whose music so much more is known, and with which it doubtless shared many traits, be it by
common inheritance or interaction.
3 Cf. for instance the wholesale inclusion of ancient musical writings under the heading of “speculative traditions” in
the Foucaultising structure of Christensen, Th., (ed.), The Cambridge history of Western music theory, (Cambridge 2002).
S. Hagel Tunings, scales and tonal systems in antiquity 2
performed music; the investigation of pitch relations with the help of experimental
instruments; and finally the quantification of musical structures by numerical means.
Obviously the last items in this list belong to a stage of advanced reflection,
without which music-making as such may function just as well. On the contrary,
the first two are crucial ingredients of many musical cultures that make use of
pitched instruments, and the third can be essential in passing on the respective
skills from master to pupil. Correspondingly, we ought not to expect to find the
origins of the more basic issues in the historical record: music is so much older
than writing, and the same is very probably also true for almost all kinds of
instruments in question. On the other hand, even with the utmost caution to avoid
the pitfalls of progressivism, one cannot but recognise the development of new
ways of analysing musical structures, which went hand in hand with philosophical
debate about the nature and value of music, in the Greek world of especially the
5th and 4th centuries BC.4
The most direct way of conceptualising pitch structures is related to the tuning of
instruments. Particularly important in this context are stringed instruments, since
their pitches are comparatively easy to adjust, and because here the entire tuning
process (in the case of lyres, harps, fretless lutes and lutes with movable frets) or a
vital part of it (in the case of lutes with fixed frets) falls within the domain of the
performer. Stringed instruments are also a priori likely to favour the development
of standardised pitch relations between their individual strings. This is because
strings – at least well-produced strings5 – tuned to resonant intervals can enhance
the instrument’s sound by establishing an augmented oscillatory regime through
sympathetic resonance, even if only one of the strings in a resonance-coupled
group is set in motion by the player. This becomes especially clear from
comparison with finger-hole-equipped wind instruments with a single duct such as
flutes: here the individual playable notes do not interfere with each other, so that
the presence or absence of resonant intervals between them does not influence
their sound. Moreover, finding the location for finger holes from which resonant
intervals can be produced without considerable effort is not at all a straightforward
task. So it is no wonder if our earliest evidence for strings implies resonance-based
tunings, while single-duct wind instruments in the archaeological record often
expose arrangements of finger holes (especially equally spaced) that seem to
preclude scales of a similar type. A special case, however, is represented by wind
instruments of two or more ducts sounding simultaneously, such as the double
reed-pipes,which were eminently important in ancient Greek music but also
featured prominently in the Near East as well as in Egypt. Here it is a priori likely
that different notes produced from the single pipes were meant to blend
harmoniously, and Greek literary sources explicitly confirm the employment of
4Cf. Barker, A., The science of harmonics in Classical Greece, (Cambridge 2007).
5I.e., strings of uniform characteristics throughout, so that the actual harmonics produced are very close to multiples
of the basic frequency. For the awareness of the superiority of smooth strings of even thickness, cf. Ptolemy,
Harmonics 1.8.
S. Hagel Tunings, scales and tonal systems in antiquity 3
consonances, 6 and the common musical settings that include strings and double
pipes outside the Greek area seem to imply that at least certain Near Eastern pipes
shared in the pitch systems that determined the tuning of the lyres.
The production of such instruments probably involved copying models that had
worked out well, or also the application of rules of thumb in the first place, perhaps
followed by reworking the details of individual holes afterwards.7 In any case, the
preconditions for the evolution of a resonance-guided tuning are much worse even
in wind instruments of this kind than in stringed instruments: whereas a new set of
finger holes always requires building an entire new instrument, experiments with
lyre strings are quick and easy. All in all, it is therefore highly probable that
resonance-based tonalities first evolved on lyres and harps and were transferred to
pipes and flutes only subsequently (and often only partially).
The employment of certain strings, notes or fingerings in certain contexts, the
establishment of resonance, and the construction of instruments necessarily lead to
the formation of related concepts in the brains of those engaged in such actions.
Whenever these skills are shared, whether in performing together or, especially, in
passing them on in a master-pupil relation, those concepts are likely to be
expressed through language: hence, specific terms are coined for distinct ideas. We
shall see below, how different these may be from culture to culture, even if related
to superficially similar underlying structures. When musical terminology is finally
set down in writing, we must anticipate to find primarily such practice-related
terminology. Only after considerable time of theoretical scrutiny, can we expect
musical discourse to develop more abstract schemes, designed to grasp facts only
implicit in the practitioner’s art; and, even then, it may take centuries before the
language of theory is fully emancipated from its immediately practical origins – if it
ever is.
Once emerging music theory is pursued for its own sake, it may be felt that certain
questions, especially regarding quantification, cannot be answered on the basis of
existing instruments. Consequently, experimental instruments are constructed for
specific purposes, most often for measuring pitch relations. The best-known is the
canon (in its simplest form: monochord), which Greek theorists employed from
the 4th century BC on, and which was perfected by Ptolemy in the 2nd century
AD.8 The study of intervals by means of the monochord (or, in fact, the physical
properties of any instrument) leads to their numerical description as ratios. This
approach coexisted with another description, which was doubtless closer to the
language of practicing musicians: concatenating intervals perceived as similar, leading
to expressions like ‘ditone’ and ‘tritone’ (which yield to a calculation in terms of
ratios just as well), but also ‘semitone’ or ‘third of a tone’ (which cannot be
expressed as ratios and thus led to a conflict of theories).
6 Cf. e.g. Barker, A, “Heterophonia and poikilia: Accompaniments to Greek melody“, in Gentili, B. / Perusino, F.,
(eds.), Mousike. Metrica ritmica e musica greca, (Pisa/Roma 1995), pp. 41–60; Hagel, S., “Calculating auloi – the Louvre
aulos scale”, in Hickmann, E. / Eichmann, R., (eds.), Studien zur Musikarchäologie 4 = Orient-Archäologie 15 (2004),
pp. 373–390.
