n, you
have got to invalidate the conclusion, for some other inference has
been drawn, and not the one which was inevitable.

And at present, indeed, in order that the case might be more easily
understood, we have brought forward an example pregnant with a
manifest and an enormous error; but it often happens that an error
when stated obscurely is taken for a truth; when either you do not
recollect exactly what admissions you have made, or perhaps you have
granted something as certain which is extremely doubtful. If you have
granted something which is doubtful on that side of the question which
you yourself understand, then if the adversary should wish to adapt
that part to the other part by means of inference, it will be
desirable to show, not from the admission which you have made, but
from what he has assumed, that an inference is really established; in
this manner:--"If you are in need of money, you have not got money. If
you have not got money, you are poor. But you are in need of money,
for if it were not so you would not pay attention to commerce;
therefore you are poor." This is refuted in this way:--"When you said,
if you are in need of money you have not got money, I understood you
to mean, 'If you are in need of money from poverty, then you have
not got money;' and therefore I admitted the argument. But when you
assumed, 'But you are in need of money,' I understood you to mean,
'But you wish to have more money.' But from these admissions this
result, 'Therefore you are poor,' does not follow. But it would follow
if I had made this admission to you in the first instance, that any
one who wished to have more money, had no money at all."

XLVIII. But many often think that you have forgotten what admissions
you made, and therefore an inference which does not follow
legitimately is introduced into the summing up as if it did follow; in
this way:--"If the inheritance came to him, it is probable that he
was murdered by him." Then they prove this at considerable length.
Afterwards they assume, But the inheritance did come to him. Then the
inference is deduced; Therefore he did murder him. But that does
not necessarily follow from what they had assumed. Wherefore it is
necessary to take great care to notice both what is assumed, and what
necessarily follows from those assumptions. But the whole description
of argumentation will be proved to be faulty on these accounts; if
either there is any defect in the argumentati