476
DOC.
465
FEBRUARY
1918
for the
mailing
of which I
thank
you heartily.
This section
was
not printed
in
the
journal
Die
Naturwissenschaften,
as
Dr.
Berliner writes
me.[3]
It
coincides
with
my
views
so thoroughly
that
I
am
almost
astonished
that
someone
else
has
written
it.
I
concur
thoroughly
with
the author
especially
in his
criticism of Mach
which, despite
its
very friendly tone,
is
really quite
sharp
in content.[4] Above
all,
I
emphasize
the
following
statements
(on p.
62
of the
booklet): “Now,
it
is
possible,
however,
to describe the
same
factual
findings by
various
judgment
systems,
consequently,
various theories
can
exist for which
the
criterium of
truth
applies
in
the
same way,
which
thus
all
can
take
care
of the observations
to
the
same
degree
and lead to
the
same
predictions.
Various notation
systems
are
simply
assigned
to
the
same
objective reality,
and various forms
of
expression convey
the
same
factual
findings. Among
all
the
possible approaches
that contain the
same
core of
truth in such
a way, one of
them has
to
be
the
simplest;
and that
we give
preference
to
this
one
in
particular is
not
based
merely
on
practical
economy, on
a
kind
of
mental
convenience
(as
has sometimes
been
thought),
but
is
logically
grounded
in
that
the
simplest theory
contains
a
minimum
of
arbitrary
factors."[5]
It
is
exactly
in this
sense
that
I
spoke
of
“home-baked
logic”
in
my
letter
to
you,
which
could be
lacking
in
a
mathematical
theory
because
mathematics
is,
as
you know,
in
the
position
to
make
arbitrary
definitions
to
suit
itself.[6]
I
could
not
possibly
think
you capable
of
a
gross logical error,
as some
philosophers
evidently
have done in
assessing your
old
theory
of
relativity;[7]
it
would
not
be
in
keeping
with
my opinion
of
you.
When
one
says,
“the
Earth
rotates,”
a
large group
of
phenomena-Foucault’s
experiment,
the
trade
winds,
the
deflection
of
a
falling stone, etc.,
etc.-are summarized
most
briefly
in
this
judgment.[8]
Naturally, according
to
your
principle
of
the
transformability
of
the fundamental
equations,
it
is
possible
to
adopt
a
coordinate
system
in which
the Earth
is
at
rest,
but then the
scheme of
spatial
space-time
order in
which
the
objects
are
placed
must be
assigned
certain
characteristic
guv's
which,
to
use
Schlick’s
expression,
contain
“arbitrary
elements”
[willkürliche Momente]
since
they
are
not determined
by
the
presence of
real
bodies;
thus, although
the
representation
describes
the
phenomena correctly,
it
now
becomes
unnecessarily complicated
and
one
must therefore
reject
it.[9]
Hence
the
expression
“the
Earth
rotates”
will
always
retain
its reasonable
content
even
despite
general
transformability.
I
would
like
to
demonstrate this from
another
angle as
well.
Take
a
field of
an
electron
at
rest.
According
to
the
principle
of
relativity
of linear
motions,
through
a
linear
transformation
I
can
immediately
obtain
from it
a new
integral
of
fundamental
equations
which
represents
the
“uniformly moving
electron.”
We
all know of
what
significance
this has been for
experimental physics
in
particular,
which
certainly
deals with real
factual
findings.
Now,
is
it
possible
also
to
derive
the
integral
that
represents
the
rotating
electron from
the
general
transformability