7 On possible traces of fine-tuning finger holes cf. Landels, J.G., “Fragments of auloi found in the Athenian agora”,
It is important for our topic that the two types of description presuppose a
different instrumental background. The perception-based approach goes well with
any type of instrument, but particularly well with those kinds in which the physical
parameter(s) that define the pitch cannot be ascertained by eye. The most typical
example is the lyre, where all strings are roughly of equal length, their pitch
depending on thickness and tension, the latter measurable only by means of
sophisticated devices. Following the model of the lyre, experimental instruments
may be conceived which merely provide a large set of tuneable strings.9 In contrast,
ratio-based theory can only emerge on the basis of instruments in whose
construction the ratios are visible, at least in good approximation. Here the classic
example is the fretted lute – but surprisingly, the ancient evidence never associates
ratio theory with this instrument type. In Greece, it seems to have been absent or at
least thoroughly marginalised in the era when ratio theory emerged (i.e. before the
late 4th century). Thus the original model for the novel way of describing intervals
was probably provided by the placement of finger holes in wind instruments,
perhaps adopted from their makers’ guidelines.10 Obviously this, in turn, demands
resonance-oriented wind instruments.
Pitch systems
All the ‘tunings’ and ‘scales’ we are dealing with here are heptatonic. Even if they
are not strictly heptatonic – we will discuss such examples later in Greek context –
they cannot be understood but against a heptatonic context.11 With such a universal
assertion, and in the face of the fact that serious misunderstandings about the
matter have turned up in relation to the Near Eastern sources, we must be careful
to make entirely clear what ‘heptatonic’ means. In principle the initial definition of
Wikipedia says it all: “A heptatonic scale is a musical scale with seven pitches per
octave”12 – though we might favour the less ambiguous wording “where the octave
is divided in seven steps”. Anyway, everything depends on the octave, and on the
idea that pitches exceeding the octave somehow merely ‘repeat’ those within it, in
contrast to musical systems that do not attribute a special status to the octave, or in
where the extension of a scale beyond its range produces notes that do not stand
an octave apart from an already established note. In such systems, the notion of
‘-tonism’ has little or no value. On the other hand, there are musical systems, in
which the octave is crucial, but divided into another number of steps – for instance,
five in ‘pentatonic’ music.
9 Cf. speculations on the nature of the ‘epigoneîon’ and the musical theorists mentioned in Plato’s Republic (531a) and
the Pap. Hibeh 13: West, M.L., “Analecta Musica”, Zeitschrift für Papyrologie und Epigraphik 92-1992, pp. 1–54: p. 21;
West, M., Ancient Greek music, (Oxford 1992), pp. 78–79.
10 In principle, approximate ratios can also be retrieved from the lengths of the various air columns in a pan-pipe.
Here, however, the tuning was apparently done by ear (partly filling tubes of similar length with wax), which would
not readily occasion an interest in resulting air column lengths, while the placement of a finger hole is generally prone
to being measured out beforehand.
11 Cf. Franklin, J.C., “Diatonic music in Greece: a reassessment of its antiquity”, Mnemosyne, 55 (2002), pp. 669–702.
erroneous mixture of conceptions start with the second sentence of the article.
S. Hagel Tunings, scales and tonal systems in antiquity 5
13 For discussion of some particular points and a bibliography to this section, cf. Hagel, S., “Is nīd qabli Dorian?
Tuning and modality in Greek and Hurrian music”, Baghdader Mitteilungen 36 (2005), pp. 287–348.
14 Colburn, J., “A new interpretation of the Nippur music-instruction fragmens”, Journal of Cuneiform Studies 61 (2009),
pp. 97–109, provides thought-provoking arguments for additional terms for seconds. I am not convinced, though,
partly because their derivation seems too artificial to have arisen in musical practice, partly because it requires an
application of octave invariance that is ultimately self-contradictory (in dichord names that are quoted as fifths, one
must apply octave doubling to pin down the corresponding fourth – just because the supposed derivation does not
work with octave-invariant dichords).
15 For the moment, assuming, that they are a fourth and a fifth in a given tuning, as is usually the case. Each of the
pairs gives a fourth and a fifth in six out of seven tunings, and two tritones in the seventh.
16 In the ‘Retuning Text’ UET VII 74, the notion of the octave surfaces more directly, when notes one octave apart
about the number and actual distribution of a’s, c’s and e’s. Here, by the way, our
modern musical terminology nicely shows how little systematic coverage of
possible conceptions we may expect from any historical system: in the case of
chords we use primarily abstract notions (‘major chords’, ‘minor chords’ etc.),
which are readily qualified in relation to the system of scales (‘G major chord’), but
cumbersome to pin down in ultimately practical terms (‘G major chord in second
inversion, with an additional root note an octave below’). In contrast, our intervals
are pitch distances, not primarily thought about in terms of harmonic usage, and
hence not subject to the process of inversion. Ancient cuneiform dichords share
characteristics of both: they are, so to speak, two-note chords, which gives the
designation ‘dichord’ a fine second level of meaning.
Thus the concept of dichords is evidently independent of the arrangement of notes
within the octave (or, actually, also beyond the octave, since the model instrument
has nine strings), but classifies all the sufficiently concordant intervals as belonging
either to the fifth/fourth type or the third/sixth type. This is clearly an element of
abstraction. But this abstraction is not of a kind that would go beyond the purely
practical domain. The employment of dichord names (in combination mainly with
numbers) for a kind of musical notation in Ugarit shows that practical music was in
some way perceived as based on dichords. Their employment in this context can
hardly be interpreted otherwise than in terms of a basic ‘intervallic harmony’, which
guided the performers of the accompaniment, while the melody of the song was
known and thus needed not be notated. 17 The concept is not entirely unlike a
modern notation of guitar chords, apart from the fact that the latter are usually
placed above the lines of text, while the Ugaritic notation, in a less helpful way, is
written in a separate block beneath it.18
Given the recognition of the octave as establishing quasi-identical notes, the usage
of ‘abstract’ dichords enables performers to identify the appropriate ‘harmony’ on
their instruments, regardless of individual ranges: this is especially vital for
orchestras where diverse instruments like harps, lyres and pipes play together,
which may produce matching notes in octave relations but not easily similar pairs
of notes. Here every musician would internalise the dichords as applicable to their
specific instruments. In some cases, like harps and lyres, this would be
straightforward, linked to the individual strings just as on the model instrument,
albeit starting and stopping on different pitches, according to the specific string
lengths. Pipes would presumably play in one ‘tuning’ only – but it would have been
possible to provide sets of pipes in different tunings, playing dichords of the same
name with identical fingering. This, it must be noted, is presently pure speculation.
