Professional Documents
Culture Documents
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
Κατατομή κανόνος
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
DIVISION OF THE CANON
!Much has been written about the meaning of πληγη in ancient, espe
cially Pythagorean acoustical theory. See the Introduction, pp. 48-49.
Regarding the translation of φθόγγος as “note,” See the Introduction,
pp. 49-52.
3Literally, “that which is needed, or proper.” I have subsequently trans
lated το ίίου as “the desired pitch” since that seems to be the intention of
the remark. Even more precise might be η τάστς ΙκομΙνη.
4From here to the end of the Introduction, the treatise states the Funda
mental Principle of Consonance (FPC) of the Pythagoreans. See the Intro
duction, pp. 52-54.
115
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
116
μου λόχψ λέγεται προς αλληλα- ώστε καί τούς φθόγγους, άναγκαίον έν
αριθμοί) λόγφ λε'χεσθαι προς άλλήλονς. των δε αριθμών οί μεν έν πολλα
πλασίουl λόγψ λέγονται· οί δε έν έπιμορίφ· οί ΐ* έν ετπμερεϊ* ώστε καί
τούς φθόγγους αναγκαίου έν τοΐς τοιούτοις λόγοις λέγεσθαι προς
5 άλληλους. τούτων δε οί μεν πολλαπλάσιοι καί έπιμόριοι ένι όνόματι
λέγονται πρός άλληλους. γινώσκομεν δε καί των φθόγγων τούς μέν
συμφώνους όντας· τούς δε όιαφώνους· καί τούς μεν συμφώνους μίαν
κρασιν την έζ άμφοΐν ποιοϋντας- τούς δε όιαφώνους ου. τούτων όντως
έγόντων εικός τούς συμφώνους φθόγγους έπειδη μίαν την έξ άμφοΐν
10 ποιούνται κρασιν τής φωνής είναι των έν ένι όνόματι προς άλληλους
λεγομένων αριθμών* ήτοι πολλαπλασίους όντας ή έπιμορίους.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
117
another by numerical ratio; and thus with notes, it is necessary that they be
related to one another in numerical ratio. Of numbers, some are related in
multiple5 ratio, some in supeiparticular, and some in superpartient; and thus
it is necessary that notes be related to one another in ratios such as these. Of
these, then, the multiples and the superparticulars are related to one another
by one name. And we know of notes that some are consonant, some disso
nant; and the consonant notes make a single blend from both notes, but dis
sonant notes do not This being the case, it is reasonable that consonant
notes, since they make a single blend of sound from both notes, are related
numerically to one another in one name, since they are either multiple or
superparticular.6
Original from
Digitized by
UNIVERSITY OF MICHIGAN
118
α
’Εάν διάστημα πολλαπλάσιον διί σνντιθο/ miff τι διάστημα, και
αύτδ πολλαπλάσιον ίσται. ϊστω διάστημα τδ βγ καί ίστω πολλα-
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
119
l.7
If a multiple interval twice combined makes some interval, this itself
will be multiple. Let BG be an interval, and let B be a multiple of G;8 and let
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
120
1 η κατασκευή in marg. Up I γ (sec.)]T sup. Un. Za2 (τρίτον [?] del.) I τον
β]τδ β codd. praeter PcPd I τον δ]τό δ OaOd I δ]Δ sup. lin.Za2 (τέταρτον
del.) I ό διορισμός φημι Up II 2 και om. FrJanMg I Utt. del. post και Pe I τον
δ]τό δ codd. I επει]ή άπόδίζις επει Up II 2-3 πολλαπλάσιοί]-!/ BbFrLdMm
mu(proer. UpVn)NnVdVjVoVv -p Vc II 3 μετρή Mm I τον β (ρτ.)]τοΰ β Pf
I ό β] om. Vo ό y Mg II 4 τον (pr.) om. Vi I μντρέί δ γ και τον JanMg I γ
τόν]γ προς τον Υο(πρδςΊ del.)LdMm^pos in ras. )zaVdΝ\(πρδς sup. Un.)
Vv I τον δ (ίβσ.)]τό δ Bb II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
121
9In Vj following αρα» two pages of the Εισαγωγή αρμονική (see Karl
von Jan, Musici scriptores graeci [Leipzig: B. G. Teubner, 1895; reprint ed.,
Hildesheim: G. 01ms, 1962], pp. 198.13-204.17) are inserted. These two
pages should follow p. 317 of Vj, where they fill the lacuna in the Εισ
αγωγή αρμονική. Ld follows Vj.
10Analysis
17 Ίτρότασις 118.2-3
η εκθεσις 118.3-120.1
ο διορισμός 120.1-2
η κατασκευή does not exist.
η απόδειζις 120.2-4
τό συμπέρασμα 120.4
Exposition
Β = aG, where α is an integer
G:B::B:D
Specification
D = bG, where b is an integer
Proof
B/G=a
D/B=B/G
D/B =a
D/aG = a
D/G = a2
Since a is an integer, a2 is an integer
Therefore
a2 * b, and D/G = b
Conclusion
D = bG
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
122
β.............................................
