Kozi Asano, Feb.

1998
1
Mathematics and Dialectic
in Platos Republic VI-VII
1
Kozi Asano
1. Introduction
In a previous paper on the Simile of the Line in Platos Republic Book VI, I
have argued that the visible (horaton) represents the sensible in general and the
opinable, that eikasia means taking an image for its original, that the equality of the
middle two segments (DC and CE) of the line is an unintended consequence of the
mechanism of the Simile of the Line, and that the objects of dianoia are mathematical
intermediates (Asano 1997A).
2
If my arguments there are accepted, they will support a
traditional view of the Simile, according to which the Simile first aims to illustrate the
relation between the visible and the intelligible by the relation between images and their
originals, and then to distinguish mathematical intermediates from Forms within the
intelligible objects.
3
A major, related issue that was not discussed in my previous paper (1997A) is
how mathematics and dialectic are different in their methods. We know that they are
concerned with different kinds of objects, intermediates and Forms; however, it is
mostly the differences in methods that Plato tells us about mathematics and dialectic in
the Simile of the Line and the subsequent pages of Book VII. On this ground some
interpreters have tried to eliminate the difference of objects between mathematics and
dialectic in favor of the differences in methods.
4
So I want to discuss in this paper the
3
For a traditional view of the Simile of the Line, see Hardie: 4965.
3
Since the present paper is a sequel to the previous paper, please read Asano 1997A,
too.
1
This paper is originally written as a chapter of my Ph. D. dissertation submitted to
the University of Texas at Austin in December, 1997.
Kozi Asano, Feb. 1998
2
mathematicians and philosophers methods as understood by Plato in the Republic.
5
In the rest of the introduction, I shall outline what Plato says about the differences
between mathematics and dialectic, and the problems involved in it; after that I shall
discuss those problems in the following sections.
In the Simile of the Line, Plato tells us that mathematics and dialectic differ in
two ways, in these words:
in one part of it [mathematics] a soul, using as images the things that were
previously imitated, is compelled to investigate on the basis of hypotheses and
makes its way not to a beginning but to an end; while in the other part
[dialectic] it makes its way to a beginning that is free from hypotheses;
starting out from hypothesis and without the images used in the other part, by
means of forms themselves it makes its inquiry through them.
6
(Republic
510b49)
The first feature of mathematics here stated is its use of images, and what it uses as
images is the things that were previously imitated. The second feature is that mathe-
matics investigates from hypotheses to an end. By contrast dialectic does not use
images, and it investigates from hypotheses to a beginning. What Plato means by those
differences is still quite unclear, as indicated by Glaucons response (510b10).
7
There
are four major, overlapping problems. First what are the hypotheses, and how are they
used? Second what does it mean to investigate from hypotheses to an end? or to a
beginning? Third what does it mean for mathematics to use images, and why does it do
so? Fourth why is it mathematics that is chosen here as a prelude to dialectic? These
7
Glaucons words I dont sufficiently (ouch hikans) understand are a rhetorical
understatement, which actually means I dont understand at all. Cf. a genuine case
of partial negation I understand although not adequately (hikans ou) which comes
after some elucidation of Platos meaning at 511c3.
6
I shall use Blooms translation of the Republic (with minor stylistic changes), unless
otherwise noted.
5
Emphasis will be put on the mathematicians methods.
4
Cf. Robinson: 192201; N. Cooper: 659; and N. White 1979: 1846; 1976: 968,
10911. It would also be possible to include here Cross & Woozley: 23038. For a
criticism of their views, see Asano 1997A.
Kozi Asano, Feb. 1998
3
four problems I shall discuss in the following sections.
2. The Hypotheses
In this section I shall discuss the problem of hypotheses. About the mathe-
maticians hypotheses Plato says:
the men who work in geometry, calculation, and the like hypothesize
(hupothemenoi) the odd and the even, the figures, three forms of angles, and
other things akin to these in each kind of inquiry, and make them hypotheses
(hupotheseis). (510c26)
The examples of hypotheses here are the odd and the even, the figures, three forms of
angles.
8
The word hupothesis
9
(hypothesis) derives from hupotithmi
10
(hypothesize), and a hypothesis is something that is hypothesized. Now hupotithmi
means: to place under, to lay down.
11
Thence hupothesis can mean a foundation, or
an assumption.
12
So mathematicians lay down the odd and the even, the figures, and
three forms of angles, and make them hypotheses. Scholars disagree, however, on what
sort of things those hypotheses are. For example, Archer-Hind (102) thinks that a
hypothesis is a definition. Robinson and Annas think that it is any kind of
proposition.
13
Cornford and Hardie think that it is an existential proposition.
14
12
For a detailed study of the origin of the meaning of hupot i t hmi and
hupothesis, see Robinson: 93100. There is an interesting and important ambiguity
in the concept of hupothesis. Whether something becomes a foundation or an assump-
tion in ones theory depends on ones attitude to it: if one is sure of, and content with,
what is laid down, it is a foundation (from ones own point of view); while if one is
not sure of, or content with, what is laid down, it is an assumption.
11
Cf. Liddell and Scott.
10
The dictionary form of hupothemenoi.
9
The singular form of hupotheseis.
8
The figures will include the square itself and the diagonal itself (510d78), and the
one (524e1, 524e6525a2, 526a24) is most likely another example of hypothesis in
calculation.
Kozi Asano, Feb. 1998
4
Crombie and Hare think that it is a thing (or concept).
15
In support of the definitional view of hypotheses, Plato says that calculation asks
what the one itself is (524e6525a2), and he seems to provide the mathematicians
definition of the one as something that contains no parts within itself and is each
equal to every other one (526a24). But on the other hand, Plato makes it clear that
mathematicians do not give an account (logos) of their hypotheses (510c67, 533c23,
534b46), and that it is the function of dialectic to try to grasp what each thing is
(533b23), which is logos of its being (534b34). This is another important difference
between mathematics and dialectic: mathematics does not give an account while dialectic
does. This, however, raises a further question: what it is to give an account (logon
didonai).
To give an account (logon didonai) is commonly taken either as giving a
definition of something or as giving a proof of something. The view of giving an
account as giving a proof will naturally go with the propositional view of hypotheses
because what one gives a proof of is a proposition, and not a thing.
16
On the other
hand, if giving an account means giving a definition, then the mathematicians do not give
definitions, while the dialecticians do. This is incompatible with the definitional view of
hypotheses, according to which the mathematicians do lay down definitions as hypo-
theses. So whichever meaning it may take to give an account, the definitional view of
hypotheses is most likely wrong.
17
That may not be the case, however. For, I suggest,
17
Perhaps the definitional view can be compatible with the propositional view of
16
Cf. C. C. W. Taylor: 195, 197; and Annas: 287.
15
Crombie, Vol. 1: 113; Hare: 224, 27; and also Raven: 154. Guthrie, too, thinks
that what is laid down is a thing (509); however, he is a little ambiguous on this
problem just as I would like to be (510, 525).
14
Cornford 1932: 65; and Hardie: 60. Also Nettleship: 252; Ross: 51; Gosling 1973:
10912; and N. White 1976: 98.
13
Robinson: 99101, 152; Annas: 28790; and also Cross & Woozley: 2468. C. C.
W. Taylor, too, thinks that what is laid down is some sort of proposition, including a
definition (1945, 1989).
Kozi Asano, Feb. 1998
5
there can be two levels of definitions, mathematically adequate ones and dialectically
adequate ones, so that the mathematicians answer the question of what a thing is on one
level but not on the other.
18
Although it is very difficult to state exactly what mathe-
matically adequate and dialectically adequate definitions are, it is possible to give a
rough idea of what they are. Before we go on to that difficult task, however, let me
discuss the propositional view of hypotheses first.
The propositional view of hypotheses is wrong in one sense but right in another.
As a general theory of hypotheses, what is laid down in a discussion is surely
something of which we can say either that it is true or that it is false.
19
But what Plato
actually says is hypothesize the odd and the even, the figures, three forms of angles,
and other things akin to these (510c35), and he does not mention any proposition
whatsoever. Scholars have suggested propositions that Plato might have had in mind.
For example, A. E. Taylor suggests about the odd and the even, a proposition that all
numbers are integers.
20
According to him, Plato believed that the mathematicians
hypothesis that all numbers are integers is false, because there are irrational numbers;
and that it can be shown by dialectic. A. E. Taylors suggestion is based on his
understanding (84) that destroying the hypotheses at Book VII 533c8 means to
disprove the hypotheses. This is another phrase that needs interpretation. In the
passage where the phrase is found, Plato says about dialectic:
only the dialectical way of inquiry proceeds in this direction, destroying the
hypotheses (ts hupotheseis anairousa), to the beginning itself in order to
make it secure.
