Aristotle
_Ubi_ or from the Category _Passio_; for neither of them furnishes
characteristics essential to the subject. All _Passio_ when
intensified to a certain degree destroys the essence of the subject
and removes it from its own appropriate species; but the differentia
is inseparable from its subject; accordingly, nothing by virtue of
which the subject is called [Greek: a)lloi=on] can be a true
differentia. If the definition sins against this rule, it will be
open to question.[295]
[Footnote 295: Ibid. b. 31-p. 145, a. 12: [Greek: o(ra=n de\ kai\ ei)
to\ e)/n tini diaphora\n a)pode/dôken ou)si/as; ou) dokei= ga\r
diaphe/rein ou)si/a ousi/as tô=| _pou_ ei)=nai.--pa/lin ei) to\
pa/thos diaphora\n a)pode/dôken.--a(plô=s d' ei)pei=n, kath' o(/sa
a)lloiou=tai to\ e)/chon, ou)de\n tou/tôn diaphora\
e)kei/nou;--a(plô=s ga\r ou)k a)lloiou/metha kata\ ta\s diaphora/s.]]
12. If the subject be relative, its true differentia ought to be
relative also; thus, science or cognition is a _relatum_, and
accordingly its three differentiæ--theoretical, practical,
constructive--are all _relata_ also.[296] The definition must conform
to this; and it must also, in cases where the relative subject has
more than one correlate, declare that correlate which is the ordinary
and natural one, not any other which is rare and realized only on
occasion.[297] You must watch to see whether this condition is
observed; and also whether the correlative enunciated in the
definition is the one strictly proximate. Thus, if the definition
given of prudence be, It is an excellence of man or an excellence of
the soul, this will not be a good definition. It ought to be--an
excellence of the rational department of the soul; for it is through
and by reason of this department that both man and soul are
denominated prudent.[298]
[Footnote 296: Ibid. a. 13.]
[Footnote 297: Ibid. a. 19-26.]
[Footnote 298: Topic. VI. vi. p. 145, a. 28-32. [Greek: prô/ton ga\r
tou= logistikou= a)retê\ ê( phro/nêsis; kata\ ga\r tou=to kai\ ê(
psuchê\ kai\ o( a)/nthrôpos phronei=n le/getai.]]
13. When the definiend is given as an affection or lasting condition
of some subject, you must examine whether it really resides or can
reside (as by nature it ought to do) in the subject to which it is
referred in the definition. If it cannot, the **definition is
untenable; and this mistake is sometimes made, the producing
conditions of a phenomenon being confounded with the phenomenon
itself, or _vice versâ_.[299] Thus, some persons have
defined sleep--incapacity of sensible perception; doubt--equality of
contrary reasonings; pain--breach of continuity violently made in
parts of the organism which naturally grow together. Now sleep does
not reside in perception, nor doubt in reasonings. Sleep is that
which produces or occasions incapacity of sensible perception; doubt
is a state of mind produced by equality of contrary reasonings.[300]
This will be a _locus_ for arguing against the definition.
[Footnote 299: Ibid. b. 11: [Greek: to\ poiou/menon ei)s to\
poiêtiko\n ê)\ a)na/palin sumbai/nei tithe/nai toi=s ou(/tôs
o(rizome/nois.]]
[Footnote 300: Ibid. a. 33-b. 20.]
14. Another _locus_ is, when the definiend has direct bearing and
reference to something different from what is enunciated in the
definition. Thus, if the respondent defines justice--a power tending
to make equal distribution, you may remark hereupon, that the just
man is he who is deliberately resolved to make equal distribution,
not he who has the power to do so. If this definition were allowed,
the justest man would be he who has the greatest power of so
distributing.[301]
[Footnote 301: Ibid. vii. p. 145, b. 34-p. 146, a. 2.]
