1
20
1

https://s3.useast1.amazonaws.com/sjcdigitalarchives/original/45709523ddccdeb799588a8848abb06e.pdf
35c2cecd076ac303d83387b733342e0d
PDF Text
Text
The St. John’s Review
Volume XLVIII, number one (2004)
Editor
Pamela Kraus
Editorial Board
Eva T. H. Brann
Frank Hunt
Joe Sachs
John Van Doren
Robert B. Williamson
Elliott Zuckerman
Subscriptions and Editorial Assistant
Audra Price
The St. John’s Review is published by the Ofﬁce of the Dean,
St. John’s College, Annapolis: Christopher B. Nelson,
President; Harvey Flaumenhaft, Dean. For those not on the
distribution list, subscriptions are $10.00 for one year.
Unsolicited essays, reviews, and reasoned letters are welcome.
Address correspondence to the Review, St. John’s College,
P Box 2800, Annapolis, MD 214042800. Back issues are
.O.
available, at $5.00 per issue, from the St. John’s College
Bookstore.
©2004 St. John’s College. All rights reserved; reproduction
in whole or in part without permission is prohibited.
ISSN 02774720
Desktop Publishing and Printing
The St. John’s Public Relations Ofﬁce and the St. John’s College Print Shop
�2
THE ST. JOHN’S REVIEW
3
Contents
Classical Mathematics and Its Transformation
St. John’s College, Annapolis
June 56, 2004
Foreword.......................................................................5
Harvey Flaumenhaft
Why We Won’t Let You Speak Of the Square
Root of Two..................................................................7
Harvey Flaumenhaft
The Husserlian Context of Klein’s
Mathematical Work.....................................................43
Burt C. Hopkins
Words, Diagrams, and Symbols: Greek and
Modern Mathematics or “On the Need to Rewrite
the History Of Greek Mathematics” Revisited............71
Sabetai Unguru
A Note on the Opposite Sections and Conjugate
Sections In Apollonius of Perga’s Conica.....................91
Michael N. Fried
Viète on the Solution of Equations and the
Construction Of Problems.........................................115
Richard Ferrier
�4
THE ST. JOHN’S REVIEW
5
Foreword
Harvey Flaumenhaft
On the weekend of June 56, 2004, the Annapolis campus of
St. John’s College hosted a conference on the topic of classical mathematics and its transformation. This is a pivotal topic
in the program of instruction here at St. John’s, where classes
consist in the discussion of great books ancient and modern,
and where approximately half of the entirely prescribed curriculum may be classiﬁed as either mathematics or heavily
mathematical natural science.
The topic was made truly pivotal at St. John’s by Jacob
Klein, who joined the faculty in the second year of the New
Program (193839) and became Dean eleven years later. Just
a few years before arriving at St. John’s, Klein had published
Die Griechische Logistik und die Entstehung der Algebra (later
translated by another Dean—Eva Brann—as Greek
Mathematical Thought and the Origin of Algebra). In a letter
written just a few weeks after arriving at the College, Klein
said, “It is almost unbelievable to me that all the things that
occupied me for years, that is, the whole theme of my work,
are realized here. The people don’t do quite right, very much
is superﬁcial and they are not quite right about certain fundamentals. But it….is clear that I am in the right spot.”
Klein’s presence, and his book because of his presence, had a
deepening and correcting effect on the life of this college, and
through this college, on intellectual life far beyond our halls.
The effects of his work pervade the conference papers, which
are printed as this issue of The St. John’s Review.
The conference took place through the generosity of the
Andrew W Mellon Foundation, as the culminating event
.
under a large grant supporting faculty study groups led by
me, over several years, on the topic of the conference. The
Harvey Flaumenhaft is Dean at the Annapolis Campus of St. John’s College.
�6
THE ST. JOHN’S REVIEW
grant also supported groups on the same topic on the Santa
Fe campus, with participation by faculty members from
Thomas Aquinas College. Grateful acknowledgment is due to
Sus3an Borden, the college’s Manager of Foundation
Relations in Annapolis, who did a great deal of work, under
very severe time constraints, to set up the conference.
7
Why We Won’t Let You Speak
of the Square Root of Two
Harvey Flaumenhaft
Part One
Progress and Preservation in Science
We often take for granted the terms, the methods, and the
premises that prevail in our own time and place. We take for
granted, as the startingpoints for our own thinking, the outcomes of a process of thinking by our predecessors.
What happens is something like this: Questions are asked,
and answers are given. These answers in turn provoke new
questions, with their own answers. The new questions are
built from the answers that were given to the old questions,
but the old questions are now no longer asked. Foundations
get covered over by what is built upon them.
Progress can thus lead to a kind of forgetfulness, making
us less thoughtful in some ways than the people whom we go
beyond. We can become more thoughtful, though, by attending to the thinking that is out of sight but still at work in the
achievements it has generated. To be thoughtful human
beings—to be thoughtful about what it is that makes us
Harvey Flaumenhaft is Dean at the Annapolis campus of St. John’s College. This
paper was delivered in the form of a Fridaynight formal lecture at St. John’s
College in Annapolis on 27 August 1999.
The ﬁrst part of it is adapted from the “Series Editor’s Preface” in the
volumes of Masterworks of Discovery: Guided Studies of Great Texts in Science
(New Brunswick, NJ: Rutgers University Press, 19931997).
The last part is adapted from the “Introductory Note on Apollonius” in
Apollonius of Perga, Conics: Books IIII, new revised edition, Dana Densmore,
ed. (Santa Fe, NM: Green Lion Press, 1998).
The three parts placed between them are adapted from the manuscript
How Much and How Many: The Euclidean Foundation for Comparisons of Size
in Classical Geometry (forthcoming from Green Lion Press), which is part of the
larger manuscript Insights and Manipulations: Classical Geometry and Its
Transformation—A Guidebook: Volume I, Starting up with Apollonius; Volume
II, From Apollonius to Descartes (also forthcoming).
�8
THE ST. JOHN’S REVIEW
human—we need to read the record of the thinking that has
shaped the world around us, and still shapes our minds.
Scientiﬁc thinking is a fundamental part of that record,
but it is a part that is read even less than the rest, largely. That
is often held to be because of the opinion that books in science, unlike those in the humanities, simply become outdated: in science, the past is held to be passé.
Now science is indeed progressive, and progress is a good
thing; but so is preservation. Progress even requires preservation: unless there is keeping, our getting is little but losing—
and keeping takes plenty of work.
Precisely if science is essentially progressive, we can truly
understand it only by seeing its progress as progress. This
means that our minds must move through its progressive
stages. We ourselves must think through the process of
thought that has given us what we would otherwise thoughtlessly accept as given. By refusing to be the passive recipients
of readymade presuppositions and approaches, we can avoid
becoming their prisoners. Only by actively taking part in discovery—only by engaging in rediscovery ourselves—can we
avoid both blind reaction against the scientiﬁc enterprise and
blind submission to it.
We and our world are products of a process of thinking,
and truly thoughtful thinking is peculiar: it cannot simply
outgrow the thinking it grows out of. When we utter deceptively simple phrases that in fact are the outcome of a complex development of thought—phrases like “the square root
of two”—we may work wonders as we use them in building
vast and intricate structures from the labors of millions of
people, but we do not truly know what we are doing unless
we at some time ask the questions which the words employed
so casually now were once an attempt to answer.
The education of a human being requires learning about
the process by which the human race gets its education, and
there is no better way to do that than to read the writings of
those masterstudents who have been the masterteachers.
FLAUMENHAFT
9
Part Two
Manyness: The Classical Notion of Number
The ﬁrst great teachers of the West, and subsequently the rest
of the world, were the classical teachers who wrote in Greek
some twoandahalf millennia ago. In their language the
word for things that are learned or learnable is mathêmata,
and the art that deals with the mathêmata is mathêmatikê—
the English for which is “mathematics.”
In mathematics, the ﬁrst and fundamental classic work is
the Elements of Euclid. From its Seventh Book we learn that
a number is a number of things of some kind. When there are
more than one of something, their number is the “multitude”
of them. It tells how many of them there are. Suppose, for
example, that in a ﬁeld there are seven cows, four goats, and
one dog. If we count cows, then the cow is our unit, the cow
being that according to which any being in the ﬁeld will count
as one item; and there is a multitude of seven such units in
the ﬁeld. If we count animals, however, then the animal is our
unit; and now there is a multitude of twelve units. Of cows,
there are seven; but of animals, there are twelve.
Of dogs as dogs, there is no reason to make a count (as
distinguished from including them in the count of animals).
There is no reason to count dogs, for the ﬁeld does not contain dogs. It contains only one dog, and one dog alone does
not constitute any multitude of dogs. If what we were thinking was “There is only one dog in the ﬁeld,” we might say
“There is one dog in the ﬁeld”; but if we were not thinking
of that single dog in relation to some possible multitude of
dogs, we would simply say “There is a dog in the ﬁeld.” Of
pigs, there is no reason to make a count, for the ﬁeld does not
contain a single pig, let alone a multitude of pigs.
If there had earlier been six horses in the ﬁeld, which
were then lent out, the ﬁeld might now be said to lack six
horses; but while six is the multitude of horses that are lacking from the ﬁeld, there is not any multitude of horses in the
ﬁeld.
�10
THE ST. JOHN’S REVIEW
If we chop up one of the cows, arranging to divide its
remains into four equal pieces, and we take three of them,
then we have not taken any multitude of cows. We have taken
merely three pieces that are each a fourth part in the equal
division of what is a heap of the makings for beef stew.
And so when we have several items, each of which counts
as one of what we are counting, then what we obtain by
counting is a number. The numbers, in order, are these: two
units, three units, four units, and so on. The numbers, we
might even say, are these: a duo, a trio, a quartet, and so on.
To us, nowadays, that seems strange: a lot seems to be lacking.
“One” does not name a multitude of units: although a
unit can be combined with, or be compared with, any number of units, a unit is not itself a number—so “one” does not
name a number. “Zero” also does not name a number, for
there is no multitude of units when there are no units for a
multitude to be composed of. Neither does “negativesix”
name any multitude, although “six” does; “six” names a number, but there is no number named “negativesix”—and hence
there is no number named “positivesix” either.
Although “three” names a number of units, and so does
“four,” “threefourths” does not name a number of units—
for if the unit is broken up, then, being no longer unitary, it
ceases to be a unit. The Latin word for breaking is “fraction.”
A fraction is not a number. (Of course we can, if we wish,
take a new unit—say, a fourth part of a cow’s equally divided
carcass—and then take three of these new units.)
As for “the square root of two,” it is not a number;
indeed, as we shall soon see, it is not even a fraction. And neither is “pi” a number.
To say it again, the numbers are: two units, three units,
four units, and so on—taking as many units as we please.
That is what Euclid tells us; and, after telling us, Euclid goes
on to speak of the relations among different numbers.
We learn, for example, that because ﬁfteen “measures”
sixty—that is to say, because a multitude of ﬁfteen units taken
FLAUMENHAFT
11
four times is equal to a multitude of sixty units—ﬁfteen is
called “part” of sixty, and sixty is called a “multiple” of ﬁfteen.
But ﬁfteen is not “part” of forty, because you cannot get
a multitude of forty units by taking a multitude of ﬁfteen
units some number of times. That is to say: the multitude of
ﬁfteen units does not “measure” forty units. Both ﬁfteen units
and forty units are, however, measured by ﬁve units; and ﬁve
units, which is the eighth part of forty units (the part obtained
by dividing forty units into eight equal parts), is also the third
part of ﬁfteen units, so ﬁfteen units is three of the eighth parts
of forty units. This is to say that ﬁfteen units, which is not a
“part” of forty units, is “parts” of it. Indeed, a smaller number, if it is not part of a larger number, must be parts of it,
since both numbers must—just because they are numbers—be
measured by the unit.
After that, Euclid goes on to tell us about sorts of numbers, and the ways in which they are related, numbers such as
those that are “even,” or “odd,” or “eventimeseven,” or
“eventimesodd,” or “oddtimesodd,” or “prime,” and so
on.
Numbers are thus of interest not only with respect to
their relations of size, but also with respect to their relations
of kind. Even and odd are distinguished by divisibility into
two equal parts (or by odd’s differing from even by a unit),
and the more complex terms here are distinguished with
respect to “measuring by a number according to another
number”—which nowadays would be stated thus: “dividing
by an integer so as to yield another integer as quotient without a remainder.”
Euclid thus deﬁnes some kinds of numbers by employing
notions such as being greater than or equal to or having a certain difference in size. However, he also sorts numbers into
kinds by referring to shape. What might lead to doing that?
We might use a dot to represent a unit, and then represent a number by dots in a line. For example, four would be
represented as it is at the top of Chart 1. We might then rep
�12
THE ST. JOHN’S REVIEW
resent in a shape the number produced by taking a number
some number of times. For example, the number produced
when four is taken three times would be represented as it is
in the middle of the chart, where we see that the numbers
four and three are “sides” of the “plane” number twelve; and
similarly, six and four are “sides” of the “plane” number
twentyfour. But the number twentyfour can also be represented as a solid, as it is at the bottom of the chart: the numbers four and three and two are “sides” of the “solid” number twentyfour.
So Euclid speaks of numbers that are “sides,” “plane,”
and “solid,” and also of the multiplication of a number by a
number, and then of the sorts of numerical products of such
multiplication—such as the numbers called “square” and
“cube,” whose factors are called “sides.”
But he does not represent numbers by dots in lines. To use
dots would force us to pick this or that number. To signify not
this or that number, but rather any number at all, it is more
convenient to use bare lines, without indicating how many
dots are carried on each line, although it might be confusing
that we then also have to represent by a line the unit.
FLAUMENHAFT
13
On the other hand, since the unit does have a ratio to any
number, as does any number to any other number, and since
the product of multiplying any number by any other number
is also a number, it might make everything more convenient
if we represent as lines not just the numbers that are “sides,”
as well as any “side” that is a unit, but also all the ﬁgurate
numbers that we can produce. After all, though the sort of
number called a “square” number may be somehow different
from the sort of number called a “cube,” it is nonetheless true
that any “square” number has a ratio to any “cube” number
(since any number has a ratio to any other number), whereas
a square cannot have a ratio to a cube (since they are not magnitudes of the same kind). If we go along with speaking in this
way, however, we must take care to keep in mind that when
it is numbers that are called “sides” or “squares” or “cubes”
we are not engaged in speech about ﬁgures; we are using ﬁgures of speech. Several of the books of Euclid’s Elements deal
with what would nowadays be called “number theory” rather
than “geometry.”
The numbers in the classical sense—the multitudes two,
three, four, and so on—tell us the result of a count, and
nowadays we often speak as if those numbers, the countingnumbers, are merely some of the items contained in an
expanded system that we call “the real numbers.”
Whenever we put numbers of things together we get
some number of things, and whenever we take a number of
things a number of times we get some number of things; but
only sometimes can we take a given number of things away
from a given number of things, and even when we can, only
sometimes can we get some number of things by doing so. For
example, we cannot take seven things away from ﬁve things;
and although we can take six things away from seven things,
we will not have a number of things left if we do, but only a
single thing. It is also only sometimes that we can divide a
given number of things into a given number of equal parts. A
multitude of ten things can be divided into ﬁve equal parts,
but a multitude of eight things cannot. In dividing a many
�14
THE ST. JOHN’S REVIEW
ness, there is constraint from which we are free in dividing a
muchness.
We often measure a muchness to say just how much of it
there is compared with some other muchness: having chosen
some unit of muchness, we count how many times such a unit
must be taken in order to be equal to the muchness that we
are measuring. A line, for example, is a kind of a muchness:
there is more line in a longer line than in a shorter line. If two
lines are not of equal size, one of them is greater than the
other. Lines are magnitudes. If we measure them, we obtain
multitudes that tell their sizes—their lengths. But magnitudes
are not multitudes. Whereas the size of a magnitude has to do
with how much of it there is, the size of a multitude has to do
with how many units it is composed of.
Not only magnitudes, but multitudes as well are often
measured: we measure these by using other multitudes to
measure them. Thus, we say that twelve is six taken two
times, or is twice six. We say, moreover, that eight is twice the
third part of twelve, or—to abbreviate—that eight is twothirds of twelve. “Twothirds” does not name a multitude
(unless we say that we are merely treating a third part of
twelve as a new unit) but “twothirds” is nonetheless a
numerical expression. If we therefore begin to treat it like the
name of a number, we are on the road to devising a system of
fractions. “Twothirds” of something is less than one such
something, but it is not fewer. Twothirds are fewer than
threethirds—but to say this, we must have broken the unit
into three new and smaller units. The numerator of a fraction
is a number or else it is “one.” The denominator is also a
number: it is the number of parts that result from a division
(although we will allow a denominator as well as a numerator to be “one.”) With a fraction, we divide into equal parts
and then we count them. A fraction is not itself a number (a
multitude), but it may be the counterpart of a number.
“Twelvesixths” is not the number “two”; but it is a counterpart, among the fractions, of the item “two” among the numbers. Another counterpart of the number “two” is the fraction
FLAUMENHAFT
15
“eightfourths.” Those counterparts are really the same counterpart differently expressed, since they are both equal to the
fraction whose numerator is the number “two” and whose
denominator is “one.”
As with division, so with subtraction. In taking things
away, we may get a numerical expression that we can treat
like a number in certain respects. When we have twelve
horses and eight of them are taken away, there is a remainder
of four horses. But if eleven are taken away, then there are
not horses left; only a single horse remains. And if twelve
horses are taken away, then not even a single horse is left:
none remain. But twelve horses can be compared with a single horse: the twelve are to the one as twentyfour are to two.
“Twentyfour” and “two” are numbers, and “twelve” is also a
number—so “one” is like a number in being comparable with
a number as a number is. Moreover, “one” can be added to a
number as a number can be, and “one” is sometimes what we
reply when asked how many of some kind of thing there are.
“One” is therefore a numerical word even if it is not a number. But in that case, so is “none”: if twelve horses are taken
away, leaving not a single horse, and we are asked how many
remain, we can then say “none.” Now suppose that Farmer
Brown owes fourteen horses to Farmer Gray, but has in his
possession only ten horses. Farmer Gray takes away the ten.
How many horses does Farmer Brown now own? None. But
he might be said in some sense to own even less than none,
for he still owes four. If we did say that, then we would have
to say that he owns four less than no horses, or that fourteen
less than ten is four less than none. But then we would be
counting, not horsesowned, but horseseitherownedorowed. We would be letting the payment of fourhorsesowed
count as wealth equal to nohorseseitherowedorowned.
Again, as with the fractions, we have numerical expressions
that we are beginning to treat like numbers. We are on the
road to devising a system of socalled “rational numbers,”
which includes “negative” items as well as such nonnegative
items as “zero” and “one” in addition to fractions (which
�16
THE ST. JOHN’S REVIEW
include the counterparts of those multitudes that were ﬁrst
called “numbers”—namely, two, three, four, and so on).
We must say “counterparts” because the multitude
“three,” for example, while it corresponds to, is not the same
as “the positive fraction ninethirds” or “the positive integer
three.” We have not brought about an expansion of a number
system by merely introducing additional new items alongside
old ones; rather, we have made a new system which contains
some items that correspond to the items in the old system, but
that differ from the old items in being items of the very same
new sort as are the new items that do not have correspondents in the old system. For example: although the number
“three” is a multitude, “positive three” is not a multitude; it
is an item deﬁned by its place in a system where, among other
things, “positive six” takes the place belonging to the item
that is the outcome of such operations as multiplying “negative two” by “negative three.”
We have just taken a look at part of the road that leads
from numbers in the classical sense to what we nowadays are
used to thinking of as numbers. We have looked at some steps
on the road to the socalled “rational numbers,” the socalled
numbers that have something to do with ratios. But the system of what we nowadays are used to thinking of as numbers
is the system of “real numbers,” only some of which are items
that have counterparts among the items in the system of
“rational numbers.” The system of what we nowadays are
used to thinking of as numbers is replete with “irrational
numbers,” some of them “algebraic” (such as “the square root
of two”) and some of them “transcendental” (such as “pi”).
Classically, numbers are multitudes, and there may be
some numberratio that is the same as the relation in size of
some one magnitude to another. But there does not have to
be a numberratio that is the same. If there did have to be
such a ratio of numbers for every ratio of magnitudes, then
there would not have been a reason for devising a system of
“real numbers.” The reason for taking the step from “rational
FLAUMENHAFT
17
numbers” to “real numbers” has to do with the difference
between multitudes and magnitudes.
The difference between how we can speak about multitudes (that is, numbers in the classical sense) and how we can
speak about magnitudes classically manifests itself in the
statements with which Euclid begins his treatment of magnitudes in the Fifth Book of the Elements and his treatment
afterwards of numbers in the Seventh Book.
Although Euclid says what number is, he does not say
what magnitude is. Examples of magnitudes can be found in
his propositions, however. After the Fifth Book of the
Elements demonstrates many propositions about the ratios of
magnitudes as such, these are used by later books of the
Elements to demonstrate propositions about such magnitudes
as straight lines, triangles, rectangles, circles, pyramids, cubes,
and spheres. Such magnitudes correspond to what we nowadays call lengths, areas, and volumes. Weight is yet another
sort of magnitude. Though weights are not mentioned in the
Elements, what Euclid says there about magnitudes generally
is applied by Archimedes to weights in particular.
At the beginning of the Fifth Book, Euclid deﬁnes ratio
for magnitudes; but at the beginning of the Seventh Book he
does not deﬁne ratio for numbers. There in the Fifth Book he
also deﬁnes magnitudes’ being in the same ratio, and only
after doing that does he deﬁne magnitudes’ being proportional. But here in the Seventh Book he does not deﬁne numbers’ being in the same ratio: he goes directly into deﬁning
numbers’ being proportional, which he needs to do in order
to deﬁne the similarity of numbers that are of the sorts called
“plane” or “solid.”
The Greek term translated as “proportional” is analogon.
After a preﬁx (ana) meaning “up; again,” the term contains a
form of the word logos. This is the Greek term translated as
“ratio.” Logos is derived from the same root from which we
get “collect” (which is what the root means); and in most
contexts, it can be translated as “speech” or as “reason.”
Proportionality is, in Greek, analogia (from which we get
�18
THE ST. JOHN’S REVIEW
“analogy”), a condition in which terms that are different may
be said to carry or hold up again the same articulable relationship.
Later in the Elements, in the enunciation of the ﬁfth
proposition of the Tenth Book, for example, Euclid does use
the term “same ratio” in speaking of ratios that are numerical. In the deﬁnitions with which the Tenth Book begins,
Euclid says that lines which have no numerical ratios to a
given straight line are called “logosless”—that is, alogoi. This
Greek term for lines lacking any articulable ratios to a given
line is translated (through the Latin) as “irrational.”
Such a line and the given line to which it is referred cannot both be measured by the same unit, no matter how small
a unit we may use to try to measure them together. They are
“without a measurement together”—that is, asymmetra. This
Greek term is translated (through the Latin) as “incommensurable.” Because magnitudes of the same kind can be incommensurable, magnitudes are radically different from multitudes, and so we must speak of them differently.
Part Three
Muchness Not Related Like Manyness: Incommensurability
Let us turn now to the classic example of incommensurability: let us consider the relation between the side of a square
and its diagonal.
If a square’s diagonal were in fact commensurable with its
side, then the ratio of the diagonal to the side would be the
ratio of some number to some other number. But it cannot be
that. Why not? Because if it were, then it would have two
incompatible properties: one property belongs to any ratio
whatever which a number has to a number, and the other
property follows from that particular ratio which a square’s
diagonal has to its side. Let us convince ourselves that there is
such a contradiction.
First, let us look at that property which belongs to any
ratio whatever that is numerical. It is this: any numerical ratio
whatever must either be in lowest terms already, or be
FLAUMENHAFT
19
reducible to them eventually. Consider, for example, the ratio
that thirty has to seventy. Those two numbers have in common the factors two and ﬁve; so, dividing each of them by
ten, we see that the ratio that thirty has to seventy is the same
as the ratio that three has to seven. The ratio of three to seven
is the ratio of thirty to seventy reduced to lowest terms. If a
ratio of one number to another number were not reducible to
lowest terms, then the two original numbers would have to
contain an endless supply of common factors, which is impossible: any number must sooner or later run out of factors if
you keep canceling them out by division.
This property of every ratio that a number has to a number cannot belong to the special ratio that a square’s diagonal
has to its side. To see why this is so, consider a given square.
If you make a new square using as the side of the new square
the diagonal of the given square, then the new square will be
double the size of the original square.
Chart 2 shows that doubling. In the lefthand portion of
the chart, the diagonal of a square divides it into two triangles, and it takes four of these triangles to ﬁll up a new square
which has as its own side that diagonal. In the righthand portion, the chart also shows what happens if we now make
another new square, using as this newest square’s own side a
half of the diagonal of the original square. The original
square, now being itself a square on the newest square’s diagonal, will itself be double the size of this newest square—
since this newest square is made up of two triangles, and it
takes four of these triangles to make up the original square.
�20
THE ST. JOHN’S REVIEW
Let us suppose that it were in fact possible to divide a
square’s diagonal as well as its side into pieces that are all of
the same size. Let us count the pieces and say that K is the
number of pieces into which we have divided the diagonal,
and that L is the number of pieces into which we have divided
the side. By taking as a unit any of those equal pieces into
which we have supposedly divided the diagonal and the side,
we have measured the two lines together: the ratio which the
number K has to the number L would be the same as the ratio
which the square’s diagonal has to its side.