21
(533c79)
20
A. E. Taylor: 812. Cf. Robinson: 103; and Cross & Woozley: 247.
19
Cf. Robinson: 936, 99; and C. C. W. Taylor: 195.
18
The idea of two levels of understanding is quite common; but to take them as two
levels of definitions (accounts of what F is) is not so common. For similar ideas of
mathematically (scientifically) adequate and dialectically adequate accounts of F, see
Murphy: 174, 178; Crombie, Vol. 1: 11819, 1245, 130; and Reeve: 5960, 729.
hypotheses; that is, if the proposition meant is of a form the definiendum is the
definiens.
Kozi Asano, Feb. 1998
6
Anaire
22
means to take away, especially, to rescind laws and customs;
23
and so
ts hupotheseis anairousa means to destroy the hypotheses.
There is, however, no evidence that Plato thought the mathematicians hypo-
theses to be false.
24
He says that mathematicians lay hold of something of what is
(533b7) and dream about what is (533b8c1). These expressions are to be taken
positively as they point to the superiority of mathematics over the other arts concerned
with becoming (533b38). Dreaming is, according to Plato, believing an image to be its
original (476c67). Just as there are three types of image-original relationships,
between the objects of eikasia and pistis, between the objects of opinion and knowledge,
and between the objects of mathematics and dialectic, there are three types of dreaming:
eikasia, opinion, and mathematics. To take a case of opinion, people who believe that to
pay back ones debts is just, are dreaming about justice insofar as they do not know the
definition of justice. But it does not follow that their opinion is wrong. In the same
way, even though mathematicians dream about what is, it does not follow that their
hypotheses are false.
Plato also says that the objects of mathematics are, given a beginning, intel-
ligible (511d2). So he does not mean that the mathematicians hypotheses are false.
What he means is that the mathematicians do not know what they dream about
(511d12, 533c1). The reason for this is that they use hypotheses and, leaving them
untouched, are unable to give an account of them (533c13).
25
By contrast, we can
25
More accurately speaking, mathematicians do not know what they dream about,
24
I do not mean that Plato thought that mathematicians never made an error, but I
believe that he thought that most of the mathematics at his time was genuinely true.
23
Cf. Liddell and Scott.
22
The dictionary form of anairousa.
21
It is not clear what the grammatical subject and object of make it secure
(bebaistai) are. But most likely, the subject is the dialectical way of inquiry, and
the middle voice of the verb is the direct reflexive middle, in which case the object is
the same as the subject, rather than the indirect reflexive middle, in which case the
object would be either the hypotheses or the beginning itself.
Kozi Asano, Feb. 1998
7
infer, dialectic destroys the hypotheses by giving an account of them.
26
This means that
when dialectic gives an account of the hypotheses, they lose their hypothetical
character.
27
Another proponent of the propositional view of hypotheses, Robinson, suggests,
for example about the three forms of angles, a proposition that every plane angle is
either right or obtuse or acute.
28
Certainly it is perfectly possible that Plato had some
such proposition in mind, since there is nothing in the text which excludes that
proposition. But there is nothing in the text which leads us to it, either.
Where there is no hint as to what specific proposition Plato had in mind about
the three forms of angles, a more plausible suggestion is an existential proposition that
there is a right angle, etc. As C. C. W. Taylor has pointed out, in the Parmenides
(136a4c5) Plato seems to regard as equivalent the expressions hupotithesthai ti
[hypothesize something] and hupotithesthai ti einai [hypothesize that there is some-
thing] (198). Similarly in Republic Book VI (507b27), Plato seems to regard as
equivalent the expressions setting down (tithentes) something and assert that there
is something. Again when Plato writes in Book X that we . . . set down (tithesthai)
some one particular form (596a67), he means that they assert that there is some one
particular form. Indeed, what philosophers do at the beginning of their investigation is
hypothesizing that there is a fair, itself by itself, a good, a large, and all the rest
28
Robinson: 103. He is followed by Cross and Woozley: 247.
27
Cross and Woozley: 248. See also Ross: 57; Robinson: 1612; Annas: 288; and
Reeve: 77.
26
Cornford, who distinguishes mathematical dialectic and moral dialectic, suggests
that the former destroys its hypotheses by giving a proof of them while the latter
destroys its hypotheses in the sense of amending or abolishing (1932: 86); and he is
followed by Guthrie: 525. But I do not think that there are two kinds of dialectic with
different procedures, although I acknowledge that there are two branches of dialectic
which share the same methods.
because they do not know their hypotheses (533c15), which in turn is because they
cannot give an account of them (534b46).
Kozi Asano, Feb. 1998
8
(Phaedo 100b57).
29
Just as philosophers introduce Forms, mathematicians introduce
mathematical objects. That is, they assume that there are Forms or mathematical objects.
For, unless there are such objects, it would be nonsense to talk about them (cf.
Nettleship: 252). So if the mathematicians hypothesis is ever a proposition, it would be
an existential proposition affirming the existence of the objects of mathematics.
I have suggested in the last paragraph that Plato regarded as equivalent the
expressions hypothesize something and hypothesize that there is something.
Granted that they are equivalent, which one is primary for Plato? Platos use of the
construction of hupotithesthai with a direct object suggests that what is hypothesized is
primarily a thing. This seems quite reasonable; because it is a thing that makes an
existential proposition true, and an existential proposition does not make a thing exist.
So Crombies and Hares view of mathematicians hypothesis as a thing seems the
closest to Platos view. This, however, excludes neither a propositional view nor a
definitional view of hypotheses.
30
My view is that the three elements, a thing, a prop-
osition, and a definition, are all involved in a hypothesis.
31
First, one cannot hypothesize a thing F without affirming an existential prop-
osition that there is F, at least if one is doing a science about reality. Mathematics as
well as dialectic is such a science (nosis)
32
for, according to Plato, the things that are
hypothesized by mathematicians as well as dialecticians, are emphatically beings (onta
or ousia), more real than any ordinary things around us.
33
So mathematicians cannot
33
This is a view which Plato and mathematicians share about the objects of mathe-
matics (cf. 525d5526a7).
32
This nosis at 534a2 is a general term that includes both nosis and dianoia in the
Simile of the Line (511d8).
31
Cf. my argument for the unity of the existential and predicative uses of being in
Asano 1994: 212.
30
The principal defect of Crombies and Hares view is that they do not take seriously
the existence of the things hypothesized in mathematics, and that especially Hare (27)
tends to equate those things with concepts.
29
This is my translation based on Gallops.
Kozi Asano, Feb. 1998
9
hypothesize, for example, the square without affirming an existential proposition that
there is the square. Further, one cannot hypothesize a thing F, nor an existential prop-
osition that there is F, without having any idea of what F is. So mathematicians have to
have some idea of what those objects are which they hypothesize; otherwise, they could
not know what in the world they are hypothesizing.
34
It is connected with Platos and
mathematicians interest. Suppose that mathematicians hypothesize the square or that
there is the square. They do not just hypothesize the existence of a name the square.
What Plato and mathematicians are interested in, is the thing called the square and its
nature. Thus Plato constantly asks the question of what a thing is, and it is asking this
question that lifts mathematicians as well as philosophers above the world of opinion
(524c10526b3, 533b13).
Now we can go back to the question of what it is to give an account. The kind
of account which dialecticians give is an account of what a thing is (533b23, 534b36).
Accordingly, the account which mathematicians fail to give is an account of what their
hypotheses are. So to give an account means to give a definition of something.
35
Nevertheless, mathematicians must know somehow what their hypotheses are, in order
to hypothesize them. That is where my suggestion comes in that there are two levels of
definitions.
In another paper (1996) I have argued that there are two arguments for Forms,
the Argument from Conflicting Appearances (ACA) and the One over Many Argument
(OMA). They reflect the two aspects of the Socratic question what is F? or the two
ways in which it can be taken. First, it can be taken as what is F and never not-F,
36
and
second, as what is that which is common to F things and makes them F.
37
Socrates
37
In this analysis, that is the subject and what is a subjective complement.
36
In this analysis, what is the subject and F is a predicate.
35
This will give an additional support to the view of a hypothesis as a thing, of which
one can, or cannot, give a definition.
34
Cf. the paradox of inquiry in the Meno (80d5e5).
