A description and analysis are given of a "speed meter'' for monitoring a c
lassical force that acts on a test mass. This speed meter is based on two m
icrowave resonators ("dual resonators"), one of which couples evanescently
to the position of the test mass. The sloshing of the resulting signal betw
een the resonators, and a wise choice of where to place the resonators' out
put waveguide, produce a signal in the waveguide that (for sufficiently low
frequencies) is proportional to the test-mass velocity (speed) rather than
its position. This permits the speed meter to achieve force-measurement se
nsitivities better than the standard quantum limit (SQL), both when operati
ng in a narrow-band mode and a wideband mode. A scrutiny of experimental is
sues shows that it is feasible, with current technology, to construct a dem
onstration speed meter that beats the wideband SQL by a factor 2. A concept
is sketched for an adaptation of this speed meter to optical frequencies;
this adaptation forms the basis for a possible LIGO-III interferometer that
could beat the gravitational-wave standard quantum limit h(SQL), but perha
ps only by a factor 1/xi=h(SQL)/h less than or similar to 3 (constrained by
losses in the optics) and at the price of a Very high circulating optical
power-larger by xi(-2) than that required to reach the SQL.