17 Of course I do not want to imply the existence of a fixed melody; if carried out by a solo singer, it may in fact have
been largely improvised, while still following traditional patterns.
18 For the persistent pattern ‘dichord – number – dichord – number…’ of the Hurrian hymns, cf., as a random
example from the internet, the following notation of a guitar accompaniment to a modern hymn, a sequence ‘chord
– number – chord – number’, also given separately from the text: “D x2 G x1 D x3 A7 x2 D x2 G x1 D x2 A7 x1 D
x1” (‘Amazing Grace’, retrieved from http://www.chordie.com/chord.pere/www.guitaretab.com/d/daniel-thomas/
124540.html on 24 Aug. 2010). Similar notation for instrumental preludes, interludes and postludes only, e.g. at
http://www.cowboylyrics.com/tabs/green-pat/rain-in-lafayette-9249.html, retrieved on 24 Aug. 2010: “D (2) G (1)
D (3) G (1) D (4) A (3) D (6) A (3) D (6) G (1)” etc.
S. Hagel Tunings, scales and tonal systems in antiquity 8
Lutes may have been tuned by shifting their frets, again resulting in identical finger
positions for the same dichords across the various tunings. All this would demand
little ‘theoretical’ effort. Likely, the development of what appears to have been a
unified system was based on the recognition of one ‘master instrument’ at a certain
point in history; but the meaning of most of the dichord names is so obscure that a
complex evolution involving several instrument types synchronously or in
succession cannot be ruled out: the nine-string model might also be a projection of
a pre-existing practice onto a particular instrument.
Finally, dichords are linked to tunings. There are seven possible fifth/fourth
relations within the seven notes of an octave, and there are also seven possible
diatonic tunings of its range (‘octave species’). The cuneiform system ingeniously
associates the two, by the fact that each fifth/fourth dichord may be thought of
‘generating’ one tuning. This is done by attributing a sense of direction to the
intervals, defining fifths as falling and fourths as rising within the tuning process.
So when tuning a lyre or harp, one starts from a particular string and proceeds in
alternating falling fifths and rising fourths. The resulting tuning is called by the
same name as the first dichord that was tuned, i.e. the one starting from the initial
string. Again, this is eminently practical: if the performer must tune his instrument
to, say, išartu, she or he need not have memorised anything like a pattern of tones
or semitones, but merely know the dichord of the same name – a piece of
knowledge, as we have seen, that lay at the core of their art anyway (cf. Figure 1).
1
1
2 2
3 3
4
4
5
5
6 6
7
7
8
8
9 9
non-tuned row of strings
dichord išartu = 2 6
tuning išartu
tuning kitmu
tuning retuning
Often, however, the instrument would already be tuned to a particular scale, while
the instrumentalist needed quickly to change to another one. Here again, one might
follow the same principle: start from the eponymous dichord and check the
sequence of fifths and fourths. At one place you will encounter a tritone instead of
S. Hagel Tunings, scales and tonal systems in antiquity 9
a pure interval. Correct this, and the following ones, until you have completed
seven steps and would return to the initial string: there you are. At first glance this
procedure appears straightforward and efficient – but it has a severe drawback. If
you apply it throughout, you always establish the required intervals by tuning
downwards, decreasing the pitch and the string tension. After merely tuning forth
and back between two scales once, the instrument is pitched a semitone lower than
it was before. Consequently the sound quality is noticeably diminished. A very few
more retuning processes, and the lowest strings cease to sound altogether, not to
mention the fact that the vocalists voice but serious complaints. In practice, it is
therefore inevitable to ensure the maintenance of an overall ‘standard’ pitch
throughout the processes of retuning. This cannot be done in a reasonable way
except by confining the retuning process within strict boundaries. How can this be
achieved? Possibly by the precept not to tune beyond a given scale in one direction,
and if it is not possible to reach the target scale in this way, to proceed in the
opposite direction instead, tuning by rising fifths and falling fourths instead. Such a
precept seems fine enough when we analyse the tunings by means of modern
diagrams. In practice, it involves keeping track of the tuning reached at whatever
stage. Now this is certainly not a problem for any musician who knows his or her
sequence of seven tunings/dichords just as we are familiar with the circle of fifths.
All the same, the ancients seem to have adopted a still much simpler rule, which
implicitly entails exactly such a restriction as required: the central string is never
retuned. 19 On the nine-stringed model instrument, this limits the sequence, as
shown in Figure 2, between qablītu and išartu. Thus the most straightforward way to
establish a new tuning out of any other – and even without knowing the present
one – was to start from the eponymous dichord and proceed in falling fifths and
rising fourths, as described above, until arriving at the central string. If all the
intervals encountered so far were pure and needed no retuning, one would go back
to the initial string and repeat the process, but in rising fifths and falling fourths,
again until the central string is reached.
19 The limits of the retuning process are inferred from the fact that in UET VII 74 the process stops at išartu in one
direction, to be reversed afterwards. The starting and end points are broken off, but it seems beyond doubt that all
seven tunings were covered and more than likely that no semitone-shifted išartu would have appeared at the other
end (in practice, one would need each tuning clearly defined, if only for the sake of unequivocal communication
between instrumentalists). Since išartu and qablītu are not singled out but for their association with a central-string-
based tuning (qablītu starting from this string, both being the last tuneable without changing it in either direction), it
is very probable that their marginal position is determined by the central-string rule. Conversely, it would be an odd
incidence to find the central-string rule as the side effect of some unknown reason to have qablītu and išartu as the
limits, of all the equally possible seven choices – note especially that the starting position is not determined by the
first or last string, or by the dichord appearing first in the list of CBS 10996.i.