Kozi Asano, Feb. 1998
10
interlocutors in Platonic dialogues often take his question in the first way; for example,
to the question what is courage? Laches answers that he is a man of courage who
does not run away, but remains at his post and fights against his enemy (Laches
190e56), and to the question what is beauty? Hippias answers that a beautiful
maiden is a beauty (Hippias Major 287e34).
38
Although Socrates rejects those
answers by saying that they are not the kind of answer he wants, they are not entirely
wrong answers, either.
39
For they are attempts to state paradigmatic cases of F
(courage and beauty respectively) that the interlocutors believe are F and never not-F.
40
40
Laches, Hippias, Euthyphro, and Theaetetus (but perhaps not Meno) are aware that
Socrates is asking for not just any cases, but solid and exemplary cases of F: solid in
the sense that they are the last things to be not-F (cf. Protagoras 330d79, Hippias
Major 288b13), and exemplary in the sense that by comparing with them one can
determine if other cases are F (cf. Euthyphro 6e57, and Republic 472c4d7). Thus
believing that they have provided such cases of F, they, excepting Theaetetus, hold a
strong conviction of the truth of their statements (Laches 190e6; Hippias Major 287e4,
288a35; and Euthyphro 6d5).
Nehamas (1975) has suggested a new reading of the answers Socrates interlocutors
give to Socrates questions in the early dialogues, which is similar to Goslings view on
ta polla kala in the Republic (for Goslings view, see Asano 1993: 127-25; and 1994:
17-20). According to Nehamas, Socrates interlocutors answer the question what is F
by telling not concrete examples (particulars) but many accounts of F. I do not agree
with him. Certainly it is true that the interlocutors answers are not always expressed in
terms of concrete examples. This is clear in the Meno 71e172a2 and Theaetetus
146c7d3 where the examples given by Meno and Theaetetus are not concrete
examples but they are kinds of virtue and knowledge respectively. (I do not mean that
the Theaetetus is an early dialogue.) But on the other hand Laches, Hippias and
Euthyphros answers are naturally read as referring to particular people and actions.
Consequently the moral to draw is that it is insignificant whether the examples are
concrete or general. Whether the examples are concrete or general, Socrates com-
plaint is the same that those examples do not answer his questions. Next the inter-
locutors answers should better be taken as examples rather than accounts of F. The
39
The passages where Socrates complains or suggests that his interlocutors did not
answer his questions are Laches 190e89, Hippias Major 289c3d5, Euthyphro
6d911, Meno 72a78, and Theaetetus 146d4e10.
38
So do Euthyphro in the Euthyphro 5d8e2, Meno in the Meno 71e172a5, and
Theaetetus in the Theaetetus 146c7d3. The translation of the Laches and Hippias
Major is Jowetts in Hamilton and Cairns.
Kozi Asano, Feb. 1998
11
This aspect of the Socratic question is captured by ACA, which infers from the premise
that perceptible objects which appear F also appear not F, the conclusion that there must
be an F itself which is incapable of appearing not F.
41
Mathematical objects (inter-
mediates) are what is generated by this argument in Republic Book VII
(523b9526a7).
42
Thus the mathematicians who hypothesize mathematical objects
43
answer the Socratic question what is F? in one way.
44
Second, Socrates in Platonic dialogues most often explains his question,
focusing on its second aspect. He says, for example, to Laches, what is that common
thing which is the same in all these cases of courage? (Laches 191e1011) and to
Meno:
45
Even if the virtues are many and various, yet at least they all have some
common form which makes them virtues. That is what ought to be kept in
view by anyone who answers the question, What is virtue?
46
(Meno 72c6d1)
46
The translation of the Meno is Guthries in Hamilton and Cairns, except that I have
changed the translation of eidos (72c7) from character to form.
45
Also Euthyphro 6d911, and Theaetetus 148d57.
44
Although the mathematicians answered the Socratic question what is F? they did
not answer Socrates questions about courage, beauty, piety, or virtue for a simple
reason that they were interested in mathematics and not in ethics.
43
By contrast, Laches, Hippias, Euthyphro, and Theaetetus, although they understood
this aspect of the Socratic question correctly, failed to come up with the right objects
that are F and never not-F. This point is especially clear in the Hippias Major
289a1d2.
42
For an argument for this claim, see Asano 1996: 90-92.
41
See Asano 1996: 84.
only answer that may seem to be accounts is Menos at Meno 71e172a2, but even in
the Meno Menos answer at 73d974a6 is clearly examples and not accounts of virtue.
Laches and Euthyphro certainly describe people and actions in general terms, but the
descriptions are made as the referring expressions and not intended as accounts of
courage or piety. Further Hippias beautiful maiden (Hippias Major 287e4) and
Theaetetus geometry (Theaetetus 146c8) can hardly pass for accounts of beauty or
knowledge. Then why do Socrates interlocutors confuse universals and particulars?
They do not. They tell paradigmatic cases of F not because they confuse those cases
with F but because those cases are heuristic devices for helping Socrates intuit the F
itself.
Kozi Asano, Feb. 1998
12
This aspect is captured by OMA, which infers from the premise that many things are
(called) F in the same sense, the conclusion that there is some one thing, the F itself,
apart from those F things.
47
As I have argued in another paper, both ACA and OMA
are necessary for the introduction of Forms.
48
Thus the dialecticians who introduce
Forms answer the Socratic question in both the first and the second ways.
49
This
means that they can give, or at least try to give, an articulate account (definition) of F that
explains why all F cases are F.
By comparison, the mathematicians who introduce a mathematical F by ACA,
formally define it as an F that is never not-F,
50
and can only show what F is by
ostensive definition, that is, by pointing to its images. They would say, for example,
Look at this. This is what I mean by the square.
51
51
It would be extreme to claim that the mathematicians of Platos time never defined a
mathematical term. What matters is: did they give definitions of the most fundamental
concepts in mathematics; and how deep or superficial were those definitions? Cf. the
dispute between Hare (258) and C. C. W. Taylor (200202): neither of them is
completely convincing. See, however, Aristotles testimony in the Metaphysics (Book
I, Ch. 5, 987a1927). His description there, although obscure, suggests that the
Pythagoreans defined terms in a superficial (987a22) way that resembles the defect
of the mathematical definitions that answer the first aspect of the Socratic question; for
it suggests that they defined the term double as 2 that is double, which is a circular
definition. For the defect of the mathematical definitions, see the preceding note 50.
50
This is a circular definition because the definiens an F that is never not-F contains
the term F to be defined. The mathematicians definition of the one as something
that contains no parts in itself (526a4) falls in this pattern, too, because no parts
(morion . . . ouden) means no plurality of parts, which is equivalent to never not-
one.
49
The self-predication of Forms is a murky problem, but Plato apparently believed
that the Form of F is F (Protagoras 330c3e1, Symposium 210e6211a5, and Parmen-
ides 132a68).
48
See Asano 1996: 88-9.
47
The principal text of this argument is Republic Book X, 596a67. For a precise
analysis and interpretation of it, see Asano 1996: 84-7. The separation of the F itself
from the many F things is a distinctly Platonic development of the Socratic question
(cf. Aristotle, Metaphysics XIII. 4. 1078b3033).
Kozi Asano, Feb. 1998
13
So the mathematicians and the dialecticians, although they answer the Socratic
question on different levels, both hypothesize certain objects and make them
hypotheses. They do not use the hypotheses in the same way, however. Plato tells us:
Mathematicians take it that (hs) they know their objects (510c6) . . . and that
those objects are clear to all.
52
(510d1)
They leave the hypotheses untouched. (533c2)
Dialecticians make the hypotheses not principles (archs) but really
hypothesesthat is, steppingstones and springboards. (511b56)
They destroy the hypotheses. (533c8)
From these quotes we can infer two things: first, mathematicians make the hypotheses
principles;
53
and second, they do not know that they do not know their objects,
54
whereas a dialectician would know if he/she did not know the hypotheses. The second
point is a case of Socratic wisdom/ignorance,
55
and it is related to the first point, which
shows the two possible and very different meanings of hupothesis.
56
Lacking a
piece of Socratic wisdom, mathematicians are quite content with mathematical objects,
which they believe are clear to all, and they dont think it worthwhile to give any
further account of those objects to themselves or others (510c67). In their attitude to
mathematical objects and Forms, they are exactly like the sight-lovers with respect to
many beautiful things and Beauty itself (cf. 476b4c7, 479e15): both are dreaming in
that they take an image for its original. Thus the mathematicians regard and use
hupotheseis (what are laid down) as a foundation or principles on which the
mathematical sciences should be based. However, the dialecticians who look not so
much to images as to their originals, regard and use hupotheseis
57
merely as
57
Although the proper objects of the dialectical knowledge (nosis 511d8, epistm
56
Cf. note 12 above.