Έαυ διάστημα δι? συυτίβΐυ y | /3 | δ τό δλον ποιβ πολλαπλάσιον δ_1
η I LV και αυτό «τται πολλαπλάσιου. ίστω διάστημα τό βγ και yeyc-
νησθω ώς ό γ πρός τον β ούτως ό β πρός τον δ· καί έστω ό δ τοϋ γ πολ-
5 λαπλάσιΟΓ φημι και τον β τοϋ γ ειυαι πολλαπλάσιου. επει γαρ ό δ του γ
πολλαπλάσιό? εστι. μέτρα αρα ό γ τον δ. Ιμαθον δε ότι ίου ωσιν αριθμοί
ανάλογου όποσοιουν ό δε πρώτο? τον Ισγατον μέτρα» και τους μεταξύ
μέτρησα· μέτρα αρα ό γ τον β. πολλαπλάσιος αρα ό β τοϋ γ.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
123
2.
If an interval twice combined, c.g., G I Β I D, makes a multiple whole,
e.g., 418 116, this itself will be multiple. Let BG be an interval, and let it
have been as G was to B similarly B was to D. And let D be a multiple of G.
I say that B is a multiple of G. For since D is a multiple of G, therefore G
measures D. I know that if—however many numbers should be in propor
tion—the first measures the last, it will also measure those in between.
Therefore, G measures B; therefore B is a multiple of G.11
11 Analysis
η πρότασιχ 122.2-3
η 122.3-5
ό διορισμός 122.5
η κατασκευή 122.6-8
η 122.5-6, 8
τδ συμπέρασμα 122.8
Exposition
G:B::B:D
D = aG, where a is an integer
Specification
B = bG, where b is an integer
Machinery
If J:Kf ..: Kn and K^/J = c, where c is an integer,
then Kj/J = d, where d is an integer (Cf. Euclid Elements 8.7; Jean
Pena also makes this connection in his Euclidis rudimenta musices
[Paris: Andreas Wechelus, 1557))
Proof
Since D = aG
D/G = a
Therefore B/G = b
Conclusion
B = bG
Original from
Digitized by
UNIVERSITY OF MICHIGAN
124
y f t ,
ΈττιμορΖου διαστήματος μέσοι· y I fi I δ ουτβ «2 ουτ* πλάους*
ανάλογον έμπ^σοϋνται αριθμοί, έστω γαρ έπιμόριον διάστημα τό β τοϋ
γ ίλάχιστοι δί έν τφ αντφ λοχφ τοίς βγ έστωσαν οίδζ θ· ούτοι συν υπό
5 μονάδος μόνης μςτροϋνται κοινού μέτρου· I β I y αφιλι ίσον τφ θ τον ζ.
λοιπόν μόνον, έπιμόριός έστιν αρα ο δζ του θ· η υπέροχη ο δ κοινόν
μέτρον τοϋ tc δζ καί τοϋ θ έστϊ· μονας αρα δ δ* ούκ αρα έμπ^σέιται άς
τούς δζ θ μέσος ουδάς. έσται γαρ δ έμπίπτων τοϋ δζ έλάττων τοϋ δέ θ
μάζων ωστ€ την μονάδα διαιρασθαι σπ^ρ αδύνατον, ούκ αρα έμπ^σάται
10 άς τους δζ θ τις. δσοι δέ άς τούς έλαχίστους μέσοι ανάλογον
Original from
Digitized by
UNIVERSITY OF MICHIGAN
125
3.12
Neither one nor many mean numbers, e.g., G, B, D,12 13 will fall propor
tionately between a superparticular interval. For let B be the superparticular
interval of G; and let DZ© be the least numbers in the same ratio as Β.Ό.
These then, by the unit alone, are measured by a common measure, e.g.,
3:2.14 Subtract Z equal to Θ. Since DZ is a superparticular of Θ, the excess
D, the only remaining number, is a common measure of both DZ and Θ.
Therefore, D is a unit; therefore no mean will fall in between DZ©. For the
inserted mean15 is less than DZ and greater than Θ; thus the unit will be
divided, which is impossible. Therefore, nothing will fall in between DZ©.
However many16 means fall proportionately in between the least numbers.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
126
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
127
so many will also fall proportionately between those numbers having the
same ratio. None will fall in between DZ©; none will fall in between
B:G.17
17Analysis
This proposition is not neatly divisible according to the standard Euclidean
model.
η πρότασή 124.2-3
η fateris 124.3-4
6 διορισμός Does not exist, but would be: ”1 say that since B:G is a
superparticular ratio, no mean will divide B:G.”
η κατασκευή 124.4 (assumed), 124.4-5,124.10-126.2
η 124.6-10
τό συμπέρασμα 124.4-5,126.2-3
Exposition
B:G is superparticular
B:G::(D+Z).©, where (D+Z)© is in lowest terms. That (D+Z)©
exists and can be found tacitly assumes Euclid Elements 7.33
Preliminary Conclusion
(D+Z) - © = 1
Machinery
Z=©
(D+Z)-Z = D
Proof
Since (D+Z)© is superparticular and in lowest terms, D = 1
Since the unit is indivisible, therefore no integral mean divides
(D+Z)©
Machinery
If J:K are in lowest terms and L}... are means such that J:Lp ...
L^K, then if J:K::N:O, then there are Mj... Mn such that N:Mf ...