Now, disregarding for a moment just what ratio the number K has to the number L, but considering only that it is supposed to be a ratio of numbers—which, as such, must be
reducible to lowest terms—we can say that there would have
to be two numbers (let us call them P and Q) such that (1) P
and Q do not have a single factor in common and (2) the ratio
that P has to Q is the same ratio that K has to L. Since the
ratio that K has to L is supposed to be a ratio of numbers,
there cannot be any such pair of numbers as K and L unless
there is also such a pair of numbers as P and Q. (If the ratio
of K to L is already in lowest terms, then we will just let K
and L themselves be called by the names P and Q.) So now we
have P being the number of equal pieces into which the
square’s diagonal is divided, and Q being the number of such
pieces into which the square’s side is divided. And now we
will see that there just cannot be any such numbers as P and
Q because this pair of numbers would have to satisfy contradictory requirements.
The ﬁrst requirement results from the necessities of numbers; the second, from the consequences of conﬁguring lines.
Because (as has been said) the ratio of P to Q is a ratio of
numbers reduced to lowest terms, P and Q are numbers that
cannot have a single common factor; but (as will be shown)
because the ratio of P to Q is the same ratio that a square’s
diagonal has to its side, P and Q must both be numbers divisible by the number two. That is to say, P and Q are such that
they cannot both be divisible by the same number, and yet
FLAUMENHAFT
21
they also must both be divisible by the number two—a direct
contradiction.
Now we must see just why the latter claim is true. We ask:
why is it, that if any pair of numbers are in that special ratio
which a square’s diagonal has to its side, then both numbers
must be divisible by the number two?
In order to see why, we must ﬁrst take note of another
fact about numbers: if some number has been multiplied by
itself and the product is an even number, then the number
that was multiplied by itself must itself have been a number
that is even. (The reason is simple. A number cannot be both
odd and even—it must be one or the other—and when two
even numbers are multiplied together, then the number that
is the product is also even; however, when two odd numbers
are multiplied together, then the number that is the product
is not even but odd.) That said, we are now ready to see why
it is that if any two numbers are in that special ratio which a
square’s diagonal has to its side, then they must have the factor two in common.
The numbers P and Q have the factor two in common
because both of them must be even. They must both be even
because each of them when multiplied by itself will give a
product that must be double another number—and therefore
each must itself be double some number. Why?
Look at the lefthand portion of Chart 3, and consider
why P must be an even number. Because the square on the
diagonal is double the square on the side, the number produced when P is taken P times is double the number produced
when Q is taken Q times. Hence (since any number which is
double some other number must itself be even) the multiplication of P by itself produces an even number, and hence the
number P itself must be even.
And now look at the righthand portion of the chart, and
consider why Q also must be an even number. Since it has
already been shown that P must be an even number, it follows
that there must be another number that is half of P—call this
number H. But because the square on the side is double the
�22
THE ST. JOHN’S REVIEW
square on the halfdiagonal, it follows that the number produced when Q is taken Q times is double the number produced when H is taken H times. Hence (since, as we said
before, any number which is double some other number must
itself be even) the multiplication of Q by itself produces a
number that is even, and hence the number Q itself must be
even.
Since both P and Q must therefore be even numbers, they
must have the number two as a common factor, even though
they cannot have any number as a common factor. It is therefore absurd to say that there is a pair of numbers like P and
Q. And hence there cannot be a pair of numbers like that K
and L which we supposed that there could be.
If a pair of numbers did in fact have the same ratio that
the square’s diagonal has to its side, then no matter how
many times we halved both numerical terms in the ratio, both
terms that we got would have to remain even. But no pair of
numbers can be like that. Each number of the pair would
have to contain the number two as a factor in endless supply.
Such a ratio would be irreducible to lowest terms, because no
matter how many common two’s we were to strike as factors
from the terms of the ratio, there would still have to be yet
FLAUMENHAFT
23
another two in each. No number, however, can contain an
endless supply of factors.
It is therefore selfcontradictory to say that any ratio can
be a ratio of one number to another and also be the ratio of a
square’s diagonal to its side. To avoid being led into absurdity,
we must say that a square’s diagonal and its side are incommensurable.
The diagonal of a square and its side are not the only pair
of lines that are incommensurable. In the tenth proposition of
the Tenth Book of the Elements, Euclid begins to show us
how to ﬁnd many incommensurable lines. After presenting
thirteen sorts of them, he concludes the Tenth Book by saying that from what he has presented, there arise innumerably
many others.
That makes it hard to say what we mean when we say that
one pair of lines is in the same ratio as another pair of lines.
Even if we could somehow ignore the difﬁculty with the ratio
of a square’s diagonal to its side, other ones like it would still
keep popping up all over the place, since the ratio of a
square’s diagonal to its side is only one of innumerably many
ratios of incommensurable lines. And it is not only with lines
that the difﬁculty arises. Whatever the kind of magnitudes
that are being compared, whatever it is with respect to which
they are greater or smaller, comparisons of muchness are not
as such reducible to comparisons of manyness. The sizes of
magnitudes cannot always be compared by comparing counts
obtained by measuring. Magnitudes of any kind can be
incommensurable. It is true that one cube may have to
another, or one weight may have to another, the same ratio
that three has to ﬁve, for example; but the one cube may have
to the other, or the one weight may have to the other, the
same ratio that a square’s diagonal has to its side.
Magnitudes are called “incommensurable”—incapable of
being measured together—not when they are of different
kinds, but rather when they are not measurable by the same
unit even though they are of the very same kind. Two magnitudes may be of the very same kind and yet it may be true that
�24
THE ST. JOHN’S REVIEW
whatever unit measures either one of them will not measure
the other one too. By a unit’s “measuring” a magnitude, we
mean that when the unit is multiplied (that is, when it is taken
some number of times), it can equal that magnitude.
Therefore, when magnitudes of the same kind are incommensurable, one of them has to the other a ratio that is not
the same as any numerical ratio.
The demonstrable existence of incommensurability
means that comparisons of size can only be approximated by
using numbers. We can, to be sure, be ever more exact—even
as exact as we please—but if we wish to speak with absolute
exactness, numbers fail us.
For the ancient Pythagoreans, who were the ﬁrst to conceive of the world as thoroughly mathematical, knowledge of
the world was knowledge of numerical relationships. In the
world stretching out around us, the Pythagoreans saw correlations between shapes and numbers, such as we encountered
when we considered kinds of numbers, “square” numbers
and “cube” numbers, for example.
The Pythagoreans noted also that the movements of the
heavenly bodies take place in cycles. Their changes in position, and their returns to the same conﬁguration, have
rhythms related by recurring numbers that we get by watching the skies and counting the times. With numbers, the
world goes round. We are surrounded by a cosmos. (Cosmos
is the Greek word for a “beautiful adornment.”) The beauty
on high appears to us down here in numbers.
But even the qualitative features of the world of nature
show a wondrous correlation with numbers: numbers make
the world sing. It is not merely that rhythm is numerical, it is
that tone or at least pitch, is too. If a string stretched by a
weight is plucked, it gives off some sound. The pitch of the
sound will be lowered as the string is lengthened. If another
string is plucked (a string of the same material and thickness,
and stretched by the same weight) then the longer the string
the lower will be the pitch of the sound produced by plucking it. Now, what set the Pythagoreans thinking was the rela
FLAUMENHAFT
25
tion between numbers and harmony. (Harmonia is the Greek
word for “the condition in which one thing ﬁts another”; a
word from carpentry thus is used to describe music.) When
the lengths of two strings of the sort just mentioned are
adjusted so that one of them has to the other the same ratio
that one small number has to another, or to the unit, then
there is music. With the Pythagoreans’ mathematicization of
music, mathematical physics begins.
Thus not only were the sights on high seen to be an
expression of mathematical relationships, so were the sounds
down here that enter our souls and powerfully move what
lies deep down within us. Thinking that nature is a display of
numerical relationships, and that human souls are gotten into
order by attending to those relationships, the Pythagoreans
formed societies that sought to shape the thinking of the
political societies of their time by being the givers of their
laws. It is said that when the discovery of incommensurability was ﬁrst revealed to outsiders, thus making public the
insufﬁciency of number, the man who thus had undermined
the Pythagorean enterprise was murdered.
But never mind the whole wide world; even the relationships of size in mere geometrical ﬁgures cannot be understood simply in terms of multitudes. If geometry could in fact
be simply arithmeticized; if we could just measure lines
together, getting numbers which we could then just multiply
together, and thus simply express as equations all the relationships that we have to handle, then much in Euclid, in
Apollonius, and in the other classical mathematicians that is
difﬁcult to handle could be handled much more easily—and
Descartes in the seventeenth century would not have had to
undertake a radical transformation of geometry. Instead of
manipulating equations, however, we must in the study of
classical mathematics learn to deal with nonnumerical ratios
of magnitudes, and with boxes that are built from lines
devised to exhibit those ratios.
However, before we can freely deal with ratios, we must
learn what can be meant by calling two ratios the same—even
�26
THE ST. JOHN’S REVIEW
when they are not the same as some numerical ratio. Euclid
tells us that in the Fifth Book of the Elements. That is long
before his discussion of number, which does not take place
until the Seventh Book, thus raising a question about what
kind of a teacher he is: after all, is it not a principle of good
teaching that questions should be raised before answers are
presented?
Incommensurability is responsible for the difﬁculty of
Euclid’s deﬁnition of sameness of ratio for magnitudes, as
well as for the difﬁculty that modern readers encounter in the
classical presentation of relationships of size generally. Let us
now look at that deﬁnition.
Part Four
Muchness Related After All: Euclid’s Deﬁnition of Same
Ratio
We insist that even when two magnitudes are incommensurable, their ratio can be the same as the ratio of two other
incommensurable magnitudes. For example, we insist that the
ratio which one square’s diagonal has to its side is the same
ratio which another square’s diagonal has to its own side.
Sameness of ratio for magnitudes is therefore not to be
deﬁned in terms of measuring magnitudes: we could so deﬁne
it only if we could divide each of the magnitudes into pieces,
each equal to some magnitude that is small enough to be a
suitable unit, and then, when we counted up all of those small
equal pieces, we found no leftover unaccountedfor evensmaller piece of either of the magnitudes; but we cannot do
that with magnitudes that are incommensurable. Although we
can say that two ratios of magnitudes are the same as each
other if they are both the same as the same ratio of numbers,
we will not say that they are the same as each other only if
they are both the same as the same ratio of numbers.
What, then, is to be said instead? According to Euclid,
this: “Magnitudes are said to be in the same ratio, the ﬁrst to
the second and the third to the fourth, when, if any equimultiples whatever be taken of the ﬁrst and third, and any equi
FLAUMENHAFT
27
multiples whatever of the second and fourth, the former equimultiples alike exceed, are alike equal to, or alike fall short of,
the latter equimultiples respectively taken in corresponding
order.”
When Euclid says for ratios of magnitudes, what sameness is, it is very difﬁcult at ﬁrst to understand just what he
means. While his deﬁnition may be the proper departure
point on a road that we need to travel, for a beginner it seems
to constitute a locked gate.
A key to open that locked gate, however, is this fact:
while there is no ratio of numbers that is the same as a ratio
of incommensurable straight lines, nonetheless every ratio of
numbers is either greater or less than such a ratio of straight
lines. That is to say: whatever ratio of numbers we may take,
we can decide whether the ratio of one line to another is
greater or less than it.
For example, we might ask: which is greater—the ratio of
the one line to the other, or the ratio of seven to twelve? We
would divide the second line into twelve equal parts, and
then take one of those pieces in order to measure the ﬁrst
line. Suppose that this ﬁrst line, although it may be longer
than six of these pieces put together, turns out to be shorter
than seven of them. We would conclude that the ﬁrst line has
to the other line a ratio that is less than the numerical ratio of
seven to twelve.
Let us consider the ratio of lines that has been of such
concern to us, the ratio which a square’s diagonal has to its
side. And let us take the following ratio of numbers: the ratio
which that three has to two. As we have seen, the ratio of
lines that we are considering cannot be the same as any ratio
of numbers whatever; so it must be either greater or else less
than that ratio of numbers which we have taken. Which is it?
It must be less—because if it were greater (that is: if the
diagonal were greater than threehalves of the side), then the
diagonal taken two times would have to be greater than the
side taken three times. But if that were so, then a new square
whose side was the diagonal of the original square would
�28
THE ST. JOHN’S REVIEW
have to be more than double the original square, as is shown
in Chart 4.
Having thus shown that the ratio of a square’s diagonal to
its side is less than the ratio which three has to two, we could
in like manner also show that the ratio of those two lines is
greater than the ratio that, say, four has to three. And we
could go on showing, for the ratio of any pair of numbers
which we choose, that the ratio of a square’s diagonal to its
side is greater or is less than the ratio of the pair of numbers
chosen.
Indeed, although no ratio of numbers is the same as the
ratio of a square’s diagonal to its side, we can nonetheless
conﬁne this ratio of lines as closely as we please by using pairs
FLAUMENHAFT
29
of ratios of numbers, as follows. (The results are shown in
Chart 5.) Let us divide the square’s side into ten equal parts,
and take such a tenth part as the unit; then the diagonal will
be longer than fourteen of these units but shorter than ﬁfteen
of them; then let us consider what we get when we take as the
unit the side’s hundredth part, and then its thousandth. We
can go on and on in that way, eventually reaching numbers
large enough to give us a pair of numberratios that differ
from each other as little as we please; we can show it to be
greater than the eversoslightlysmaller numberratio, and
smaller than the eversoslightlygreater numberratio. So,
although the ratio of a square’s diagonal to its side cannot be
the same as any numberratio, it can be conﬁned between a
pair of numberratios that differ from each other as little as
we please.
�30
THE ST. JOHN’S REVIEW
The question that we want to answer, however, is not
when it is that ratios of magnitudes are almost the same, but
rather when it is that they are the very same—even when they
are not both the same as the same ratio of numbers. The key
to the answer, restating what we said, is this: given any ratio
of magnitudes, any ratio of numbers that is not the same as it
must be either greater than it or smaller.
That gives us the following answer: given two ratios of
magnitudes, they will be the same as each other (whether or
not they are the same as some ratio of numbers) whenever it
is true that—taking any ratio of numbers whatever—if this
ratio of numbers is greater than the one ratio of magnitudes,
then it is also greater than the other; and if it is less than the
one, then it is also less than the other.
Euclid’s deﬁnition seems much more complicated than
that, however, because it emphasizes “equimultiples.” Why
does it do that? Because that makes it simpler to compare
ratios of magnitudes with ratios of numbers: taking multiples
is a way to make the comparisons without performing any
divisions. We could in fact get a deﬁnition by using division,
but the deﬁnition would be clumsier if we did.
Let us put the notion of equimultiples aside for a
moment, and consider multiples simply, in contradistinction
to divisions into parts.
Suppose, for example, that the ratio of some magnitude
(A) to another magnitude (B) is greater than the ratio of nine
to seven but less than the ratio of ten to seven. This means
that if we divide B into seven equal parts and we use as a unit
(for trying to measure A) one of those seventh parts of B, then
A will turn out to be greater than nine of those parts of B, but
less than ten of them. If A and B happen to be incommensurable, then, no matter how many equal parts into which we
divide B—that is, no matter how small we make the Bmeasuring unit with which we try to measure A, we will ﬁnd that
we cannot divide A into pieces of that size without having a
smaller piece left over.
FLAUMENHAFT
31
So, in comparing the ratio of A to B with the numerical
ratio of some multitude m to some multitude n, it is less awkward to speak of Atakenntimes and of Btakenmtimes
than to speak of the little piece of A that may be left over
when we divide A into m pieces that are each equal to the nth
part of B—or, in other words, the little piece that may be left
over in A, no matter how small are the equal parts into which
we divide B. To make the requisite comparisons, we must
consider multiples of magnitudes, but we need not consider
parts of them; we have to multiply, but we do not also have
to divide. Instead of trying to measure magnitudes together,
by dividing them and counting the parts, we can speak merely
of multiples of the magnitudes.
What we have just now seen is this: to say that the ratio
of magnitude A to magnitude B is either the same as the
numerical ratio of m to n, or is greater or less than it, is to say
that Atakenntimes is either equal to Btakenmtimes, or is
greater or less than it. Let us take that and put it together with
what we saw earlier—which was this: to say that the ratio of
A to B is the same as the ratio of C to D is to say that whatever numerical ratio you may take (say, of m to n) this ratio
of numbers will not be the same as, or greater than, or less
than one of the ratios of magnitudes unless it is likewise so
with respect to the other one.
All that remains is for multiples of a certain sort to be
brought in—namely, “equimultiples.” Equimultiples of two
magnitudes are two other magnitudes that are obtained by
multiplying the two original magnitudes an equal number of
times. Here, as Chart 5 shows, both ratios’ antecedent terms
(namely, A and C) are each taken n times, and their consequent terms (namely, B and D) are each taken m times. In
other words, equimultiples (nA and nC) are taken of the ﬁrst
and third magnitudes (A and C); and also equimultiples (mB
and mD) are taken of the second and fourth magnitudes (B
and D). And m and n are any numbers at all: we are interested
in all the equimultiples of the ﬁrst and third magnitudes, and
also of the second and fourth ones.
�32
THE ST. JOHN’S REVIEW
Now at last we are in a position to see that Euclid’s deﬁnition is not so bewildering as might have seemed at ﬁrst
glance. There are several ways of saying what we can do with
what we have seen. In abbreviated form, they are exhibited in
Chart 7.
FLAUMENHAFT
33
Now in interpreting the chart, just remember that whenever we say “IF something, THEN something else,” that is
equivalent to saying “NOT something UNLESS something
else.” So, the ways laid out in the chart are these:
First way: Staying with ratios, we can say that two ratios
of magnitudes (call them the ratios of A to B, and of C to D)
are the same if, and only if, whatever numerical ratio we may
take (say the ratio that some number called m has to some
other number called n) the following is true: that numerical
ratio which m has to n will not be the same as, or greater
than, or less than one of the two ratios of magnitudes, unless
it is likewise so with respect to the other one.
Second way: We might want to get more explicit by making divisions into equal parts—that is, divide B into n equal
parts and do the same to D (these, B and D, being the conse
�34
THE ST. JOHN’S REVIEW
quent terms of the two ratios). Then the ratios would be the
same if the following is true: antecedent term A will not be
greater than, or less than, or equal to m of those nth parts of
its consequent term B unless the other antecedent term C is
likewise so with respect to that same number m of those nth
parts of its own consequent term D.
Third way: Someone who was willing to do all that
(namely, willing to divide magnitudes into equal parts and
then count them up and compare sizes) might insist on using
fractions to restate that as follows: A will not be greater than,
or less than, or equal to m/nths of B unless C is likewise so
with respect to that same fraction (m/nths) of D.
Fourth, and ﬁnal way: Rather than getting into the complicated business of dividing, counting divisions, and comparing sizes, even though we can brieﬂy restate it all by using
fractions, we might simply take multiples alone; and then we
would say that the ratios are the same if the following is true:
Atakenntimes will not be greater than, or less than, or
equal to Btakenmtimes unless CtakenexactlyasmanytimesasA is likewise so with respect to DtakenexactlyasmanytimesasB.
Those are several ways to determine, despite incommensurability, when we may say that the ratio of one magnitude
to another is the same as the ratio of a third magnitude to a
fourth one. The ﬁrst way formulates our initial insight, and
the ﬁnal way formulates Euclid’s deﬁnition of same ratio for
magnitudes. Euclid’s deﬁnition, like the others, covers all
ratios of magnitudes, whether or not they are the same as
ratios of multitudes; but, unlike the others, it manages to do
so by speaking simply of multitudes of magnitudes.
Magnitudes and multitudes are not the same as each
other, nor is either of them the same as ratios of them, but
magnitudes and multitudes and their ratios are all, in a way,
alike. How?
In the ﬁrst place, any two magnitudes of the same kind
are equal, or else one of them is greater than the other; and
likewise, any two multitudes are equal, or else one of them is
FLAUMENHAFT
35
greater than the other. Moreover, any magnitude has a ratio
to any other magnitude of the same kind, just as any multitude has a ratio to another multitude. Finally, any two ratios
(whatever it may be that they are ratios of, whether magnitudes or multitudes) are the same as each other, or else one of
them is greater than the other. This last fact supplies the
insight that enabled Euclid to deﬁne same ratio for magnitudes regardless of whether the magnitudes are commensurable.
Euclid’s deﬁnition can help us to understand what
enabled Dedekind several thousand years afterward to deﬁne
“the real numbers” in terms of “the rational numbers.”
Dedekind found himself in the following situation. In modern times, after proportions containing magnitudes and multitudes had given way to equations containing “real numbers,” there was still some difﬁculty in saying just what “real
numbers” were. To say much about the matter, it seemed necessary to refer not only to multitudes to which we come by
counting (that is, numbers in the strict sense) but also to magnitudes that we visualize (namely, lines). If we place the
counting numbers along a line (as in the top portion of Chart
8), it is then clear not only where to put all the “rational numbers,” including those which are fractional and those which
are nonpositive (as in the middle portion of the chart)—but
also where to put such “irrationals” as “the square root of
two.” To put “the square root of two” in its place, for example, (as in the bottom portion of the chart) just erect a square
that has a side that is a unit long, and swing its diagonal
down.
�36
THE ST. JOHN’S REVIEW
But if “the square root of two” is a number, it should be
deﬁnable without appealing to visualization. After all, we
need not visualize in order to count, or even to go on to think
up fractions and negatives. That thought troubled Dedekind
as he taught calculus in the nineteenth century, relying upon
appeals to a numberline. He ﬁnally saw how to treat “real
numbers” (the modern counterpart of Euclidean ratios of linear magnitudes) in terms of collections of “rational numbers”
(the modern counterpart of Euclidean ratios of numbers),
thus making it more plausible to speak of the “real numbers”
as really numbers.
Reading Dedekind is one way to think about what it
means to rely on a system of “real numbers.” Unless we do
think through the meaning of relying on a system of real
numbers, many of the things we take for granted in the modern world are without foundation.
But does Dedekind’s work simply represent progress
beyond Euclid’s, or are very important differences covered
over by the important similarities between what Dedekind
had his mind on and Euclid his?
Ratios relate multitudes, or magnitudes. Ratios (like multitudes) have homogeneity, and also (like magnitudes) have
continuity; but magnitudes do not have homogeneity, and
multitudes do not have continuity. Ratios themselves are
FLAUMENHAFT
37
therefore not things of the very same sort as the things that
they relate.
Ratios are not things of the same sort as quotients either.
To be sure, ratios are like quotients in that they are quantitative. Like any two magnitudes of the same sort, or any two
multitudes—or any two quotients—any two ratios (whether
of magnitudes or of multitudes) have an order of size. But
quotients, unlike ratios, are like what they are quotients of: a
quotient is itself a “real number” that is obtained through the
operation of dividing a “real number” by a “real number”; if
A and B are “real numbers,” then the quotient A/B is a “real
number” also.
Yet, although ratios and “real numbers” are things of different sorts, they do have a similarity—and not merely that
they are both quantitative. Although the magnitudes are not
homogeneous, and the multitudes are not continuous, the
ratios are both homogeneous and continuous—and so also
are “the real numbers.”
Part Five
Mathematics and the Modern Mind
In all this, we have been considering the classical handling of
numbers and lines in its difference from modern notions
which conﬂate the two, but we have not considered what it
was that led to the conﬂation. That is, indeed, an immensely
important matter, but it is too long a story to be told now. To
tell that tangled tale, one must traverse much of the road that
constitutes the Mathematics Tutorial at St. John’s College.
Along that arduous road, it is easy to be overwhelmed, however, and thus to lose sight of why it is worth our while to traverse it at all—so perhaps at least something should be said
about it now.
Euclid’s Elements prepares us for a higher study in geometry, the Conics of Apollonius. This was, for almost twoandahalf millenia, the classic text on the curves which—following the innovative terminology of Apollonius—came to be
called the “parabola,” the “hyperbola,” and the “ellipse.”
�38
THE ST. JOHN’S REVIEW
Apollonius’ conic sections were lines ﬁrst obtained upon a
plane by cutting a cone in various ways; they were then characterized by relative sizes and shapes of certain boxes formed
by associated straight lines standing in certain ratios. After
Descartes, although the names of those curves persisted, and
they continued to be called collectively “the conic sections,”
they nonetheless eventually ceased to be studied as such: they
came to be studied algebraically. But although Descartes’
Geometry may be in fact what it has been called—the greatest single step in the progress of the exact sciences—no one
could clearly see it as that without studying Apollonius, for
Cartesian mathematics has shaped the world we live in and
shapes our minds as well.
Apollonius tells his tale obliquely. He does not give us
questions, but rather gives us only answers that are too hard
to sort out and remember unless we ourselves ﬁgure out what
questions to ask. About Apollonius as a teacher we must ask
whether his work is informed by wisdom and benevolence.
Descartes did not think so. In his own Geometry, and in his
sketch of rules for giving direction to the native wit,
Descartes found fault with the ancient mathematicians.
Descartes severely criticizes them—for being showoffs.
He says that they made analyses in the course of ﬁguring
things out, but then, instead of being helpful teachers who
show their students how to do what they themselves had
done, they behaved like builders who get rid of the scaffolding that has made construction possible. Thus they sought to
be admired for conjuring up one spectacular thing to look at
after another, without a sign of how they might have found
and put together what they present.