55
Socratic ignorance in the sense of the lack of Socratic wisdom.
54
This is also confirmed by these men dont possess intelligence (noun) with respect
to the objects (511d12).
53
This is also implied by when the beginning is what one doesnt know at 533c3.
52
This is my translation.
Kozi Asano, Feb. 1998
14
assumptions that are laid down without grounds.
58
That difference of attitude to the hypotheses between the mathematicians and the
dialecticians leads them to different movements of thought. On the discussion of these
movements I shall spend the next section.
3. The Upward and Downward Movements of Thought
From the hypotheses, mathematicians go to an end while dialecticians go to a
beginning
59
(510b56). The movement of thought to an end is a downward movement
while the one to a beginning is an upward movement.
60
About the mathematicians
movement of thought Plato further says:
Beginning from the hypotheses, the mathematicians go ahead with their
exposition of what remains and end consistently at the object toward which
their investigation was directed. (510d13)
The procedure described here is quite familiar to Glaucon (510d4), and there is no
doubt that it is demonstration, a chain of deductive reasonings from a set of premises to
a conclusion. For example, the mathematicians prove the Pythagorean theorem that the
square on the hypotenuse of a right triangle is equal to the sum of the squares on the
other two sides, from the hypotheses of the triangle, the right angle, the square, etc. So
60
Cf. above (anter) 511a6, goes down (katabaini) 511b8, and going up
(anelthontes) 511d1.
59
This is a true beginning (arch) that is not hypothesized (510b7), in contrast to the
mathematicians beginnings (archai) that are merely hypotheses (511b5).
58
This is a point agreed upon by most interpreters. Cf. Robinson: 152, 156; Cross &
Woozley: 242, 245; and Annas: 277.
533e8) are Forms, the dialecticians hypotheses include mathematical objects as well as
Forms (cf. 511d12, where it is implied that the dialecticians can give an account of
the hypotheses which the mathematicians cannot give an account of). In other words,
the dialecticians recognize mathematical objects (intermediates) and appreciate what
the mathematicians are doing. This is only natural because nosis could recognize the
objects of dianoia just as pistis can recognize the objects of eikasia.
Kozi Asano, Feb. 1998
15
the downward movement here is a movement of deduction from the grounds to what
follows from them.
What is the upward movement, then? The Phaedo describes three stages of the
hypothetical method: the first is to set down a hypothesis and accept as true whatever
seems to accord with (sumphnein)
61
it (Phaedo 100a35); the second is to see if the
consequences (hormthenta)
62
of the hypothesis are mutually consistent or not
(101d46); and the third is to set up another hypothesis that will give an account of the
first hypothesis (101d67). The first stage seems what the mathematicians do: to make
hypotheses and show the truth of a certain proposition from them.
63
It is not clear
where the second stage can be assigned in the Republics scheme: it may be part of
dialectic, or part of mathematics, or both.
64
The third stage seems what the dialecticians
do,
65
and if so, it may help us understand the upward movement. According to the
Phaedo, the third stage is not essentially different from the first: in the same way
65
Cf. Ross: 53; Robinson: 171; and Cross & Woozley: 250.
64
Part of dialectic: because the second stage is conducted when the hypothesis is
challenged (ei de tis kai ta loipa, Phaedo 101d34), it can be considered as an exam-
ination of the hypothesis itself; and also it seems to resemble a Socratic elenchus
(Republic 534c13). Cf. Ross: 57; Robinson: 17071; and Cross & Woozley: 24950.
Part of mathematics: the second stage seems to be conducted before the challenge is
met (ouk apokrinaio hes, Phaedo 101d4); and because what is set down in the first
stage is not any arbitrary hypothesis but what seems the strongest hypothesis, and a
hypothesis that implies mutually inconsistent consequences cannot seem strong, the
second stage can be considered as part of the selection procedures of the strongest
hypothesis.
63
Cf. Ross: 53. It is not the case that the dialecticians do not do the first stage. But
they do not stay there.
62
The verb horma, which is the dictionary form of hormthenta, simply means:
to proceed. Thus I mean by consequences what came out of the hypothesis, not
necessarily logical consequences implied by it.
61
The meaning of sumphnein is not very clear. Cf. Robinson: 1269. Apparently
sumphnein can mean either consistent with or deducible from, and the former
meaning is more natural; but it seems to me that consistency can amount to
deducibility if Plato assumes that the hypothesis is relevant in an appropriate sense to
a proposition to be proved.
Kozi Asano, Feb. 1998
16
(hsauts, Phaedo 101d7) as the first, the third stage makes another hypothesis and
shows the truth of the first hypothesis from it. This process of a hypothesis turning
into what is to be accounted for by another hypothesis, continues as long as a new
hypothesis can be challenged (Phaedo 101d8). Here another hypothesis is not an
alternative hypothesis that replaces the first one but a prior hypothesis that can account
for the first one.
66
So the upward movement seems a movement of hypothesizing from
a consequence to what accounts for it.
Going back to the Republic, we can say that the difference between the mathe-
maticians downward movement and the dialecticians upward movement lies in the
direction of thought: the former goes from the hypotheses to what they can account for
whereas the latter goes from the hypotheses to what can account for them. Probably the
best commentary on the upward movement would be the following passage in Russells
Introduction to Mathematical Philosophy (1):
instead of asking what can be defined and deduced from what is assumed to
begin with, we ask instead what more general ideas and principles can be
found, in terms of which what was our starting-point can be defined or
deduced.
Although Russells words describe the upward movement very well, we should also be
aware of a few differences between Plato and Russell.
What Russell had in mind was an ideal of axiomatization, that is, the reduction
of mathematics into a few axioms, and logical rigor in deduction was utmost for him.
We do not know, however, how logically rigorous Plato was in making a dialectic move
upward. First, as I have argued above, a hypothesis for Plato is not strictly a proposition
but primarily a thing, and yet the hypothesis of a thing can include more than one
proposition, for example, that there is F and that F is so and so.
67
Second, Plato does
67
Cf. Robinson: 132; and Cross & Woozley: 250.
66
The Phaedo, too, uses a metaphor of above (anthen, 101d8) to describe a prior
hypothesis.
Kozi Asano, Feb. 1998
17
not make explicit everything that is necessary for the deduction of a consequence.
When he says that a consequence follows from a hypothesis, he is tacitly relying on
some of our standing assumptions, including the formal principles of logic.
68
Third,
most importantly, the upward movement for Plato is more than making a higher
hypothesis that accounts for a consequence. In the downward movement, whatever
follows from a hypothesis is regarded as true; but in the upward movement not
everything that can account for a consequence is made a hypothesis. Making a higher
hypothesis involves a selection of the best one among the higher hypotheses (cf.
Phaedo 101d78). Best (beltist) here does not simply mean strongest from a theo-
retical point of view but also supremely good in a moral sense. What can you do if
several hypotheses are equally strong? This is where intuition comes in.
Suppose that several hypotheses are equally strong in that they can all account
for a consequence well. Clearly then it is not sufficient for a hypothesis to be able to
account for a consequence, because there are several such hypotheses. A selection must
be made, but according to what criterion? For Plato this criterion is one of goodness.
The dialecticians must have a capacity of intuiting the goodness of things, choosing the
best one among hypotheses, and finally reaching the Form of the Good as the
foundation of all knowledge.
69
That is why the upward movement is essentially
different from the downward movement.
70
70
After the description of the hypothetical method, the Phaedo gives us a warning not
to mix up the discussions of a hypothesis and its consequences (101e12). Aristotles
testimony is equally clear:
For Plato, too, was right in raising this question and asking, as he used to
do, are we on the way from or to the first principles? There is a
difference, as there is in a race-course between the course from the judges
69
The hypothesis the dialecticians reach at the very end of the upward
movementalso called a beginning that is free from hypotheses (510b7) and the
beginning of the whole (511b7)is almost certainly the Form of the Good
(517b8c1, 532a5b2).
68
Cf. Robinson: 1323, 168; and Cross & Woozley: 250.
Kozi Asano, Feb. 1998
18
There is another element in the upward movement. Since the dialecticians
account for the hypotheses, which are things, probably by fewer hypotheses and finally
by one hypothesis, the upward movement seems to involve some kind of generalization:
the higher a hypothesis is, the more general it is.