Mj>:O (Euclid Elements 8.8; Pena notes die reliance here on the
Elements)
Conclusion
Since there are no means between (D+Z)©, then there are no means
between B:G
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
128
δ
Έαν διάστημα μη πολλαπλάσιον διί (Tvvrc&fj, το δλον ουτ€ πολλα
πλάσιον εσται ούτε έπιμόριον. έστω γαρ διάστημα μη πολλαπλάσιον0 τδ
βγ και γεγεντ/σβω ως δ γ προς τον β δ β πρδς τον δ. λέγω δτι δ δ του γ
5 ουτ( πολλαπλάσιος ovre έπιμόριός έστιν έστω γαρ πρώτον δ δ τοΰ γ
πολλαπλάσιος· ούκοΰν έμάθομαι δτι εάν διάστημα δις συντεθέν τδ δλον
ποιτ) πολλαπλάσιον, και αυτά πολλαπλάσιόν εστιν. εσται αρα δ β τοΰ γ
πολλαπλάσιος, ουκ ην δέ· αδύνατον αρα τδν δ τοϋ γ είναι πολλα
πλάσιον. άλλα μην ούδ’ έπιμόριον. έπιμορίου γαρ διαστήματος μέσος
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
129
4.
If a non-multiplc interval is twice combined, the whole will be neither
multiple nor superparticular. For let there be a non-multiplc interval, BG,
and let it have been that as G was to B so B was to D. I say that D is neither
i multiple nor a superparticular of G; for in the first place, let D be a multi
ple of G; accordingly, we learned that if an interval twice combined makes a
multiple whole, then this itself is multiple. Therefore, B will be a multiple of
G. But it was not; therefore it is impossible for D to be a multiple of G. Yet
it is truly not a supcrparticular.18 For no mean falls proportionately between
18Od breaks off with εττι-, although it appears the text originally
continued.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
130
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
131
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
132
e
’Εαρ διάστημα διί συντίθίν τό δλου μη ποιη πολλαπλάσιον, I Γ | β |
δ ούδ* αύτό ίσται πολλαπλάσιου. έστω yap διάστημα τό βγ και ycyt-
νησ&ω ώί ό γ πρόί τον β' ο β προς τον δ και μη Έστω ό δ τοΰ γ πολλα-
5 πλάσιΟΓ λ<χω δτι ούδί ό β τοϋ γ ίσται πολλαπλάσιοί* ci yap Έστιυ ό β
τοϋ γ πολλαπλάσιο?· ίσται και ό δ του y πολλαπλάσιοί* ούκ Έστι όΈ. ούκ
αρα ό β τοϋ γ Έσται πολλαπλάσιοί.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
133
5.
If an interval twice combined does not make a multiple whole, e.g., IG I
Β ID 21 then it will not be multiple. For let BG be an interval, and let it have
been as G was to B so B was to D. And let D be a non-multiple of G. I say
that B will not be a multiple of G; for if B is a multiple of G, then D will be
a multiple of G, which is not the case. Therefore, B is not a multiple of G.22
21 This diagram is similar to the one that appears in the text with Propo
sition 3 (124.2; see n. 13) and may indicate an earlier diagram or example
that is no longer preserved in the manuscripts. In simple terms: let B be a
mean between G and D.
^Analysis
η πρότασή 132.2-3
η Ικθίσι,ς 132.3-5
ό διορισμός 132.5
η κατασκευή 132.5-6
η άπόδ€^ΐί 132.5-6
τό συμπέρασμα 132.6-7
Exposition
G:B::B:D
- 3a(D/G = a) where a is an integer
Specification
- 3b(B/G = b) where b is an integer
Machinery
Proposition 1
Proof
Suppose 3b(B/G = b)
Since G:B::B:D, and from Proposition 1
3a(D/G = a), which contradicts the assumption in the exposition
Conclusion
- 3b(B/G = b)
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
134
1 Θεώρημα έκτων UpVn II 2 Τ deest FrVv I δύο om. za sup. Un. Mm2[?] II 3
ημιολίου ... επιτρίτου]επιμορίου καί ημιολίου Za (τρίτου sup. Un.) επι-
τρίτου μορίου καί ημιολίου VeVw I εκ (sec.) om. EcPcVm I επί τρίτου Mm
I τοϋ δζϊτή δζ Pb δζom. jh II 4 ήμιόλιοί ... δζom. CbPcPf I διπλάσιου
BbJanMg II 5 CbFrMnAnuNn(K quaes.JPcPfVc3 β za I τδν γ]τδν γλ
FrLdMm^muNnVc^VdiX rfe/.)Vj(X del.)No τδ γλ CbPf τδ γδ Pc τδν γδ Vp
I ίπόΐίπι Εο(-π€ΐ in marg.)FrNn* (corr. Nn2) I τοϋ δζοηχ. Ne II 6 ήμιόλιο?]
και ίκ τοϋ τριπλού καί επιτρίτου τδ τετραπλάσιον ομοίως δείκνυτατ καί
εκ τοϋ τετραπλού και επιτετάρτου τδ πενταπλοϋν και ad ομοίως. add.