Descartes also suggests, however, that they did not fully
know what they were doing. They did not see that what they
had could be a universal method. They operated differently
for different sorts of materials because they treated materials
for operation as simply objects to be viewed. Hence they
learned haphazardly, rather than methodically, and therefore
they did not learn much. Mathematics for them was a matter
FLAUMENHAFT
39
of wonderful spectacle rather than material for methodical
operation. The characteristic activity of the ancient mathematicians was the presentation of theorems, not the transmission and application of the ability to solve problems.
They had not discovered the ﬁrst and most important
thing to be discovered: the signiﬁcance of discovery. They
had not discovered the power that leads to discovery and the
power that comes from discovery. They were not aware that
the ﬁrst tools to build are tools for making tools. They were
too clever to be properly simple, and too simple to be truly
clever. They were blinded by a petty ambition. Too overcome
by their ambition, they could not be ambitious on the greatest scale.
Were the ancient mathematicians as teachers guilty of the
charges set out in that Cartesian critique? Were they guilty of
the obtuseness of which Descartes in his Geometry accused
them—were they guilty of the desultory foolingaround and
disingenuous showingoff of which Descartes had accused
them earlier, in his Rules? You cannot know unless you study
them.
In any case, the study of Euclid and Apollonius gives
access to the sources of the tremendous transformation in
thought whose outcome has been the mathematicization of
the world around us and the primacy of mathematical physics
in the life of the mind. Scientiﬁc technology and technological science have depended upon a transformation in mathematics which made it possible for the sciences as such to be
mathematicized, so that the exact sciences became knowledge
par excellence. The modern project for mastering nature has
relied upon the use of equations, often represented by graphs,
to solve problems. When the equation replaced the proportion as the heart of mathematics, and geometric theoremdemonstration lost its primacy to algebraic problemsolving,
an immense power was generated. It was because of this that
Descartes’ Geometry received that accolade of being called
the greatest single step in the progress of the exact sciences.
To determine whether it was indeed such a step, we need to
�40
THE ST. JOHN’S REVIEW
know what it was a step from as well as what it was a step
toward. We cannot understand what Descartes did to transform mathematics unless we understand what it was that
underwent the transformation. By studying classical mathematics on its own terms, we prepare ourselves to consider
Descartes’ critique of classical mathematics and his transformation not only of mathematics but of the world of learning
generally—and therewith his work in transforming the whole
wide world.
It is in the study of Apollonius on the conic sections that
the modern reader who has been properly prepared by reading Euclid can most easily see both the achievement of classical mathematics and the difﬁculty that led Descartes and his
followers to turn away from it.
It all has to do with ratio, and with notions of number
and of magnitude. For Apollonius, as for Euclid before him,
the handling of ratios is founded upon a certain view of the
relation between numbers and magnitudes. When Descartes
made his new beginning, almost two millenia later, he said
that the ancients were handicapped by their having a scruple
against using the terms of arithmetic in geometry. Descartes
attributed this to their not seeing clearly enough the relation
between the two mathematical sciences. Before modern readers can appreciate why Descartes wanted to overcome the
scruple, and what he saw that enabled him to do it, they must
be clear about just what that scruple was. Readers must, at
least for a while, make themselves at home in a world where
howmuch and howmany are kept distinct, a world which
gives an account of shapes in terms of geometric proportions
rather than in terms of the equations of algebra. For a while,
readers must stop saying “ABsquared,” and must speak
instead of “the square arising from the line AB”; they must
learn to put ratios together instead of multiplying fractions;
they must not speak of “the square root of two.”
Mathematical modernity gets under way with Descartes’
Geometry. By homogenizing what is studied, and by making
the central activity the manipulative working of the mind,
FLAUMENHAFT
41
rather than its visualizing of form and its insight into what
informs the act of vision, Descartes transformed mathematics
into a tool with which physics can master nature. He went
public with his project in a cunning discourse about the
method of well conducting one’s reason and seeking the truth
in the sciences; and this discourse introduced a collection of
scientiﬁc tryouts of this method, the third and last of which
was his Geometry.
For those who study Euclid and Apollonius in a world
transformed by Descartes, many questions arise: What is the
relation between the demonstration of theorems and the solving of problems? What separates the notions of howmuch
and howmany? Why try to overcome that separation by the
notion of quantity as represented by a numberline? What is
the difference between a mathematics of proportions, which
arises to provide images for viewing being, and a mathematics of equations, which arises to provide tools for mastering
nature? How does mathematics get transformed into what
can be taken as a system of signs that refer to signs—as a symbolism which is meaningless until applied, when it becomes a
source of immense power? What is mathematics, and why
study it? What is learning, and what promotes it?
With minds that are shaped by the thinking of yesterday
and of the days before it, we struggle to answer the questions
of today, in a world transformed by the minds that did the
thinking. We will proceed more thoughtfully in the days
ahead if we have thought through that thinking for ourselves.
Our scientiﬁc past is not passé.
�42
THE ST. JOHN’S REVIEW
43
The Husserlian Context of
Klein’s Mathematical Work
Burt C. Hopkins
I have to begin my remarks with the admission of my ignorance about their ultimate topic, which is neither Edmund
Husserl’s nor Jacob Klein’s philosophy of mathematics nor,
for that matter mathematics itself, but numbers. I do not
know what numbers are. To be sure, I can say and read, usually with great accuracy, the numerals that indicate the prices
of things, street addresses, what time it is, the totals on my
pay stubs, the typically negative balance in my checkbook at
the end of each month, and so on. Moreover, I know how to
count and calculate with them, though usually with less accuracy no matter how much I try to concentrate on what I am
doing.
But if I am asked or try to think about what they are
when, for instance, I say my address is six hundred ﬁftythree
Bell Street, or that my house has two bedrooms, or that I only
have twenty bottles of wine left, it is obvious to me that I do
not know what I am talking about. Of course, I know that my
address refers to my abode, that my bedrooms are the rooms
in the house with beds, and that my wine is something I drink
solely for its medicinal purposes. But the six hundred ﬁftythree, the two, or the twenty, what are they? I do not know.
Likewise, if I am asked or try to think about what the
numbers are with which I calculate, when for instance I think
that because I need onehalf pound of steak per person to
feed each of my dinner guests, that I expect ﬁve guests, and
that therefore I need two and a half pounds to feed my guests,
Burt Hopkins is Professor of Philosophy at Seattle University, and is Secretary
of the International Circle of Husserl scholars. He is the author of a book,
forthcoming from the University of Wisconsin Press, entitled Edmund Husserl
and Jacob Klein on the Origination of the Logic of Symbolic Mathematics:
An Inquiry into the Historicity of Meaning.
�44
THE ST. JOHN’S REVIEW
or when I think that the price of two and a half pounds of
steak, at eight ninetynine per pound is twentytwo dollars
and fortyseven and a half cents, I know well enough what
dinner guests, steak, and dollars and cents are. But the one
half and the eight ninetynine with which I calculate, and the
two and a half and twentytwo and fortyseven and a half that
are the results of my calculations, what are they? I do notknow.
Some among us will say and, indeed have already said for
a long time, that numbers, any numbers, are amounts of
things, speciﬁcally, amounts of whatever it is that we count in
order to answer the question how many of the things in question there are. In my examples above, to talk or think about
two as the amount of my bedrooms, twenty as the amount of
my bottles of wine, ﬁve as the amount of my dinner guests,
eight ninetynine as the amount of money it takes to buy a
pound of steak, twentytwo and fortyseven and a half as the
amount of money it takes to buy two and a half pounds of
steak, certainly does seem to make sense. But what about my
house address? Is six hundred ﬁftythree the answer to the
question of how many of my house? Or is the one half pound
of steak needed to feed each dinner guest the answer to the
question how many of pounds of steak? The numbers here do
not straight away appear to be telling us how many houses my
house is, since the answer to that question is that it is not
many at all but one; neither do numbers tell us how many
pounds of steak are needed to feed each dinner guest because
each does not need many pounds at all but only a half a
pound. And, to complicate things deliberately, let us consider
what it means, numerically speaking, if I have a negative balance in my checkbook at the end of the month. Whatever the
amount of the negative number, let us hypothetically say it is
an even negative one hundred six dollars, that certainly does
not seem to provide the answer to the question of how many
dollars and cents I have.
For the moment, however, let us push these concerns
aside and assume that a number really is the amount of some
HOPKINS
45
thing. Moreover, let us suppose that the amount of something
or its number ﬁrst makes sense to us when we count more
than one of something. Finally, let us suppose that any group,
that is, any collection of more than one of something, no matter how big, has a number, which is to say, an exact or deﬁnite amount that answers the question how many with
respect to the things in the group, and that this can be arrived
at by counting. Does this really answer the question what a
number is? Or, more precisely, does this really answer the
question what numbers are? I say numbers because when we
count we always use more than one number to arrive at the
exact amount of something, even though once we arrive there
we conclude the count by saying a single number.
What, then, are the numbers two, three, seven, or, if I
have a lot of enumerative stamina, ﬁve hundred eighteen, that
I say when, having nothing better to do, I count the grains of
sand I have decided to put into separate piles? Or, what are
the numbers two, three, seven, ﬁve hundred eighteen, that I
say when I count the crystals of salt that I use to duplicate the
numbers of grains of sand I previously put into piles? The
things counted and therefore their amounts are not the same;
that is to say, an amount of grains of sand and an amount of
salt crystals are different things. But are their numbers the
same? Is the two that is the amount of grains of sand the same
two that is the amount of crystals of salt?
If number is supposed really to be the amount of something, and if the stress is placed on the of something, then
number, as its (the something’s) amount, would be nothing
more or less than the two grains of sand, or, the two crystals
of salt, both of which, being different somethings, would also
be different numbers. This situation would be illustrated were
I to say that there are a number of people I do not know in
the room. If I proceeded to count them, whatever number I
came up with would be not just an exact amount of anything
whatever, but precisely the exact amount of people I do not
know in the room.
�46
THE ST. JOHN’S REVIEW
On the other hand, if the stress were placed on the how
many of the amounts in question, then two would indeed be
how many grains of sand and crystals of salt there are in my
smallest piles of each, and therefore their amounts and hence
numbers would indeed be the same. This situation would be
illustrated were I to say all dogs, if they are naturally formed,
have the same number of legs, namely, four. The exact number is the same, though the legs are not, because, silly as it
sounds to say this, different dogs have different legs.
Now some might want to put an end to this whole line of
inquiry, but especially to my last question about whether
numbers of different things are different or the same, by saying that the obvious answer is that what numbers are are
abstract concepts. Hence they are really ideas, ideas that we
can relate to different things when we want to count or want
to apply the results of our calculations. But we do not have
to. Thus I can add 4 to 6 and get 10. I can multiply 10 times
10 and get 100, and so on, without having to think or answer
anybody’s question about 4 of what, or 6 of what, or 10 of
what, and so on. Just as it makes perfect sense to say that 2
plus 2 is 4, it does not make any sense to say 2 plus 3 is 4,
because everybody who can count knows 2 plus 3 is 5.
Perhaps. But perhaps not. And this is where Husserl and
then eventually Klein come in. But ﬁrst Husserl. It is generally known that Edmund Husserl, the German philosopher
who, as the founder of the socalled phenomenological movement in philosophy, was responsible for one of the two dominant approaches to philosophy in the last century (the other
being socalled analytic philosophy), was originally a mathematician. Known likewise is that his ﬁrst book, published in
1891, is titled Philosophy of Arithmetic.1 However, the contents of this book are not so well known, because, among
other reasons, soon after its publication both its author and
Gottlob Frege had some very critical things to say about
them.
For our purposes, however, the contents of the book are
more important than their criticism of them. From beginning
HOPKINS
47
to end, the book concerns the answer to the question whether
numbers are really abstract concepts that make perfect sense
without saying or thinking what they are numbers of; or
whether to make sense as numbers, numbers have to be spoken or thought about as being numbers of something. Husserl
began his answer to this question by ﬁrst distinguishing
between two kinds of numbers, one of which he called
“authentic” and the other “symbolic.” We will discuss the latter ﬁrst, because even though, as its name suggests, it is the
less authoritative kind of number, it is also the one with
which we are usually more familiar. A number is symbolic in
Husserl’s understanding when the number and the sign used
to designate it are indistinguishable. For example: 3. Most of
us have no doubt been taught or learned that this is the number three rather than what it really is, which is a number sign
or numeral. Indeed, even though most of us are also aware of
other numerals, for instance, Roman numerals, my suspicion
is that when we see such numerals we immediately interpret
what they really mean in terms of our numeral system, a system that was actually invented by the Arabs. A symbolic number, then, is a number that most of us—with or without thinking about it—identify with the sign that we either write or
read.
Most of us, that is, unless we have thought about the fact
that the signs used to designate numerals, and therefore these
numerals themselves, are based on what were originally and
still remain conventions, even if they are no longer recognized as such. Different conventions mean different numerals, as we have just seen. But what about the numbers? Do
different numerals mean different numbers? Many when
faced with this question conclude no. Their thinking here is
that numbers remain the same, however different the numerals that express them, because numbers are really concepts.
Hence, no matter what numeral or word is used to express
the number three, ‘three’ is a concept that remains the same
because it is not identical with some numeral (which is subject to change) or with a word from some language (which is
�48
THE ST. JOHN’S REVIEW
subject to variability). Moreover, for just this reason number
is a properly abstract concept: it remains identical with itself
no matter what sign or word is used to express it.
Despite the sophistication of this view of number, it is
dead wrong according to Husserl, because if numbers were
really abstract concepts in this sense, then the most basic
operation of arithmetic—addition—becomes unintelligible.
For instance, if in adding the number two to the number two,
what we are really adding is the abstract concept of the number two to another (!) abstract concept of the number two,
then arriving at their putative sum, the abstract concept
‘four’, becomes a great mystery. Most obviously, there is the
problem of how a concept that is supposed to remain identical can nevertheless change into another concept that is also
supposed to remain identical. That is, the abstract concept
‘two’ is not supposed to be able to change as words and signs
can and do, but to remain what it is, namely, the number two.
Yet precisely this supposition has to be abandoned if talk of
adding the concept of two to the concept of two to get the
concept of four is to make sense, since when the number two
is added to the number two the sum is not two number twos
but the number four.
Indeed, it is precisely this consideration that led Husserl
to the realization that authentic numbers are not abstract concepts. This is, admittedly, a difﬁcult thought. What, then, are
they, these authentic numbers, if they are not abstract concepts? Husserl’s answer is that they are the deﬁnite amounts
of deﬁnite things that have been grouped together by the
mind. What kind of things? Literally any kind. What deﬁnite
amounts? Pretty much the ﬁrst ten, namely two, three, four,
ﬁve, six, seven, eight, nine, ten. Why are zero and one not
among these? Because they are not deﬁnite amounts, which in
the case of zero is obvious, while in the case of one is less so
but still fairly obvious, since one is not an amount; it is not
many. What has the mind to do with authentic numbers?
Plenty for Husserl, since not only do authentic numbers ﬁrst
show up in counting, but also, only those deﬁnite things can
HOPKINS
49
be counted that have been grouped together by it to compose
what Husserl, and as we shall also see, ancient Greek philosophers and mathematicians, called a multiplicity. Moreover—
and this is the most important consideration for our purposes—an authentic number really is something that manifestly cannot be found in either the reality of the deﬁnite
things that are counted or in any relationship among them.
This last point requires closer scrutiny. If we ask how is it
that the ﬁrst authentic number, two, is able to register a deﬁnite amount of deﬁnite things as ‘two’, in the sense that in
saying or thinking the number two, the things in question are
recognized as being exactly two with regard to their number,
we are then asking about something that is different from the
things that this number numbers as two. Husserl, following a
long philosophical and mathematical tradition, refers to what
is asked about in this question as the unity of this number. In
asking it, we are asking how it is that distinct things, in this
case two of them, can nevertheless be brought together as
precisely this, namely, the two that is articulated by the number two. This allimportant question about the unity of
authentic numbers becomes more explicit when we consider
Husserl’s reason for thinking that only the ﬁrst ten or so deﬁnite amounts can be authentic numbers. Husserl’s reason is
deceptively simple: the mind can only apprehend—all at
once—each of the deﬁnite items that are numbered in a number when these things do not exceed ten. When the amount
of deﬁnite things in a multiplicity of things exceeds ten, each
one of them cannot be apprehended all at once by the mind
as when, for instance, it counts thirty of them.
These considerations, do not provide us yet with
Husserl’s answer to the question of the unity of authentic
numbers, but only address what is at stake in it. What is at
stake is that in registering the deﬁnite amount of deﬁnite
things, an authentic number is bringing them together in a
way that cannot be explained by each of the things so brought
together, no matter whether these are considered by themselves or as each relates to the other things. Each considered
�50
THE ST. JOHN’S REVIEW
by itself cannot explain it for the simple reason that authentic numbers are always amounts of something that is more
than one. Considering the relations among things cannot
explain it either—they may be side by side, or on top of, or
bigger and smaller than one another, and so on, because none
of these relations is even remotely numerical. Each deﬁnite
thing, as one thing, can only be registered as having a number by being brought together with other deﬁnite things, each
of which is also one, a bringing together which does not
apprehend the things brought together singly but precisely as
all together.
Husserl’s explanation of how the mind does this is as simple as it is remarkable. The mind combines things into groups
or collections in such a way that what is grouped or collected
forms a whole that is different from the group members or
collected items, even though as the whole of just these members or items, it is clearly related to them. For example, in a
row of trees, a gaggle of geese, or a ﬂock of birds, what is
named by the row, gaggle, and ﬂock is the whole Husserl has
in mind, a whole that cannot be separated from what it is a
whole of any more than it can be identiﬁed totally with it.
Authentic numbers for Husserl are also comprised of wholes
like this, so that the whole of the number three is clearly
related to each of the things it registers as three, without,
however, its being totally identical with them. Two distinct
but related things are involved for Husserl here. One is the
fact that there are groups and collections of this kind, and the
other is that they have their origin in the mind. The talk
about the mind’s involvement in originating groups and collections of things does not mean that these things are only ﬁgments of the mind. Husserl only mentions the mind to
explain something very speciﬁc, the fact that the unity, the
whole of authentic numbers that we have been talking about,
is neither an abstract concept nor something that can be
explained by the deﬁnite things it registers as to their exact
number. Once this is recognized, and only once it is recognized, are we then in a position to understand why Husserl
HOPKINS
51
tried to explain this unity—this being a whole—of authentic
numbers by the way the mind organizes and grasps things as
groups and collections.
Husserl is very speciﬁc about this. The mind considers
each thing that it groups or collects as something that belongs
to the whole of whatever it is grouping and collecting. In the
case of authentic numbers, it considers each as something, a
certain one, without attending in the slightest to any other
qualities that belong to what it collects. This is the case
because unlike other groups and collections, which are
groups of something speciﬁc, for example, geese or trees,
authentic numbers are wholes of quite literally anything
whatever. The moons of Jupiter, Homer’s psyche, the city of
Annapolis, etc., can be collected together and the collection
registered as an authentic number—so long, of course, as the
amount in the collection does not get too big. The process of
forming authentic numbers, as well as other kinds of groups
and collections, is expressed in language according to Husserl
by the word “and,” although both this process and the wholes
it generates are for him most decidedly not anything linguistic. Hence the formation of a group, such as the group or
whole of students in a room, comes about when one student
and one student and one student and one student, and so on,
are collected by the mind. Note well, however, that in this
example not just anything can be grouped together, but only
what belongs to the whole that is being grouped, namely, students. Likewise, the formation of a collection, the whole of
which is the red objects in a room, comes about when one of
any kind of red object and one of any kind of red object and
one of any kind of red object, and so on, are collected by the
mind. Again, as in the whole that is a group, not just anything
can be collected. Finally, the formation of a collection, the
whole of which is its number, comes about when something
and something are collected, and then the process is stopped.
More precisely, when the process of collecting is stopped
after something and something are collected, the ﬁrst authentic number, two, is the result. Likewise, when something and
�52
THE ST. JOHN’S REVIEW
something, and something are collected, the authentic number three is the result, and so on, until the limit of authentic
numbers, ten, comes about.
Authentic numbers, then, are just such collections of
something, which is to say, of anything that can be considered
as simply one thing, without any other qualities or determinations of it being relevant to its belonging to a collection
that is numbered. Since its being just one is the only thing relevant here, Husserl follows a long tradition and refers to
these collected ones as “units.” Because authentic numbers
are amounts of units, Husserl can explain on their basis what
the understanding of numbers as abstract concepts cannot,
namely, addition. Adding two and two involves the combination of ‘one unit and one unit’ with ‘one unit and one unit’,
and hence, there is no mystery here, since ‘one unit and one
unit, added to ‘one unit and one unit’, yields ‘one unit and
one unit, and one unit, and one unit’, which is the authentic
number four.
Before considering what Husserl thinks happens to numbers when their number exceeds ten and they are no longer
authentic, let us step back for a minute. I have been talking
now for some time about something I have acknowledged my
ignorance of, namely, what numbers are. After considering
two common views of them—one that considers them to be
amounts of something and the other that considers them to
be abstract concepts—I have turned our attention to the contents of Husserl’s book on the philosophy of arithmetic. The
very question it attempts to answer is which of these two
views of number is correct. So far I have pointed out that
Husserl begins this investigation by distinguishing between
authentic and symbolic numbers, both of which we have now
discussed in some detail. Indeed, at this point it might seem
that Husserl’s answer to the question of whether numbers are
amounts of something or abstract concepts is pretty obvious,
since authentic numbers can explain what abstract conceptual
numbers cannot, namely, the basic operations of arithmetic.
However, things are not that simple because when we con
HOPKINS
53
sider numbers greater than ten, Husserl thinks that they
become inauthentic, as they cannot be authentic. They cannot
be authentic, it will be recalled, because the mind cannot
grasp more than ten things all at once. Symbolic numbers on
Husserl’s understanding are therefore also in this sense inauthentic.
Even though they are inauthentic, however, symbolic
numbers are not inferior, mathematically speaking, to authentic ones for Husserl. On the contrary, because they deal with
numbers larger than ten, they come in handy any time calculation with large numbers is required, since without them, we
would be reduced to counting units when we calculate with
such numbers, which no doubt would be both tedious and
time consuming. At the time Husserl wrote his ﬁrst book,
mathematicians and philosophers wanted an explanation
how it was possible to do what no one denies can be done: to
calculate with inauthentic numbers, which are symbolic and
so in some sense are abstract concepts. In the ﬁrst ten chapters of The Philosophy of Arithmetic, Husserl attempted to
prove that authentic numbers and symbolic numbers are logically equivalent because each refers to the same objects,
speciﬁcally, the collections of more than one unit that authentic numbers register the ﬁrst ten amounts of. He argued that
authentic numbers do this directly and symbolic numbers
indirectly. When Husserl reached chapter eleven, however, he
realized something that shook him to his depths, quite literally. (He later recounted a decadelong depression that
ensued as a result.) He realized not only that symbolic numbers did not refer to the same objects as authentic ones—to
collections of units—but also that the basic operations with
quantities that are known, what he called general arithmetic,
could only be explained on the basis of the very opposite of
what he had argued in the ﬁrst ten chapters. General arithmetic only makes sense if the numbers it uses are symbolic in
the sense I discussed above, wherein the sign and the numerical concept are identical so that a number is interpreted to be
the sign that we write or read—what Husserl called “sense
�54
THE ST. JOHN’S REVIEW
perceptible” signs. Husserl also realized he had no idea of
how this is possible. As he described it some two decades
later, “how symbolic thinking is ‘possible’, how...mathematical...relations constitute themselves in the mind...and can be
objectively valid, all this remained mysterious.”2
Thus we can say that, while recognizing that some numbers are clearly deﬁnite amounts of units and others are
abstract, symbolic concepts because they do not refer to such
units, Husserl came to see that he did not know what either
of them really is. Husserl eventually thought he could solve
this mystery by explaining the manipulation of symbolic
numbers, or, more precisely, number symbols, as well as all
mathematical symbols, in terms of what he called “the rules
of a game,” rules that were invented not by mathematics but
by logic. Mathematics thus came to be understood by Husserl
as a branch of logic. Moreover, Husserl thought that all the
rules invented by logic have their foundation in concepts that
are true of other concepts, these other concepts, in turn,
being true of anything whatever, that is, anything that can be
experienced and therefore thought of as an individual object.
As a consequence, Husserl’s eventual explanation of how
symbolic mathematics is possible, in The Crisis of European
Sciences, was really not so far from his failed ﬁrst attempt at
explanation. To be sure, his later explanation does not characterize number symbols as referring to the same objects that
authentic numbers do, namely to units, but it did trace the
truth of the logic that invented the rules for manipulating
them to a basis in individual objects. This, we shall soon see,
is a problem if we follow Jacob Klein’s mathematical investigations, which I am going to suggest can best be followed by
ﬁrst considering their Husserlian context.
The bulk of Klein’s mathematical investigations are contained in his Greek Mathematical Thought and the Origin of
Algebra, which as many of you know was originally published
in German as two long articles in 1934 and 1936 and translated into English in 1968 by St. John’s tutor and former
Dean of the College, Eva Brann.3 The historical nature of the
HOPKINS
55
topic announced by its title might initially seem to be far
removed from what we now know is the systematic nature of
Husserl’s topic in Philosophy of Arithmetic. However, one
does not need to read very far in Klein’s book to discover that
he understands the key to the historical investigation of his
topic to lie precisely in the distinction between symbolic and
nonsymbolic numbers. Speciﬁcally, he expresses the view
from the start not only that the symbolic number concept is
something that makes modern, algebraic mathematics possible, but also, that symbolic numbers were entirely unknown
to ancient Greek mathematicians and philosophers. The very
point of departure for Klein’s investigation of the origin of
algebra is therefore informed by his view that unless the nonequivalence of ancient Greek numbers and modern symbolic
numbers is recognized, the change in the nature of the very
concept of number that took place with the transformation of
classical mathematics into modern mathematics in the sixteenth century will go unrecognized.