71
Now Plato says:
the person who is capable of an overview (sunoptikos) is dialectical while the
one who isnt, is not. (537c7)
The Greek word sunoptikos comes from sunora which means: to see together,
to see all at once.
72
When one sees things together, what is seen is their community
(koinnia), relationship (suggeneia), and kinship (oikeiots) with one another (531d13,
537c2). This means that the dialecticians who can see together see the one in the
many.
73
Now dialectic includes the downward movement as well as the upward
movement. Plato says:
When dialectic has grasped the beginning of the whole [in the upward
movement], it keeps hold of what follows from this beginning and in such
fashion goes back down again to an end.
74
(511b78)
Moreover mathematics has its own upward movement, by which the mathematicians rise
up from the visible objects to the intelligible objects.
75
So both dialectic and mathe-
74
Literally translated, the subject of the sentence is the logos itself with the power of
dialectic. The part keeps hold of what follows from is borrowed from Grubes
translation.
73
Cf. Phaedrus 265d34; Parmenides 132a23; Symposium 210a4211d1; Adam:
137; and Robinson: 162.
72
Cf. Liddell and Scott.
71
Moreover it seems that the higher a hypothesis is, the grander it is, because the Form
of the Good, which is the highest hypothesis, is the greatest study (505a2), the
principle of the whole (511b7), and makes it possible for the other things to exist as
they are as well as to be known (509b610). Cf. Cross & Woozley: 251.
to the turning-point and the way back. (Nicomachean Ethics I. 4.
1095a321095b1)
Kozi Asano, Feb. 1998
19
matics include upward and downward movements, in spite of the impression my
discussion might have given to the effect that the upward movement is limited to
dialectic and the downward movement to mathematics.
76
The difference between math-
ematics and dialectic lies in the level of their achievement; that is, the mathematicians rise
up only, for example, from the numbers in these heard accords (531c12) to
numbers themselves (525d6) while the dialecticians can see those mathematical
numbers together to move up to higher and more general hypotheses.
Seeing that dialectic includes both the upward and downward movements, some
interpreters took them as the synthesis and division described in the Phaedrus
(2656).
77
Plato says in the Phaedrus (265c8266b1) that the two procedures of
dialectic are the synthesis of several species into one genus and the division of one
genus into several species. In the Symposium (210a4211d1), too, the ascent to Beauty
itself is a process of generalization. Thus it seems natural that the dialecticians upward
movement, in the Republic, to the Form of the Good is also a process of generalization,
especially since Plato does not clearly separate the beautiful and the good (452e12;
Symposium 201c45, 204d5e4). I wonder, however, if the downward movement is a
division in the Republic. The mathematicians downward movement is anything but a
division. Since the dialecticians, just like the mathematicians, make a hypothesis to
account for a consequence, the downward movement should be the same for both of
them, at least in intent:
78
to account for the consequence from the hypothesis. So we
78
Although it is doubtful whether the dialecticians downward movement is so
77
Robinson: 163. According to him, Zeller, Heinrich Maier, and Rodier took such a
view. See also Cornford 1932: 74; and Murphy: 175.
76
The impression is justified insofar as we consider the movements from the mathe-
maticians hypotheses.
75
This upward movement allows some tacit generalization, because one mathematical
object can account for many visible instances of it. Although mathematical objects are
not generated by the one over many argument and there are many mathematical
objects of one and the same kind, any one of them can account for many visible
instances of it.
Kozi Asano, Feb. 1998
20
cannot apply the synthesis and division in the Phaedrus as they are to the upward and
downward movements of dialectic in the Republic.
79
4. The Use of Images
Next I want to discuss the other feature of mathematics, i.e., the use of images.
First let us hear from Plato how mathematicians use images:
they use visible forms besides and make their arguments about them, not
thinking about them but about those others that they are like. They make the
arguments for the sake of the square itself and the diagonal itself, not for the
sake of the diagonal they draw, and likewise with the rest. These things
themselves that they mold and draw, of which there are shadows and images
in water, they now use as images, seeking to see those things themselves that
one can see in no other way than with thought. (510d5511a1)
This is a procedure familiar enough:
80
the images mathematicians use are figures drawn
on paper,
81
and counters; and while speaking about them, they actually think about the
intelligible objects of which the visible figures and counters are only images. But why
do they use visible images, and what does it mean that they do so? These questions are
connected with another question: what is the relationship between the two features of
mathematics, that is, the use of images and the use of hypotheses?
82
Do mathemati-
cians use images because they rely on hypotheses? Or do they rely on hypotheses
because they use images?
82
There should be some connection between the two features. For, if they were
independent from each other, there would be something else besides mathematics and
dialectic which would have only one of those features; but there is no hint for there
being anything else. Cf. Robinson: 1545.
81
Although the ancient Greeks did not have paper, we usually draw figures on paper.
80
As indicated by Glaucons response, What you say is true (511a2). See also
526e6527a4.
79
Cf. Adam: 2, 174; Robinson: 1635; and Cross & Woozley: 257.
rigorous in form as the mathematicians downward movement, that is, a demon-
stration.
Kozi Asano, Feb. 1998
21
There is a passage that may suggest by its order of presentation the first view
that mathematicians use images because they rely on hypotheses:
83
a soul in investigating mathematics is compelled to use hypotheses, and does
not go to a beginning because it is unable to step out above the hypotheses.
And it uses images . . . (511a36. Emphasis added.)
Similarly in 510c2511a1, Plato first explains the mathematicians use of hypotheses
and then their use of images, and the word he uses to express the latter is pros-chrntai
(use something beside). The prefix pros, which means addition, can suggest that the
use of images is derivative of the use of hypotheses. But we should wonder by what
logic the mathematical use of hypotheses leads to the use of images.
To the last question, an answer is given by Sinaiko: according to him (160,
1623), the hypotheses are assumptions which should be checked against empirical
data, and that is why the use of hypotheses requires the use of visible images. But
mathematical hypotheses are not proved or disproved by their visible exemplifications.
For example, even if the sum of the three angles of a triangle drawn on paper is found
181 by exact measurement, it does not prove that a mathematical theorem that the sum
of the three angles of a triangle is 180 is false. So Sinaikos view is wrong in
assimilating mathematics to empirical sciences.
84
Another answer to the question,
suggested by Robinson (155) for the sake of Jacksons view, is: since the hypotheses
have not yet been proved certain, they need to be made plausible by visible images. But
this suggestion is hardly plausible, either. For mathematicians are confident that the
hypotheses are clear to all (510d1).
85
85
Indeed, the two textual considerations adduced for the first view in the previous
paragraph are very weak. The mere order of presentation does not say anything
84
Cf. Sinaiko: 159. It is a difficult question where the natural sciences should be
assigned in the four-stage scheme of the Simile of the Line in the Republic; but most
likely they belong to the sphere of opinion (pistis). For different views, see Nettleship:
24951; and Raven: 1589.
83
This view was taken by Jackson (1445) and Burnet (229). See also Ross (512)
and Robinson (155).
Kozi Asano, Feb. 1998
22
On the other hand, there is also a passage that suggests the second view that
mathematicians rely on hypotheses because they use images:
86
a soul, using (chrmen) as images the things that were previously imitated, is
compelled to investigate on the basis of hypotheses . . . (510b45)
In this passage, where Plato first introduces the two features of mathematics, the
participle chrmen is naturally taken as giving a reason why a soul is compelled to
investigate on the basis of hypotheses.
87
Again at 511a4, mathematicians are said to be
compelled to use hypotheses. Compelled by what? The reason is given: because
they are unable to step out above the hypotheses(511a56).
88
Then why are they
not able to do so? The quote (511a56) is immediately followed by the mention of the
use of images (511a67): this can give an impression that they are not able to step out
above the hypotheses because they use images.
89
Some textual considerations for the second view that mathematicians rely on
hypotheses because they use images, may seem rather weak, but others confirm it. The
participle chrmen may simply express a concurrent action, or perhaps a means for the
use of hypotheses. That a negative expression is followed by a positive expression does
not necessarily mean that the latter gives the reason for the former, because the latter can
be a consequence of the former. There is, however, one point we should insist on:
89
Although in the grammatical structure the conjunctive particle de (but, and) at
511a6 connects the mention of the use of images (511a68) with the mention of the
use of hypotheses (511a34) while the intervening 511a56 is an addition in paren-
theses.
