Aristotle
[Footnote 312: Topic. VI. ix. p. 147, b. 4-25.]
24. When the definiend is a Privative Term, the definition given
ought to enunciate that which it is, and that of which it is the
privation; also that subject in which it resides naturally and in the
first instance. In defining ignorance, the definition must enunciate
not privation only, but privation of knowledge; nor will this be
sufficient unless it be added that the privation of knowledge is in
the rational department of the soul ([Greek: e)n tô=| logistikô=|]).
Privation of knowledge in the soul or in the man, will not suffice;
because neither of these subjects is that in which the attribute
resides in the first instance: the rational department of the soul
must be named by itself, as being the primary subject of the
attribute. If the definition be wanting in any of these conditions,
you will have an argument for impeaching it.[313]
[Footnote 313: Ibid. b. 26-p. 148, a. 2.]
25. A term that is privative in form may sometimes be used in the
sense of mere negation, not in that of privation. If this term be
defined generally by privation, the definition will not include the
merely negative sense, and will therefore be impeachable. The only
general explanation attainable is that by pure negation, which is
common both to the negative and the privative. Thus, if the
respondent defines ignorance--privation of knowledge, such privation
can be predicated only of subjects whose nature it is to have
knowledge or who might be expected to have it: such privation cannot
be predicated of infants, or of inanimate objects like stones. To
include these, ignorance must be explained as the mere negation or
non-existence of knowledge; the definition thereof by privation is
inadequate.[314]
[Footnote 314: Ibid. p. 148, a. 3-9: [Greek: o(ra=n de\ kai\ ei) mê\
legome/nou kata\ ste/rêsin sterê/sei ô(ri/sato, oi(=on kai\ e)pi\
tê=s a)gnoi/as do/xeien a)\n u(pa/rchein ê( toiau/tê a(marti/a toi=s
mê\ kat' a)po/phasin tê\n a)/gnoian le/gousin.]
Waitz says in note, p. 503:--"Sensus loci hic est. Peccant qui per
privationem ignorantiam definientes non eam ignorantiam definire
voluerunt quæ est [Greek: kat' a)po/phasin], sed eam quæ est [Greek:
kata\ dia/thesin]." Compare Analyt. Poster. I. xvi. p. 79, b. 23.]
26. If you are debating with one who holds the Platonic doctrine of
Ideas, you will note whether any definition that he may give fits not
only the definiend itself but also the Idea of the definiend. Thus,
Plato in defining animal introduces mortality as a part of his
definition;[315] but mortality cannot be predicated of the Idea or
generic animal--the Self-animal; therefore, you will have an argument
against his definition. In like manner, if any active or passive
attribute is brought into his definition, you will object that this
cannot apply to the Ideas; which are avowedly impassive and
unchangeable.[316]
[Footnote 315: Topic. VI. x. p. 148, a. 15: [Greek: oi(=on ô(s
Pla/tôn o(ri/zetai to\ thnêto\n prosa/ptôn e)n toi=s tô=n zô/|ôn
o(rismoi=s.]
This may perhaps allude to Plato's manner of speaking of [Greek:
zô=|a] in Sophistes, p. 246, E., p. 265, C.; Timæus, p. 69, C.]
[Footnote 316: Topica, VI. x. p. 148, a. 14-22. [Greek: a)pathei=s
ga\r kai\ a)ki/nêtoi dokou=sin ai( i)de/ai toi=s le/gousin i)de/as
ei)=nai].]
27. Another _locus_ for counter-argument is, where the definiend is
Equivocal or Analogous, while one and the same definition is made to
apply to all its distinct meanings. Such a definition, pretending to
fit all, will in reality fit none; nothing but an univocal term can
come under one and the same definition. It is wrong to attempt to
define an equivocal term.[317] When its equivocation is not obvious,
the respondent will put it forward confidently as univocal; while you
as assailant will expose the equivocation. Sometimes, indeed, a
respondent may pretend that an univocal word is equivocal, or that an
equivocal word is univocal, in the course of the debate. To obviate
such misconception, you will do well to come to an agreement with him
prior to the debate, or to determine by special antecedent reasonings
what terms are univocal or equivocal; for at that early stage, when
he does not foresee the consequence of your questions, he is more
likely to concede what will facilitate your attack. In the absence of
such preliminary agreement, if the respondent, when you have shown
that his bad definition will not apply universally, resorts to the
pretence that the definiend, though really univocal, is equivocal,
you will press him with the true definition of the part not included
under his definition, and you will show that this true definition
suits also for the remaining parts of the definiend. You will thus
confute him by showing that, upon his original hypothesis, it must
follow that there are two distinct definitions for the same
definiend--the bad one which he has given, and the true one which you
have constrained him to admit.[318] Perhaps, however, the term which
he has undertaken to define may be really equivocal, and therefore
indefinable; nevertheless, when you have shown the insufficiency of
his definition, he may refuse to admit that the term is equivocal,
but will deny a portion of its real meaning. You will then remind him
that, as to the meaning of names, we must recognize tradition and
custom without presuming to disturb it; but that, when we combine
these names in our own discourse, we must beware of those
equivocations which mislead the multitude.[319]
[Footnote 317: Ibid. a. 23-37: [Greek: e)/ti ei) tô=n kath'
o(mônumi/an legome/nôn e(/na lo/gon a(pa/ntôn koino\n
a)pe/dôken.--a)ll' ou)de\n ê(=tton, ei) o(poterôsou=n pepoi/êken,
ê(ma/rtêken.]