S. Hagel Tunings, scales and tonal systems in antiquity 10
išartu
kitmu
degrees of
qablītu
pītu
embūbu
nīš tuḫri
nīd qabli
bottom notes
tunings
eponymous dichord, from which a given tuning starts
tritone interval within a given tuning, to be corrected by re-tuning
Figure 2: The ancient Near-Eastern (re)tuning cycle20
On the nine-stringed model instrument, therefore, called ğišzà.mí in UET VII, the
retuning sequence appears defined by the notion of a central string – whose
centrality of course depends on the number nine. 21 The same nine strings are
attested in Nabnītu 32 i, where their names are given: here the fact that they are
numbered inwards from both directions demonstrates that the notion of the fifth
string as a centre was deeply ingrained in the system. It seems that a nine-stringed
instrument played a crucial role in musical practice (at least music-educational
practice) long enough that not only its string names would become the basis of a
canonical description, but also the particular retuning boundary between išartu and
qablītu that depends on its nine strings.22 The name of the dichord and tuning qablītu,
“middle”, which start from the fifth string, evinces that the nine-string standard is
20 For perfect symmetry within the transcription to modern notes, one would identify the central string with d, thus
making pītu the natural scale and attributing three flats to qablītu and three sharps to išartu. I have used a instead,
which makes the diagram directly comparable to those of ancient Greek music, where the ‘middle’ string, mésē, a fifth
below the highest, is best equated with modern a. If zà.mí string lengths roughly corresponded to those of typical
Greek citharas, this would also result in approximately correct pitch.
21 If a similar sequence were to be built not on nine strings, but merely on the seven that are necessary to establish
the heptatonic scales without octave doublets, dropping the two lowest strings, the fourth string would be central. If
it is kept constant, the sequence would start at nīd qabli and end with nīš tuḫri.
22 For possible evidence for nine strings in the Hurrian hymns cf. Hagel, “Is nīd qabli Dorian”, p. 320. In the (roughly)
Neo-Babylonian text YBC 11381, the nine strings still appear in what seems to be a very performance-oriented
context, labelling nine benedictions in turn (cf. Payne, E.E., “A new addition to the musical corpus”, in: S.C.
Melville/A.L. Slotsky (eds.), Opening the tablet box. Leiden/Boston (2010), pp. 291–300).
S. Hagel Tunings, scales and tonal systems in antiquity 11
23 Cf. Arndt-Jaemart, J., “Zur Konstruktion und Stimmung von Saiteninstrumenten nach den mesopotamischen
Keilschrifttexten”, Orientalia 61.4 (1992), pp. 425–447: p. 435. Lines 12–16 would then mean: “išartu – enfeebling
išartu – kitmu – enfeebling kitmu – enbubu…”; on this interpretation, the much discussed question whether the si ip-
qualified terms refer to tunings or dichords actually loses its point: both are destroyed at once, the manipulation of
the tuning being carried out on its eponymous dichord.
24 The term is less than felicitous because it implies a non-reentrant standard, within which ‘starting points’ of new
seven would already have been part of such a heptatonic music culture in a very
early stage.
The names attached to the dichords and scales have puzzled scholars, and many of
them so far escaped convincing explanation.25 This is only to be expected. Their
meaning must not be sought in terms of the ready-made system, let alone in its
abstract graphical representations as stars or whatever, not even if these are found
in the ancient record.26 Did they originate as dichords or as tunings? The terms for
the thirds/sixths class of course only exist as dichords, but this does not decide the
case, since they may be later extensions: after all, two of them are seemingly derived
from ones of the fifths/fourths type, albeit in an apparently unsystematic manner:
titur qablītu and titur išartu. The conceptional path from dichord to the tuning it
generates is doubtless smoother than vice versa. On the other hand, the term
embūbu, “double pipe”, might make more sense as a scale.27 The actual development
probably followed much more convoluted trails than most of us would like to
imagine.
Anyway, we must never forget how little we know about the music whose basics
were communicated by the excavated system of terms. We know nothing about the
absolute pitch, until the zà.mí is identified and its (approximate) string lengths are
known.28 We do not know how the system was applied to other instrument classes,
particularly instruments with more than nine strings. We must not forget that the
dichord system perhaps applied to only a very small segment of ancient Near
Eastern music-making. Finally, we must remember that tunings are not modes. Any
of the seven tunings may have been associated with a particular mode defined by
typical starting, focal and final notes, prominent dichordal ‘harmonies’ and their
interplay.29 Or there may have been more than one way of putting each tuning to
modal use.30
Octave-enhanced resonance
The enneachordal nature of the system, which seemed to go beyond the ephemera
of an eventual codification, allows us at least to speculate about some essential
25 Cf. Krispijn, Th.J.H., “Musik in Keilschrift. Beiträge zur altorientalischen Musikforschung 2”, in Hickmann, E. et
al., (eds.), Studien zur Musikarchäologie 3 = Orient-Archäologie 10, pp. 465–479: 471.
26 For modern stars, cf. Vitale, R., “La musique suméro-accadienne. Gamme et notation musicale”, Ugarit Forschungen
pipes of a pair is hardly viable: embūbu covers a fifth, but no known ancient pair of pipes has so large an interval
either between the highest or between the lowest notes of each pipe. Octave-invariance, it is true, would turn embūbu
to a fourth, such as is well attested for Greek auloi; but this would entail that the pipes exceeded the range of the
nine strings by just a semitone at the lower end or a tone at the higher end. So it would become very difficult to
understand from a musical viewpoint why the instruments’ ranges would not have been aligned outright, so that, for
instance, the highest finger holes would have corresponded to strings 1 and 4.
