Cockcroft and Lameijer (rather an elderly copy now) tell us this:
"The term ‘narrow channel’ is not easily defined. In deciding whether a particular stretch of water is or is not a narrow channel the Courts take into account the evidence as to the way in which seamen usually navigate the locality and the advice given by the Elder Brethren. A narrow channel need not be of any particular length and does not necessarily terminate at the last of the buoys or objects marking the channel. The narrow channel rule has been held to apply to the passage between two piers and to 100 metres (yards) outwards beyond the objects marking a harbour entrance. It was held not to apply to a recommended route between two buoys where vessels could have gone outside them in safety. Passages approximately 2 miles wide have sometimes been considered narrow channels. In considering the passage between Duncansby Head and the Skerries in the Pentland Firth (Anna Salen-Thorshovdi, 1954) Mr Justice Willmer said: For myself, I certainly see difficulties in applying the ‘narrow channel’ rule to a passage which is nearly four miles wide. I should hardly have thought that ‘narrow’ was the word to use for this passage, for it is not a particularly narrow passage. In the Faith I-Zndependence (US Court, 1992) the passage between buoys at the entrance to Delaware Bay, approximately 1.2 miles wide, was held not to be a narrow channel but it was held that good seamanship and prudent navigation require that every vessel keep to starboard if safe and practicable. Rule 9 will apply to any narrow channel connected with the high seas which is navigable by seagoing vessels unless there is an inconsistent local rule. It does not apply to lanes of traffic separation schemes although such lanes may be relatively narrow."
The bit I've bolded seems to support both sides of the argument.
There is nothing in the MAIB report to suggest that the skipper of Phoenix would have known what it meant.