88
Precisely speaking, this is given as a reason why mathematicians do not go to a
beginning; but in the context, using hypotheses and not going to a beginning mean the
same thing.
87
Cf. Ross: 52; and Robinson: 1545.
86
Most interpreters take this view. Cf. Ross: 52; Murphy: 172 and 177; Robinson:
156; Cross & Woozley: 2445; N. White 1976: 96, 11112, note 43; and Annas: 278.
definite: what is presented later can be more fundamental than what is presented first.
The use of the prefix pros may mean a mere addition without implying any relation
between what is added and what it is added to.
Kozi Asano, Feb. 1998
23
mathematicians are compelled to use hypotheses. This clearly means that the use of
hypotheses comes from what compels mathematicians to do so; and the only thing that
we can think of as compelling them to do so is the other feature of mathematics, i.e., the
use of images.
Perhaps we can better understand this from the nature of the two features. The
use of images is not only a feature distinguishing mathematics from dialectic, but more
importantly it is what distinguishes mathematics from opinion. One should not forget
this simply because it is obvious. It is, as a matter of fact, what has elevated mathematics
above opinion, and is related to mathematics upward movement, that is, from the visible
figures and counters to the intelligible objects. The use of hypotheses, on the other
hand, is related to mathematics downward movement, that is, from the hypothesized
objects to whatever follows from them. So the use of images must come before the use
of hypotheses just as the upward movement should naturally precede the downward
movement. For unless mathematicians go up to the intelligible objects first, they cannot
go down from them at all. Thus the use of images is more fundamental, and it is some-
thing that determines what mathematicians can do and cannot do afterwards.
Then how does the use of images compel mathematicians to rely on hypotheses
and to be unable to step out above them? It has to do with the other aspect of the use of
images. Images both lift up and keep down people. We have already seen that images
lift up mathematicians to the intelligible objects; but the other effect of images is that the
mathematicians understanding of intelligible objects is dependent on images and
cannot get rid of a constraint imposed by them. The constraint I mean is the spatiality
of images. Since the visible figures and counters are spatial, mathematicians imagine
their intelligible objects to be spatial, too. For example, just as a triangle on paper is
triangular, having a certain shape and size, the mathematicians triangle is also triangular,
having a ceratin shape and size. The only difference is that the former is visible whereas
the latter is not visible with the eye, yet imaginable with the minds eye (cf. Crombie,
Kozi Asano, Feb. 1998
24
Vol. 1: 130).
The last point is directly related to the nature of mathematical objects and mathe-
matics as a study about those objects and their relations. As I have argued above,
90
mathematicians define a mathematical object F as what is F and never not-F. Thus there
is no doubt that, for example, a mathematical triangle is a triangle. Moreover, it is not
just a triangle but an idealized triangle, which, being abstracted from matter, never fails to
be a triangle in any way. That is why mathematical objects cannot be seen with the eye
but can be seen with the minds eye. This way of conceiving mathematical objects is
not necessarily wrong: at least the ancient Greeks took mathematics to be a science
about space and spatial entities (cf. Cross & Woozley: 23940), including units,
numbers, two-dimensional and three-dimensional figures.
91
We can now answer the questions we have set at the beginning of this section.
Mathematics uses images, because they are fairly close approximations to the intelligible
objectsclose enough to let mathematicians intuitively grasp those objects. For they
are spatial images of spatial objects. So using the images is a good method for the
study of those objects. What this means is that mathematicians intelligible objects are
of the same nature as the visible figures and counters in an important way: the former is
just as spatial as the latter.
92
Since mathematicians understand their intelligible objects
on the basis of images, they believe that those objects are clear to all (510d1) who
see the images.
92
To borrow Crombies words, the objects of mathematics are non-empirical, yet
semi-empirical. Cf. Crombie, Vol. 1: 110, 118, 1246.
91
Arithmetic and algebra are not exceptions, for the ancient Greeks understood them
in a geometrical manner, representing units and numbers as dots and their collections.
Cf. Ross: 4950. Interestingly, in a later work Timaeus (48e249a6, 50c7d2,
51e6b5) Plato makes space one of the three natures of the universe, along with Forms
and becoming, although the relationship between space of the Timaeus and mathe-
matical objects is totally obscure (for one thing, space is formless while mathematical
objects are not so).
90
In the later part of Section 2.
Kozi Asano, Feb. 1998
25
Further, mathematical objects, being spatial, are individuated by space; that is,
there can be the same things, for example, the same units, here and there.
93
Seeing
those units, dialecticians would wonder and ask what it is that makes them all units. But
to ask that question is beyond the capacity of mathematicians, because they believe that
mathematical objects are clear to all, and dont think it worthwhile to give any
further account of them to themselves or others (510c67). That is why they cannot
step out above the hypotheses but have to rely on them.
94
By contrast, dialecticians do not use images at all: what they deal with are Forms
only (510b79, 511c12). Since Forms are each unique (479a4, 507b6, and 596a6)
95
and cannot be individuated by space or time, they are neither spatial nor temporal (cf.
Symposium 210e6211a4). So they are radically different from their images in space
and time. Thus the images are more misleading than instructive for the cognition of
Forms, and that is why dialecticians do not use images; instead, the only thing they rely
on is the power of discourse (511b4).
5. Mathematics as the Prelude to Dialectic
The last question I want to discuss is the most difficult one in this paper.
Mathematics is singled out as the prelude to dialectic, which ultimately aims at the Form
of the Good (532a5b2). The question is: how does mathematics have anything to do
with ethics? It should be noted at the beginning that since the textual evidence on this
matter is scanty, the following discussions cannot avoid being quite speculative.
As we have already seen, mathematics has the power to lift up the mind from the
95
For the uniqueness of Forms, see Asano 1996: 84-92.
94
Theoretically speaking, there is nothing that prevents one and the same person from
becoming a mathematician and a dialectician at once (cf. 531e23). That is why
mathematics can be a good prelude to dialectic; but insofar as people are mathe-
maticians, they do not ask dialectical questions.
93
For the plurality of mathematical objects, see Asano 1996: 88-92.
Kozi Asano, Feb. 1998
26
visible to the intelligible objects. Thus there is at least one reason why mathematics is
singled out as the prelude to dialectic: it is a training of logical thinking.
96
That is the
whole reason, according to one interpretation (Cross & Woozley: 2546). There is no
direct link between mathematics and ethics; that is, there is nothing particularly
mathematical about ethics, and nothing particularly moral about mathematics. However,
mathematics is just useful as a preliminary training for dialectic in that it turns ones
mind from the world of becoming to the world of being. But that does not seem to be a
reason sufficient for explaining Platos insistence on mathematics in his educational
program for the future rulers (521c1531e1). For example, does he not find any
significance in the other feature of mathematics, i.e., the use of hypotheses?
According to another view, mathematics, because of its hypothetical method, is a
model for any serious science.
97
The hypothetical method consists in constructions of
premises, deductions of conclusions, a clear distinction between conclusions and
premises, and an open confession of premises.
98
Those are the characteristics of
98
Robinson: 177. In Annas version of this view, mathematics serves as a model
because of its hypothetico-deductive procedure, which achieves an understanding of a
rationally organized system where it is clear what is basic, what derived, and how each
result depends on what has gone before (Annas: 28990).
97
Robinson: 153, 1778; and Annas: 2724, 28790.
96
What Nettleship calls mental gymnastic (270). The necessity of a preliminary
training for philosophy is voiced in the Parmenides, too:
That is because you are undertaking to define beautiful, just, good,
and other particular forms, too soon, before you have had a preliminary
training. I noticed that the other day when I heard you talking here with
Aristoteles. Believe me, there is something noble and inspired in your
passion for argument, but you must make an effort and submit yourself,
while you are still young, to a severer training in what the world calls idle
talk and condemns as useless. Otherwise, the truth will escape you.
(135c8d6)
Although the preliminary training meant in the Parmenides is different from mathe-
matics, it is possible that the Republic conceives mathematics as the preliminary
training in a similar manner.
Kozi Asano, Feb. 1998
27
mathematics which the future rulers should learn so that they could pursue ethics in a
similar methodical manner. Thus, according to this view, there is nothing particularly
moral about mathematics; but dialectic is mathematical in method and when it reaches
the Form of the Good, it will be ordine geometrico demonstrata (demonstrated in
geometrical order) (Annas: 288, 290).