28 On strings and pitches cf. Hagel, S., Ancient Greek music: A new technical history, (Cambridge 2009), pp. 88–92.
29 I have tried to unravel some of the modal characteristics of the Hurrian hymns in “Is nīd qabli Dorian”, pp. 307–
338.
30 After all, the assumption of a 1:1 relation between tunings and modes might be bolstered by the fact that songs are
found classified by tuning (KAR 158 VIII, see below) – but this is far from conclusive, since tunings are readily
classified, while modes, which need not even exist as explicit concepts in the musicians’ minds, are difficult to pin
down.
S. Hagel Tunings, scales and tonal systems in antiquity 13
ingredients of modality. Of nine notes, the two highest are duplicated at the lower
octave; and we have seen that their functional equivalence was taken for granted.
This means that when dichords are played on such an instrument, several of these
can avail themselves not of two but of three notes by including the octave
counterpart of one boundary note. In this way, some dichords can be executed
with significantly higher resonance than the others. This is particularly important in
the case of fifths and fourths, where, due to the simple frequency ratios involved,
the resulting three-note dichords blend especially well. Four possible tri-dichords of
this sort exist, involving either strings 1 and 8, coupled with either 4 or 5, or 2 and
9, together with either 5 or 6. Not all of these, however, are available in every
tuning: wherever the tritone starts from string 1 or 2 downwards, the
corresponding set is not among the concords (cf. Figure 3).
1 1
2 2
3 3
4 4
5 5
6
6
7 7
8 8
9 9
embūbu kitmu
Figure 3: Examples for octave-enhanced resonance in Near-Eastern enneachords
Even so, no tuning has fewer than three octave-enhanced fifths/fourths, so that
none is seriously impaired as regards its modal capabilities. So it is no wonder that
we find all tunings in actual use, in the song list of KAR 158 VIII (cf. Table 1). The
comparison of kitmu with pītu, the former sporting fewer octave-enhanced dichords
yet covering four times as many songs, might indicate that three octaves divided by
fourths/fifths were definitely felt sufficient to support a full-fledged modal scale.
Admittedly, though, we do not know whether the canonical nine played a practical
role at the time (ca. 12th/11th cent. BC) and the place (Aššur) where the list was
compiled, so that the value of its testimony in this respect is necessarily limited.
S. Hagel Tunings, scales and tonal systems in antiquity 14
songs in
high fourths high fifths
KAR 158 VIII
išartu 1 2 23
kitmu 2 1 17
embūbu 2 2 24
pītu 2 2 4
nīd qabli 2 2 ?
nīš tuḫri 1 2 ?
qablītu 2 1 ?
Table 1: Octave-enhanced resonance and number of songs in KAR 158 VIII
In any case, wherever the gamut of a ninth was prominent, we must assume that it
imprinted its characteristics upon the mode or modes associated with each tuning.
In particular, we expect that the highest and lowest two notes enjoyed a prominent
status, and similarly the central note: apart from those mentioned, string 5 is the
only one that can take part in two octave-enhanced sets of the fifth/fourth type.
Here we have an additional musical explanation for the structural significance of
this central string of nine. Strings 4 and 6, which take part in no more than one,
may still have had some importance – perhaps serving as contrastive foci.
Conclusion
On balance, as far as I can see, every single feature that ancient Near Eastern
musical texts exhibit is intimately associated with instrumental practice. There is no
evidence for wider-reaching abstraction. Scales are conceived as tunings, and
terminology as far as understood gives no hint about any fine-tuning that would
deviate from the ‘Pyhtagorean’ scales arising from pure fifths and fourths.31 Beyond
the mere codification of practical aspects, so far we only have a Late Babylonian
heptagram (CBS 1766) whose points are labelled as string names in such a way that
the circumference takes us through the seven steps of the octave, while the drawing
of the star reproduces the fifths and fourths (depending on whether counting
clockwise or anticlockwise) of the tuning and retuning process.32 This is doubtless
an ingenious representation of musical facts in diagrammatical form, testifying to a
profound understanding of the implications of the tuning process within an octave-
invariant paradigm.
31 There have been speculations about certain texts reflecting fine tuning processes (Kilmer, A. / Tinney, S., “Old
Babylonian Music Instruction Texts”, Journal of Cuneiform Studies 48 (1996), pp. 49–56; Kilmer, A.D. / Tinney, S.,
“Corrections to Kilmer/Tinney ‘Old Babylonian Music Instruction Texts’, JCS 48 (1996)”, Journal of Cuneiform Studies
49 (1997), p. 118; Smith, J.C. / Kilmer, A.D., “Laying the Rough, Testing the Fine (Music Notation for CBS 10996)”,
in Hickmann, E. / Eichmann, R., (eds.), Studien zur Musikarchäologie 1 = Orient Archäologie 6 (2000), pp. 127–144;
Kilmer, A. / Peterson, J, “More Old Babylonian music-instruction fragments from Nippur”, Journal of Cuneiform
Studies 61 (2009), pp. 93–96.), but these have little explanatory value (cf. Colburn, “New interpretation”, pp. 97–109),
and have also not established any definite musical meaning. My own speculations about a possible tuning with pure
thirds in the Hurrian hymns are based on dichord statistics and do not suppose a theoretical formulation behind or
beyond the practice (cf. n. 29 above).
32 Horowitz, W., “A Late Babylonian tablet with concentric circles from the University Museum (CBS 1766)”, Journal
of the Ancient Near Eastern Society 30 (2006), pp. 37–53; Waerzeggers, C. / Siebes, R., “An alternative Interpretation of
the seven-pointed star on CBS 1766 (Horowitz, JANES 30)”, N.A.B.U. 2007 pp. 43–45.
S. Hagel Tunings, scales and tonal systems in antiquity 15
On the other hand, we have no evidence for abstract terms for intervals classified
by size, such as ‘fifth’, ‘fourth’, ‘tone’ or ‘semitone’.33 As a consequence, we know of
no explicit interval arithmetic of the sort of ‘an octave consists of a fifth plus a
fourth’. No doubt the idea as such was represented in the musicians’ mind, but the
seeming absence of specific terminology would have precluded theoretical
formulation. Lacking interval arithmetic of the ‘straightforward’ sort, finally, there
was hardly any point from which a description of pitch relations in terms of ratios
would have started. Surely, the Babylonians would have had at their command the
mathematical means to deal with pitch arithmetic – but so far there is no evidence
that they employed them in that way. And indeed it would be very surprising if
such evidence ever came to light, given the absence of prerequisite key concepts in
the tablets that have so far been identified as treating musical matters.