Klein’s thought here can be made clearer by closely considering precisely how he characterizes the difference
between the ancient Greek numbers and the modern symbolic ones. Ancient Greek numbers or arithmoi are manifestly
not abstract concepts, but rather beings that determine deﬁnite amounts of deﬁnite things. In contrast, symbolic numbers
are characterized by him to be abstract concepts that do not
refer to anything deﬁnite except that which is referred to by
their senseperceptible signs. When we consider the fact that
Husserl articulated the difference between authentic and
symbolic numbers in precisely these terms, the resemblance
between Klein’s and Husserl’s view of the distinction between
symbolic and nonsymbolic numbers is striking. So striking is
it in fact that some have drawn the conclusion that Husserl
should be given precedence in this matter, as either the source
or the major inﬂuence on Klein’s formulation of the distinction in question.4 These matters, however, are not so simple.
To begin, it is important to keep in mind that Husserl
sought in Philosophy of Arithmetic to demonstrate the logical
�56
THE ST. JOHN’S REVIEW
equivalence of nonsymbolic and symbolic numbers, whereas
Klein’s book begins with the insight that this is impossible, an
insight, as we have seen, Husserl arrived at only reluctantly.
Moreover, subsequent to his ﬁrst book Husserl explains the
relationship between symbolic and authentic numbers as a
matter of logic. In his Logical Investigations (1900) and
Formal and Transcendental Logic (1927), the relationship is
explained as a matter of the translation of logical truths into
rules for the manipulation of symbols. Logical truths for
Husserl are rooted in concepts that are true of other concepts, other concepts that, in turn, can be traced back to
truths that are rooted in individual objects. Husserl calls the
rules established on the basis of this logic the “rules of a
game,” because even though they permit calculational operations on symbols that yield mathematically correct results,
these operations and hence the very process of symbolic calculation have nothing to do with insight into concepts that
pertain to the objects to which they are ultimately related,
and therefore nothing to do with real knowledge. In a word,
Husserl thought he resolved the issue of symbolic and
authentic numbers by exposing symbolic calculation to be a
“technique” whose cognitive justiﬁcation can only be provided on the basis of the conceptual knowledge of individual
objects that logic provides.
Klein, however, thought otherwise. He thought that it is
impossible to explain the sharp distinction between nonsymbolic and symbolic numbers on the basis of the knowledge—
logical or any other kind—of individual objects. He conducted an historical investigation into how an aspect proper
to ancient Greek arithmoi was transformed into modern symbolical numbers and concluded that the objects referred to by
each kind of number are fundamentally different. The objects
referred to by arithmoi are deﬁnite, which is to say, individual. The objects referred to by symbolic numbers are indeﬁnite. They refer neither to individuals nor to their qualities
and are therefore indeterminate. Individual objects, both
Husserl and Klein agree, are objects that we encounter in our
HOPKINS
57
experience of the world and thus, they can be pointed to.
Moreover, their qualities, which either some (general qualities) or all (universal qualities) individual objects share, are
qualities that, even though they are not individual, can nevertheless be spoken about in connection with the individual
objects that we do encounter in the world. For instance, we
can point to dogs and cats, each one of which is therefore
individual. We can also talk about their general or universal
qualities and relate these to the individual dogs and cats that
we encounter in the world. Indeterminate objects, on the
contrary, can never be encountered in our experience of the
world, and therefore they cannnot be pointed to. Husserl and
Klein also agree that because their objects are not determinate
in this very precise sense, symbolic numbers cannot have a
direct relationship to any individual objects in the world or to
their qualities. Finally, both Husserl and Klein agree that symbolic numbers themselves, and not their indeterminate
objects, are nevertheless encountered in the world: namely,
they are encountered as the senseperceptible signs that, we
mentioned earlier, many of us interpret as numbers.
Where Husserl and Klein disagree, or more accurately,
where Klein would have had to express his departure from
Husserl’s understanding of the relationship between symbolic
and nonsymbolic numbers, had he chosen to do so, has to do
with the possibility of theoretically clarifying the philosophical meaning of symbolic numbers. Husserl thought this could
be done—indeed, he thought he did it in his two books on
logic. Klein did not think it could be done. In fact, as we shall
see, Klein’s mathematics book explains why the very attempt
to clarify theoretically the philosophical meaning of symbolic
numbers and mathematical symbolism generally is doomed to
failure. It is so doomed because all the concepts available to
provide such a theoretical clariﬁcation, without exception,
only make sense when the philosophers or anyone else using
them are talking about individual objects and their qualities.
Klein explains why this is the case by establishing a fundamental difference in what he calls the “conceptuality”5 of
�58
THE ST. JOHN’S REVIEW
the concepts that belong to ancient Greek and modern science. He uses this term to articulate both the way in which
the concepts proper to these respective sciences are structured and the status of their relationship to the nonconceptual realities of both the mind and the world. Klein thinks
that despite the continuity discernable in the technical vocabulary of ancient Greek mathematics and philosophy and their
modern counterparts, the mathematical and philosophical
signiﬁcance of each of the words in it is nevertheless radically
divergent for the ancients and moderns. Thus in his view the
fundamental signiﬁcance of words like knowledge, truth,
concept, form, matter, nature, energy, number, and so on is
completely different for the ancient Greeks and the moderns
because of the differences in the respective conceptualities of
each. In other words, Klein thinks that these conceptualities
shape the meaning of words, rather than the other way
around, that is, rather than the signiﬁcance of words shaping
the structure of the conceptualities.
Klein locates the key example of the shift from the
ancient Greek to the modern conceptuality in the transformation the concept of number undergoes in the sixteenth
century. Prior to Viète’s invention of the mathematical symbol, the concept of number according to Klein always only
meant a deﬁnite amount of deﬁnite things, a meaning that
was established by the ancient Greeks and that remained
operative in both European mathematics and the Europeans’
everyday praxis of counting and calculation until the invention of algebra. Klein claimed in his book—but did not elaborate—that this transformation is paradigmatic for the conceptuality that structures the modern consciousness of the
world.
Before considering in more detail Klein’s account of this
exemplary transformation of the conceptuality of number, a
few words about the potentially misleading talk of the “concept” of number are in order. Klein engages in such talk when
his investigation is comparing what he refers to as the different “number concepts” of ancient Greek and modern mathe
HOPKINS
59
matics. It is potentially misleading because for Klein the most
salient difference between in these number concepts is
located in the fact that the ancient Greek “concept” of number is precisely something that is not at all a concept, but a
being, while the modern “concept” of number is precisely
something that is not a being but a concept. The object of the
word “concept” in these comparative contexts is, I think,
clearly the “conceptuality” of ancient Greek and modern
numbers, which means it would be a mistake to attribute to
Klein in such contexts the thought that in ancient Greek and
modern mathematics numbers are concepts, albeit different
in kind. Klein, however, also talks about the ancient Greek
“arithmosconcept” (arithmosBegriff or AnzahlBegriff) and
the modern “numberconcept” (ZahlBegriff), which again is
potentially misleading, for the same reasons. Yet here, too,
careful consideration again discloses that such talk always
occurs within the context of his comparison of what he presents as the different ancient Greek and modern characterizations of numbers, only one of which formulated them as concepts.
Before elaborating Klein’s account of these different characterizations, I want to raise and then answer one more question. From what perspective was Klein able to compare the
ancient Greek and modern numbers? Klein, after all, was neither ancient nor Greek but, by his own admission, thoroughly
modern. How, then, was he able nevertheless to get sufﬁcient
distance from the presuppositions that inform his modernity,
from his modern outlook and consciousness, such that he
could encounter and investigate what, again by his own
admission, are the radically different presuppositions of the
ancient Greeks?
I think the answer to this question can be found in the single reference in Klein’s published work to Husserl’s
Philosophy of Arithmetic.6 It occurs in “Phenomenology and
the History of Science,” an article Klein wrote for a memorial
volume of essays on Husserl’s phenomenology published in
1940, two years after Husserl’s death. In this article Klein did
�60
THE ST. JOHN’S REVIEW
not at all hesitate to articulate the philosophical signiﬁcance
of his 1934 and 1936 investigations of ancient Greek mathematics and the origin of modern algebra in terms of Husserl’s
last writings, which were published in 1936 and1939.7 In
these writings Husserl traces the cause of the crisis of
European sciences to their failure to grasp properly the scope
and limits of scientiﬁc methods that are rooted in modern
mathematics for understanding the nonphysical, which is to
say human spiritual reality together with the world of its
immediate concerns. I will come back to this theme at the end
of my remarks. I want to focus for now on Klein’s reference
to what he characterizes in this article as Husserl’s “earliest
philosophical problem,” namely “the ‘logic’ of symbolic
mathematics.” He asserts, “The paramount importance of
this problem can be easily grasped, if we think of the role that
symbolic mathematics has played in the development of modern science since the end of the sixteenth century.” Klein concludes his remarks on “Husserl’s logical researches” by saying
that these researches “amount in fact to a reproduction and
precise understanding of the ‘formalization’ which took place
in mathematics (and philosophy) ever since Viète and
Descartes paved the way for modern science.”
Klein’s qualiﬁcation that Husserl’s logical researches
“amount . . . to” both a “reproduction” and “precise understanding” of the “formalization” in mathematics initiated by
Viète contains the key to my answer to the question. It indicates that Klein, but not Husserl, was aware of the historical
signiﬁcance of these researches. The result of the “formalization” in mathematics referred to here by Klein concerns the
“indeterminacy” of the object of mathematical symbols that I
called attention to earlier, the absence of any direct reference
to both individual objects and their general and universal
qualities that is the mathematical symbols’ most salient characteristic. Because Klein thought Husserl’s logical researches
in Philosophy of Arithmetic amount in fact to the reproduction and precise understanding of the historical genesis of this
formalization, it would follow from this that Klein under
HOPKINS
61
stood that work’s investigation of the relationship between
authentic and symbolic numbers to mirror what his own
research presents as the relationship between the ancient
Greek arithmos and the modern symbolic number.
This being the case, the question of image and original
suggested by my mirror metaphor arises, namely, did
Husserl’s investigations mirror Klein’s or Klein’s Husserl’s?
Here I think chronology is relevant, which would suggest that
what enabled Klein to encounter the presuppositions of both
his own modern as well as the ancient Greek conceptuality
was Husserl’s reluctant discovery in Philosophy of Arithmetic
that the formal conceptual status of symbolic numbers cannot
be rendered intelligible in terms of authentic numbers. By
saying this, however, I want to emphasize in the strongest
terms possible that I am not suggesting what some others
have suggested, namely, that what makes Klein’s comparison
of the conceptuality of the ancient Greek and modern numbers possible is his projection of Husserl’s “concepts” of
authentic and symbolic numbers back into the history of
mathematics.8 On the contrary, I want to suggest and then
develop a much more subtle and more radical claim.
Husserl’s failure provided Klein with the guiding clue that
enabled him to trace and illuminate certain historical dimensions of that very failure. Klein detected an historical transformation of nonsymbolic numbers: an aspect of their conceptuality was transformed into symbolic numbers. This discovery resulted in a more deﬁnitive philosophical account of
both kinds of numbers. Moreover, Klein discovered something of which Husserl had not the slightest inkling, namely,
that the formal conceptuality of symbolic numbers, and symbolic conceptuality in general, cannot be made intelligible on
the basis of theoretical concepts traceable to ancient Greek
science. And that discovery highlighted something more ominous: built into symbolic cognition is the misguided selfunderstanding that evaluates its own cognitive status as the
everincreasing perfection of ancient Greek science’s theoretical aspirations.
�62
THE ST. JOHN’S REVIEW
Klein’s account of the origination of algebra is really the
story of the simultaneous invention of the mathematical symbol, the symbolic formulae it makes possible, and the resultant novel art of symbolic calculation. It is a story with a such
a complex array of intricate subplots and such a diverse cast
of characters that it is not always easy to follow the action,
especially given its approximately twothousandyear time
span. It is also a difﬁcult story to follow, since not only are
there really no good and bad guys to identify with or to dislike, but its beginning as well as its ending is obscure. The
action takes place, for the most part, in the realm of pure
beings and pure concepts, a realm that is invisible to the eyes
and in which all the actors are likewise invisible and therefore, with some justiﬁcation, referred to by many as
“abstract.” Despite its otherworldly aura, it is a story well
worth trying to follow because what it is about is the origin
of the mistaken identity of the very technique that has
enabled mathematical physics and the technology spawned
from it quite literally to transform the world. Unraveling the
plot is an exercise in discovering the true identity of this technique and, in the process, rediscovering something essential
about our relation to the world that the events surrounding
the technique’s origination continue to make it easy for us to
forget.
Turning now to the story: we have already seen that nonsymbolic and symbolic numbers are key players in Klein’s
tale. Guided by our discussion of Husserl’s account of them,
we are in a position to see what it means to say with Klein
that the nonsymbolic numbers of the ancient Greeks are not
concepts, let alone abstract concepts: being deﬁnite amounts
of deﬁnite things, arithmoi were initially understood not only
to be inseparable from things but also to be what is responsible for the wellordered arrangement proper to all their parts
and qualities. In other words, they were understood by the
Pythagoreans as the very being of everything that is. To be is
to be countable, and because to be countable each thing has
to be one, the one was very important, as was the odd and the
HOPKINS
63
even, since whatever is counted ends up being odd or even.
The one or the unit (monas), as something without which
counting is impossible, is therefore the most basic principle
(archê) of arithmos. The odd and the even, which order the
arithmos of everything countable, insofar as it has to be one
or the other, manifest the ﬁrst two kinds (eidê) of arithmoi.
Moreover, since the even can be divided without ever arriving at a ﬁnal arithmos, while the odd cannot be divided
evenly at all, because a one is always left over, these two kinds
are understood, respectively, as unlimited and limit.
It is important to note here three things, according to
Klein: (1) the arithmoi are inseparable from that which is
countable; (2) the most basic principle as well as the kinds of
arithmoi are not themselves arithmoi. In other words, they
are not numerical if by numerical we understand, as the
ancient Greeks did, number to be a deﬁnite amount of deﬁnite things; and (3) the arithmoi, being inseparable from what
is countable, are not abstract entities, and because they are
different from both their most basic principle and their kinds,
they are not even remotely “conceptual,” assuming for the
moment that it is even appropriate to use this adjective to
refer to the archê and eidê of arithmoi. It is important to note
these three things because, on Klein’s telling, no matter how
much the ancient Greek characterization of the mode of
being of arithmoi changes in what become, in Plato and
Aristotle, the two paradigmatic ways of its understanding it,
these three things about the arithmoi remain constant. In
Klein’s words, “All these characterizations stem from one and
the same original intuition [Anschauung], one oriented to the
phenomenon of counting.”9
While the discovery of incommensurable magnitudes
brought to an end the Pythagorean dream of a world in which
being counted was identical with being measured, Plato’s
positing of the invisible, indivisible, and therefore sensibly
pure mode of being of the archê of arithmoi brought into
being another dream, the dream of what Klein calls an arithmological ordering of the eidê responsible for arithmoi as
�64
THE ST. JOHN’S REVIEW
well as anything else that is a being. As Klein relates it, Plato
realized that the Pythagorean account of the archê of arithmoi, the one or the unit, as something that is quite literally
inseparable from the sensible things that are countable on its
basis, presented an obstacle to understanding the true relation
of arithmoi to the soul when it counts. This is the case
because Plato must have noticed that even before, the soul
counts each one of the deﬁnite things that it perceives, it
already has some understanding of the arithmoi it employs in
arriving at the arithmos of such things. According to Klein,
Plato must have thought this possible because prior to counting sensible things the soul has available to it arithmoi that are
deﬁnite amounts of intelligible (noêta) units, intelligible in the
sense that they cannot be seen with the eyes, cannot be
divided like the things seen with the eyes, and cannot be
unequal like the things seen by the eyes. Just like the
Pythagoreans’ sensible arithmoi, these intelligible arithmoi
are either odd or even, though unlike the Pythagorean eidê,
those belonging to intelligible arithmoi are likewise intelligible, and thus cannot be seen with the eyes.
Now it has to be stressed here that absolutely nothing is
either abstract or general about Plato’s intelligible arithmoi.
They are not abstract because they are not lifted off anything.
They are not general because they are precisely deﬁnite
amounts of deﬁnite things, albeit in this case the things are
noeta. Moreover, they are not general because, just like the
Pythagorean arithmoi, they are not concepts: each arithmos is
a deﬁnite whole, the unity of which is exactly so and so many
intelligible units. What allows intelligible numbers to be used
in the counting of anything whatever is what Plato’s Socrates
never tired of pointing out to his interlocutors, namely, that
the true referents of our speech, in counting off amounts of
things or in anything else, are not sensible but intelligible
beings. Thus in counting what the soul is really aiming at
when it counts off in speech, the deﬁnite amounts of deﬁnite
sensible things, things that in being counted are treated as sensible units, are deﬁnite amounts of intelligible units. Indeed, it
HOPKINS
65
is precisely this state of affairs that allows the soul to count
anything that happens to be before it, since the true units of
its counting are not those that can be seen but precisely those
that can only be thought.
Plato’s way of explaining how the availability of intelligible arithmoi to the soul enables it to count anything whatever
means that these intelligible units are manifestly unlike the
units in Husserl’s authentic numbers. Plato’s intelligible units
explain the ability of arithmoi to count anything whatever
because they, and not the “whatever,” are each arithmos’ true
referent. In other words, for Plato—and for that matter, for
all the ancient Greeks including Aristotle10—there is no such
concept of any thing, or any object, or any being whatever.
Such a concept, as we have seen, is explained by Husserl in
terms of an abstracting activity of the mind that is so powerful it is powerful enough to create a concept so general that
literally anything whatever (Etwas überhaupt) can “fall under
it.” This is to say, for Husserl the mind is able to create a formal concept that has absolutely no determinate reference to
any individual thing in the world or to the general and universal qualities of such things. Klein’s point, and in my judgment the point is the fulcrum upon which the story told in his
math book pivots, is that until Viète invented algebra, the
power behind this abstraction—what Klein calls a symbol
generating abstraction—was something that the world had
never seen before.
Husserl’s concept of the units in authentic numbers is
therefore modern, which is something Klein was not only no
doubt aware of, but it is also no doubt the reason why Klein
silently passed over in silence Husserl’s investigations of the
logic of symbolic mathematics. Indeed, Klein’s book also
bypassed any reference to Husserl’s concept of intentionality
when he articulated the difference between the conceptuality
of nonsymbolic and symbolic numbers in terms of the mediaeval concept of intentionality, and, again, no doubt it was for
the same reason: a part of the composition of Husserl’s concept of intentionality was already determined by the very for
�66
THE ST. JOHN’S REVIEW
mality that Klein was investigating the origin of, in the shift
from nonsymbolic to symbolic numbers. Once this is realized, Klein’s use of the medieval concepts of ﬁrst and second
intentional objects, rather than Husserl’s concepts of straightforward and categorial intentional objects, to talk about the
difference in the mode of being of nonsymbolic and symbolic numbers makes perfect sense.
The transformation of an aspect of the ancient Greek
arithmoi into the modern, symbolic numbers, however, does
not make perfect sense. That is, Klein’s account of the shift in
the referent of ancient Greek and modern symbolic numbers
is something that does not, and indeed cannot, render theoretically transparent the philosophical meaning of either the
shift or the different numbers in question. The characterization of ancient Greek arithmoi in terms of their direct
encounter with either sense perceptible objects encountered
in the world or with intelligible objects encountered in the
soul—what Klein reports the medievals called ﬁrstintentional objects—does not explain what such arithmoi are, in
the precise sense of how it is that the different arithmoi render intelligible the different deﬁnite amount that characterizes each arithmos. Indeed, Klein never suggests that the concept of a ﬁrstintentional object can do this. Likewise, the
characterization of symbolic numbers as having their referent
in the mind’s conception, a conception the medievals called
the object of a second intention, does not render theoretically
perspicuous what symbolic numbers are, either. To characterize symbolic numbers as pertaining to that aspect of arithmoi
that concerns the “how many” of something, while at the
same time no longer pertaining to its exact determination that
each arithmos brings about, does not clarify theoretically
what a symbolic number is. In other words, pointing out that
the conceptuality of symbolic numbers disregards both the
units of the something whose deﬁnite amount it is the
province of arithmoi to determine and the exact amount of
these units that each arithmos registers, does not explain
what a symbolic number is—and neither does Klein’s account
HOPKINS
67
of how more than this shift is involved in its conceptuality.
Klein afﬁrms that what is required in order for us to be dealing with a symbolic number is that the secondintentional
mode of being of the “how many,” which brings into the
world for the ﬁrst time a formal concept because it is now
shorn of any connection to either ﬁrstintentional objects or
the exact determination of their amount, be expressed in a
senseperceptible sign that is grasped by the mind as the object
of a ﬁrst intention.
What Klein’s talk of objects of ﬁrst and second intentions
accomplishes is to call attention to something that nobody
else in the twentieth century had seen, namely, that the invention of symbolic cognition represents nothing less than a
reversal of the premodern relationship between concepts
and objects: what were concepts for the ancient Greeks are
now objects and what were objects for them are now concepts. Modern “theoretical thinking,” being symbolical, is
thus necessarily blind to this reversal. Ancient “theoretical
thinking,” not being symbolical, is likewise necessarily blind
to it. It does not follow from this, however, that thinking per
se must remain blind to it. On the contrary, for a thinking that
is on its guard against falling victim to the most shameful
ignorance, that is, to thinking it knows what, in truth, no
mortal can know, not only is the shift in conceptuality articulated by Klein something that can be seen, but once seen, it
is something that the soul’s phronesis can never forget.
If we had more time, I would continue my remarks by
calling attention to what I think is the key to Klein’s account
of how something like a symbol generating abstraction was
able to come into the world, namely on the basis of Viète’s,
Stevin’s, Descartes’s, and Wallis’s formulating the method of
an art that permits calculation with what the ancient Greeks
characterized not as arithmoi but as their eidê. And, indeed, I
would call attention to the fact that, for Klein, with this not
only do the true objects of mathematics become conceptual,
that is, formal, but also, such concepts at the same time
become numerical. Finally, I would call attention to the par
�68
THE ST. JOHN’S REVIEW
allel Klein draws between the breaking of the bounds of the
intelligibility proper to the logos that is the result of Plato’s
formulation of the eidê of beings as having an arithmological
structure and the similar breaking of these bounds by Viète’s
numerical formulation of the symbolic calculation with the
species of numbers in symbolic cognition,11 and then ask the
following question: Do both of these attempts to comprehend beings theoretically transcend the limits of what can be
spoken of intelligibly for the same simple reason, namely, that
their theories presuppose a knowledge of what no mortal can
claim in truth really to know, namely that most wondrous gift
of the gods to humans—one and number?
Before I conclude, I would like to return very brieﬂy to
Klein’s memorial essay on Husserl that I mentioned earlier
and to the matter of his articulation in that essay of the philosophical meaning of his mathematical investigations in terms
of Husserl’s of last writings, the socalled crisis texts. It is
important to draw attention here to the chronology of Klein’s
mathematical investigations and Husserl’s last writings,
because it is only in these writings that Husserl connects the
two themes that had already informed Klein’s earlier investigations of the history of mathematical concepts. Prior to
1936, when the second part of Klein’s investigations were
published, Husserl therefore had not yet recognized what
Klein had already recognized, and indeed investigated extensively, guided as I have suggested by Husserl’s ﬁrst—and
failed—investigation of the relationship between nonsymbolic and symbolic numbers. Klein recognized the connection
between the philosophical meaning of mathematical concepts
and the history of their origination.
Notes
Edmund Husserl, Philosophie der Arithmetik, ed. Lothar Eley,
Husserliana XII (The Hague: Nijhoff, 1970), 245; English translation: The Philosophy of Arithmetic, trans. Dallas Willard
(Dordrecht: Kluwer, 2003).
1
HOPKINS
69
Edmund Husserl, Introductions to the Logical Investigations, ed.
Eugen Fink, trans. Philip J. Bossert and Curtis H. Peters (The
Hague: Martinus Nijhoff, 1975), 35. German text, “Entwurf einer
‘Vorrede’ zu den ‘Logischen Untersuchungen’ (1913),” Tijdschrift
voor Philosophie (1939): 106133, here 127.
2
Jacob Klein, Greek Mathematical Thought and the Origin of
Algebra, trans. Eva Brann (Cambridge, Mass.: M.I.T. Press, 1969;
reprint: New York: Dover, 1992). This work was originally published in German as “Die griechische Logistik und die Entstehung
der Algebra” in Quellen und Studien zur Geschichte der
Mathematik, Astronomie und Physik, Abteilung B: Studien, vol. 3,
no. 1 (Berlin, 1934), pp. 18–105 (Part I); no. 2 (1936), pp.
122–235 (Part II). Hereinafter referred to as GMTOA.