Mmza del. Mm I β]βλ Mm3muvc (Vc3 praet. Ld λ del. VdVj) I ίστιυ FrOa
PbUpJanMg I του δί]τόυ δί Pd I ras. post πάλιν Vc I ίπαίκαι Bb ό μεν jh
om. vc II7 ό (pr.)]rj CbjhPcPfVc3 (in ras.) om. Bb I ίσται za I δ]δκ Mm3mu
δ< δκ vcfVc3 κ del. VdVj)Vv I δ<]<δ LdVdVj I τεταρτημόριον]τεταρτομό-
ριόν εστι Ee(ante corr.)Mm(-o- Mm3)Oa(an/e corrjmu (praet. UpVn)
τεταρτημόριων ίστι UpVn τεταρτημόρον Vc II 8 δίδκ Mm3muvc (Vc3 κ
del. LdVj) I β]βλ MnAnuvc (Vc3 praet. FrNn λ deL LdVdVj) II 8-9 δζ
ημισυ ... β τοϋ om. mu II 9 β (ρτ.)]βλ Mm3muvc (β in marg, λ in ras.
Vc3 λ del. LdVdVj) I δ]δκ Mm3vc (Vc3) I β]βλ MnPvc (Vc3 λ in ras. Nn) I
βγ]β ία I δ (scc.)]6k Mm3muvc(Vc3 κ del. LdVj) II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
135
6.23
The duple interval is put together from the two greatest superparticulars,
from the sesquialter and from the sesquitertian. For let BG be a sesquialter
of DZ, and let DZ be a sesquitertian of Θ. I say that BG is a duple2324 of Θ.
For I subtracted Z equal to Θ and G equal to DZ Accordingly, since BG is a
sesquialter of DZ,25 B therefore is a third part of BG, a half of DZ Again
since DZ is a sesquitertian of Θ, D is a fourth part of DZ and a third part of
Θ. Accordingly, since D is a fourth part of DZ and B a half of DZ, therefore
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
136
τοβ βγ ίκτον μέρος έστίν. ην & δ δ τοϋ θ τρίτον μέρος· δ αρα βγ τοϋ θ
διπλάσιά? ίστιν.
1 τοΰ (pr.) om. Bb I & om. Up I δ]δκ Mm3mu(«5 Vn)vc (Vc3 κ del.
LdVdVj) II 2 διπλάσιο? ίστί. za II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
137
Original from
Digitized by
UNIVERSITY OF MICHIGAN
138
1 *E deest Fr I τοΰ om. mu II2 ημώΚιος cerri za I ό α (ter.) om. za sup. tin.
Mm (Mm2?) I a (ter.) om. EcVmVp II 3 τδν β]τοϋ β Pf I τδν ημισνν]τδν
ημασυ BaEcOa(a/ire corr.)PbVm τδ ημισ Vn τδ ημών UpJanMg I δύο]/3 za
I a sup. Un. rub. Mm2Mu om. za I €ΐσι τοϊί β τρωϊ za II 4 y (ρτ.)]δ vc
τρίτον Mm*1 (τρίτον del. γ sup. Un Mm2) I γ (sec.)]i/y ante corr. VeVw I
τδν τρίτον mu (praet. UpVn) II 5 οί β (pr.)Joi δύο za I ίσοι «σι (ρτ.ψίσα/
Ισοι VeVw I κίσι (pr.)]tlaiv Pb I τ€τταρσι]τ€ταρσι CbEe(anre corr.)Mm
mu(praer. KaPb)Oa(anre corr. )PePfza τέταρσιυ Pb τ<σσαρσι Ka I yjay za
I Tpeis in ras. Nn I β ($£θ.)]δύο CbPePf II 6 δυσΙ]δύο codd. I τ€τταρσι]τ€-
ταρσι CbEe(ante corr.)Mmmu(praet. Ka) τεσσαρσι Kaza rirpacri PcPf I
apa δ α]ό apa δ a Vc(o [pr.] Vc3) ό apa a \c(praet. Vc)Mg I Taorfurow Bb
II7 apa om. Bb II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
139
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
140
C
'Εκ του διπλάσιου διαστήματος καί ημιολίου τριπλάσιον διάστημα
γίνεται. Έστω γαρ ό μέν α τοΟ β διπλάσιΟΓ ο δί β τοϋ γ ημιόλιος- λεχω
δτι ό α τοϋ γ ίστϊ τριπλάσιος- νηά γαρ ό α τοϋ β ίστι διπλάσιος, ό αρα a
5 ίσος ίστϊ δυσϊ τοΐς β· πάλιν ίπά ό β τοϋ γ ίστϊν ημιόλιος, αρα ό β εχα
τδν γ καί τδ ημισυ αύτοϋ. δύο αρα οι β ίσοι είσι τρισϊ τοΐς γ- δύο δί οι β
ίσοι άσϊ τφ α. και δ α αρα ίσος ίστϊ τρισϊ τοΐς γ τριπλάσιος αρα ίστϊν ό
ατοϋγ.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
141
7.