33 It has been conjectured that pismu in Nabnītu 32 i 11 means ‘octave’. If this is correct, nevertheless we should not
take it as ‘octave’ in the primary modern sense of an interval of a particular size if that concept seems absent
otherwise.
34 Cf. e.g. West, Ancient Greek music, pp. 219–223.
35 Philolaus, fr. 6a = Nicomachus, Encheiridion 9, p.252.17–22 Jan: “The size of harmonía is a fourth [syllabá] and a fifth
[di’ oxeiân]. And the fifth is larger than the fourth by 9:8. For from hypátē to mésē there is a fourth, from mésē to n tē a
fifth, from n tē to trítē a fourth, from trítē to hypátē a fifth.” The term harmonía here evidently refers to the octave, but
not simply as an interval but apparently in the sense of a fixed size structured internally so as to become a
harmonious whole.
S. Hagel Tunings, scales and tonal systems in antiquity 16
high-pitched’. This must also have originated in lyre tunings, where the interval
from the central note to the top note, i.e. that spanning the high-pitched half, was
in fact a fifth (cf. below with Figure 4). In this original sense, the notion
corresponds to a dichord. But Philolaus no longer restrains the term to a particular
location, but applies it also to the lower range of the strings, in the abstract sense of
‘a fifth’. Here we seem to witness the transformation of dichordal thinking to the
abstraction of interval sizes. Why the interval of a tone is called by this name, tónos
‘stretching, pitch’, is not very clear, but it is probably a sign that diatonic tunings
played a chief role.36 For smaller intervals, the term díesis, ‘letting through’, was used,
which is thought to derive from the practice of half-covering finger holes on wind
instruments. Scales, finally, were apparently analysed in terms of their constituent
intervals, or as consisting of fourths and tones.
Aristoxenus’ fully-developed system is strictly heptatonic. Its coherence is
guaranteed by fifths and fourths in a sophisticated way, although it was by no
means restricted to a ‘Pythagorean’ diatonic, but included scales with interval sizes
ranging from quartertones to undivided ditones. Even so, any pure instance of a
scale would have seven pitches within the octave – if less, this would count as a
conscious ‘omission’ on the part of the composer; if more, this would result from
modulation, a mixture of more than one scale.
Heptatonism, it seems, was indeed the paradigm of Greek music as far as we can
trace it back; even if it was probably often subjected to modifications of just the
sort Aristoxenus had in mind. I will come back to this point in a moment.
The remains of archaic Greek double pipes usually sport finger holes that are most
easily interpreted in terms of heptatonism. The holes are too evenly spaced to yield
tones and semitones, but also definitely not equidistant. Almost all of the extant
items are however made from bone, so we cannot exclude the possibility that pipes
made of more perishable materials might have looked quite different. In any case,
some kind of non-Pythagorean diatonic, perhaps close to modern bagpipe scales,
seems to have dominated much of archaic piping. Later sources imply that such
pipes would have been used for music with microtonal divisions also, probably by
half-stopping certain finger holes.
Lyres
At the same period, lyres almost universally have seven strings, both in the
iconography and in literary testimony. This would go well with heptatonism, and
one could imagine a full (re-)tuning cycle on seven strings, very similar to the Near
Eastern evidence. Yet none of the sources reflects such a state. Already in about
the 3rd century BC, the question how the seven-stringed lyre had been tuned was
subject to discussion and speculation 37 – but the more informed testimonies,
including a single one from the 5th century, are unanimous in placing an octave
between its outermost strings. The early one is the Philolaos quotation which we
have already come across: he identifies the octave and the fourths and fifths that
structure it with reference to lyre string names, as shown in Figure 4. The fact that
the so-called ‘third’ string is tuned to a note a fourth from the top proves that
reference is here made to the typical seven-stringed lyre, still the standard
instrument of education in the period in question. To our eyes, and to those of
Aristoxenus, this is a ‘gapped’ tuning; more neutrally it can be described as
heptatonic at the bottom but pentatonic at the top.
nḗtē, ‘bottommost’ e'
trítē, ‘third’ b
mésē, ‘middle’ a
hypátē, ‘topmost’ e
Figure 4: The seven-stringed lyre according to Philolaus
Philolaus does not refer to some specifically tuned lyre;38 he talks about harmonía as
such, and when he comes to prove his proposition by reference to lyre strings,
there is clearly no idea that any of those mentioned might ever be tuned differently.
From Philolaus’ fragment we can therefore conclude, despite its brevity, that at his
time the seven-stringed lyre was invariantly tuned to a ‘gapped’ octave, structured
in fifths and fourths by two symmetric notes in the centre, and further divided by
one remaining string in the upper half, and two in the lower. The particular pitches
of these would define the tuning (there were of course more than one). As a matter
of fact, these characteristics are corroborated by much later evidence that evidently
stands in unbroken historical continuity with the lyre of Philolaus’ time: the cithara
tunings as described in Ptolemy’s Harmonics still feature exactly the same structured
octave (although now fully heptatonic). Surprisingly, the citharist of the 2nd
century AD concert hall still refrained from following the retuning cycle to a point
where the basic symmetry of the octave would be broken (cf. Figure 5; we need not
discuss here the question why retuning actually stops one step short of this point
in the case of upwards tuning).
38That he talks about a lyre at all emerges from the fact that the term trítē, ‘third’ string/note, shifted to the third step
in a full heptatonic octave before it was generalised as a note name, independent of instrument.