It may seem odd at first sight that mathematics is a model of dialectic, when
Plato contrasts them by saying how mathematics is inferior to dialectic. Certainly the
hypothetical method as practiced by mathematicians could not be a model of dialectic,
because they practiced it defectively in the two ways for which Plato criticized them: the
use of images, and the consequent, dogmatic acceptance of hypotheses. According to
the above view, however, if purified of those defects, the hypothetical method can
become the method of dialectic (Robinson: 1778). So dialecticians will use the
hypothetical method without using images or leaving the hypotheses unquestioned.
This view is sensible: it duly recognizes both how dialectic is similar to
mathematics and how it is not so. We can hardly doubt that both mathematics and
dialectic use the hypothetical method or that the one does not have the two defects for
which the other is criticized. Nevertheless, the view does not seem to explain the
intermediate status of mathematics well. Certainly dialectic uses the hypothetical
method, but the hypothetical method is not particularly mathematical. Then why should
the future rulers study mathematics? It is not because mathematics is valuable in itself;
but simply because it was the only branch of knowledge familiar to Plato which had
advanced in using the hypothetical method (Annas: 289). So what really matters is the
hypothetical method, according to this view, and if a curriculum of the hypothetical
method was worked out, mathematics could be dispensed with.
But what other disciplines could Plato have thought of as the preliminary study
to dialectic? As he says, all the other arts are directed to human opinions and desires,
or to generation and composition, or to the care of what is grown or put together
Kozi Asano, Feb. 1998
28
(533b36). This suggests that mathematics is chosen as the preliminary study not
because its hypothetical method is unique but because its objects are unique and
uniquely exempt from change. If so, mathematics cannot be replaced by other curricula
of the hypothetical method.
According to still another view, there is something good about the objects of
mathematics, and the study of them is the first step towards understanding the Good.
99
In a weaker version of the view, mathematical objects are good because they are the
elementary principles of being and everything in the world is ultimately a manifes-
tation of the divine intelligence (Nettleship: 270). This version, however, does not
distinguish mathematics from the other sciences: mathematics is only the elementary
part of the sciences; and that is why mathematics comes first. But, as we have seen just
above, Plato does not conceive any other science but mathematics as the prelude to
dialectic.
100
In a stronger version of the view, mathematical objects are good in a special
way.
101
They are good because they are spatial images of the Forms, which are good
things.
102
Although the visible objects are also images of Forms, they are material
images, and hence they are muddled, i.e., show conflicting appearances. Unlike the
material images, the mathematical objects are pure and clear. Moreover, every Form has
mathematical images. For example, 4 may be an image of justice.
103
Hence to study
103
The thought behind this example of the Pythagorean origin may be that 4 is the first
and typical square and that a square is an expression of reciprocity (Crombie, Vol. 1:
125). Although it may sound like a fancy, it is possible that Plato may have had such
a mathematical understanding of Forms. Plato was clearly influenced by the Pythag-
oreans (cf. Aristotle, Metaphysics I. 6. 987a2931).
102
Because it is the Form of the Good that makes the other Forms be what they are
(509b78), they would be good things in spite of Platos occasional mention to
negative forms such as ugliness (475e9), injustice and evil (476a4). See Santas for the
relation between the Form of the Good and the other Forms.
101
Crombie, Vol. 1: 1246. See also Gosling 1973: 101.
100
Neither does he conceive the natural sciences as part of dialectic.
99
Nettleship: 26970; and Crombie, Vol. 1: 1246 and 1323.
Kozi Asano, Feb. 1998
29
mathematics is to lay hold of something of what is (533b7). I find this version
attractive. For it seems to explain well the unique position of mathematics in Platos
educational program for the future rulers.
What then is so good about the mathematical objects? I think that we tend to
view mathematics as something unrelated to ethics because we view it only as something
true. But for Plato mathematics is not only true but equally beautiful. He writes of the
beauty of the sciences in the Symposium (210c67):
And next, his attention should be diverted from institutions to the sciences, so
that he [the initiate] may know the beauty of every kind of knowledge
(epistmn kallos).
104
This is a stage in the ascent of eros between the one in which the initiate loves the
beauties of the soul, laws and institutions, and the one in which he/she loves the Beauty
itself. This stage might be fairly regarded as corresponding to the mathematical
education in the Republic (cf. Cornford 1950A: 1267). Thus the major reason why the
future rulers should study mathematics is that it is beautiful, or we might say elegant.
Since Plato does not clearly separate the beautiful and the good, mathematics has a
moral significance for him.
Lastly there is a view, which seems to go further in this direction (Gosling 1973:
10007, 11719). According to Gosling, there is a distinction among the five branches
of mathematics, i.e., arithmetic (522c5526c8), plane geometry (526c8527c11), solid
geometry (528a9e2), astronomy (528e3530c5), and harmonics (530c5531c8), in
Platos educational program. Arithmetic and geometry are pure mathematics which is
non-evaluative: they are concerned with the odd, the even, the figures, etc., but not with
the good and the bad. Pure mathematics is necessary but not sufficient as the
preliminary study to dialectic. Astronomy and harmonics on the other hand are
honorable mathematics which involves more directly the notion of the good. For
104
The sciences are also called mathmata (211c6). The translation of the Symposium
is Joyces in Hamilton and Cairns.
Kozi Asano, Feb. 1998
30
they are concerned with good and bad proportions and arrangements (Gosling 1973:
104). This means that they can tell us, for example, at what speed a star should move
(cf. 529d14), and what proportion can make a harmonic sound (cf. 531c3). Thus
according to Gosling, at least astronomy and harmonics are already moral in that they
study not simply mathematical structures of things but proper, i.e., good structures of
things in mathematical terms (Gosling 1973: 101, 103). That may seem to make it
easier to understand why mathematics can be a good preparation for the study of the
Form of the Good.
Gosling assimilates astronomy and harmonics to practical disciplines such as
medicine and navigation. Plato, however, distinguishes mathematics and vulgar arts
(533b36). Then what is the point of distinguishing astronomy and harmonics from the
vulgar arts? It is, according to Gosling, to distinguish the proper sciences from the
vulgar arts (1973: 106): first the vulgar arts are conditioned by immediate human needs
while the proper sciences are detached from them, and second the vulgar arts concentrate
on observable properties of things while the proper sciences develop a precise mathe-
matical account (cf. 529a9530c1, 530e5531c4). The role of pure mathematics is to
raise the vulgar arts to the proper sciences, specifically to help develop a theoretical
interest, and to give precision to disciplines (Gosling 1973: 11718).
Goslings view is a bit complicated as it divides mathematics into two sorts: pure
mathematics (arithmetic and geometry) and honorable mathematics (astronomy and
harmonics). Pure mathematics is only a tool for honorable mathematics, and preference
is given to the latter (Gosling 1973: 118). But, even though that might be an interesting
development of Platos thought in his later years, there is simply no textual support for
it in the Republic. First the Republic does not distinguish the two sorts of mathematics.
Second, arithmetic and geometry are far more prominent than astronomy and
harmonics: in the Simile of the Line there is no mention of astronomy or harmonics;
and in Republic Book VII astronomy as the study of the motion of what has depth
Kozi Asano, Feb. 1998
31
(528e1) is conceived as a natural extension of solid geometry, and Plato says that
astronomy and harmonics should be studied in the same way in which geometry is
studied, i.e., by using as images what is visible or audible, and using hypotheses
(529c7530c1, 530e7531c4).
105
So Plato, assimilating astronomy and harmonics to pure mathematics, seems to
be turning them into a priori sciences rather than empirical ones. This suggests that the
truths of astronomy and harmonics as Plato understood them, are necessitated by their
logic as much as those of arithmetic and geometry. Then are astronomy and harmonics
really concerned with the good and the bad as Gosling claims? What Plato says is:
1. [In astronomy] the true movements in which the really fast and the really
slowin true number and in all the true figuresare moved . . . must be
grasped by argument and thought, not sight. (529d15)
2. [One cannot] grasp the truth about equals, doubles, or any other proportion
(summetria) in the visible things. (529e5530a1)
3. As for the proportion (summetria) of night to day, of these to a month, of a
month to a year, and of the rest of the stars to these and to one another . . .
[the stars are not] always the same [in movement] but deviate . . . for they are
connected with the body and are visible. (530a7b3)
4. [People who practise harmonics in a wrong way] seek the numbers in these
heard accords (sumphnia) and dont rise to the problems, to the consid-
eration of which numbers are concordant (sumphnos) and which not, and
why in each case. (531c14)
The first three quotes simply state that the movements and proportions astronomy is
concerned with are intelligible ones and not visible ones. Further the second quote
actually refers to a geometer, and summetria is a term in geometry.