Aristotle here cites and censures the definition of life given by a
philosopher named Dionysius; he remarks that life is an equivocal
term, having one meaning in animals, another and a different one in
plants. Dr. Whewell has remarked that even at the present day a good
definition of life is matter of dispute, and still a desideratum with
philosophers.
Mr. John S. Mill adverts, in more than one portion of his 'System of
Logic' (Bk. IV. ch. iii. s. 5, p. 222, seq.; Bk. V. ch. v. s. 8, p.
371), to the mistake and confusion arising from attempts to define
Equivocal Terms. "The inquiries of Plato into the definitions of some
of the most general terms of moral speculation, are characterized by
Bacon as a far nearer approach to a true inductive method than is
elsewhere to be found among the ancients, and are, indeed, almost
perfect examples of the preparatory process of comparison and
abstraction; but, from being unaware of the law just mentioned, he
often wasted the powers of this great logical instrument on inquiries
in which it could realize no result, since the phenomena, whose
common properties he so elaborately endeavoured to detect, had not
really any common properties. Bacon himself fell into the same error
in his speculations on the nature of heat, in which he evidently
confounded, under the name hot, classes of phenomena which had no
property in common."--"He occasionally proceeds like one who seeking
for the cause of hardness, after examining that quality in iron,
flint, and diamond, should expect to find that it is something that
can be traced also in hard water, a hard knot, and a hard heart."]
[Footnote 318: Topic. VI. x. p. 148, a. 37, seq. [Greek: e)pei\ d'
e)/nia lantha/nei tô=n o(mônu/môn, _e)rôtô=nti me\n ô(s sunônu/mois
chrêste/on, au)tô=| d' a)pokrinome/nô| diairete/on_. e)pei\ d'
e)/nioi tô=n a)pokrinome/nôn to\ me\n sunô/numon o(mô/numo/n phasin
ei)=nai, o(/tan mê\ e)pharmo/ttê| e)pi\ pa=n o( a)podothei\s
lo/gos,--prodiomologête/on u(pe\r tô=n toiou/tôn ê)\ prosullogiste/on
o(/ti o(mô/numon ê)\ sunô/numon, o(po/teron a)\n ê)=|; _ma=llon ga\r
sugchôrou=sin ou) proorô=ntes to\ sumbêso/menon_.]
These counsels of Aristotle are remarkable, as bearing on the
details, and even the artifices, of dialectical debate.]
[Footnote 319: Topic. VI. x. p. 148, b. 16-22. [Greek: r(ête/on pro\s
to\n toiou=ton o(/ti tê=| me\n o)nomasi/a| dei= chrê=sthai tê=|
paradedome/nê| kai\ parepome/nê| kai\ mê\ kinei=n ta\ toiau=ta,
e)/nia d' ou) lekte/on o(moi/ôs toi=s polloi=s.]]
28. If the definiend, of which a definition is tendered to you, is a
compound, you may subtract from this definition the definition of one
of the parts of the definiend, and then examine whether the remainder
will suit as a definition of the remaining part of the definiend. If
the remainder should not suit, this will show that the entire
definition tendered is not tenable. Thus, if the definiend be a
finite straight line, and if the definition tendered be, It is the
boundary of a finite plane, of which (boundary) the middle covers or
stands in the way of the extremities; you may subtract from this
definition the definition of a finite line, viz., the boundary of a
plane surface having boundaries, and the remainder of the definition
ought then to suit for the remainder of the definiend. Now the
remainder of the definiend is--straight; and the remainder of the
definition is--that of which the middle covers or stands in the way
of the extremities. But these two will _not_ suit; for a line may be
straight, yet infinite, in which case it will have neither middle nor
extremities. Accordingly, since the remainder of the definition will
not suit for the remainder of the definiend, this will serve as an
argument that the entire definition tendered is not a good one.[320]
[Footnote 320: Topic. VI. xi. p. 148, b. 23-32.]
If the definiend be a compound, and if the definition contain no
greater number of words than the definiend, the definition must be
faulty; it will be nothing better than a substitution of words. Still
more faulty will it be, if it substitutes rare and strange words in
place of others which are known and familiar; or if it introduces a
new word which signifies something different from that which it
replaces.[321]
[Footnote 321: Ibid. b. 32-p. 149, a. 13.]
The definiend, being compound, will contain both a generic and a
differential term. In general, the generic term will be the better
known of the two; yet sometimes the other is the better known.
Whichever of the two is the better known, the definer ought to choose
that, if all that he aims at is a mere substitution of one name in
place of another. But, if he aims at something more or at the
substitution of an explanatory proposition in place of a name
(without which there can be no true definition), he ought then to
choose the differentia in preference to the genus; for the definition
is produced for the purpose of imparting knowledge, and the
differentia, being usually less known than the genus, stands most in
need of extraneous help to cognition.[322] When the definition of the
differentia has thus been tendered, you will examine whether it will
be equally suitable for any other definiend also. If it be, you have
an argument against the goodness of the definition. For example, the
definition of odd number tendered to you may be--number having a
middle. Here, since number is common both to the definiend and to the
definition, having-a-middle is evidently put forward as the
equivalent of odd. But this cannot stand as equivalent to odd; since
various other subjects which are not odd (such, for example, as a
body or a line), nevertheless have a middle. Since, then, we see that
having-a-middle would be suitable in defining definiends which are
not odd, it cannot be admitted, without some qualifying adjunct, as a
good definition of odd. The adjunct annexed must declare in what
sense middle is intended, since it is an equivocal phrase.[323]
[Footnote 322: Ibid. p. 149, a. 14-28.]
[Footnote 323: Ibid. a. 29-37.]