S. Hagel Tunings, scales and tonal systems in antiquity 18
e’
d
c ,
c
b
a
g ,
g
f ,
f
e
hypértropa
parypátai
iástia
trítai
lýdia
Whence this restriction to only a part of possible tunings? I think the answer is
closely related to the ‘gapping’, which testifies to a preference for having an octave
on the instrument to having ‘all notes within it’ available. Just as we have speculated
for the Near Eastern tunings, here also an octave between the outermost strings
meant the possibility to play octave-enhanced ‘dichords’. 39 With seven strings,
however, there is only one octave and therefore are only two respective dichords of
the fifth/fourth kind (as compared to four on the nine-stringed zà.mí). The loss of
one of these would have done substantial damage to the harmonic capabilities of
the instrument, even to the resonance of its sound. Naturally it would have been
regarded too high a cost for the additional modal variety gained by retuning one of
the central strings. So it is perfectly understandable why one preferred to stay with
a mere three retunable strings (until more strings were added).
Perhaps it was this restriction that set the Greeks on the path of pursuing
microtonal variety. Unfortunately, we know little of the practical background of
this part of theory: when Aristoxenus’ describes various shades of tunings, it is
generally hard to tell to which instrument these would apply. But at least sometimes
the lyre is a very likely candidate.40
4th century AD, mostly vocal repertoire. Where sufficient portions of the melody
are extant, we are able to identify modal characteristics such as focal notes or
structurally important intervals. Furthermore, some of the songs can be identified
as lyre-accompanied, and if so, the key in which they are notated may betray their
tuning, thanks to Ptolemy’s account.41 Where we have more than one piece in the
same key from the same context, it is possible to inspect the relation between
modality and tuning.
Luckily, a few citharodic pieces, probably from a kind of beginner’s lyre textbook,
have come down to us in the manuscript tradition. The four items whose notation
survives (at least to a great part) are all notated in the Lydian key and can be
associated with Ptolemy’s tuning lýdia. Nevertheless, they do not all exhibit similar
modal characteristics. One short Invocation is clearly centred on an axis between E
and A, corresponding to the old octave-enhanced ‘dichord’ with mésē on the archaic
lyre, while the fourth between G and D provides a harmonic contrast.42 In sharp
opposition, a Hymn to Nemesis is almost exclusively focused on G, supported by D,
while other notes only rarely acquire prominence.43 Other evidence suggests that
both these modal options were common.
In Greek lyre music, therefore, tuning and modality were interrelated, but not
associated in a one to one relation. As so often, the reality of music defies simple
categories. Whether the same would be true for Near Eastern songs, or whether
there each tuning might have implied a mode, at least at some places in some
periods, we will probably never know.
Auletic ‘modes’
The scales of archaic aulos music are so far mostly names for us. Later sources talk
about Dorian, Phrygian and Lydian pipes;44 this may or may not result from learned
reconstruction. Some kind of Dorian double-pipe layout may be reconstructed
from Aristoxenus’ remarks,45 but on the other hand, the remains of instruments so
far defy classification.
Towards the end of the 5th century BC, we learn, the development of Greek music
took a huge leap with the invention of modulating pipes, i.e. auloi including
keywork that made it possible to open different sets of holes. Such an enterprise
requires the analysis of pipe scales that had previously existed side by side, and
mapping them onto each other (perhaps with inevitable modifications). In fact the
business of scale analysis seems to have been well underway in the first half of the
century. The Roman-imperial writer Aristides Quintilianus famously mentions
irregular scale forms, ‘harmoníai’, which he associates with Plato’s time, and
41 Even if a piece was written down during Ptolemy’s lifetime, it is clear that we cannot assume that the microtonal
details described by him would have applied to the intended performance; for the following argument, however, we
need not assume more than that tunings were associated with keys, and as such constant for a given artist at a given
period. This, I think, is more than reasonable.
42 Pöhlmann, E. / West, M.L., Documents of Ancient Greek Music, (Oxford 2001), nr. 24.
probably thinks to be even older (cf. Figure 6).46 These can be argued to derive
from an early systematisation, at some time around 500 BC, that was related to
aulos design.47 Thus they are probably still in some way not very far removed from
early classical ‘modes’ – though unfortunately once more reduced to scales, whose
tonal hierarchies we can only guess.
e e
d d
c c
b b
a a
g g
f, ,
f
f f
E E
D D
C C
B B
quartertone grid
Lydian
Dorian
Phrygian
Iastian
Mixolydian
Syntonolydian
When extra strings were added, these did not so much enlarge the ambitus, but
provide additional, modulating notes. This would lead to the later canonical
‘chromatic’ with its sequence of two semitone steps within the fourth.
On the other hand, there was the simple aulos, playing a more pipe-like variant of
diatonic, but also microtones that would later be codified as ‘the enharmonic genus’.
Sometimes auloi and lyres played together, ensuring a certain amount of
compatibility.
At about this stage, three things of great consequence came to pass. Firstly,
musicians began to give accounts for the scales they used, developing a terminology
in which to speak about intervals and larger systems. Important musical vocabulary
that is still in use was coined, starting from the word ‘music’ itself. Obviously in
close connection with such endeavours, signs were conceived for a first melodic
notation, at first mostly related to the aulos, but soon embracing lyre music, too.
Thirdly, it was discovered that concordant intervals relate to quantitative physical
differences that are describable in ratios between small integers. The clue might
have been taken from the art of positioning finger holes on pipes or flutes; it seems
to have been corroborated by experiments of different kinds as well – though
apparently not yet on a monochord, or with strings at all.48 The discovery had the
greatest influence on Pythagorean philosophy, consequently attaching itself to the
founder of the movement. It spawned a separate, mostly speculative, branch of
music theory, which readily emancipated itself from the banalities of actual music-
making to contemplate hypothetical cosmic harmonies.
With the invention of two kinds of mechanism, auloi grew larger and more versatile.
