tion--that what we call a natural law is merely an arbitrary
selection made by ourselves of certain among natural processes. The
discussion would not be affected by this view, because the argument is
really based upon the existence of a cosmos as distinguished from a chaos;
and therefore it would be rather an intensification of the argument than
otherwise to point out that, for the maintenance of a cosmos, natural laws,
as conceived by us, would be inadequate. And this seems a fitting place to
make the almost superfluous remark, that throughout this present essay I
have used the words "Natural Law," "Supreme Law-giver," &c., in an
apparently unguarded sense, merely in order to avoid needless obscurity.
Fully sensible as I am of the misleading nature of the analogy which these
words embody, I have yet adopted them for the sake of perspicuity--being
careful, however, never to allow the false analogy which they express to
enter into an argument on either side of the question. Thus, even where it
is said that the existence of Natural Law points to the existence of a
Supreme Law-maker, the argument might equally well be phrased: The
existence of an orderly cosmos points to the existence of a disposing mind.

[26] First Principles, pp. 27-29.

[27] It may be here observed that this quality of indefiniteness on the
part of such reasoning is merely a practical outcome of the theoretical
considerations adduced in Chapter V. For as we there saw that the ratio
between the known and the unknown is in this case wholly indefinite, it
follows that any symbols derived from the region of the known--even though
such symbols be the highest generalities which the latter region
affords--must be wholly indefinite when projected into the region of the
unknown. Or rather let us say, that as the region of the unknown is but a
progressive continuation of the region of the known, the determinate value
of symbols of thought varies inversely as the distance--or, not improbably,
as the square of the distance--from the sphere of the known at which they
are applied.

[28] _i.e._, illegitimate in a _relative_ sense. The conclusion is
legitimate enough in a _formal_ sense, and as establishing a probability of
some _unassignable_ degree of value. But it would be illegitimate if this
quality of indefiniteness were disregarded, and the conclusion supposed to
possess the same character of actual probability as it has of formal
definition.

[29] In order