From a duple interval and a sesquialter, a triple interval comes about
For let A be a duple of B, and B a sesquialter of G. I say that A is a triple of
G, for since A is a duple of B, therefore A is equal to two B*s. Again, since
B is a sesquialter of G, therefore B holds G and a half of it Therefore, two
B’s are equal to three G’s; and two B’s arc equal to A. Therefore, A is equal
to three G’s. Therefore, A is a triple of G.29
^Analysis
η πρότασή 140.2-3
η extent 140.3
δδωρισμόί 140.3-4
η xararxevq 140.6
η amte&t 140.4—7
τδ σνμττίρασμα 140.7-8
Exposition
A = 2B
B = G+G/2
Specification
A = 3G
Machinery
2B = 3G from exposition
Proof
From exposition and machinery
Conclusion
A = 3G
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
142
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
143
8.30
If a sesquitertian interval is subtracted from a sesquialter interval, the
sesquioctave remainder is left behind. For let A be a sesquialter of B and G
a sesquitertian of B; I say that A is a sesquioctave of G. For since A is a ses
quialter of B, therefore A holds B and a half of it. Therefore, eight A’s31 are
equal to twelve B’s. Again, since G is a sesquitertian of B, G therefore holds
B and a third of it Therefore, nine G’s are equal to twelve B’s. Twelve B’s
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
144
οί β ίσοι ίίσίυ οκτώ τοίί α· οκτώ αρα οί α ίσοι άσϊυ ίυυία τοΐς γ ό αρα α
Ισος ίστϊ τφ γ και τφ όχδόφ αυτού, αρα ο α του γ ίστϊν ίπόγδοος.
1 οί β]οί δυο CbPcPf I τοΐς α]τοί$ δ ante corr. Mm (α sup. Un. Mm12?) I
άσϊν (sec.)]doi EcPcZa I <ίσι του ίυυία Ec I y]/3 za γ in ras. Mm2 I ό apa
α]ό a apa zaJanMg ό αρα ό a BaDaEeJIOaPdViVp I a (ter.) sup. Un. Mm2 II
2 τφ γ]τό γ FrNn I όγδωω Ecl όα apa UpVnMg I ίττόγύοος ίστί za II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
145
are equal to eight A’s. Therefore, eight A’s are equal to nine G’s. Therefore,
A is equal to G and an eighth of iL Therefore, A is a sesquioctave of G.32
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
146
e
τά ίττόγδοα διαστήματα μάζουά ίστι διαστήματα? iiw διπλάσιου,
έστω γαρ ας αριθμός ό α, και τοβ μίν α ίπόγδοος «στω ό β- τοϋ δβ β
ίιτογόοος ό γ τοϋ δί γ ίττογόοος ό Jr τοϋ δί δ εηόγόοος δ ε τοϋ € ί ττόγδοο?
5 b C τοϋ ζ ίττόγόοος ό η· λίγω ότι ό η τοϋ α μάζων ίστϊν η διπλάσιο?.
Έπ<ι ίμάθομίν έυραν ίπτα αριθμούς ίφίζής ίπογόόους άλληλωυ
ευρησθωσαν οΐαβγόίζη* καί γίνεται
ό μίν α κς μύρια βρυίτ
ό δί β κθμύρια βΊ)ιβ·
10 δ δέ γ Κγ μύρια ,αψοΓ
δ διό λζμύρια /γσμη·
ό δέ e μα μύρια βνό*
δ δί ζμζ μύρια βτ9β·
δόίηνγ μύρια /αυμα καί
Original from
Digitized by
UNIVERSITY OF MICHIGAN
147
9.33
Six sesquioctave intervals are greater than one duple interval. For let one
number be A, and let B be a sesquioctave of A, G a sesquioctave of B, D a
sesquioctave of G, E a sesquioctave of D, Z a sesquioctave of E, and H a
sesquioctave of Z; I say that H is greater than the duple of A. Since we
learned to find seven numbers consecutively sesquioctave to one another, let
A, B, G, D, E, Z, H be found; and
A becomes 262,14433 3435
B becomes 294,91235
G becomes 331,776
D becomes 373,248
E becomes 419,904
Z becomes 472,392
H becomes 531,441,36 and
33Za2 contains the scholion: υπέρ μεν διπλάσοχ τοϋ πρώτου εστϊ νβ
μύροι fioirq ύπερεχεΐ Bi αντοίν?] δ η [τό?] l/lp2'/· The scholion claims that
η exceeds the duple (i.e., 528,244) of the first (i.e., a) by 7153.
34Up writes out the seven numbers thus:
262,144—είκοσιεζ μύροι δοτχίλια εκατόν καί τεσσαράκοντα καί τεσσαρα
294,912—είκοσιεννεα μύροι, τετρακοτχίλοι, εννεακόσοι καί δώδεκα
331,ΊΊ6—τροτοωντατρία μύροι, χίλια επτακόσοι καί εβδομακονταδύο
(should be not -δύο)
373,248—τεσσαρακονταετία μύροι τροτχίλοι διακόσια καί τεσσαρακον-
ταοκτώ
419,904—τεσσαρακονταεν μύροι εννεακοτχίλοι, εννεακόσοι καί τεσσαρα
472392—τεσσαρακονταεπτα μύροι, δοτχίλοι τροικόσοι καί εννεακκόντα
δύο
531,441—πεντεκοντατρία μύροι, χίλια τετρακόσοι καί τεσσαρακονταεν.
35The sampi (D), indicating 900, was regularly copied as a tau in the
Renaissance codices, both here and at line 12. Earlier symbols included /I
and T. Unfortunately, τ indicates 300, and this caused considerable confu
sion in the Renaissance codices. Only Vc1, Mm1, Za, and Vv have a sym
bol other than tau for 900 at line 9 ( Λ). Mm1 and Vv repeat this symbol at
line 12, whereas Za has a “broken cross” (/k). Vw and Ve have ψ instead
of ε for 900 at line 12. Vc has T for 900 at line 9 but otherwise does not dis
tinguish between the symbols for 900 and 300.