15. Again, the definition will be assailable, if the definiend admits
graduation of More or Less, while that which is enunciated in the
definition does not admit it, or _vice versâ_; also, if both of them
admit graduation, but the variations of the two are not corresponding
and concomitant. The defining phrase ought to be identical in
signification with the term defined.[302] If both of them agree in
reference to some common correlate, but one is to this in the
relation of more while the other is in the relation of less, the
definition is faulty.[303]
[Footnote 302: Ibid. p. 146, a. 3-12. [Greek: ei)/per dê\ tau)to/n
e)sti to\ kata\ to\n lo/gon a)podothe\n tô=| pra/gmati.]
Here we have a principle of Concomitant Variations analogous to that
which is so well unfolded, as one of the Four Inductive Methods, in
Mr. J. S. Mill's 'System of Logic.' See Book III. ch. viii. sect. 6.]
[Footnote 303: Topic. VI. vii. p. 146, a. 6-20: [Greek: e)/dei d'
a)mpho/tera ma=llon tô=| au)tô=| u(pa/rchein, ei)/per tau)ta\ ê)=n],
&c.]
16. Again, you will be able to object, if the definition enunciate
references to two distinct correlates, severally or alternately:
_e.g._, The beautiful is that which affords pleasure either through
the eye or through the ear; _Ens_ is that which is capable either of
suffering or acting. You may show that, according to this definition,
beautiful and not beautiful, or that _Ens_ and Non-Ens, will coincide
and be predicable of the same subjects.[304]
[Footnote 304: Topic. VI. vii. p. 146, a. 21-32.
The definition here given of _Ens_ appears in the Sophistes of Plato,
p. 247, E. The definition of the beautiful ([Greek: to\ kalo/n])
appears in the Hippias Major of Plato (p. 298, E, seq.), where it is
criticized by Sokrates.]
17. When the definition is tendered, you ought to examine and define
its own terms, which, of course, profess to enunciate genus and
differentia of the definiend.[305] You will see whether the
definitions of those defining terms are in any way inapplicable to
the definiend.
[Footnote 305: Ibid. a. 33-35.]
18. If the definiend be a _Relatum_, the definition ought to
enunciate its true correlate, or the true correlate of the genus to
which it belongs. You must examine whether this is done, and whether
the correlate enunciated be an ultimate end, as it ought to be
(_i.e._ not merely a means towards something ulterior). If the
correlate enunciated is a generation or a process, this will afford
you an argument against the definition; for all generation or process
is a means towards some ulterior end.[306]
[Footnote 306: Ibid. viii. p. 146, a. 36-b. 19. This is a subtle
distinction. He says that desire must be defined (not desire of _the
pleasurable_, but) desire of _pleasure_: we desire _the pleasurable_
for the sake of _pleasure_. He admits, however, that there are cases
in which the argument will not hold: [Greek: schedo\n ga\r oi(
plei=stoi ê(/desthai ma=llon bou/lontai ê)\ pepau=sthai ê(do/menoi;
ô(/ste to\ e)nergei=n ma=llon te/los a)\n poioi=nto tou=
e)nêrgêke/nai.]]
19. The definition ought not to omit any of the differentiæ of the
definiend; if any be omitted, the real essence is not declared. Here
then is a defect in the definition, which it is your business always
to assail on its defective side.[307] Thus, if the definiend be a
_relatum_ corresponding, not to some correlate absolutely but, to
some correlate specially quantified or qualified, the definition
ought to enunciate such quantification or qualification; if it does
not, it is open to attack.
[Footnote 307: Ibid. b. 20: [Greek: pa/lin e)p' e)ni/ôn ei) mê\
diô/rike tou= po/sou, ê)\ poi/ou, ê)\ pou=, ê)\ kata\ ta\s a)/llas
diaphora/s,--a)polei/pôn ga\r diaphora\n ê(ntinou=n ou) le/gei to\ ti
ê)=n ei)=nai; _dei= d' a)ei\ pro\s to\ e)ndee\s e)picheirei=n_.]]