3
See the following: Hiram Caton, who claims that “Klein projects
Husserl back upon Viète and Descartes,” (Studi International Di
Filosophia, Vol 3 (Autumn, 1971): 222226, here 225; J. Phillip
Miller, who writes “Although Husserl’s own analyses move on the
level of a priori possibility, Klein’s work shows how fruitful these
analyses can be when the categories they generate are used in studying the actual history of mathematical thought,” (Numbers in
Presence and Absence [The Hague: Martinus Nijhoff, 1982], 132;
Joshua Kates’ account is more circumspect, as he notes “[i]t is difﬁcult to capture adequately . . . how much of Klein’s understanding
of Greek number is already to be found in Husserl, despite the
important differences between them,” (“Philosophy First, Last, and
Counting: Edmund Husserl, Jacob Klein, and Plato’s
Arithmological Eidê,” (Graduate Faculty Philosophy Journal, Vol.
25, Number 1 (2004), 6597, here 94.
4
This is the literal translation of the word in question here,
“Begrifﬂichkeit,” which is rendered for the most part as “intentionality” in the English translation GMTOA. Because of this, the point
I make below about the “concept of number” and “number concepts” will be more familiar to readers of Klein’s original German
text.
5
Jacob Klein, “Phenomenology and the History of Science,” in
Philosophical Essays in Memory of Edmund Husserl, ed. Marvin
Farber (Cambridge, Mass.: Harvard University Press, 1940),
143–163; reprinted in Jacob Klein, Lectures and Essays, ed. Robert
6
�70
THE ST. JOHN’S REVIEW
B. Williamson and Elliott Zuckerman (Annapolis, Md.: St. John’s
Press, 1985), 65–84, here 70.
Edmund Husserl, “The Origin of Geometry,” in The Crisis of
European Sciences and Transcendental Phenomenology, trans. David
Carr (Evanston, Ill.: Northwestern University Press, 1970). The
German text was originally published in a heavily edited form by
Eugen Fink as “Die Frage nach dem Ursprung der Geometrie als
intentionalhistorisches Problem,” Revue internationale de
Philosophie I (1939). Fink’s typescript of Husserl’s original, and signiﬁcantly different, 1936 text (which is the text translated by Carr)
was published as Beilage III in Die Krisis der europäischen
Wissenschaften und die transzendentale Phänomenologie. Eine
Einleitung in die phänomenologische Philosophie, ed. Walter
Biemel, Husserliana VI (The Hague: Nijhoff, 11954, 21976).
Edmund Husserl, “Die Krisis der europäischen Wissenschaften und
die transzendentale Phänomenologie. Eine Einleitung in die
phänomenologische Philosophie,” Philosophia I (1936), (the text of
this article is reprinted as §§ 1–27 of the text edited by Biemel).
7
8
See note 4 above.
9
GMOT, 54.
Aristotle’s dispute with Plato over the mode of being of the arithmoi studied by the discipline of mathematics was about the origin
of the units that they are the deﬁnite amounts of, and not whether
theoretical arithmoi are deﬁnite amounts of units.
10
“As Plato had once tried to grasp the highest science ‘arithmologically’ and therewith exceeded the bounds set for the logos (cf.
Part I, Section 7C), so here [in Viète’s invention of symbolic calculation] the ‘arithmetical’ interpretation leads to . . . the conception
of a symbolic mathematics,” the implication being, of course, that
such a conception exceeds the same bounds as did Plato’s attempt
to grasp dialectic in terms of the arithmoi eidetikoi (GMOT, 184).
11
71
Words, Diagrams, and Symbols:
Greek and Modern
Mathematics or “On the Need
To Rewrite The History of
Greek Mathematics” Revisited
Sabetai Unguru
Mademoiselle de Sommery, as Stendhal tells us in De
l’Amour, was caught “en ﬂagrant delit,” i.e., in ﬂagranti, by
her lover, who was shaken seeing his “amante” bedding down
another man. Mademoiselle was surprised by her lover’s
angry reaction and denied brazenly the event. When he
protested, she cried out: “Oh, well, I can see that you no
longer love me, you would rather trust your eyes than what I
tell you.”
The historian of mathematics should behave like
Mademoiselle’s lover: believe his eyes and not what mathematiciansturnedhistorians tell him about the texts he studies. There are optical illusions, it is true, but they are to be
preferred to the illusionary mental constructs of the mathematical historians. A text is a text is a text. Moreover,
metaphors aside, not everything is a text and it behooves the
cultural critic, and surely the historian, to relate only to written records as texts. Furthermore, texts are about something
deﬁnite and not every conceivable interpretation suits them
all. The Conica is about conic sections not about women. It
deals with three (or four) kinds of lines obtained by cutting a
Sabetai Unguru is Professor Emeritus and former Director of the Cohen
Institute for the History and Philosophy of Science and Ideas at the
University of TelAviv. He is the author of a controversial study with the title
“On the Need to Rewrite the History of Greek Mathematics,” as well as of a
twovolume introduction to the history of mathematics. He is (with Michael
Fried) the coauthor of Apollonius of Perga’s Conica: Text, Context, Subtext.
�72
THE ST. JOHN’S REVIEW
conic surface with a plane, lines baptized by Apollonius as
parabolas, ellipses, hyperbolas, and opposite sections. It does
not deal with a classiﬁcation of women according to their
degree of perfection, as seen by a male chauvinist pig, hiding
his true intentions behind the cloud of mathematical jargon.
Texts are made of words, sometimes accompanied by
illustrations; in the case of Greek geometry, the texts are
made of words and diagrams, C’est tout. And the diagram is,
in a very deﬁnite sense, the proposition. The words accompanying it serve to show how the diagram is obtained; they
provide us the diagram in statu nascendi. But in principle it
would be possible to supply the absent words to an extant
diagram, though I doubt the possibility of understanding a
reasonably sophisticated geometrical text in the absence of
diagrams. If there is a manipulative aspect to Greek geometry,
and I think there is, it resides in the construction, the bringing into being of the diagram, while the steps of the process
are supplied by the words accompanying the diagram, typically in the kataskeue (construction).
There are no true symbols in a Greek mathematical text.
What looks like symbols to the untrained modern eye are
actually proper names for identifying mathematical objects.
They are not symbols, and cannot be manipulated, as algebraic symbols are. Even in Diophantus, which is a late and
special case, his socalled symbols are actually verbal abbreviations, making his Arithmetica an instance of syncopated, not
symbolic, algebra, in Nesselmann’s tripartite division (Die
Algebra der Griechen).
In modern, postCartesian mathematical texts, on the
other hand, there are words and diagrams and symbols, but
the actual necessity of the ﬁrst two ingredients is minimal,
serving heuristic, pedagogical, and rhetorical needs that can
be dispensed with, leaving the text in its symbolic nakedness.
It is no exaggeration to see modern mathematics as symbolic,
while ancient mathematics cannot be seen historically in symbolic terms. This being the case, interpretations of ancient
mathematical texts relying on their symbolic transmogriﬁca
UNGURU
73
tion are inadequate and distorting. They are ahistorical and
anachronistic, making them unacceptable for an understanding of ancient mathematics in its own right.
An ancient text—mathematical, philosophical, literary, or
whatever—is the product of a culture foreign to ours, whose
concerns, values, aims, standards, ideals, etc., are, as a rule, as
alien to ours as can be. Though, inescapably, all history is retrospective history, the approach described and decried by
Detlef D. Spalt, in his Vom Mythos der mathematischen
Vernunft (1981), as Resultatismus, or, “Orwellsche 1984Geschichtsschreibung für den grossen Bruder Vernunft,” that
takes its bearings and criteria from what it sees as the modern
outcome of a lengthy, linear, and necessary evolution of concepts and operations, looking always back at the past in light
of its modern offspring, is necessarily a highly distorting
approach, since it adopts unashamedly the perspective of the
present to (in this order) judge and understand the past.
Taken at face value, Percy W Bridgman’s statement that the
.
past has meaning only in terms of the present is simply not
true. The historian’s stance is rather the opposite: the present
has meaning only in terms of the past. Prima facie and on
principled grounds therefore, an interpretation of any written, reasonably extensive document belonging to an ancient
culture that results in its totally unproblematic and absolute
assimilation to our own is suspect. That this is so is, more or
less, acceptable when it comes to cultural artifacts other than
mathematical ones, which seem to enjoy the privilege of perdurablility and universality. The immunity from cultural
speciﬁcity that mathematical truths command stems from the
prevailing view that their outward appearance—their packaging, as it were—and their purely mathematical content—the
packaged merchandise, as it were—are neutral, unrelated,
and mutually independent items. It is a calamitous view and
the root of all evil in the historiography of mathematics.
But there is another entrance into an ancient text, mathematical or not, one that does no violence to it, that does not
break the inviolable unity of form and content and then enter
�74
THE ST. JOHN’S REVIEW
victoriously through the shambles it created, with the claim
that the text is now understood; it is rather an ingress that
accepts willingly and respectfully the unity of the text without pulverizing its inseparable aspects, bringing to its understanding both critical acumen and full acceptance of its outward appearance. This is the historical approach. The mathematical and historical approaches are antagonistic. Whoever
breaks and enters typically returns from his escapades with
other spoils than the peaceful and courteous caller.
Let me be more speciﬁc. Faced with an ancient mathematical text, the modern interpreter has an initial choice.
First is the mathematical approach. It consists of two steps:
(1) try to ﬁnd out how one would do it (solve the problem,
prove the proposition, perform the construction, etc.) and
then (2) attempt to understand the ancient procedure in light
of the answer to step (1). Instead of this preeminently mathematical approach, however, the modern interpreter can
refuse to decipher the text by appealing to modern methods,
using for its understanding only ancient methods available to
the text’s author. This is the historical approach. Needless to
repeat, the spoils of interpretation differ according to the two
approaches followed. The longstanding traditional approach
has been the mathematical, though in the last three decades
or so the historical approach is gaining increasingly more and
more ground and, at least in the domain of ancient Greek
mathematics, seems to be now the prevailing one. What
seems certain is that in practice no compromise is possible
between the mathematical and historical methodological
principles. Adopting one or the other has fateful consequences for one’s research, effectively determining the nature
of the results reached and the tenor of the inferences used in
reaching them.
What I am saying, then, is that despite the numerous and
varied styles of writing the history of mathematics throughout the centuries, it is possible to group all histories of mathematics into two broad categories, the “mathematical” and
the “historical.” The former sees mathematics as eternal, its
UNGURU
75
truths unchanging and unaffected by their formal appearance,
and sees the mathematical kernel of those truths as being
independent of their outward mode of expression; the latter
denies this independence and looks upon past mathematics as
an unbreakable unity between form and content, a unity,
moreover, that enables one to grasp mathematics as a historical discipline, the truths of which are indelibly embedded in
changing linguistic structures. It is only this approach that is
apt to avoid anachronism in the study of the mathematics of
other eras.
To make this a historical talk, what is needed are speciﬁc,
historical examples, supporting the preceding generalizations. I have offered numerous such examples in my published work, most recently in the book I published with
Michael Fried, Apollonius of Perga’s Conica: Text, Context,
Subtext (2001). However, I would like now to enrich my
offerings by drawing illustrations from historical sources less
drawn upon in the past. One such source is Euclid’s Data.
In one of the attacks launched against a notorious article
of 1975, “On the Need to Rewrite the History of Greek
Mathematics,” Hans Freudenthal argues that, had its author
been aware of the existence of the Data, he “would never
have claimed there were no equations in Greek geometry.”
For Freudenthal, and not only for him, the Data is a “textbook on solving equations.” He summarizes the 94 propositions contained therein in a succinctly and strikingly epigrammatic statement: “Given certain magnitudes a, b, c and
a relation F(a, b, c, x), then x, too, is given.” But the fact
remains that Greek geometry contained no equations. One
cannot ﬁnd even one equation in the entire text of the Data.
Proof (as the Hindu mathematician would say): “Look!”
Unless one has at his disposal the algebraic language and the
capacity to translate into it, it is impossible to sum up this little treatise of rather varied content as offhandedly as
Freudenthal has done, Indeed, had Euclid at his disposal
Freudenthal’s functional notation, it is rather easy to infer
�76
THE ST. JOHN’S REVIEW
that he would not have needed 94 propositions to get his
point across.
Each case in Euclid’s Data is unique, having its own
method of analysis, and none is subsumable under or
reducible to other cases, though, of course, later propositions
rely on earlier ones. Thus, “The ratio of given magnitudes to
one another is given” (proposition 1) and “If a given magnitude have a given ratio to some other magnitude, the other is
also given in magnitude” (proposition 2)—to use perhaps the
simplest illustration possible—are not for Euclid both
instances of “Given a, b, c and y=F(a, b, c, x), x is also given,”
but are two different problems, interesting in their own right,
having their own solutions. Of course, Freudenthal’s description is mathematically correct. Historically, however, it is
wanting. Heath is much more to the point when he says:
The Data…are still concerned with elementary
geometry [my italics], though forming part of the
introduction to higher analysis. Their form is that
of propositions proving that, if certain things in a
ﬁgure [my italics] are given (in magnitude, in
species, etc.), something else is given. The subjectmatter is much the same as that of the planimetrical books of the Elements, to which the Data are
often supplementary.
This is what the Data is, not a textbook on solving equations, but a treatise presenting another approach to elementary geometry—other than that of the Elements, that is.
As an example of this characterization of the Data, I shall
present proposition 16, in the new translation of C. M.
Taisbak [Euclid’s Data or The Importance of Being Given,
(Copenhagen, 2003)]:
If two magnitudes have a given ratio to one
another, and from the one a given magnitude be
subtracted, while to the other a given magnitude
be added, the whole will be greater than in ratio to
the remainder by a given magnitude (p. 75).
77
UNGURU
First we must clarify the meaning of the expression “greater
than in ratio...by a given.” Deﬁnition 11 of the Data reads:
A magnitude is by a given greater than in ratio to a
magnitude if, when the given magnitude be subtracted, the remainder has a given ratio to the
same. (p. 35)
The meaning of this deﬁnition is, according to Taisbak, as follows (p. 57): “M is by the given G greater than the magnitude
L which has to N a given ratio;” in other words, M=L+G
and L:N is a given ratio.
Back to the proof of prop. 16:
Let two magnitudes AB, CD have a given ratio to one
another. From CD let the given magnitude CE be subtracted
and to AB let the given magnitude ZA be added.
Then, the whole ZB is greater than in ratio to the remainder
DE by a given magnitude.
Z
C
A
E
H
D
B
Now, since the ratio AB:CD is given and AH can be obtained,
by Dt. 4, from AH:CE::AB:CD, i.e., AH:CE is also given, it
follows that CE is also given. Hence, AH is given (by Dt. 2).
But AZ is given; therefore the whole ZH is given (by Dt. 3).
Since AH:CE::AB:CD, the ratio HB:ED of the remainders is
also given, by V 19 and Def. 2, i.e., HB:ED::AB:CD. But HZ
.
is given; therefore, ZB is by a given greater than in ratio to
ED (Def. 11), Q.E.D.
No algebra appears here, and although the language of
givens, the idiosyncratic concept “by a given greater than in
ratio” and the sui generis concatenation of inferences burden
the understanding, the proposition is clear and rather simple.
It is graspable as it stands, without any appeal to foreign
�78
THE ST. JOHN’S REVIEW
tools. And yet, Clemens Thaer, in his Die Data von Euklid
(1962), proceeds as follows to its clariﬁcation:
Let a:b=k , then the proposition claims that if x=ky, then
(x+c)=k(yd)+(c+kd), which is, of course, correct mathematically, but this blatantly algebraic procedure betrays the
Data. It is a betrayal since ratios in Greek mathematics are
not real numbers, i.e., the initial substitution a:b=k is not
kosher. Though there is some controversy about the status of
ratios in the Elements, with respect to their being two, or
fourplace relations, their status in the Data is uncontroversial: a ratio P:Q is an individual item “however impalpable. A
and B are magnitudes, most often (and least problematically)
understood to be line segments; one may think of them as
positive real numbers, that is as lengths of line segments, while
remembering that the Greek geometers could not think like
that, for want of such numbers” (Taisbak, Euclid’s
Dedomena, p. 32). Taisbak claims that distorting “clariﬁcations” like the one above characterize all of Thaer’s algebraizations. He goes on to say:
About the following four theorems (Dt 1720) he
maintains that they prove that all linear transformations, {ax+b a, b ℜ} form a group (‘dass die
ganzen linearen Substitutionen einer
Veraenderlichen eine Gruppe bilden’). I am not sure
I understand what he means to say, and the Data
certainly does not help me,—so probably Euclid
would not understand either. (p. 77)
Let us take our next example from Archimedes. Against
Freudenthal’s assertion, Archimedes’ works are not
“instances of algebraic procedure in Greek mathematics.”
Heath’s edition of The Works of Archimedes (1897) is “in
modern notation.” It is faithful only to the disembodied
mathematical content of the Archimedean text, but not to its
form. And this is crucial. If one abandons Archimedes’ form
and transcribes his rhetorical statements by means of algebraic symbols, manipulating and transforming the latter, then
UNGURU
79
clearly “the algebraic procedure” appears. But this procedure
itself is not “in Greek mathematics.” It is a result, as
Freudenthal himself states it, of “replacing vernacular by artiﬁcial language, and numbering variables by cardinals, a quite
recent mathematical tool.” Indeed! Archimedes’ text is
anchored securely in the terra ﬁrma of Greek geometry. If one
is not willing to compress wording, to replace “vernacular”
by artiﬁcial language, to introduce variables and number
them by cardinals, and to apply all the other technical tricks
which are “quite recent mathematical tools,” then
Archimedes’ proof of Proposition 10 of Peri Helikon is geometric, not algebraic. This was discerned in a curious way
even by Heath, who justiﬁed his algebraic procedure and the
use of the symbols , “in order to exhibit the geometrical character of the proof” (p. 109, my italics).
Dijksterhuis himself in his Archimedes said: “In a representation of Greek proofs in the symbolism of modern algebra it is often precisely the most characteristic qualities of the
classical argument which are lost, so that the reader is not sufﬁciently obliged to enter into the train of thought of the original.” So let us oblige ourselves to enter into the train of
thought of the original by having a look at the 10th proposition of Peri Helikon.
I shall give you the full enunciation and then set at its side
Heath’s algebraic variant:
If any number of lines, exceeding one another by
the same magnitude, are set one after the other,
the excess being equal to the smallest, and if one
takes other lines in the same number, each of
which is equal to the magnitude of the greatest of
the ﬁrst lines, [then] the squares on the lines equal
to the greatest, augmented by the square on the
greatest, and by the rectangle the sides of which
are the smallest line and the sum of all lines
exceeding one another by the same magnitude are
equivalent to thrice the sum of the squares on the
�THE ST. JOHN’S REVIEW
80
lines which exceed one another by the same
amount.
The enunciation is geometrical and it is accompanied by simple, linear diagrams.
And here is Heath’s enunciation:
If A1, A2, A3, ..., An be n lines forming and ascending
arithemetical progression in which the common
difference is equal to A1, the least term, then
(n+1) An2+A1 (A1+A2+...+An)=3(A12+A22+...+An2)........
…the result is equivalent to
12+22+32+...+n2=n(n=1)(2n+1)
6
The proof is not difﬁcult, but very long (more than three
pages in Mugler’s edition of Archimedes’ works, if one
includes the porism), and I think we can dispense with it, but
not before pointing out the fact that it lies squarely within the
realm of traditional Greek geometry, relying on Elements 2.4,
it is true, which, as is well known, belongs to the socalled
geometric algebra, but which, as I have argued elsewhere, is
strictly geometric; additionally, the porism relies on Elements
6.20.
As in the enunciation, Archimedes formulates consistently
his statements in terms of lines squares, and rectangles, which
he manipulates à la Grecque, and his diagrammatic notation
is not at all perspicuous to a modern eye, making his transformations opaque, or at least cloudy, to a mind spoiled by
the easy mechanics of algebraic manipulations and their
immediate visual transparency. This makes following his elementary inferences quite difﬁcult and almost forces upon the
reader recourse to algebraic notation. Such a procedure, how
UNGURU
81
ever, easy, pellucid, and revealing as it is, is not legitimate historically.
What I have said about the proposition applies in its
entirety to the porism following it, out of which Heath makes
two corollaries!
Let us, again, limit ourselves to the enunciations of
Archimedes and Heath, which substantiate our characterization. Archimedes ﬁrst:
It is manifest from the preceding that the sum of
the squares on the lines equal to the greatest is
inferior to the triple of the sum of the squares on
the lines exceeding one another by the same
magnitude, because, if one adds to it [the former]
some [magnitudes], it becomes that triple, but
that it is superior to the triple of the second sum
diminished by the square on the greatest line,
since what is added [to the ﬁrst sum] is inferior
to the triple square of the greatest line. It is for
this [very] reason that, when one describes similar
ﬁgures on all the lines, both on those exceeding
one another by the same magnitude, as well as
on those which are equal to the greatest line, the
sum of the ﬁgures described on the lines which
are equal to the greatest line is inferior to the
triple of the sum of the ﬁgures described on the
lines exceeding one another by the same magnitude, but it is superior to the triple of the second
sum, which is diminished by the ﬁgure constructed
on the greatest line, because similar ﬁgures are in
the same ratio as the squares [on their sides].
�THE ST. JOHN’S REVIEW
82
Now Heath:
Cor. 1. It follows from this proposition that
nAn2<3(A12+ A22+...+An2), and also that
nAn2<3(A12+ A22+...+An12).
Cor. 2. All the results will equally hold if similar
ﬁgures are substituted for squares.
The differences between Archimedes and Heath are blatant.
Faithfulness to the Archimedean way of doing things
demands, therefore, the rejection of an edition of his works
“edited in modern notation.” There is no escape for the historian but to take texts at their face value.
One last example I shall draw from Diophantus’s
Arithmetica. Diophantus is an exception in the long tradition
of Greek mathematics, which is largely geometric. Living
probably in the third century A.D., his preserved work is an
instance, the only one of its kind in Greek mathematics, of
what we call algebra. It is a sui generis algebra, however, in
which the notations are properly nonexistent, except for the
unknown, arithmos, which is itself most likely an abbreviation, and a few abbreviations for powers of the unknown and
for the unit, monas. That is all. This is what makes his rhetorical algebra syncopated, according to Nesselmann’s classiﬁcation. It is an algebra in which there is a total lack of true,
operative symbols, including symbols for operations, relations, and the exponential notation, with the exception of a
symbol for subtraction, in which the exceptional skill of
Diophantus enables him somehow to overcome with great
dexterity the builtin drawbacks of his Arithmetica. This
Greek algebra, however, is most emphatically not geometric
algebra. If anything, it is conceptually a far away relation to
what has been traditionally called Babylonian algebra, perhaps not even this, in light of the recent researches of Jens
UNGURU
83
Hoyrup, who identiﬁes the geometrical roots of the
Babylonian recipes [Lengths,Widths, Surfaces, (Springer,
2002)].
Whatever it is, it is largely a collection of problems, to be
solved by means of skilful guesses and less by systematic
methods (though one also ﬁnds methods, for example, a general method for the solution of what we call determinate
equations of the second degree and another for double equations of the second degree), involving speciﬁc known numbers. It is not a book of propositions to be proved, but rather
of exercises to be solved, leading mostly to simple determinate and indeterminate equations, by means of which one
ﬁnds the required numbers, which are always rational, quite
often nonintegral. Still, despite what I just said, one ﬁnds in
the treatise also some “porisms” and other propositions in the
theory of numbers, which proved themselves inﬂuential in
the history of the theory of numbers, especially in Fermat’s
work. Thus, Diophantus knew that no number of the form
8n+7 can be the sum of three squares, and that for an odd
number, 2n+1, to be the sum of two squares, n itself must not
be odd, which means that no number of the form 4n+3 or
4n1 can be the sum of two squares.
Now, some simple illustrations from the Arithmetica, to
get the ﬂavor of the true Algebra der Griechen:
Problem 1. To divide a given number into two
numbers, the difference of which is known.
Let the given number be 100, and let the difference be 40 monads; to ﬁnd the numbers.
Let us assume that the smaller number is 1 arithmos; hence, the greater number is 1 arithmos and
40 units. Therefore, the sum of the two numbers
becomes 2 arithmoi and 40 units. But the given
sum is 100 units; hence, 100 units are equal to 2
arithmoi and 40 units. Let us subtract the like
from the like, that is, 40 units from 100 and, also,
�THE ST. JOHN’S REVIEW
84
the same 40 units from the 2 arithmoi and 40
units. The two remaining arithmoi equal 60 units
and each arithmos becomes 30 units.
Let us return to what we assumed: the smaller
number will be 30 units, while the greater will be
70 units, and the validation is obvious.
And here is Heath’s faithful version of the solution:
Given number is 100, given difference 40.
Lesser number required x. Therefore
2x+40=100
x=30.
The required numbers are 70, 30.
Problem 2. Diophantus:
It is necessary to divide a given number into two
numbers having a given ratio.
Let us require to divide 60 into two numbers in
triplicate ratio.
Let us assume that the smaller number is 1 arithmos; hence, the greater number will be 3 arithmoi,
and thus the greater number is thrice the smaller
number. It is also necessary that the sum of the
two numbers be 60 units. But the sum of the two
numbers is 4 arithmoi; hence 4 arithmoi are equal
to 60 units, and the arithmos is therefore 15 units.
Hence, the smaller number will be 15 units, and
the greater 45 units.
Heath:
Given number 60, given ratio 3:1.
Two numbers x, 3x. Therefore x=15.
The numbers are 45, 15.
UNGURU
85
Finally, an example of what is called indeterminate analysis of
the third degree:
Problem 4.8. Diophantus:
To add the same number to a cube and its side and
make the same.