36The manuscripts that preserve the arithmetic Division contain numer
ous corrections of the large numbers involved in lines 8—14 of Proposition 9.
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
148
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
149
37F. 396 in Vw was tom before the scribe executed the treatise. It reads: pel ζου.
38This primitive diagram occurs in the text in Mm1 and za. In all other
manuscripts that contain diagrams, line 148.2 takes the form of a diagram. It
is omitted in manuscripts that do not contain diagrams. All of the diagrams
that accompany this proposition give at best a crude picture of this essen
tially numerical demonstration.
Analysis
η πρότασίΛ 146.2
η 146.3-5
ό διορισμός 146.5
η κατασκευή 146.6-7
η άττόδ^ΐί 146.7-14
το συμπέρασμα 148.1
Exposition
B:A::9:8
G:B::9:8
D:G::9:8
E:D::9:8
Z:E::9:8
H:Z::9:8
Specification
H>2A
Machinery
If b:a is in lowest terms where a and b are integers, then for any inte
ger n, there exists the series x^x^: .. .Xp where x is an integer and
XniXn.fzbia (Euclid Elements 8.2; Pena notes this dependency in his
translation)
Proof
Let n=7, b=9, a=8
xn_6 = a<nl) =262,144 = A
Xn-5 = b(^)a(n-2) =294,912 = B
= b<n*5>a<n-3) =331,776 = G
Xn.3 = b<n4>a(n^ =373,248 = D
= bin3>a(n·5) =419,904 = E
x^ = b<n-2)a(^) =472,392 = Z
= b<n-i> =531,441 = H
Conclusion
H>2A
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
150
ι
τό δια πασών διάστημά ίστι πολλαπλάσιον. ίστω yap νητη μίυ
ύπ^ρβολαίων ό α* μ^ση δ< ό β· προσλαμβανόμενος δε ό γ· τό αρα ay
διάστημα διί δια πασών δν, ίστι σύμφωνον, ήτοι ουν ίπιμόριον ίστιν η
5 πολλαπλάσιον* ίπιμόριον μίν ow ούκ ίστιν* ίπιμορίου γαρ διαστήματος
μίσος ovdcir ανάλογον ίμπίπτα* πολλαπλάσιον αρα ίστιν* ίπ€ΐ ουν δύο
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
151
10.
The diapason39 interval is multiple. For let the nete hyperbolaion be A,
the mese be B, and the proslambanomenos G. Therefore, the interval AG,
being a double diapason, is consonant40 Then it is either superparticular or
multiple. It is not superparticular, for no mean falls proportionately between
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
152
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
153
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
154
ΙΛ
Τό δια τεσσάρων διάστημα και τό δια π«ΗΈ inirepov ίπιμόριόν ίστι·
€στω yap νητη μίν συνημμένων ό α* μίση & ο β· υπάτη δ< μισών ό γ τό
άρα αχ διάστημα διτ δια Τ€σσάρων δν ίστι διαφωνον ούκ αρα ίστι πολ-
5 λαπλασιον* ίπό ουν δύο διαστήματα ισα τα αβ βγ συντίθέντα το δλον
μη ποιτ) πολλαπλάσιον, ούδί αρα τό αβ ίστι πολλαπλάσιον* και ίστι
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
155
ll43
With respect to the diatessaron interval and the diapente, each is super
particular. For let the nete synemmenon be A, the mese be B, and the hypate
meson G; therefore the interval AG, being a double diatessaron, is disso
nant Therefore, it is not a multiple.44 Since two equal intervals, AB, BG,
joined together do not make a multiple whole, neither therefore is AB a mul
tiple;45 but it is consonant and therefore superparticular. The same proof
Original from
Digitized by
UNIVERSITY OF MICHIGAN
156
σύμφωνον* εττιμόριον αρα. η αυτή bi απόόειζις και ίπι τοϋ δια πεντε. τό
δια πασών διάστημά ίστι διπλάσιαν. Άείζαμεν γαρ αυτό πολλαπλάσιον*
ούκοΰν ήτοι ίιπλίσιον ίστϊν η μεϊζον η διπλάσιαν, άλλ’ ίπό έδιίέαμιν
τό διπλάσιαν διάστημα ίκ δύο των μεγίστων επιμορίων συγκείμενον
5 ώστε εΐ εσται. τό δια πασών μεϊζον όιπλασίου ου σνγκείσεται εκ δύο
μόνων επιμορίων άλλ' εκ πλκόνων. συγκειται bi εκ δύο σύμφωνων
διαστημάτων εκ τε τοϋ δια πεντε καί δια τεσσάρων ούκ αρα εσται τό δια
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
157
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
158
πασών μάζον διπλάσιου· διπλάσιου αρα. άλλ’ επειδή τδ δια πασών ίστι
διπλάσιον τό δε διπλάσιον εκ τών μεγίστων έπψορίωυ δύο συνεστηκε·
και τό δια πασών αρα έζ ημωλίου και επιτρίτου σννέστηκν ταΰτα γαρ
μέγιστα. σννέστηκ* γαρ έκ τοΰ δια πεντε και εκ τοΰ δια τεσσάρων, διττών
5 επιμορίων τό μεν άρα δια πεντε επειδή μίίζόν εστιν ημιόλιον αν ειψ τό
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
159
tessaron. Therefore, the diapason will not be greater than duple; therefore it
is duple.47 Yet since the diapason is duple and the duple is joined together
from the two greatest superparticulars, the diapason is therefore joined
together from a sesquialter and a sesquitertian, for these are greatest. For the
diapason is joined together from the diapente and from the diatessaron,
which are superparticular, therefore, the diapente, since it is the greater,
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
160