20. Suppose that the definiend is one of the appetites, relative to
an _appetitum_ as correlate, a mode of the good or agreeable. You
will take notice whether the definition given thereof enunciates the
correlate as only an apparent mode of good: if it does not, you have
a _locus_ for attacking it. But if it does, and if the definer be one
who believes in the Platonic Ideas, you may attack him by showing
that his definition will not square with that doctrine. For the
definition as so given will not suit for the ideal or generic
appetite--the Self-appetite; which correlates with the ideal or
generic good--the Self-good. In this no distinction is admissible of
real and apparent: a Self-apparent-good is an absurdity.[308]
[Footnote 308: Topic. VI. viii. p. 146, b. 36-p. 147, a. 11. [Greek:
e)a\n de\ kai\ a)podô=| to\ ei)rême/non, e)pi\ ta\ ei)/dê a)kte/on
to\n tithe/menon i)de/as ei)=nai; ou) ga/r e)stin i)de/a phainome/nou
ou)deno/s, to\ d' ei)=dos pro\s to\ ei)=dos dokei= le/gesthai, oi(=on
au)tê\ e)pithumi/a au)tou= ê(de/os kai\ au)tê\ bou/lêsis au)tou=
a)gathou=. ou)k e)/stai ou)=n phainome/nou a)gathou= ou)de\
phainome/nou ê(de/os; a)/topon ga\r to\ ei)=nai au)to\ phaino/menon
a)gatho\n ê)\ ê(du/.]
Compare Plato, Parmenides, pp. 133-134, where this doctrine that if
the _relatum_ be an Idea (_sensu Platonico_), the _correlatum_ must
also be an Idea, is enunciated and pushed to its consequences:
[Greek: o(/sai tô=n i)deô=n pro\s a)llê/las ei)si\n ai(/ ei)sin,
au)tai\ pro\s au(ta\s tê\n ou)si/an e)/chousin, a)ll' ou) pro\s ta\
par' ê(mi=n ei)/te o(moiô/mata ei)/te o(/pê| dê/ tis au)ta\
ti/thetai], &c.--[Greek: au)tê\ de\ despotei/a au)tê=s doulei/as
e)sti\n o(/ e)sti], &c. (133, C-E.)]
21. Again, suppose that the definiend is a habit or disposition. You
will examine how far the definition fits as applied to the individual
person who has the habit; and how far it fits when taken in
comparison with subjects contrary or congeneric. Every such
definition, if good, implies in a certain way the definition of the
contrary: he who defines cognition furnishes by implication the
definition of ignorance.[309]
[Footnote 309: Topic. VI. ix. p. 147, a. 12-22.]
22. Or suppose the definiend to be a generic _relatum_, and the
definition to enunciate its generic correlate. You must call to mind
the specific terms comprehended under these two generic terms, and
observe whether they fit on to each other respectively. If they do
not, the definition is faulty.[310]
[Footnote 310: Ibid. a. 23-28.]
23. You will farther examine whether the Opposite of the definition
will serve as definition to the Opposite of the definiend, as the
definition of half is opposite to the definition of double; thus, if
double is that which exceeds equality, half is that which is exceeded
by equality. The like is true of Contraries: if the profitable be
that which is productive of good, the hurtful will be that which is
productive of evil or destructive of good. If, on trying the
contraries, you find that this will not hold, the definition
originally given will be found unsatisfactory.[311] In defining the
privative contrary of any term, a man cannot avoid enunciating in the
definition the term of which it is the privative: but he is not
allowed to define the term itself by means of its privative. To
define equality--that which is contrary to inequality, is improper.
You will require him at once to define inequality; and his definition
must be--the privation of equality. Substitute this definition of the
term inequality, in place of that term itself, in the above-named
definition of equality: and the last definition will then run as
follows: Equality is that which is contrary to the privation of
equality. Here the definiend is enunciated as a part of the
definition of itself; a proof that the original definition--Equality
is the contrary of inequality--is itself wrong.[312]
[Footnote 311: Topic. VI. ix. p. 147, a. 29-b. 4.
We most remember that Aristotle, classifying _Relata_ as one species
under the genus _Opposita_, treats double and half as _Opposita_,
_i.e._ _Relative-Opposita_. I have already said that I think this
classification improper, and that _Opposita_ ought to be ranked as a
species under the genus _Relata_.]