It is interesting to see how someone's mistake if inculcated enough can become dogma. Simply by looking at the muscles, as was pointed out to me, the actions are apparent. As the obliques attach on the temporal aspect of the orbit, when they contract, the orbit will rotate against the force around its axis. In other words, the eye will abduct. How clinicians missed this for apparently a long time is beyond me. However, the British have a theory:
Our final conclusions, and what is now our policy for instruction on this muscle, is as follows. The superior oblique, acting in isolation, turns the eye down and out. However, if it was tested clinically by the patient being asked to look down and out, its action could be mimicked by the combined action of inferior and lateral recti. This is particularly so as the inferior rectus acts most effectively when the eye is abducted (looking laterally). Thus if the patient is asked to look down and in, these muscles are excluded and the problem is solved. Essentially we are testing the ability of the superior oblique to look downwards.Ah ha! So it was the ophthalmologists' fault! Well, well. Mystery solved!
The confusion, which I hope the above clarifies, is compounded by some ophthalmologists being so used to testing the muscle by asking the patient to look down and in that they have forgotten that the isolated action is down and out.