106
The only quote
that might suggest the value-ladenness of the objects of harmonics is the last one,
106
Summetria is a noun form of summetros, which means: commensurate with. Cf.
Theaetetus 147d56, 148b1.
105
The use of problems (530b6) refers to geometrys use of hypotheses to solve the
problems. Cf. Burnet (222): Simplicius [in de caelo] . . . tells us that Plato, who held
that the movements of the heavenly bodies must be regular, propounded it as a
problem to the mathematicians of the Academy to find on what hypotheses (tinn
hupotethentn) their apparent irregularity could be explained so as to save the
appearances.
Kozi Asano, Feb. 1998
32
especially the term sumphnos. Yet, here again, what matters for Plato seems not so
much agreeable sounds as an agreeing of numbers in some mathematical character; and
why such and such numbers are concordant would be explained from, among others, the
hypothesis of concordance. So, after all, astronomy and harmonics are not more value-
laden than arithmetic and geometry. All of them are beautiful and good, but none is
more so than others.
6. Conclusion
In this paper I have discussed the four major problems concerning the
differences and relations between mathematics and dialectic. First, I have argued, the
hypotheses of mathematics are mathematical objects. The hypotheses at the same time
involve the definitions of those objects as understood by the mathematicians, and
existential and other propositions concerning those objects. What distinguishes mathe-
maticians and dialecticians is their respective attitude to the hypotheses: the former
accept them as obvious principles while the latter try to account for them further.
Second, the upward and downward movements of thought are generally distinguished
by the direction in explanation. The upward movement goes from something to what
can account for it, and the downward movement from something to what can be
accounted for by it. However, those movements are found in both mathematics and
dialectic. Third, mathematicians use images as an aid because the objects of their study
are spatial entities, to which the spatial images can be a fairly close approximation. And
that expedient method is what compels mathematicians to rely on hypotheses. Fourth,
mathematical studies are the only appropriate prelude to dialectic because they are a
priori sciences and their objects, being pure images of good Forms, are uniquely
beautiful and good.
Kozi Asano, Feb. 1998
33
Bibliography
Adam, James. The Republic of Plato. Vol. 2. 2d ed. Cambridge: Cambridge University
Press, 1963.
Allen, R. E., ed. Studies in Platos Metaphysics. London: Routledge and Kegan Paul,
1965.
Annas, Julia. An Introduction to Platos Republic. Oxford: Clarendon Press, 1981.
Anton, J. P., and Preus, A., eds. Essays in Ancient Greek Philosophy, Vol. II. Albany,
NY: State University of New York, 1983.
Archer-Hind, R. D. The Phaedo of Plato. 2d ed. 1894. Reprint, New York: Arno Press,
1973.
Asano, Kozi. [1993]. Degrees of Reality in Plato: Part I. Aichi (Philosophy) Vol. 10:
131118.^(,T,910,13111})
. [1994]. Degrees of Reality in Plato: Part II. The Hannan Ronshu (Journal
of Hannan University) Humanities & Natural Science Vol. 30, No. 2 (Sep.):
1734.||(,||p= /,,9309`
,1994'),1734})
. [1996]. Two Arguments for Forms in Plato: Conflicting Appearances and
One over Many. The Hannan Ronshu (Journal of Hannan University)
Humanities & Natural Science Vol. 32, No. 1 (June): 7996. ||(,
||p= /,,9339,199b),799b
})
. [1997A]. The Simile of the Line in Platos Republic VI. Sapientia (The
Eichi Uni-versity Review) No. 31: 20734.T('`,9
31,30734})
. [1997B]. A Study of Platos Metaphysics in the Republic. Ph.D. diss.,
University of Texas, Austin.
Bambrough, Renford, ed. New Essays on Plato and Aristotle. London: Routledge &
Kegan Paul, 1965.
Barnes, Jonathan, ed. The Complete Works of Aristotle. Vol. 2. Princeton: Princeton
University Press, 1985.
Bloom, Allan, trans. The Republic of Plato. 2d ed. New York: Basic Books, 1991.
Burnet, J. Greek Philosophy. London: Macmillan, 1914.
Cooper, Neil. The Importance of dianoia in Platos Theory of Forms. Classical
Quarterly 16 (1966): 659.
Cornford, F. M. [1932]. Mathematics and Dialectic in the Republic VIVII. Mind
Kozi Asano, Feb. 1998
34
41: 3752, 17390. Reprinted in Allen 1965: 6195. (Cited in the latter
pagination.)
. [1950A]. The Doctrine of Eros in Platos Symposium. In his 1950B:
6880. Re-printed in Vlastos: 11931. (Cited in the latter pagination.)
. [1950B]. The Unwritten Philosophy and Other Essays, edited by W. K. C.
Guthrie. Cambridge: Cambridge University Press.
Crombie, I. M. An Examination of Platos Doctrines. 2 Vols. London: Routledge &
Kegan Paul, 1962.
Cross, R. C., and Woozley, A. D. Platos Republic: A Philosophical Commentary.
London: Macmillan, 1964.
Gallop, David, trans. Plato: Phaedo. Oxford: Clarendon Press, 1975.
Gosling, J. C. B. [1960]. Republic V: ta polla kala etc. Phronesis 5: 11628.
. [1968]. Doxa and Dunamis in Platos Republic. Phronesis 13: 11930.
. [1973]. Plato. London: Routledge and Kegan Paul.
Grube, G. M. A., trans. Platos Republic. Indianapolis: Hackett Publishing, 1974.
Guthrie, W. K. C. A History of Greek Philosophy IV Plato, the Man and his
Dialogues: Earlier Period. Cambridge: Cambridge University Press, 1975.
Hamilton, E., and Cairns, H., eds. The Collected Dialogues of Plato. Princeton:
Princeton University Press, 1963.
Hardie, W. F. R. A Study in Plato. Oxford: Clarendon Press, 1936.
Hare, R. M. Plato and the Mathematicians. In Bambrough: 2138.
Jackson, Henry. On Platos Republic VI 509d sqq. Journal of Philology 10 (1882):
13250.
Kraut, Richard, ed. The Cambridge Companion to Plato. Cambridge: Cambridge
University Press, 1992.
Liddell, H. G. and Scott, R. A Greek-English Lexicon. 9th ed. Oxford: Clarendon Press,
1968.
Murphy, N. R. The Interpretation of Platos Republic. Oxford: Clarendon Press, 1951.
Nehamas, Alexander. Confusing Universals and Particulars in Platos Early
Dialogues. Review of Metaphysics 29 (1975): 287306.
Nettleship, R. L. Lectures on the Republic of Plato. 2d ed. London: Macmillan, 1901.
Raven, J. E. Platos Thought in the Making: A Study of the Development of his
Metaphysics. Cambridge: Cambridge University Press, 1965.
Reeve, C. D. C. Philosopher-Kings: The Argument of Platos Republic. Princeton:
Princeton University Press, 1988.
Robinson, Richard. Platos Earlier Dialectic. 2d ed. Oxford: Clarendon Press, 1953.
Kozi Asano, Feb. 1998
35
Ross, David. Platos Theory of Ideas. Oxford: Clarendon Press, 1951. Reprint,
Westport, CT: Greenwood Press, 1976.
Russell, Bertrand. Introduction to Mathematical Philosophy. 2d ed. Reprint, New York:
Dover Publications, 1993.
Santas, Gerasimos. The Form of the Good in Platos Republic. In Anton and Preus:
23263.
Sinaiko, H. L. Love, Knowledge, and Discourse in Plato: Dialogue and Dialectic in
Phaedrus, Republic, and Parmenides. Chicago: University of Chicago Press,
1965.
Taylor, A. E. A Note on Platos Republic VI 510c25. Mind N. S. 43 (1934): 814.
Taylor, C. C. W. Plato and the Mathematicians: an Examination of Professor Hares
Views. Philosophical Quarterly 17 (1967): 193203.
Vlastos, Gregory, ed. Plato: A Collection of Critical Essays. Vol. II, Ethics, Politics,
and Philosophy of Art and Religion. Garden City, NY: Anchor Books, 1971.
Reprint, Notre Dame, Indiana: University of Notre Dame Press, 1978.
White, Nicholas P. [1976]. Plato on Knowledge and Reality. Indianapolis: Hackett
Publishing.
. [1979]. A Companion to Platos Republic. Indianapolis: Hackett Publishing.
(This paper appeeared in Sapientia (Eichi University) 32 (1998): 117-42.)