Eleven
properties are given. The punishment must be (1) 'variable,' that is,
capable of adjustment to particular cases; and (2) equable, or
inflicting equal pain by equal sentences. Thus the 'proportion' between
punishment and crimes of a given class can be secured. In order that the
punishments of different classes of crime may be proportional, the
punishments should (3) be commensurable. To make punishments efficacious
they should be (4) 'characteristical' or impressive to the imagination;
and that they may not be excessive they should be (5) exemplary or
likely to impress others, and (6) frugal. To secure minor ends they
should be (7) reformatory; (8) disabling, _i.e._ from future offences;
and (9) compensatory to the sufferer. Finally, to avoid collateral
disadvantages they should be (10) popular, and (11) remittable. A
twelfth property, simplicity, was added in Dumont's redaction. Dumont
calls attention here to the value of Bentham's method.[413] Montesquieu
and Beccaria had spoken in general terms of the desirable qualities of
punishment. They had spoken of 'proportionality,' for example, but
without that precise or definite meaning which appears in Bentham's
Calculus. In fact, Bentham's statement, compared to the vaguer
utterances of his predecessors, but still more when compared to the
haphazard brutalities and inconsistencies of English criminal law,
gives the best impression of the value of his method.

Bentham's next step is an elaborate classification of offences, worked
out by a further application of his bifurcatory method.[414] This would
form the groundwork of the projected code. I cannot, however, speak of
this classification, or of many interesting remarks contained in the
_Principles of Penal Law_, where some further details are considered. An
analysis scarcely does justice to Bentham, for it has to omit his
illustrations and his flashes of real vivacity. The mere dry logical
framework is not appetising. I have gone so far in order to illustrate
the characteristic of Bentham's teaching. It was not the bare appeal to
utility, but the attempt to follow the clue of utility systematically
and unflinchingly into every part of the subject. This one doctrine
gives the touchstone by which every proposed measure is to be tested;
and which will give to his system not such unity as arises from the
development of an abstract logical principle, but such as is introduced
into the physical sciences when we are able