Let the number to be added be 1 arithmos and the
side of the cube be a certain amount of arithmoi.
Let this amount be 2 arithmoi and it follows that
the cube is 8 cubic arithmoi.
Now if one adds 1 arithmos to 2 arithmoi, they
become 3 arithmoi, while if one adds it to the 8
cubic arithmoi, they become 8 cubic arithmoi and
1 arithmos, which are equal to 27 cubic arithmoi.
Let us subtract 8 cubic arithmoi, and it follows
that the remaining 19 cubic arithmoi will become
equal to 1 arithmos. Let us divide all by the arithmos, and 19 squared arithmoi will be equal to
1 unit.
But 1 unit is a square, and if 19, the amount of
square arithmoi, were a square, the problem would
be solved. But the 19 squares ﬁnd their origin in
the excess by which 27 cubic arithmoi exceed 8
cubic arithmoi; and 27 cubic arithmoi are the cube
of 3 arithmoi, while 8 cubic arithmoi are the cube
of 2 arithmoi. But the two arithmoi are taken by
hypothesis, and 3 arithmoi exceed by one the
amount taken arbitrarily as the side. Therefore, we
are led to ﬁnding two numbers which exceed one
another by one unit, and the cubes of which
exceed one another by a square.
Let one of those numbers be 1 arithmos, and the
other 1 arithmos plus 1 unit. Hence the excess of
their cubes is 3 square arithmoi plus 3 arithmoi
plus 1 unit. Let us set this excess equal to the
square the side of which is 1 unit less 2 arithmoi,
�THE ST. JOHN’S REVIEW
86
and the arithmos becomes 7 units. Let us return to
what we supposed, and one of the numbers will be
7 and the other 8.
Let us now return to the original question and
assume the number to be added to be 1 arithmos,
and the side of the cube to be 7 arithmoi. This
cube will be then 343 cubic arithmoi. Hence, if
one adds the arithmos to each of these last numbers, one will become 8 arithmoi and the other
343 cubic arithmoi plus 1 arithmos. But we
wanted this last expression to be a cube the side of
which is 8 arithmoi; hence 512 cubic arithmoi are
equal to 343 cubic arithmoi plus 1 arithmos, and
1
the arithmos becomes 13.
Returning to the things we assumed, the cube will
343
7
1
be 2197, the side 13, and the number to be added 13.
I assume you could follow, at least in outline,
Diophantus’s solution procedure, though this is not essential
for my main purpose. (For those of you who could not follow
after all the details, which, as I said, is not really necessary, a
glance at Heath’s, or Ver Eecke’s, Diophantus would clarify
matters.) My purpose is to compare Diophantus to his modern editors. This time, instead of Heath, I shall take, however,
Nesselmann.
Nesselmann:
It is necessary that x3+y=(x+y)3, i.e., 3x2+3xy+y2=1.
Solving for y, we get y=
1
2
2
[3x±√43x ].
4mn
Putting 43x2=(2=m x)2, one gets x= 3n2+m2.
n
Hence, y=
6mn±(m23n2)
3n2+m2
. Taking only the + sign,
87
UNGURU
for y to be positive, it is necessary that m23n2>6mn,
or,
m2 6 m
n
n2
>3, or, ( m 3)2>12, or ( m >3+√12.
n
n
Diophantus’s solution corresponds to m=7, n=1.
Now this is historical faithfulness!
Time to conclude. Historians must take the past seriously.
For historians of science and mathematics, this means taking
texts seriously. How does one do this? By reading them as
they are, in their nakedness, as it were, in the language in
which they were written, without shortcuts and transmogriﬁcation, resulting in their translation into scientiﬁc and mathematical languages which became historically available only
long after they were written. It is, therefore, crucial that the
form of those texts remain inviolable. Without this, anachronism, i.e., historical misunderstanding, becomes rampant and
the resulting interpretation is misinterpretation. As Benjamin
Farrington put it,
History is the most fundamental science, for there
is no human knowledge which cannot lose its scientiﬁc character when men forget the conditions
under which it originated, the questions which it
answered, and the function it was created to serve.
A great part of the mysticism and superstition of
educated men consists of knowledge which has
broken loose from its historical moorings.
In 1975 an article appeared in the Archive for History of
Exact Sciences, arguing for the need to rewrite the history of
Greek mathematics. In one of its many footnotes, namely the
one numbered 126, attention is called to an important book,
Greek Mathematical Thought and the Origin of Algebra. The
footnote in question contains a parenthesis, saying:
[L]et me urge those readers who have a choice and
wish to read [this] highly interesting study to refer
back to the original German articles: somehow the
�88
THE ST. JOHN’S REVIEW
pomposity, stufﬁness, and turgidity of the author’s
style are better accommodated by the Teutonic
cadences than by the more friendly sounds of the
perﬁdious Albion.
This was a nasty and uncalled for remark, attributable to the
author’s hubris.
How pleasantly and embarrassingly surprised must he
have been, then, when, a few weeks after the article’s appearance in the Archive, a postcard from the slighted author of the
book (to whom no reprint was sent!) arrived, saying: “Dear
Dr.,… Thank you very much for your important article in the
Archive.…” In character, you may say. Indeed.
As a belated, and highly inappropriate atonement for that
unsavory footnote, its author would like to ﬁnish this lecture
with a highly pertinent and lengthy quotation from an article
written by the great scholar he so carelessly slighted:
[T]he relation between ancient and modern
mathematics has increasingly become the focus
of historical investigation… Two general lines of
interpretation can be distinguished here. One—
the prevailing view—sees in the history of science
a continuous forward progress interrupted, at
most, by periods of stagnation. On this view, forward progress takes place with ‘logical necessity,’
accordingly, writing the history of a mathematical
theorem or of a physical principle basically means
analyzing its logic. The usual presentations,
especially of the history of mathematics, picture
a rectilinear course; all of its accidents and
irregularities disappear behind the logical
straightness of the whole path.
The second interpretation emphasizes that the
different stages along this path are incomparable.…it sees in Greek mathematics a science totally
distict from modern mathematics.…Both interpretations, however, start from the presentday
UNGURU
condition of science. The ﬁrst measures ancient by
the standard of modern science and pursues the
individual threads leading back from the valid theorems of contemporary science to the anticipatory
steps taken towards them in antiquity.… The
second interpretation strives to bring into relief,
not what is common, but what divides ancient and
modern science. It too, however, interprets the
otherness of ancient mathematics...in terms of the
results of contemporary science. Consequently, it
recognizes only a counterimage of itself in ancient
science, a counterimage which still stands on its
own conceptual level.
Both interpretations fail to do justice to the true
state of the case. There can be no doubt that the
science of the seventeenth century represents a
direct continuation of ancient science. On the
other hand, neither can we deny their differences…above all, in their basic initiatives, in
their whole disposition (habitus). The difﬁculty is
precisely to avoid interpreting their differences
and their afﬁnity onesidedly in terms of the new
science. The issues at stake cannot be divorced
from the speciﬁc conceptual framework within
which they are interpreted.…
We need to approach ancient science on a basis
appropriate to it, a basis provided by that science
itself. Only on this basis can we measure the
transformation ancient science underwent in the
seventeenth century—a transformation unique
and unparalleled in the history of man.…
This modern consciousness is to be understood
not simply as a linear continuation of ancient
επιστημη, but as the result of a fundamental
conceptual shift which took place in the modern
era, a shift we can nowadays scarcely grasp.
89
�THE ST. JOHN’S REVIEW
91
The name of the man who wrote these lines is, as I am sure
you know, Jacob Klein. Yehi Zichro Baruch! May his memory
be blessed!
A Note on the Opposite
Sections and Conjugate Sections
in Apollonius of Perga’s Conica
90
Michael N. Fried
Introduction
To a careful reader of Apollonius of Perga’s Conica, the difference between Apollonius’ view of conic sections and ours
ought to be evident on nearly every page of the work. Yet this
has not always been the case. Indeed, it has not always been
easy to persuade readers that there really is something Greek
about Apollonius’ mathematics. Partly to blame is the seductive power of the algebraic framework in which we study
conic sections today and in which historians of mathematics
have interpreted the book in past years.1 Viewed algebraically,
for example, all of Book 4 of the Conica—the book concerning the number of points at which conic sections can meet—
can be reduced to a single proposition, namely, that a system
of two quadratic equations in two unknowns can have at
most four solutions. This tremendous power allowing one to
obtain results corresponding to ones in the Conica makes it
all too easy to think that Apollonius’ own text, in effect, can
be bypassed and replaced by an updated algebraic version of
it. However, there are things in the Conica that are refractory
to this kind of modernization of the text and show its essentially geometric character. One of these, surely, is that most
peculiar entity in the Conica, the opposite sections; it is they
that I shall turn my attention to in this note. In particular, I
want to show something about status of the opposite sections
in the Conica and show, among other things, why their status,
Michael Fried teaches mathematics education at BenGurion University of the
Negev. He is the author of a translation, with introduction, of Book Four of
the Conics of Apollonius, and is the coauthor (with Sabetai Unguru) of
Apollonius of Perga’s Conica: Text, Context, and Subtext.
�THE ST. JOHN’S REVIEW
92
why their place in the Conica, is different from that of the
conjugate sections.
The Absence of the Opposite Sections from the Modern View
The reason why looking at the opposite sections is a good
way to refocus on Apollonius’ text, rather than on an algebraic reconstruction of it, is simply that the opposite sections
do not exist in modern mathematics. For us, there is only the
hyperbola. Our view of the hyperbola, on the other hand, is
that it is a curve consisting “of two open branches extending
to inﬁnity.”2 That we are not bothered by one curve consisting of two is directly related to our deﬁning the hyperbola, as
we do the other conics, by means of an equation. Deﬁning it
this way, the hyperbola becomes merely a set of points given
by coordinates, say the Cartesian coordinates (x, y), satisfying
an equation such as this:
x2  y2
=1
a2 b2
Hence, there is nothing shocking when we discover that this
set of points consists of two disjoint subsets, one containing
points whose x coordinates are less than or equal to –a
(where a is taken to be a positive real number) and one containing points whose x coordinates are greater than or equal
to +a; the only relevant question to ask is whether the coordinates of a given point satisfy or do not satisfy the given
algebraic relation.3 The equation, in this view, tells all; it contains all the essential information about the object; in a sense,
the equation of the hyperbola is the hyperbola (see Fried &
Unguru, 2001, pp. 102103; Klein, 1981, pp. 2829). To
speak about opposite sections in addition to the hyperbola is
unnecessary because they are not distinguished by different
equations.
FRIED
93
The Opposite Sections in the Conica
The point of view above has also been the point of departure
for an older, but still much listened to, generation of historians of mathematics. Zeuthen (1886) and Heath (1921,
1896), leaders of that generation, had no doubt that
Apollonius’ achievement was in the uncovering and elucidation of the equations of the conic sections and that
Apollonius understood the conic sections in an algebraic
spirit. For them, Apollonius’ view was the modern view. Thus
Zeuthen writes:
An ellipse, parabola or hyperbola is here [in Book
1] planimetrically determined as a curve which is
represented by the equation (3), (1) or (2) [the
Cartesian equations for the ellipse, parabola, and
hyperbola, respectively] in a system of parallel
coordinates with any angle between the axes.
Thus, apart from the determination of the position
of these curves, they seem to depend on three constants, namely, that angle [between the axes, or,
equivalently, the ordinate angle], p [latus rectum],
and a [the length of the diameter] (Zeuthen, 1886,
p. 67).
And Heath, echoing Zeuthen, writes:
Apollonius, in deriving the three conics from any
cone cut in the most general manner, seeks to ﬁnd
the relation between the coordinates of any point
on the curve referred to the original diameter and
the tangent at its extremity as axes (in general
oblique), and proceeds to deduce from the relation, when found, the other properties of the
curves. His method does not essentially differ from
that of modern analytical geometry except that in
Apollonius geometrical operations take the place
of algebraic calculations (Heath, 1896, p. cxvi)
�94
THE ST. JOHN’S REVIEW
As for why Apollonius notwithstanding referred to both
opposite sections and hyperbolas, Heath remarks that, “Since
[Apollonius] was the ﬁrst to treat the doublebranch hyperbola fully [emphasis added], he generally discusses the hyperbola (i.e., the single branch) [Heath’s emphasis] along with
the ellipse, and the opposites [Heath’s emphasis], as he calls
the doublebranch hyperbola, separately” (Heath, 1921,
p.139). Zeuthen simply says that, as a matter of terminology,
“What [Apollonius] calls an hyperbola is always only a hyperbolabranch” (Zeuthen, 1886, p. 67). But these remarks are
hardly satisfying. First, while it is most likely that Apollonius
was truly the ﬁrst to treat the “‘doublebranch hyperbola”’—
that is, the opposite sections—fully, the opposite sections
were not so completely unfamiliar to his contemporaries that
he had to continually remind them of their existence; indeed,
in the preface to Book 4 he refers to the opposite sections as
if they were known and discussed, at least by Conon of Samos
and Nicoteles of Cyrene. Second, although Zeuthen’s remark
is not meant to explain Apollonius’ use of both terms, it still
begs the question: if what Apollonius called the hyperbola is
truly represented by the Cartesian equation and therefore, is
the modern doublebranched hyperbola (as Zeuthen unambiguously implies throughout his book), why cause confusion
by referring to a branch of the hyperbola as the hyperbola
and the two branches together as something else, namely, the
opposite sections? Zeuthen himself never speaks of “opposite
sections” but only of the “two (or “corresponding”) branches
of the hyperbola” the “whole hyperbola” (vollständige
Hyperbel), and, most often, simply the “hyperbola.” In this
way, he and his followers, including Heath, merely push aside
as if nugatory the obvious fact that in the Conica, from start
to ﬁnish, there are hyperbolas and there are opposite sections.
The opposite sections (hai tomai antikeimenai) are quite
prominent in the work. They ﬁgure in 24 of the 53 propositions in Book 2; 41 of the 56 propositions in Book 3, and 38
of the 57 propositions in Book 4. In all these propositions,
Apollonius treats opposite sections as something apart from
FRIED
95
hyperbolas. A sign of this is that the opposite sections are separated from the hyperbola in the enunciations of propositions. For example, the statement of proposition 3.42 is: “If
in an hyperbola or ellipse or circumference of a circle or opposite sections [emphasis added] straight lines are drawn from
the vertices of the diameter parallel to an ordinate, and some
other straight line at random is drawn tangent, it will cut off
from them straight lines containing a rectangle equal to the
fourth part of the ﬁgure to the same diameter.” 3.44 reads, “If
two straight lines touching an hyperbola or opposite sections
[emphasis added] meet the asymptotes, then the straight lines
drawn to the sections will be parallel to the straight line joining the points of contact.” If one knew nothing about the
opposite sections, these statements would suggest that the
hyperbola and the opposite sections were as different from
one another as the hyperbola is different from the ellipse.5
The proofs of these two particular propositions make little
distinction between the hyperbola and opposite sections. In
3.44 a separate diagram for the opposite sections is needed to
bring out a case that applies to the two sections together, but
in 3.42 not even that is needed, despite Heiberg’s insistence
on adding an extra diagram for the opposite sections anyway
(ﬁg. 1). Nothing from the logical rigor would be lost if
Apollonius referred only to the opposite sections in these
propositions. So even where the logic does not require it,
Apollonius makes a point to separate, in words verbally, the
hyperbola from the opposite sections.
�THE ST. JOHN’S REVIEW
96
Fig. 1: Conica 3.42 (Heiberg)
Α
Ζ
Μ
Η
Θ
97
FRIED
Fig. 2: Diagram for Conica 1.14
H
Ε
N
Ξ
Δ
Κ
P
O
Θ
Γ
Α
Y
K
Λ
Β
T
Μ
Η
Ζ
Θ
Ε
Δ
Κ
Β
A
Λ
Γ
Of course this is not to say there is no connection
between the opposite sections and the hyperbola, for the central property of the opposite sections is that they are composed of two hyperbolas. Apollonius proves this in the
proposition that introduces the opposite sections, proposition 1.14:
If the vertically opposite surfaces are cut by a
plane not through the vertex [see ﬁg. 2], the section on each of the two surfaces will be that which
is called the hyperbola [emphasis added]; and the
diameter of the two sections will be the same
straight line; and the straight lines, to which the
straight lines drawn to the diameter parallel to the
straight line in the cone’s base are applied in
square, are equal; and the transverse side of the
ﬁgure, that between the vertices of the sections, is
common. And let such sections be called opposite
(kaleisthôsan de hai toiautai tomai antikeimenai)
Π
E
Z
B
M
Λ
Σ
Γ
Δ
This is also the way he refers to the opposite sections
later, especially in Book 4, where he consistently speaks of
“an hyperbola and its opposite section.”7 The genitive in
those phrases suggests that the opposite section belongs to the
given hyperbola, that is, every hyperbola has its very own
opposite section.
The most striking instance in which Apollonius refers to
the opposite sections as two hyperbolas is in his construction
of the opposite sections in 1.59. It is worth reviewing how
this construction is carried out. Apollonius asks, speciﬁcally,
for the following:
�THE ST. JOHN’S REVIEW
98
Fig. 3: Diagram for Conica, 1.59
Δ
A
E
B
H
Γ
K
Z
Θ
Given two straight lines perpendicular to one
another, ﬁnd opposite [sections], whose diameter
is one of the given straight lines and whose vertices are the ends of the straight line, and the lines
dropped in a given angle in each of the sections
will [equal] in square [the rectangles] applied to
the other [straight line] and exceeding by a [rectangle] similar to that contained by the given
straight lines.8
Let BE and BΘ be the given lines and let H be the given
angle (see ﬁg. 3). Choosing BE to be the transverse diameter
(plagia) and BΘ to be the upright side of the ﬁgure (orthia),
Apollonius says to construct an hyperbola ABΓ, adding, that
“This is to be done as has been set out before (prosgegraptai).” For the latter, he has in mind 1.5455 where he shows
how an hyperbola is constructed having a given diameter,
upright side, and ordinate angle. Next he says: let EK have
been drawn through E perpendicular to BE and equal to BΘ,
and draw an hyperbola ΔEZ having BE as its diameter, EK its
upright side, and H its ordinate angle, again, presumably,
relying on 1.5455. With that, he concludes: “It is evident
(phaneron dê) that B and E [i.e., ABΓ and ΔEZ] are opposite
[sections] and they have one diameter and equal upright
sides.” I shall return to the question whether it truly is evident
that B and E are opposite sections, but for now what is
important to understand is that Apollonius constructs the
FRIED
99
opposite sections by constructing one hyperbola and then
another.
Hence the opposite sections are two hyperbolas but not
any two hyperbolas; they are two hyperbolas somehow
linked together; they belong to one another the way identical
twins do.9 In a sense, then, Apollonius, as I understand him
and as Zeuthen understands him, begins with the hyperbola;
however, we disagree about the direction in which he proceeds. Whereas Zeuthen begins with the hyperbola as the
twobranched curve given by the Cartesian equation and then
focuses on the single branch, which Apollonius calls the
hyperbola, my view is that Apollonius begins with the single
connected curve, which he calls the hyperbola, and then
investigates the opposite sections in terms of it. Although
Ockham’s razor would probably fall on the side of the latter
view, we need to get a better grasp of this peculiar situation
wherein two welldeﬁned curves, the two hyperbolas, become
together a distinct geometrical entity, the opposite sections.
To do this, let us consider the ways in which two conic sections are juxtaposed in the Conica, that is to say, let us consider how Apollonius treats pluralities of curves.
Pluralities of Curves
Surely, the juxtapositions of conic sections are to be found in
Book 4 of the Conica. The very point of that book is to investigate the ways in which conic sections can come together—
whether they touch, whether they intersect, at how many
points can they touch, at how many they can intersect, and so
on. The variety of conﬁgurations Apollonius considers can be
seen in the diagram for, say, 4.56:
�THE ST. JOHN’S REVIEW
100
Fig.4: Diagram for Conica 4.56
Δ
Δ
Γ
Γ
Γ
FRIED
101
straight lines drawn to either of the curved lines parallel to
some straight line…and I call that straight line the upright
diameter which, lying between the two curved lines, bisects
all the straight lines intercepted between the curved lines and
drawn parallel to some straight line.” An illustration for the
upright and transverse diameter can be seen in ﬁg. 5 below,
which is not a diagram from Apollonius’ text.
Δ
Fig. 5 The upright and transverse diameters
Δ
Γ
Δ
Δ
Γ
Γ
As discussed elsewhere,10 one of the fundamental facts
explored in Book 4 is the ability of conic sections to be placed
arbitrarily in the plane, as Euclid postulates for circles and
lines—a fact partially justifying Book 4’s inclusion into what
Apollonius terms a “course in the elements of conics.” But it
is this arbitrariness that makes the question of the plurality of
conic sections in Book 4 the exact opposite of the one we are
trying answer about the opposite sections; for the opposite
sections are not thrown together; they belong together.
The ﬁrst indication that a pair of curves may be associated
with one another as the opposite sections is given by
Apollonius in the deﬁnitions at the start of Book 1. Among
these deﬁnitions are those for the transverse and upright
diameters (diametros plagia and diametros orthia). Having
deﬁned the diameter of a curved line, Apollonius continues,
“Likewise, of any two curved lines (duo kampulôn grammôn)
lying in one plane, I call that straight line the transverse diameter which cuts the two curved lines and bisects all the
These deﬁnitions prepare us for the transverse diameter
of the opposite sections introduced in the same proposition
in which the opposite sections themselves are deﬁned, 1.14;
that the two hyperbolas making up the opposite sections have
a common diameter is one of the things that links them
together. Interestingly enough, the upright diameter, which is
the diameter explicitly deﬁned for two curves and that which
is most clearly applicable to the opposite sections,11 is rarely
used by Apollonius: it is used twice in Book 2 (2.37, 38) and
once in Book 7 (7.6); and even in those cases Apollonius is
quick to identify it as one of a pair of conjugate diameters
(suzugeis diametroi), which are good for both two curves and
one. So while Apollonius provides for the possibility of distinct curves linked together, he seems almost to avoid the
idea, at least with regards to the opposite sections.
�THE ST. JOHN’S REVIEW
102
The Conjugate Sections
But the opposite sections are not the only curves linked
together in the Conica. What Apollonius calls the conjugate
sections (hai tomai suzugeis) consist of four hyperbolas, or, to
be precise, two pairs of opposite sections.12 The very word
Fig.6: Conjugate Sections
K
H
Γ
B
A
Δ
Λ
Θ
suzugeis refers to a couple yoked together, a married pair.13
What links them together? First, by the deﬁnition given them
in 1.60, the diameter of the one pair is the conjugate diameter of the other, that is, if the diameters of AB and ΓΔ are D
and d, respectively, then lines drawn in AB parallel to d will
be bisected by D, and vice versa; moreover, D is equal in
square to the rectangle contained by d and the latus rectum
with respect to d (this rectangle being the “ﬁgure” or eidos of
the opposite sections), while d is equal in square to the rectangle contained by D and the latus rectum with respect to D.
Second, as Apollonius shows in 2.17, “The asymptotes [lines
HΘ and ΛΚ in ﬁg. 6] of conjugate opposite sections are common.” This is also shown for the opposite sections in 2.15;
these shared asymptotes are certainly a crucial uniﬁer not
only of the conjugate sections but also of the two opposite
sections themselves.14
Compared to the arbitrary juxtapositions of conic sections presented in Book 4, the opposite sections and the conjugate sections seem to be quite similar in that their component curves are bound by means of asymptotes and diameters.
Indeed, Apollonius often treats the opposite sections and conjugate sections in analogous propositions for instance: 2.41
states, “If in opposite sections two straight lines not through
FRIED
103
the center cut each other, then they do not bisect each other”;
and 2.42, states that “If in conjugate opposite sections two
straight lines not through the center cut each other, they do
not bisect each other.” Yet this close connection between the
opposite sections and conjugate sections raises the question
about why the opposite sections nevertheless enjoy a status in
the Conica that the conjugate sections do not? Why, in particular, are the conjugate sections never mentioned in conjunction with the other conic sections as are the opposite sections? Why are only the opposite sections included in the
clique, parabola, hyperbola, ellipse, and opposite sections?15
For this, we must return to the beginning of the Conica, for
the answer has very much to do with beginnings.
The Genesis of the Opposite Sections
The Conica opens not with the deﬁnition of the conic sections, but with the deﬁnition of the conic surface (kônikê
epiphaneia) from which arises the ﬁgure of the cone. Here is
Apollonius’ deﬁnition:
If from a point a straight line is joined to the circumference of a circle which is not in the same
plane with the point, and the line is produced in
both directions, and if, with the point remaining
ﬁxed, the straight line being rotated about the circumference of the circle returns to the same place
from which it began, then the generated surface
composed of the two surfaces lying vertically
opposite one another, each of which increases
indeﬁnitely as the generating straight line is produced indeﬁnitely, I call a conic surface, and I call
the ﬁxed point the vertex…..And the ﬁgure contained by the circle and by the conic surface
between the vertex and the circumference of the
circle I call a cone.