δί δια τεσσάρων ίπίτρπον. φανερόν 6η δτι και τό δια πεντε και τό δια
πασών τριπλάσιόν εστατ εδείζαμεν γαρ δτι ck διπλάσιου διαστήματος
καί ημιολίαν τριπλάσιον διάστημα γίνεται· ώστ< καί τό δια πασών και τό
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
161
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
162
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
163
and the diapente are triple.49 The double diapason is quadruple.50 Therefore,
it was shown in what ratios each of the consonances51 holds the bounding52
notes to one another. It remains necessary to go over the interval of a whole
tone, that it is sesquioctave. For we learned that if a sesquitertian interval is
subtracted from a diapente interval, the sesquioctave remainder is left
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
164
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
165
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
166
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
167
a duple interval; therefore the diapason is less than six tones.55 The diatessa
ron is less than two tones and a half tone, and the diapente is less than three
tones and a half tone. For let the nete diezeugmenon be B, the paramese be
G, the mese be D, and the hypate meson Z. Accordingly, GD is a tone inter
val. BZ» a diapason, is less than six tones. The remainders, therefore—BG
and ZD being equal—are less than five tones. Thus, there are less than two
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
168
1 ίν τφ om. UpVn I ίΚαττόν ίστι δύο Pb I δ]οΰ ante corr. CbPf accent. Vc3
δν ΜπΛηυ II 2-170.1 τεσσάρων ... πίντ* om. Ee II 2 δ<)τ€ UpVn I ojoy
Mm(-v quaes. Mrr^Jmu accent. Vc3 II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
169
tones and a half tone in BG, which is a diatessaron. BD, which is a diapente,
is less than three tones and a half tone.56 The tone will not be divided into
Original from
Digitized by
UNIVERSITY OF MICHIGAN
170
1 ό τόνος]θ&ί>ρημα έκτον και δέκατου ό τόνος UpVn 16. ό τόνος Jan ι$’. ό
τόνος Mg I ισα ουτ* ας κλάω JanMg II 2 ων]όν Mm (corr. Mm1 2) αν Vc αν
BbCb(sup. /in.)NnPf I €πιμορίου]ί7τιμορίων JIPd I μέσοι]ό μέσης CbPf
μέσος Bb μέ* LdVd(de/.)Vj μέσοι in marg. Vc3 om. MuVv I ovrc πλάους
om. Bb II
Original from
Digitized by Goc-gle UNIVERSITY OF MICHIGAN
171
two equal tones nor into more.57 For it was shown as being superparticular,
and neither many means nor one falls proportionately between a superpartic
ular. Therefore, the tone will not be divided into equal intervals.58
57Certainly ίσα ovre ds ttAcud would be more precise here (see Musici
scriptores graeci, p. 161.8): “the tone will not be divided into two equal
intervals nor into more.” See also 170.3. The codices are unanimous, how
ever, in giving ίσου? and irXdovs.
58By “equals” or “equal intervals,” the treatise equates these intervals
with the integral, geometric means of Proposition 3. In a geometric sense,
the interval of a tone can be divided precisely in half, but that is not the
intention of this radically numerical, Pythagorean document
Analysis (lines 170.1-3)
η πρότασή 170.1
η έκθισιε does not exist,
ό διορισμός does not exist
η κατασκευή 170.1-3
η απόΐκιζις does not exist
τό συμπέρασμα 170.3
Machinery
Proposition 3
Proposition 11 (162.3-164.2)
Conclusion
follows directly from the machinery
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
172
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
173
[Enharmonic Passage]59
The paranetai and the lichanoi will be found60 through consonances in
this way. For let the mese be B. Let a diatessaron be tightened to G, and
from G let a diapente be loosened to D. Therefore, BD is a tone. Again,
from D let a diatessaron be tightened to E, and from E let a diapente be loos
ened to Z. Therefore, ZD is a tone. Therefore, ZB is a ditone. The lichanos
therefore is Z; in like manner, the paranetai will also be found.61 The parhy-
patai and the tritai do not divide the pyknon into equal intervals. For let the
mese be B, the lichanos be G, and the hypate D; let a diapente be loosened
from B to Z. Therefore, ZD is a tone. And from Z let a diatessaron be tight-
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
174
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
175
62That GE is a tone requires some calculation. See n. 65. From this point
on, Jan (Musici scriptores graeci. pp. 162-63) and Menge (“Scripta musica,”
in Euclidis Opera omnia, vol. 8, Phaenomena et scripta musica, ed. Henri-
cus Menge [Leipzig: B. G. Teubner, 1916], pp. 176-78) go astray in their
editions of this passage. See Thomas J. Mathiesen, “An Annotated Transla
tion of Euclid’s Division of a Monochord.” Journal of Music Theory 19
(1975): 257, nn. 66-69.