The first development was reflected in the conception of a model scale with the
ambitus of two octaves, the ‘Perfect System’. It is conceived as a framework of
fourths and fifths, whose inner notes can be tuned in various ways: ‘fixed’ and
‘moving’ notes – a concept that we have seen preformed in the lyre with its
comparable framework ‘harmonía’ with tuneable strings in between. Just as Philolaus
had accounted for the lyre harmonic framework in mathematical terms, we find
traces of a similar endeavour for the two-octave framework, perhaps connected to
the famous Pythagorean Archytas.49
On the other hand, the modulating capabilities of the new generations of auloi had
required that the old scales be related to each other and, if possible, aligned at their
highest pitches. We can imagine star musicians and instrument makers brooding
over newly-devised diagrams. Later sources give echoes from what seem to have
been two different solutions to the problem. At any rate, when the 5th century BC
drew towards its end, auletes who had the financial resources to commission such
an instrument, were able to switch between different ‘keys’ during performance.
Consequently, notation had to be adapted in order to write down such ‘New Music’,
here also introducing the concept of ‘keys’. Their theoretical formulation comes
under the heading of ‘tónoi’ or ‘trópoi’. At this stage, it seems already to have been
commonplace to describe scales in terms of tetrachords, four-note units spanning a
48Cf. Hagel, “Calculating auloi”, pp. 384–385; Hagel, Ancient Greek music, pp. 176–177; 333 n.21; Creese, Monochord.
49Hagel, S., “Twenty-four in auloi: Aristotle, Met. 1093b, the harmony of the spheres, and the formation of the
Perfect System”, in Hagel, S. / Harrauer, Ch., (eds.), Ancient Greek Music in Performance, (Vienna 2005), pp. 51–91.
S. Hagel Tunings, scales and tonal systems in antiquity 22
fourth, that were concatenated in various ways, either directly or with a separating
interval of a tone in between – though it was not before Aristoxenus that exact
rules were pinned down, which kinds of tetrachord can be freely arranged in this
way. A particular ‘school’ of musical thinkers, associated with the name of
Eratocles, specially emphasised the octave, demonstrating how its different
‘species’ are generated by subsequently transferring intervals from one end to the
other. They christened those species by the names of identical or similar modal
scales. The idea was apparently of little practical importance by then; thanks to its
neat layout, however, it was later to become one of the central concepts of Western
music theory.
In the middle of the 4th century, the musicians’ view was apparently ruled by the
idea of a quartertone grid into which any scale and key could be fitted; Aristotle
would simply quote the quartertone as the unit of measure in music.50 Plato, on the
other hand, adopted the mathematical description of a tuning in pure fifths and
fourths for the structure of the world soul in his Timaeus, thus setting a standard for
later ‘Pythagorean’ (i.e. in fact Platonising) thinkers until late antiquity. From this
point on, the quest for a mathematical description of harmony in terms of ratios
would remain divided into a largely fossilising branch of Platonisers and a handful
of original thinkers who, sometimes more playfully, sometimes with more
philosophical seriousness, searched for an optimal description of musical structures
in terms of numbers.
simple auloi lyre tunings
diatonic / ‘enharmonic’ notation I diatonic / ‘chromatic’
tónoi / trópoi
Platonising
Aristoxenus
open
Towards the end of the century, Aristoxenus drew all strands of non-Pythagorean,
practice-oriented theory together and devised his inclusive system: ‘diatonic’,
‘chromatic’ and ‘enharmonic’ as three genera on an equal par, in practice realisable
Comparison
If we compare the musical cultures of the ancient Near East and the Hellenic and
Hellenistic world, they have doubtless much in common: dominant instrument
types as well as a heptatonic basis (both going back to times unfathomed). Greece
always documented a historical connection to Anatolia by calling musical styles of
central importance by the names of Lydia and Phrygia. Western Anatolia, however,
is not Mesopotamia, but was culturally much more closely linked to Greece. In any
case, both in Greek and cuneiform sources, heptatonism is combined with strong
focus on fifths and fourths as scale-generating intervals, and an awareness of the
octave as generating some sort of functional identity (and, it seems, modally
important enhanced resonance on the lyre).
When we come to the written sources, however, the difference is surprising, and
not all of it can be blamed on the different ways of preservation: continuous
manuscript tradition plus papyri versus excavated clay tablets. In the cuneiform
texts we hardly find anything that cannot be directly related to musical practice: lists
and definitions of dichords, which were apparently at the heart of the
instrumentalists’ art, sequences of tunings and retuning instructions, classification
of songs according to their tuning, finally some musical notation. In the Greek and
Hellenistic-Roman world, such types of sources form a small minority or are
altogether missing. Of course there are the documents of musical notation, some
transmitted from an ancient book, a few found inscribed on stone, the bulk from
scraps of papyrus. Apart from these, parts of the compilation known as
Bellermann’s Anonymus appear to have a strongly practical background, for
instance explaining the performance of special notational signs. There must have
been practical handbooks for musicians, presumably also for instrument builders,
but none of these were further transmitted when the music culture in which they
had made sense came to an end. The extant works are of a very different sort: part
original treatises furthering ‘harmonics’ as an abstract science, much as related to
music-making; part handbooks for the educated elite, compiling definitions from
the meaning of ‘music’ and ‘harmonics’ down to the names and mutual relations of
scales, perhaps also coupled with a more or less deep-reaching treatment of ratio
mathematics. All this mostly for a target audience that was more expected to
produce learned statements about music when invited to dinner than to exercise
musical skills themselves.
With all its intellectual penetration of pitch structures and philosophical reflection
of music’s values, then, the rise of Greek science has not protected real music from
a certain degree of social demotion. If anything, it has supported an increasing
delegation of actual music-making to professionals, upon whose skills – regardless
whether those of a trained slave or of an international star – the upper-class citizen
could always frown. An intellectual grasp of technicalities, however superficial, thus
enabled one to display the status of a ‘musical man’ (as they had used to call a
gentleman in Classical times) without actively engaging in the production of music.
The elevation of ‘real’ harmony to the realm of the inaudible, the cosmos and the
well-tuned soul, ultimately paved the way for the rejection at least of instrumental
music in upcoming monotheist mainstream doctrine: an ideological task of musical
self-discrediting that innocent dichord lists would not have accomplished.