The deﬁnition is vivid and visual owing in large part to its
motionimbued language. Motion, as is well known, is
�104
THE ST. JOHN’S REVIEW
avoided in Greek mathematics. Thus, much has been made of
the hints of motion in Elements, 1.4. But avoiding motion
seems to be a desideratum chieﬂy for propositions and
demonstrations, for motion in deﬁnitions is not all that
unusual. Besides this one from the Conica, Archimedes gives
similar kinematic deﬁnitions for conoids and spheroids in the
letter opening On Conoids and Spheroids, and Euclid himself
deﬁnes the cone, sphere, and cylinder in Book 11 of the
Elements by means of rotating a triangle, semicircle, and rectangle, respectively. The presence of motion in these deﬁnitions gives one the sense that one is witnessing the very coming to be of the objects being deﬁned. In this way, such deﬁnitions take on a mythic quality; one almost wants to see the
demiourgos turning the generating line about the circumference of the base circle.
Fig.7: The Conic Surface
Following the deﬁnitions, Apollonius proceeds in ten
propositions to develop the basic properties of sections of
cones and, in particular, to show (in 1.7) how a cone may be
cut so that the section produced will be endowed with a
diameter. When the reader arrives at 1.11, which begins the
series of four propositions deﬁning the parabola, hyperbola,
ellipse, and opposite sections, one is ready to see how these
sections arise within the cone. I do not use the word “see”
lightly. In these propositions, Apollonius dedicates considerable space both in the enunciation and in the body of the
FRIED
105
proof to describe the sequence of geometrical operations
undertaken to produce the particular sections; just as one witnesses the coming to be of the conical surface and cone, now
one witnesses the coming to be of the conic sections.
Of all the sections presented in 1.1114, the one that is
most obviously related to Apollonius’ cone is in fact the
opposite sections. The reason is clear when one considers
Apollonius’ deﬁnition of the conical surface and cone. When
Euclid, by contrast, deﬁnes a cone in Book 11 of the
Elements, he does it by describing the rotation of a right triangle about one of its legs. Hence, Euclid begins by deﬁning
the ﬁgure of the cone; moreover, Euclid’s cone, from the
start, is right, that is, its axis is perpendicular to its base, and
it is bounded. Apollonius begins by deﬁning a conic surface,
which is generally oblique (since the line from the ﬁxed point
to the center of the circle about whose circumference the generating line turns is not necessarily perpendicular to the plane
of the circle), unbounded, and, signiﬁcantly, double. The cone
arises from this surface as the ﬁgure (schema) contained by
the base circle and the vertex. Later, in proposition 1.4,
Apollonius shows that by cutting the conic surface with
planes parallel to that of the base circle any number of cones
may be produced from the conical surface. In this way, the
two vertically opposite surfaces (hai kata koruphên
epiphaneiae) of the conic surfaces give rise to cones on either
side of the vertex.
�THE ST. JOHN’S REVIEW
106
Fig. 8: Conica 1.4
Η
Δ
Ε
Θ
A
A
Α
Θ
Δ
Κ
Β
E
H
K
B
Z
Γ
Z
Κ
Θ
Γ
Δ
H
Β
Ε
Ζ
Γ
The doubleness of the conic surface and the double set of
cones that arise from it is (together with its obliqueness)
surely one of the most striking aspects of Apollonius’ deﬁnition. So, when Apollonius deﬁnes the opposite sections in
1.14, the deﬁnition is not entirely unexpected; it has been
preﬁgured in the deﬁnition of the conic surface itself. Indeed,
while the hyperbola is deﬁned in terms of the ﬁgure of the
cone, which is deﬁned by means of the conic surface, the
opposite sections are deﬁned in terms of the conic surface,
and in this sense, the opposite sections are prior to the hyperbola. It is worth observing in this connection that 1.14 is the
last proposition in the book in which the conic surface
appears, as if to say that it no longer has to appear, having
served its purpose.
But here a little more needs to be said. For while it is true
that the conic surface does not appear again in the Conica,
the cone does: it reappears in Book 6, where it is used to
carry out various constructions connected with similar and
equal conic sections, and, most importantly, at the end of
Book 1, where Apollonius constructs conic sections in a plane
having given diameters, latera recta, and ordinate angles. In
the latter constructions, Apollonius generally proceeds by
taking the given plane as the cutting plane and then constructing the cone cut by that plane so that the section pro
FRIED
107
duced is the section required. One might expect, therefore,
that when in 1.59 he set out to construct opposite sections,
he would similarly have constructed the conic surface cut by
the given plane to produce opposite sections. But, as we saw
above, this is not what he does. Applying 1.5455 in the same
sequence, he constructs one hyperbola and then another.
Furthermore, if one follows 1.5455 to the letter in constructing the hyperbolas on both sides of BE, which involves
constructing cones as I have remarked, one does not arrive at
the two vertical opposite surfaces composing the conic surface without signiﬁcantly altering the construction in 1.5455. Can Apollonius truly say, then, as he does, that “it is evident” (phaneron de) that opposite sections are produced in
1.59? I think he can. For when he produces the ﬁrst hyperbola in 1.59, he does so by producing a cone, in accordance
with 1.5455; but the cone is only a ﬁgure cut off from a
conic surface, and we know from 1.14 that when the plane of
the ﬁrst hyperbola cuts the opposite surface of the conic surface it will produce the opposite section of the ﬁrst hyperbola. What remains for Apollonius is only to produce that
second hyperbola so that it has the right diameter, latus rectum, and ordinate direction, which is precisely what he
does.16 So, although the conic surface does not appear explicitly in 1.59, it is still there implicitly. In fact, this is the case
with all the constructions at the end of Book 1: the initial
construction of the parabola, ellipse, and hyperbola, in which
the ordinate direction is right, is carried out by explicitly constructing a cone, but in the continuation of the constructions,
that is, in the cases in which the ordinate directions are not
right, the cone does not appear. In this way, I think it is wrong
to think Apollonius is trying to break away from the cone; the
conic sections are rooted in the cone. Indeed, it is like the
roots of a tree: although the roots are unseen, the leaves, even
those at the very summit of the tree, will not survive without
them.
Thus we can still say with conﬁdence that the opposite
sections have the special status that they do because of their
immediate origin in the double conic surface. They inherit
�108
THE ST. JOHN’S REVIEW
their doubleness directly from the doubleness of the conic
surface; their unity they obtain from the single plane that cuts
the surface. This geometric birth of the opposite sections sets
them apart from the conjugate sections and puts them in the
same class as the parabola, ellipse, and hyperbola.
Summary and Concluding Remarks
Apollonius’ treatment of the opposite sections tells us much
about his mathematical world—mostly because these sections
are not found outside of it. In modern mathematics there is
only the hyperbola, and the hyperbola has two branches. In
Apollonius’ geometry there is an hyperbola and there are
opposite sections. The hyperbola is produced by cutting a
cone; it has an opposite section since the conic surface from
which the cone arises extends not only below but also above
the vertex of the cone. The opposite sections are linked by
their common asymptotes and common diameter. The conjugate sections are linked similarly by common asymptotes and,
though not a common diameter, by conjugate diameters.
However, this kind of linkage seems to be less compelling for
Apollonius than the linkage arising from the single plane cutting the conic surface, for only the opposite sections, and not
the conjugate sections, are spoken of in conjunction with the
other conic sections.
Earlier I referred to Apollonius’ deﬁnition of the conic
surface as having a mythic quality. This was to give some
explanation of the motionﬁlled description constituting the
deﬁnition and to highlight the possibility that the development of the conic sections from the conic surface and the ﬁgure of the cone was, for Apollonius, nothing short of a spectacle of mathematical genesis. The use of the word “myth” in
a mathematical context is jarring for modern ears; mathematics, if to anything, should be related to logos not muthos.
But in the Greek world, it must be recalled, muthos and logos
overlapped as well as being opposed (Peters, 1967, pp. 120121), so that relationship between the two was one of tension
rather than exclusion. Cassirer has gone far to show that the
FRIED
109
mythmaking function is fundamental to human experience,
and I rather agree with him when he says:
The mythical form of conception is not something
superadded to certain deﬁnite elements of empirical existence; instead, the primary “experience”
itself is steeped in the imagery of myth and saturated with its atmosphere. Man lives with objects
only in so far as he lives with these forms; he
reveals reality to himself, and himself to reality in
that he lets himself and the environment enter into
this plastic medium, in which the two do not
merely make contact, but fuse with each other.
(Cassirer, 1946, p. 10)
Be that as it may, the use of the word “myth” in the context
of the Conica does suggest a view of Apollonius as one very
much engaged with the being of his objects and where they
come from; indeed, the distance we feel between myth and
mathematics is partly the result of an overly pragmatic view
of mathematics in which properties take precedence over origins, a view which is characteristically modern. In this paper,
I have tried to show that if we disregard visible geometrical
origins and focus only on abstract relations, such as are captured in an algebraic equation, it becomes difﬁcult to understand why the opposite sections enjoy the status they do in
the Conica as conic sections different from the hyperbola,
and why their status is different from that of the conjugate
sections—and, conversely, I have tried to show that the opposite sections bring us back to the importance of geometrical
origins in the Conica and, in so doing, allow us one key to the
character of classical mathematics.
Bibliography
Apollonius. (1891, 1893). Apollonii Pergaei quae Graece exstant
cum commentariis antiquis. Edited by I. L. Heiberg. 2 volumes.
Leipzig: Teubner.
�110
THE ST. JOHN’S REVIEW
Apollonius. (1998). Apollonius of Perga Conics Books IIII.
Translated by R. Catesby Taliaferro. Santa Fe: Green Lion Press.
Apollonius. (2002). Apollonius of Perga Conics Book IV. Translated
with introduction and notes by Michael N. Fried. Santa Fe: Green
Lion Press.
Cassirer, E. (1946). Language and Myth. New York: Harper and
Brothers.
Fried, Michael N. and Unguru, Sabetai. (2001). Apollonius of
Perga’s Conica: Text, Context, Subtext. Leiden: Brill.
Goodwin, W W & Gulick, C. B. (1992). Greek Grammar. New
. .
Rochelle: Aristide D. Caratzas.
Heath, Thomas L. (1896). Apollonius of Perga, Treatise on Conic
Sections. Cambridge: At the University Press.
Heath, Sir Thomas L. (1921). A History of Greek Mathematics,
2 vols., Oxford: At the Clarendon Press (repr., New York: Dover
Publications, Inc., 1981).
Klein, Jacob (1981). “The World of Physics and the Natural
World.” The St. John’s Review, 33 (1): 2234
Peters, F. E. (1967). Greek Philosophical Terms. New York: New
York University Press.
Sommerville, D. M. Y. (1946). Analytic Conics. London: G. Bell
and Sons, Ltd.
Toomer, G. J. (1990). Apollonius Conics Books V to VII: The Arabic
Translation of the Lost Greek Original in the Version of the Banu
Musa. 2 vols. (Sources in the History of Mathematics and Physical
Sciences, 9), New York: SpringerVerlag
Toomer, G. J. (1970). Apollonius of Perga, in DSB, 1, 179193.
Youschkevitch, A. P (1976). The Concept of Function up to the
.
Middle of the 19th Century. Archive for History of Exact Sciences
16, pp. 3788.
Zeuthen, H. G. (1886). Die Lehre von den Kegelschnitten im
Altertum, Kopenhagen, (repr., Hildesheim:Georg Olms
Verlagsbuchhandlung, 1966).
FRIED
111
Notes
The algebraic reading of the Conica as an historical interpretation
has, of course, Zeuthen (1886) as its greatest representative;
indeed, Zeuthen’s view of the Conica was completely adopted by
Heath (1896) and, to some extent, by Toomer as well (Toomer,
1990, 1970).
1
From Sommerville (1946), but, needless to say, a similar description can be found in any other textbook of analytic geometry.
2
In this connection, it is worth recalling that the sense in which
functions such as f(x) = a/x (whose graph, of course, is also an
hyperbola) are to be considered continuous or not has itself an
interesting history, one that shows again how history rarely moves
in straight lines. For Euler, “continuity” referred to the rule, so that
a function, like f(x) = a/x, governed by a single analytic equation
should called “continuous.” Although this view still echoes in our
calling the hyperbola one curve with two branches, Euler’s
approach to the continuity of functions gave away to a more geometrical view of continuity, that is, where continuity has much to
do with connectedness in the work Cauchy and Bolzano (For a
detailed discussion of these shifts in the meaning of continuity in
the history of the function concept, see Youschkevitch, 1976).
3
They are less prominent, however, in the extant later books,
Books 57 and in Book 1, where they appear in only 9 propositions.
4
The same argument also shows the difﬁculty in maintaining that,
for Apollonius, the circle was a kind of ellipse. The relationship
between the ellipse and the circle has been discussed extensively in
Michael N. Fried & Sabetai Unguru, Apollonius of Perga’s Conica:
Text, Context, Subtext, chap. 7.
5
Unless stated otherwise, all translations from Books 13 are from
R. Catesby Taliaferro in Apollonius of Perga Conics: Books IIII
(Green Lion edition).
6
For example in 4.48 (although any proposition from among 4.4154 could be taken as an example) we have in the ekthesis: estôsan
antikeimenai hai ABΓ, Δ, kai huperbole tis he AHΓ...kai tês AHΓ
antikeimene estô he E.
7
8
My translation.
�112
THE ST. JOHN’S REVIEW
For this reason, one might expect the dual to be used in referring
to the opposite sections instead of the simple plural. The dual was
used for denoting “natural pairs” (see W W Goodwin & C. B.
. .
Gulick, Greek Grammar, §§ 170, 838, 914). But, as Christian
Marinus Taisbak has pointed out to me, it seems that by Hellenistic
times “the dual had become more or less obsolete.”
9
Fried & Unguru, op. cit., chap. 3; Fried, M. N., Apollonius of
Perga: Conics Book IV, pp. xxixxvii.
10
Thus, it is not at all surprising that the ﬁgure Eutocius uses to
illustrate Apollonius’ general deﬁnition of the transverse diameter is
evidently a pair of opposite sections (Heiberg, 2.201), and not, say,
a pair of circles, which would have served as well and which I have
used above.
11
12
The conjugate sections are introduced in the last proposition of
Book 1, proposition 60. Besides 1.60, the conjugate sections appear
also in propositions 2.1723,4243; 3.1315,2326,2829, as well
as in a very striking proposition in Book 7, 7.31.
For the verb suzeugnumi, from which suzugeis is derived, Liddell
and Scott give the deﬁnitions, “to yoke together, couple or
pair…esp. in marriage.” It is worth noting that in 4.49, 50, 51
Apollonius uses the same adjective suzugeis to refer to opposite sections.
13
14
See Fried & Unguru, op.cit., pp. 123124.
In Book 4, particularly in 4.25 (which is the central theorem of
the book: “A section of a cone does not cut a section of a cone or
circumference of a circle at more than four points”) the opposite do
seem to be separated from the other conic sections. In this speciﬁc
case, however, Heath’s argument above might be correct, namely,
that the cases involving the opposite sections, as Apollonius stresses
in the preface, were new and needed to be highlighted. It may also
be that 4.25 does actually intend the opposite sections, even though
the proof for the opposite sections comes latter; that kind of nexus
is not unusual in the Conica. Whatever the case, it still remains that,
unlike the opposite sections, the conjugate sections do not appear at
all in Book 4.
15
One might argue here that he is using a result here from Book 6,
namely, that two hyperbolas having the equal and similar ﬁgures are
16
FRIED
113
equal (6.2), but this could be argued regarding all the constructions
at the end of Book 1.
�114
THE ST. JOHN’S REVIEW
115
Viète on the Solution of
Equations and the Construction
of Problems
Richard Ferrier
Introduction
I wish to do two things in this paper, ﬁrst to review the
grounds for taking François Viète as a pivotal ﬁgure in the
history of thought, and second to clarify an interesting technical aspect of his work, namely, what he called the “exegetical art.” This clariﬁcation will show Viète to be a traditional
thinker as well as a revolutionary one, by showing how he
stops short of writing symbolic formulae or solutions to problems in geometry.
The second part is much longer, and will take us through
some proofs, none terribly difﬁcult, I hope. The mathematical climax is the analysis of the inscription of a regular heptagon in a circle, an example of the exegetical art in action
taken from Viète’s work.
The Importance of Viète
According to Jacob Klein, the man chieﬂy responsible for
interest in Viète in the last 70 years, “The very nature of
man’s understanding of the world is henceforth, (that is to
say, after Viète’s work), governed by the symbolic number
concept. In Viète’s ‘general analytic’ this symbolic concept of
number appears for the ﬁrst time, namely in the form of the
species.” Now this last remark needs clariﬁcation, particularly
in its use of the term “species,” that is “form” or “eidos,” but
also, perhaps, in the notion of symbolic mathematics that Mr.
Richard Ferrier teaches at Thomas Aquinas College. He has written, but not
yet published, a study of Viète’s transformative work on the analytic art.
�116
THE ST. JOHN’S REVIEW
Klein is drawing upon (Greek Mathematical Thought and the
Origin of Algebra, p. 185). I will give a very brief account of
what I think he means.
It comes to something like this. The letters upon which
we operate when we do algebra are not signs of the same
order as the words of common speech. They do not immediately intend or signify anything. They are ciphers related to
each other by rules of connection analogous to syntax in a
spoken language. It is by those rules alone that they acquire
what “meaning” they have. That meaning is now called “syntactic” as opposed to “semantic” meaning. The equals symbol
in an equation does not mean, “is the same in magnitude or
number.” It is rather a relation between two other terms, say
A and B, which happens to be a convertible relation, so if the
ﬁrst relation holds, A = B, so also the second, B = A holds.
If you give this symbol enough of the behaviors or of the
characteristics that ordinary equality has, then what will be
valid for its use, that is, what will be consistent to say of it
once you have posited the suitable rules governing its role in
a system of such symbols, will also be true of arithmetic or
geometry. This presumes, of course, that you interpret all the
symbols as the kind of beings that arithmetic considers, relations of inequality and equality, numbers themselves, and so
forth. But prior to this interpretation, the set of symbols does
not signify numbers, magnitudes, equalities, additions, divisions, or any other determinate kind of mathematical being.
Moreover, the indetermination in the single letter symbols
gives rise to the notion of a variable, which in turn underlies
modern mathematical or formal logic. Such mathematics is
symbolic mathematics. Mr. Klein claims that the symbolic
concept of number originates in the work of François Viète,
and that this concept comes to dominate modern mathematical thought and to inﬂuence modern conceptuality altogether.
Now I do not intend to demonstrate these larger claims
or to argue for them, but rather to look into some particulars
connected with Klein’s view of Viète. First let us note this
FERRIER
117
aspect of his view: Klein says, “The symbolic concept of number appears for the ﬁrst time . . . in the form of the species.”
How can that be? How is it that the term “species” can be the
entry point for the modern notion of number, and in general
for a mode of cognition, symbolic cognition, to enter mathematics?
This is what I understand Mr. Klein to mean, and I think
he is correct: Viète found a use of the word “eidos” in a text
of solutions to number problems, an ancient text, the
Arithmetica of Diophantus, where the word is used for the
unknown number. That is, Diophantus calls the unknown a
“species.” Viète conjoined this use to a discussion of a
method of ﬁnding unknown magnitudes in geometry. This
method is called analysis. In geometrical analysis, authors
such as Archimedes and Apollonius operate on unknown or
not actually given lines and ﬁgures as though they were given.
Viète, then, brought these two together: (1) the name
“species” or “eidos” for the unknown or, as we would say,
variable; and (2) operating or calculating without making any
distinction between unknown and known quantities. But he
went further. He replaced the given quantities, numbers, or
magnitudes with more species, and so produced the ﬁrst
modern literal or nonnumerical algebra, which he called
“species logistic,” that is, computation in forms.
A very elementary example may serve to clarify this idea
somewhat. You are familiar with the problem of ﬁnding a
fourth proportional. Now if you express that in algebra you
would come up with something like this: let ‘x’ stand for the
unknown. x is to a as b is to c. In that expression the letter ‘x’
does not stand for something you now have or know. It is not
clear how you should add or multiply or in any way deal with
something that you do not yet have.
Let us ignore this mystery of operating on what is not
there for you, what is not given. Writing x is to a as b is to c,
you take the product of means and extremes, and you wind
up with x times c = a times b. Then you divide both products
by c and you have x = a times b divided by c, where a, b, and
�118
THE ST. JOHN’S REVIEW
c are the givens that will determine x, and when you have that
you have a formula for making x be given to you. Of course,
you cannot compute x unless a, b, and c are actual numbers,
say 4, 5, and 10. If these are the given numbers, then x is
given as well, and it is the number 2. Neither can you construct x, unless you have determinate magnitudes, say straight
lines, as the three givens.
If you keep the letters a, b, and c and you allow a, b, and
c to be equally indeterminate with x, then you no longer have
a deﬁnite x so much as a kind of relation between a, b, c, and
x. In algebra, we stop with this relation, calling it both the
formula and the solution to our problem. That is what happens when the particular numbers of an ordinary problem—
to ﬁnd the fourth proportional—are replaced by signs for the
possibility of ﬁnding such a number. You no longer can take
a times b and divide it by c, because a, b, and c are just as
indeterminate as the original unknown. This is what I mean
when I say that Viète used Diophantus’ term species to designate not only the unknowns but also the knowns. He called
the resulting ﬁeld of calculations and operations, and the consequences of performing them according to laws set out,
“species logistic,” that is, calculation in species.
If we take two easy steps beyond what Viète did, writing
a formula for quadratics and using two unknowns as axes in
a plane, his work leads directly to negative, irrational and
complex numbers, to the idea of a variable, and to analytic
geometry. All this stems from letting the symbol, the
“species,” stand in for the determinate number or magnitude
of classical mathematics.
Viète himself puts it this way, (he uses the term “zetetic,”
which I will discuss later). “The zetetic art does not employ
its logic on numbers, which was the tediousness of ancient
analysts, but uses its logic through a logistic which in a new
way has to do with species. This logistic is much more successful and powerful than the numerical one.” So Viète commands our attention as the originator of the modern algebraic
number concept, and the notion of mathematics as a symbolic
FERRIER
119
science not immediately about anything but the arrangements
of its own terms into systems. I say Viète commands our
attention as the originator. Now it is of course the case that
however revolutionary Viète thought he was and however
proud he was of his accomplishment, he did not see all the
things that came from it, for example, imaginary numbers, or
coordinate geometry. How did he think of his work? He considered his work largely a restoration of a way of seeking that
went back to Plato. I say he considered it largely a restoration
because there is an addition to the restored analysis that Viète
expressly claims as his own, which he calls exegetic, and it is
this part that I would like to set out next.
The principal ancient text from which Viète formed his
view of analysis is the Seventh book of the Collectio of
Pappus. Pappus distinguishes between seeking, or “zetetic,”
and providing, or “poristic” analyses. This difference answers
to the difference between the propositions in the Elements
that end “Q.E.D” and those that end “Q.E.F.,” that is,
between theorems and problems. There are two types of
analysis, since there are two types of propositions with which
geometers are ordinarily concerned. Viète, for reasons too
subtle to lay out here, read Pappus in another way, making
zetetic and poristic parts of one procedure, a procedure ordinarily applied to a problem, not a theorem. Still, he conceived
of his work as a restoration of the lost art—or perhaps the
partially lost art—of analysis, that is, as a renovation, not an
innovation. That is a striking characteristic of Viète and his
contemporaries. You ﬁnd in them a preference for the lost or
partially recovered sources in antiquity, for Democritus over
Aristotle, and when Plato is not fully available, for Plato over
Aristotle, for Archimedes over Euclid, and in general for
whatever the schoolmen did not hand on over what the
schoolmen did hand on. They sought the key to the deepest
truths in the arts and sciences in a correct restoration of the
faultily preserved sources of antiquity in preference to accepting the ones that they had received more perfectly intact from
their teachers.
�120
THE ST. JOHN’S REVIEW
This is what Viète says: “And although the ancients had
set forth a twofold analysis, the zetetic and the poristic, it is
nevertheless ﬁtting that there be established also a third kind,
which may be called ‘rhetic’, [telling] or ‘exegetic’, [showing
or exhibiting], so that there is a zetetic art by which is found
the equation or proportion between the magnitude that is
being sought, and those that are given, a poristic art, by which
from the equation or proportion the truth of the theorem set
up is examined, and an exegetic art, by which from the equation set up or the proportion there is exhibited the magnitude
itself which is being sought.” As if to emphasize his own
achievement in completing the method of analysis, he says,
“And thus the whole threefold analytical art, claiming for
itself this ofﬁce may be deﬁned as the science of right ﬁnding
in mathematics.” As to the importance of the third part,
rhetic or exegetic, he says in another place, “Rhetic and
exegetic must be considered to be most powerfully pertinent
to the establishment of the art, since the two remaining provide examples rather than rules.” He emphasizes the importance of the third part in the remarkable concluding words of
the introduction to the analytic art: “Finally, the analytical
art, having at last been put into the threefold form of zetetic,
poristic, and exegetic, appropriates to itself by rights the
proud problem of problems: To leave no problem unsolved.”
What exactly is the third part of the analytic art, the
showing or exegetic part, the part that Viète himself emphasized and considered his own invention? One way to proceed
to answer would be to read or reread the parts of Viète’s
Isagoge that touch on exegetics. If your experience is like
mine, though, you know that it is one thing to have someone
describe a procedure, especially a mathematical procedure,
another to know it from having carried it out. Accordingly, I
propose to explicate Viète’s notion of exegesis by doing some
geometry. To prepare for the example of exegetics taken from
Viète, we will look at a problem of my own.
FERRIER
121
Viète’s Exegetic Art
Analysis Bootcamp
Our practice will be the analysis of the construction of a regular polygon, the pentagon.
Problem: to inscribe a regular pentagon in a circle.