^Literally, “be set forth.” προσκώσθω would be more regular than
Ίτροκιίσθω. although either will do here. See Bb, Up, Vn, and Musici scrip-
tores graeci. p. 163.7.
64Cb stops here, at the bottom of f. 11 v, but continues with the treatise at
f. 16r. F. 12r begins with section 4 of the so-called Neapolitan excerpts
(Musici scriptores graeci. p. 412.17): Μουσική ίση (ρυθμόν) και. Μου
σική ίση appears at the top of the folio, perhaps in a second hand. F. 15v
contains the Nicomachean excerpt 2 (Musici scriptores graeci. p. 271.8-15):
τον μκδ ... τονωϊα ίιαστηματα. Pf, which is copied from Cb, stops one
word earlier but does not resume.
Original from
Digitized by Go< >gle UNIVERSITY OF MICHIGAN
176
Original from
Digitized by Go< >gle UNIVERSITY OF MICHIGAN
177
DG, which is from hypate to lichanos. Therefore, the parhypate does not
divide the pyknon into equal intervals. In like manner, neither does the
trite.65
Original from
Digitized by LnOoQie
UNIVERSITY OF MICHIGAN
178
1 Ίδν]θίώρημα ένατον καχ δίκατον τον UpVn 19. τον Jan ιθ'. τον Mg I
κατά γράψαι Cb II 2 διηρησθω]διαιράσθω KaVeVw -ηρη- in ras. Nn II 3
τίσσαρα]δ UpVeVw I κατά om. PbUpVn I βα]αβ βα BbFrNnQ’n rar.)Vc3
(βα Vc^Vo αβ CbLdVdVj I βαρύτατος Vw21 ών]ων (corr. Mm3) II
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
179
[Canon]66
To mark the canon according to the so-called immutable system.
Let there be a length of the canon, which is also AB of the string 67 and
let it be divided into four equal parts by G, D, E. Therefore, BA will be the
B = nete
G = paranete
D = paramese
Comment
From the point of adding the common interval DG (line 174.2), the
proof is clumsy and may indicate a confused transmission. The object is to
show that the parhypate does not divide the enharmonic pyknon equally.
The Pythagorean restriction to integers is assumed here, although the termi
nology does not indicate so (ίσα: equal [intervals]). Certainly the enhar
monic pyknon could be divided into two equal intervals, although they
would be irrational. The same assumption is made here as was made in
Proposition 11 (170.1-3), i.e., if an integral mean docs not exist between an
interval, it cannot be divided in half. Given the musical restriction to inte
gers by the Pythagoreans, this is a logical but unstated condition. In a strict
sense, the addition of the common interval DG is unnecessary, although the
intent may have been to compose the diatessaron from the two equal inter
vals of a tone, EG and DZ, and the common interval DG. Thus, the observa
tion that ZG = DE is parenthetical at best. The remainder of the proof argues
inexplicitly that if the parhypate divides the pyknon into equal intervals, this
division must take place somewhere in the common interval DG. This is so
by definition from the exposition. Furthermore, since EG equals DZ, the
common interval GD is the central region of the diatessaron EZ, and thus a
division of GD would entail a division of the diatessaron. Since this is
impossible (assuming the obvious, i.e., EG + GD > DZ and GD + DZ >
EG), the parhypate does not divide the pyknon into equal intervals.
^Following Mu, several—but not the most authoritative—manuscripts
have a break in the text before τον. Mm contains a mark in the margin
before τον. Vw and Ve pick up the text again at τον, and Vw2 places a
mark here. The title “Canon” does not actually occur in the Greek.
67In the margin, Up correlates divisions of the canon with notes of the
system and also counts the cuts of the canon, thus.
line word in margin
178.2 η πρώτη τομή
180.2 KOI πρώτος προσλαμβανόμενος
180.3 o apa δεύτερος υπάτη ίιάτονος
182.1 καί τρίτος μέση
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
180
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
181
Original from
Digitized by Go< >gle UNIVERSITY OF MICHIGAN
182
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
183
Original from
Digitized by
UNIVERSITY OF MICHIGAN
184
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
185
than BE. Thus, NB will be the trite hypcrbolaion, and MB the diatonic
hyperbolaion. I took a third part of NB, and I established ΝΞ equal to one of
these. Thus, SB is a sesquitertian of NB and sounds consonantly a diatessa
ron lower. SB becomes the trite diezeugmenon. Again, taking a half part of
SB, I established SO equal to one of these. Thus, OB to SB sounds a dia
pente consonance. Therefore, OB will be the parhypate meson. And I estab
lished OP equal to SO, thus PB becomes the parhypate hypaton. I surely
took GR, a fourth part of BG, thus RB becomes the diatonic meson.73
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
186
Divide GB in two at Z
Divide DB into three parts and subtract one, leaving HB
Diagram!
a_______ I______ I___ LJU_______I_______
O D Ζ H E B
Original from
Digitized by Google UNIVERSITY OF MICHIGAN
187
LPGOORDKE E B
- B
~ B nete hyperbolaion
~ N trite hyperbolaion
“ H nete diezeugmenon
- K paramese
- D mese
- O parhypste meson
- Θ hypate meson
- P parhypate hypaton
~ L hypate hypaton
A proslambanomenos
Original from
Digitized by Google UNIVERSITY OF MICHIGAN