Figure 1 Inscription of a regular pentagon in a circle
Now the analysis of a construction always starts this way:
suppose you already have what you want to make.
Let it be done, or rather as they say, “let it have been
done.” Our ﬁgure is, then, given not really but hypothetically.
ABXCY is supposed to be a regular pentagon in a circle. All
the sides are equal, equal sides subtend equal arcs, equal arcs
subtend equal angles at the circumference, and therefore
angle ACB is precisely half the angle ABC, and is also half the
angle CAB.
The triangle CAB is therefore an isosceles triangle, CA
equaling BC, with the vertex angle half the angles at the base.
Next let the angle at A be bisected by AD. The two half
angles then will each be equal to the angle ACD and the triangle ADC will also be an isosceles triangle.
�122
THE ST. JOHN’S REVIEW
Since the angle BAD is equal to the angle ACD, if a circle
be described through the three points A, D, C, the line AB
will be tangent to that circle, for the angle that a tangent
makes with a chord is equal to the angle subtended by the
same arc at the circumference. But the square on a tangent is
equal to the rectangle contained by the whole secant and the
part outside the circle, that is BC and BD, i.e., the square on
AB = rectangle BD, BC. But inasmuch as AB = AD = DC,
then the square on DC = the rectangle BD, BC.
Well, then, if I had this pentagon in this circle, I would
have this line BC cut at a point D so that the square on DC,
the greater segment, is equal to the rectangle contained by the
whole line and the lesser segment. But there is a proposition
that tells me how to do that, the eleventh proposition in the
second book of the Elements. So, let it be done and then
prove “forwards,” or synthetically, all the connections that I
just established “backwards,” or analytically, and at the end
we will really have our pentagon in a circle.
Now, that is a classical geometrical analysis and synthesis
of the sort you would ﬁnd in the second book of Apollonius,
numerous places in Archimedes, and in Pappus. I have not
used anything algebraic. But, let us indulge ourselves in some
algebra, and look at the equation here: the square on DC =
BD, BC. Calling BC ‘a’, and DC ‘x’, the equation is x2 = a(ax). “And,” you could say to yourself, “if I could only solve for
x, then I would know how long to make DC, and if I could
make DC then I could do the problem. But this is a quadratic
equation, and I can solve it.”
That is the method of analysis. I include here what came
to be called the ‘resolution.’ We began in the classical mode
by looking at the angles and ﬁnding a proportion, which we
resolved into an equality between a square and a rectangle.
When we reach this point, we can either continue in the classical mode by thinking of the equality in geometrical terms,
and do the synthesis as a construction, which would be the
answer to the problem. Or we can examine the equality alge
FERRIER
123
braically and work with it in that mode to reach a formula for
x, and consider the formula the solution of the problem.
Recall what Viète said about exegetics: it is supposed to
provide the unknown magnitude itself, to exhibit or construct
it. One would expect exegetics to provide the way for ﬁnding
such an unknown as we had in the present problem, that is,
it would seem to be some sort of procedure for cutting the
line BD at a point like D that will produce the requisite properties. One is tempted to think that the procedure intended is
to use the quadratic formula to solve for x, and then interpret
the right hand side as a series of simple geometric constructions. As we shall see later, though, this is an error. Viète’s
exegetic art is not the mechanical exploitation of the formulae of solved equations.
Analysis with Live Ammunition
Next, we will see what Viète does in a real problem that is
considerably more complex but of greater interest.
The problem is to inscribe a regular sevensided polygon,
a heptagon, in a circle. It is taken from Viète’s published work
and is a real instance of the kind of solution Viète promises in
the introduction, that is, it is a real instance, I think, of
exegetics. To solve it we will need a postulate and two lemmas.
Postulate: To draw a straight line from any point to any two
lines so that they cut off on it any possible predetermined
interval. (I have illustrated this in the next ﬁgure. See ﬁgure 2.)
�124
THE ST. JOHN’S REVIEW
Figure 2 Postulate
FERRIER
125
Figure 3 First Lemma: Trisection of an angle:
First Lemma: To trisect a given angle.
Let the given angle be DCE.
The point called the pole is the “any point” of the postulate. “To draw a straight line from any point”: so the straight
line must pass through the pole. “To draw it to any two
lines”—those two lines are the curved line and the straight
line below—“so that they cut off on it any possible predetermined interval”—the interval is up there above; it is given in
advance, predetermined, and what I would like you to grant
me in this postulate is the power to put a line through that
pole so that between those two other lines, that interval is cut
off. Now I must say, “Any possible predetermined interval”
because as is obvious to the eye, I guess, in this ﬁgure, if I
made that interval too short, I could not ﬁt the line through
the pole so as to have that interval cut off on it no matter
what. If the two given lines were, for another example, concentric circles, and if the pole were the center of the circle,
then this postulate would be of no use at all because the only
distance between those two lines on lines drawn from the
center of the circle is the difference between their radii; that
is all I could have. So I must say, “any possible predetermined
interval.” If it is possible, I would like the right to place that
line passing through the pole so as to have that distance intercepted on it. This postulate is mentioned in the Isagoge. To
the reader who has not read more of Viète than the Isagoge,
it is not clear why that is in there at all—but he asks for that
postulate for certain purposes, and I ask for it, too.
Let a circle be described with C as center and radius CD.
With pole D and interval = CD we use the postulate to insert
line ABD, so that the interval falls between the circle and the
line CE extended.
I say that angle BAC is 1/3 of angle DCE.
Since AB = BC = CD, the triangle BCD is isosceles, with
equal angles at B and D. But each of these is double the angle
at A. And the one at D together with A is equal to angle DCE.
In the case where the angle to be trisected is greater than 135
degrees, we use the equal exterior angles at B and D, which
are again each twice the angle at A, as in the second ﬁgure.
Q.E.F.
Second Lemma: In the ﬁrst ﬁgure for Lemma 1, if we construct DE=CD, the following equality holds:
The cube on AC, minus three times the solid contained by
AC and the square on AB, is equal to the solid contained by
CE and the square on AB.
�126
THE ST. JOHN’S REVIEW
Figure 4 Lemma 2: cube AC  3 times solid (AC, sq. AB) = solid (CE, sq. AB)
Looking at the ﬁgure (ﬁgure 4), if I drop perpendiculars
BI, DK, and erect a perpendicular CH and extend it, on the
vertical line FHCG I have the situation presented in the
Elements 2.5, that is, I have a straight line bisected and cut
somewhere else, and therefore the square on the half will
equal the sum of the rectangle contained by the unequal segments and the square on the segment between the bisection
point and the point making the unequal cut. And looking at
the straight lines BHD and FHG, I have the situation of 3.35,
namely two lines in a circle cutting each other. When two
lines in a circle cut each other, the rectangles contained by the
segments are equal. I will use those properties and the
Pythagorean Theorem and a few other elementary truths to
FERRIER
127
demonstrate the lemma, but the key ones are the ones I just
indicated.
The proof is given next to the ﬁgure. The square on AB,
which, since AB = BC is the same as the square on CG, is
equal by that proposition in the second book of the Elements
to the square on CH and the rectangle contained by the
unequal segments FH, HG. But the square on AB minus the
square on CH will be the rectangle which is in turn equal to
the rectangle BH, HD by the other proposition from the third
book, namely that rectangles made of the segments of two
intersecting lines in a circle are equal. Now the square on
CH, by the Pythagorean Theorem, will be equal to the difference between the square on the hypotenuse AH and the
square on AC. But since BI is a perpendicular drawn from the
vertex of an isosceles triangle, it will bisect the base AC, and,
BI being parallel to HC, the line AH will also be bisected at
B, so that the square on AH is four times the square on its
half, AB. Then the square on CH is the difference between
four times the square on AB and the square on AC. Next,
from that and the line above, namely, that the difference
between the square on AB and the square on CH is equal to
the rectangle BH HD, the square on AC minus 3 times the
square on AB will be equal to the rectangle BH HD. Now that
is almost where I want to be because I am interested in the
cube on AC, and the difference between that and—I am looking at the conclusion now—three times the solid contained by
AC and the square on AB. I have those solids with the height
removed, as it were, I just need the height AC and I have
those solids. That is my next goal.
Then BH is to HD as IC is to CK, since any two lines cut
by parallel lines are cut proportionally. And if BH is to HD as
IC is to CK, then it is also in the same ratio as their doubles
AC and CE. So that BH is to HD as AC is to CE. Then, taking the common height, BH, the square on BH is to the rectangle BH, HD as AC is to CE. But BH is equal to AB so that
the square on AB is to the difference between the square on
AC and three times the square on AB—that was what we
�128
THE ST. JOHN’S REVIEW
found equal to the rectangle BH, HD—in the same ratio,
namely as AC is to CE.
Then taking means and extremes, the solid contained by
the means is equal to the solid contained by the extremes,
namely the cube on AC, minus three times the solid contained
by AC and the square on AB is equal to the solid contained
by CE and the square on AB. Q.E.D.
I would note at this point that everything we have gone
through in these preliminaries is strictly classical, synthetic
geometry. It could all have been done by Archimedes or
Apollonius, and some of it, in fact, was done by them.
Now let us see what interest this second lemma might
hold. It seems to me the easiest thing would be to look at this
ﬁgure given below (ﬁgure 5) which removes all the middle
steps and just looks at the result.
Figure 5
If you have two isosceles triangles with equal sides p and
unequal sides q and x, then you get an equation answering to
the equality that I gave in the geometrical mode namely that
x3 – 3xp2 = qp2 or, in short, this is a geometrical conﬁguration that answers to a cubic equation. If you have ever gone
back and looked at things in Euclid, trying to express them in
equations, you will know that cubics are unusual. You almost
always get squares. This theorem answers to a cubic equation.
And that is why Viète proves it. He proves this theorem in
order to give a geometrical counterpart to that cubic equation.
If the terms in the equations are interpreted as the sides
and bases of the triangles in this theorem, then the solution
FERRIER
129
of the interpreted equation would reduce to the construction
of this ﬁgure. That is, if you could construct the triangles set
up this way, and in particular if you could give yourself x,
then you would have solved this equation, this interpreted
equation, as I would like to speak of it. To put this another
way, if x33xp2+qp2, then x can be found as the base of an
isosceles triangle whose equal sides are p and whose base
angle is 1/3 the base angle of an isosceles triangle with sides
p and base q. Now the angles of such a triangle, namely the
one with two sides p and a base q, are given. Why? Because
the whole triangle is given. If you have three sides you have
a determinate triangle, and so the angles are given. And given
two equal sides and any one angle in an isosceles triangle, the
remaining side and the remaining angles are also given.
Consequently the base of the ﬂatter triangle would be given.
Thus all we need to do to solve for x is to trisect a given
angle. This is easily done by means of our ﬁrst lemma.
To review: we have a postulate that allows us to insert a
rotating line through a given point across two lines so as to
have cut off on it any interval. We have found a ﬁgure that
answers to a certain cubic equation, namely, paired isosceles
triangles with a 3:1 base angle ratio, that answers to a certain
cubic equation, and we obtained this ﬁgure by trisecting an
angle. Those are the preliminaries.
Last, let us see how Viète inscribes a regular sevensided
polygon in a circle. This problem, incidentally, is Viète’s own
ﬁnal problem in his Supplementum Geometriae, his completion of geometry. I have chosen it not only for its intrinsic
interest but also because I think Viète, who calls the
Supplementum a work in exegetics, gave this problem as a
specimen of the workings of his exegetic art.
Problem: To inscribe a regular heptagon in a circle.
Let it have been done. (See ﬁgure 6.)
�130
THE ST. JOHN’S REVIEW
Figure 6 Inscription of a regular heptagon in a circle
And let BE be one side of the heptagon; then the angle ECB
will be 1/7 of two right angles. Then the angle at E, CEA, will
also be 1/7 because that is an isosceles triangle, EAC. Then
the angle EAB will be 2/7 because it is the sum of those. Now
from the point E let ED be drawn equal to the radius of the
circle, and where it cuts the circle at F let the line FA be
FERRIER
131
drawn to the center. Then we have two isosceles triangles of
the sort in the trisection lemma from which it follows that the
angle FAC is three times angle EDA. It is the same ﬁgure that
we had in the trisection lemma. In the triangle EAC the two
smaller angles each being 1/7 of two right angles, the remaining angle at A is 5/7 of two right angles, but the whole angle
FAC being triple the angle at D, which is itself 2/7, will be
6/7. The remaining angle FAE will be 1/7 of two right angles.
It will be equal then to the angle AEC, so that the lines FA and
EC will be parallel.
Since those two lines are parallel they cut the sides of the
triangles DEC and DFA proportionately, so DE is to DC as
DF is to DA. But, by another proposition from book 3 of the
Elements, the book of circles, the rectangle contained by the
segments of a secant, that is, the whole secant and the part
outside, DF, DE, is equal to the rectangle contained by the
segments of any other secant drawn from the same point outside. So DB, DC equals DE, DF. Then, turning that equality
into a proportion: DE is to DC as DB is to DF, but as DE is
to DC as DF is to DA, therefore DB, DF, DA are continuously
proportional. Then, in a continuous proportion the ﬁrst is to
the third in the duplicate ratio, or as the squares, on any
terms in the ratios and also any terms having the same ratio
as those terms. So DB is to DA as square DE is to square DC.
But DE was by construction made equal to the radius so DB
is to DA as square AB is to square DC. Here we stop.
How does one know when to stop in an analysis? Well,
here is one way I know in this one. You stop when you have
projected the ratios you know in various parts of the ﬁgure
down to ratios on one line. It happens in the synthetic proofs
of Apollonius, too. You take a number of ratios in the ﬁgure
and when you transform them into a proportion all on a single line, that tells you how the lengths in that line are related.
That is where we are now. We have taken the ratios that exist
through the ﬁgure because of its shape and we have turned
them into a proportion among the parts of one line. That is
also beautifully adapted for conversion into an equation.
�132
THE ST. JOHN’S REVIEW
That line, of course, is the diameter extended, DBAC, and
our proportion is DB is to DA as square AB is to square DC.
Let us turn this into species logistic, or algebra. What do
we need to construct the heptagon that we do not know? Is
it not clear that we need to know DB? If we knew DB, we
could locate point D and since DE is equal to a radius, we
could just swing a radius out from D and cut off the other end
of the side of the heptagon, and we would be done. So DB is
the unknown, and that is the only term we need to ﬁnd on
this line, on the diameter extended. So call DB ‘x’, and the
radius ‘r’. DA will then be (r+x), and DC will be (2r+x). The
proportion then may be written: x: (r+x) :: r:(2r+x). Means
and extremes can be taken, yielding the equation: x3 + 4rx2
+ 3r2x = r3.
This does not look very promising. It is a cubic, and we
have a pattern for solving a cubic, but not this cubic. Happily,
there is an algebraic gimmick that Viète invented called
“plasma,” which will do the trick. (Almost none of Viète’s
technical terms passed into common use, by the way.
Accordingly, it is like reading ancient lawbooks to read him
talking about his procedure. One of the few Viètean terms
that did last was “coefﬁcient,” but “plasma” did not.)
“Plasma” is the technique of getting rid of an unwanted term
in an algebraic expression by taking a substitute variable. In a
way you might say that is what you do when you complete
the square in solving a quadratic. You ﬁnd the right thing to
remove the middle term, the term that is just an x, and then
it is just a simple square; it is no longer a threeterm expression. Now in cubics if you could remove two terms at once,
then all you would have to do is take cube roots and you can
solve them in much the same way you solve quadratics.
Plasma gets rid of one term, a term in x or x squared, but
unfortunately only one. The substitution that you make to do
it is y = x + 4/3r, or x = y – 4/3r. Let that substitution have
7
been made, and you get the equation in y:y3 – 7 r2y= 27r3.
3
FERRIER
133
Now if we let p = r √7/3, and r/3 be q, this equation
transforms into our pattern y3 – 3yp2 = qp2. Point D gives x
and y, and hence one side of our heptagon. Since we have the
side, we can just go around the circle six more times and the
problem is done.
I omit a slight digression in Viète’s analysis here. It is
done for reasons of elegance and does not substantially
change the argument. This is the pivot point of the argument,
where we move from analysis to synthesis. Viète’s synthesis
runs through the analysis backwards: ﬁrst the analysis in the
equations and proportions, and then the analysis in the
angles, until he has a proof that the ﬁgure that he had made
in the circle is a regular sevensided ﬁgure. I will omit the synthesis.
It must be noted, however, that not only would every
ancient geometer have provided a synthesis, but that Viète,
too, gives it in the Supplementum Geometria, the text from
which this all comes. Indeed, it is all that he gives. What I
have laid out here is a reconstructed analysis following his
remarks on how to do these sorts of problems. That is, he
says that the proof will be the reverse of the analysis and that
the analyst dissimulates and does not show you everything he
did. This seems to me to be a kind of challenge to readers of
those texts to ﬁnd out what he did do. Descartes says just that
in his Géométrie. So that is what I have given you: it is a
reconstructed analysis of his problem. Like Descartes, Viète
complains that the Ancients covered their tracks and then
does the same himself!
Conclusion
A survey of our procedure is now in order. First, we did
not do something that the discussion of exegetics in the
Isagoge might suggest. I think also it is what Mr. Klein’s book
suggests, and I know it is what I thought when I ﬁrst read the
Isagoge myself. We did not, using algebra, or as Viète calls it,
species logistic, solve the equation to which we had reduced
our problem. That equation would be the ﬁrst one that had
�134
THE ST. JOHN’S REVIEW
the fours in it, and that we later on reduced by plasma to
another equation. At no point in working with that equation
did we solve for the unknown in the shape of a formula.
There is not an algebraic or analytic solution to that equation;
it is not to be found. You perhaps may wonder if I misspoke
my thought there: absolutely not. In fact, the cubic equation
involved here is the socalled “irreducible cubic,” which
means that it cannot be solved by algebra. If you try to solve
it by algebra, you involve yourself with imaginary quantities,
and every attempt you make to get rid of them is like stepping on the bump in the rug: they pop up again elsewhere.
They cannot be gotten rid of. So it is not just that he failed to
do it. He cannot do it; no one can do it.
Exegetics, then, does not mean performing the geometric
counterpart of an algebraic solution as it is laid out in a formula. It does not mean getting b±√b24ac
for example, and then
2a
going through the squaring,
adding, subtracting, and so on in geometry, with the formula
laid out for you as a kind of blueprint. This is something to
which the opening pages of Descartes’ Géométrie point. I say
“point” because I am not sure even Descartes does it; you
have to think for awhile to see whether he intends his ﬁgures
to be the simple exploitation of his operation or not. Perhaps
they are not exploitations of the quadratic formula, but independently given geometrical constructions. As is so often the
case with the foxy Descartes, he does not say enough to help
us be sure what he has in mind. It is defensible, though, in the
light of what he does say, that he intends you to have a geometric counterpart for each algebraic operation, and just to
exploit the formula directly. That is at least defensible in
Descartes. In Viète it is not so.
So what did we do if we did not do that? Well, ﬁrst we
performed an ordinary geometrical analysis of the sort
Archimedes might have done. This ended in a proportion.
That proportion was then treated as an item in logistic. We
forgot that it was about lines and just regarded it as a number
of items in a logistic relation. It was transformed until it
FERRIER
135
reached a certain form. What form? A form we knew in
advance from our ﬁrst lemma to have its geometrical counterpart in a constructable ﬁgure—that is, constructable if you
grant our postulate. Then, after we had noted certain proper
elegances pertaining to this particular problem, we proceeded
synthetically (or rather, we would have if we had gone
through all of the steps), ﬁrst constructing the requisite line
and then the whole polygon, and then, via the resulting equation and other relations, proving that it is indeed a regular
sevensided ﬁgure in a given circle. Zetetic is Viète’s name for
the ﬁrst two parts of this procedure, both the geometrical
part terminating in the proportion and the algebraic analysis
in which we reach an equation in standard form.
Zetetic ended when the equation to which the problem
had been reduced ﬁt one of a number of standard forms, our
lemma for cubics being one such form. This lemma provides
the principal exegetical part. By means of this lemma, we can
now ﬁnd the unknown, in this case the line DB. Exegetics as
a procedure is regular and synthetic. I mean those terms
strictly, that is, the construction is a standard, mechanical
exploitation of the equation that can be given by a standard
construction. The complete synthesis simply runs through the
zetetics in reverse order. If your problem reduces to a cubic
of another form (this is the same with the one we have been
examining, but with the signs on the left side of the equation
switched: 3p2x – x3 = p2q) there is, in the Supplementum,
another, similar exegetical lemma you can use to start your
construction, and the situation is quite the same for problems
such as inscribing the pentagon, reducing to quadratic equations.
Viète treats quadratics in another treatise, and it is an
extraordinarily puzzling little work. Almost everything in it is
painfully obvious, at least the ﬁrst 14 or so propositions are,
and it has the rather odd and stuffy title, which I think should
now make sense, “The Standard Enumeration of Geometrical
Results.” It is a handbook; it is like a carpenter’s manual.
When you have one of these, build in this order that, etc. He
�136
THE ST. JOHN’S REVIEW
does this for quadratic equations and for certain biquadratic
equations. Moreover, in all these cases, the proof, which follows the construction, as in Euclid, is a straightforward reversal of the geometrical and logistic analyses that preceded—
now that, I think, ﬁts Viète’s language in his introduction; it
makes sense of his language.
So exegetics in practice comes to this. It is the provision
and employment of a series of standard constructions for
ﬁnding unknown quantities when such quantities are
enmeshed in equations of various degrees, without solving
the equations. This is important, since it manifests the way
Viète stands in two camps, modern or symbolic, and ancient
or constructive. Though species logistic is of profound value
to him, he never rests in the fully analytical or symbolic solution to his equations. They must always be perfected by realization in construction or computation, artful synthetic and
nonsymbolic procedures.
I hope we may now see why Viète wrote, “Exegetic comprises a series of rules, and is therefore to be considered the
most important part of analytic, for these rules ﬁrst confer on
the analytic art its character as an art, while zetetic and poristic consist essentially of examples.” The analyst who knows
exegetics will know what the standard constructable forms in
each degree are, and he will accordingly know what to aim at
in his zetesis, in his analysis. This will in turn inform his
development of the techniques of species logistic, and ﬁnally,
if he believes, as Viète came to believe, that he has all the
exegetics that can be, he may well boast that, “The analytical
art, having at last been put into the threeform form of zetetic,
poristic, and exegetic, appropriates to itself by right the
proud problem of problems, which is to leave no problem
unsolved.”
�
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Title
A name given to the resource
<em>The St. John's Review</em>
Description
An account of the resource
<em>The St. John's Review</em><span> is published by the Office of the Dean, St. John's College. All manuscripts are subject to blind review. Address correspondence to </span><em>The St. John's Review</em><span>, St. John's College, 60 College Avenue, Annapolis, MD 21401 or via email at </span><a class="obfuscated_link" href="mailto:review@sjc.edu"><span class="obfuscated_link_text">review@sjc.edu</span></a><span>.</span><br /><br /><em>The St. John's Review</em> exemplifies, encourages, and enhances the disciplined reflection that is nurtured by the St. John's Program. It does so both through the character most in common among its contributors — their familiarity with the Program and their respect for it — and through the style and content of their contributions. As it represents the St. John's Program, The St. John's Review espouses no philosophical, religious, or political doctrine beyond a dedication to liberal learning, and its readers may expect to find diversity of thought represented in its pages.<br /><br /><em>The St. John's Review</em> was first published in 1974. It merged with <em>The College </em>beginning with the July 1980 issue. From that date forward, the numbering of <em>The St. John's Review</em> continues that of <em>The College</em>. <br /><br />Click on <a title="The St. John's Review" href="http://digitalarchives.sjc.edu/items/browse?collection=13"><strong>Items in the The St. John's Review Collection</strong></a> to view and sort all items in the collection.
Creator
An entity primarily responsible for making the resource
Office of the Dean
Publisher
An entity responsible for making the resource available
St. John's College
Identifier
An unambiguous reference to the resource within a given context
ISSN 02774720
thestjohnsreview
Contributor
An entity responsible for making contributions to the resource
St. John's College Greenfield Library
Text
A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text.
Page numeration
Number of pages in the original item.
136 pages
Original Format
The type of object, such as painting, sculpture, paper, photo, and additional data
paper
Dublin Core
The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/.
Creator
An entity primarily responsible for making the resource
Office of the Dean
Title
A name given to the resource
The St. John's Review, 2004/1
Date
A point or period of time associated with an event in the lifecycle of the resource
2004
Contributor
An entity responsible for making contributions to the resource
Kraus, Pamela
Brann, Eva T. H.
Van Doren, John
Williamson, Robert B.
Zuckerman, Elliott
Price, Audra
Hunt, Frank
Sachs, Joe
Flaumenhaft, Harvey
Hopkins, Burt C.
Unguru, Sabetai
Fried, Michael N.
Ferrier, Richard
Description
An account of the resource
Volume XLVIII, number one of The St. John's Review. Published in 2004.
Identifier
An unambiguous reference to the resource within a given context
ISSN 02774720
The_St_Johns_Review_Vol_48_No_1
Coverage
The spatial or temporal topic of the resource, the spatial applicability of the resource, or the jurisdiction under which the resource is relevant
Annapolis, MD
Publisher
An entity responsible for making the resource available
St. John's College
Language
A language of the resource
English
Type
The nature or genre of the resource
text
Rights
Information about rights held in and over the resource
St. John's College owns the rights to this publication.
Format
The file format, physical medium, or dimensions of the resource
pdf
St. John's Review