This first response was from an amateur, but a highly respected and published one:

"A Celtic origin for the Maes- element is possible but
there is one sticking point. Maeshowe is a later name
for the site, which supplanted Orkahaugr during the
earldom or later.

So, although I've thought about this long and hard
over the years, it makes no sense for a Celtic name to
supplant a Norse name, in an area where the language
was a derivative of Old Norse and particularly given
the fact that practically all our placenames have a Norse origin (although there are a few badly corrupted
elements that could, and probably, have a pre-Norse
element).

It doesn't fit the other Maes- placenames either.
Mount Maesry, for example is on the seashore."

This second response is from an academic, a Dr of Celtic studies:

"I cannot pretend to be a place-name expert, but fortunately my colleague <name removed> is an expert on Norse place-names. I've taken the liberty of copying this to him. When I hear the name Maeshowe I do think automatically of Brittonic 'maes', but Orkahaugr pretty much convinces me that the -howe part of the name is Norse -haugr. On top of that the place-name evidence from Orkney in general tells fairly emphatically against the idea of pre-Norse survivals: it would be strange for one element, maes, to survive alone.

My instinct would therefore be that maes- has a Norse origin, and has for some
reason come to replace orka- as the 'specific' element denoting this particular
'haugr'. It would be interesting to know when maes- is first attested there."

I am awaiting responses from two further, similarly qualified, experts in Celtic/Pictish history, and one expert in pre-Viking Orkney.

I do agree that it would be most unusual for a Celtic name to replace a Norse one, especially given that Orkney is an area where the locals would have been speaking a language closely related to Old Norse at the time. However, I do feel that the Celtic definition of the name seems to be too neat a physical description of the sites to be a pure coincidence. My own thoughts are that an older Celtic based name persisted in oral use among the locals throughout the "viking" period (probably while the site was referred to in writing, and "officially" as Orkahaugr) and eventually came to outright replace (or merge with) the Norse name, or maybe even that the Norse on the island for some reason just started to use an older Celtic element in their own name for the places.

I don't like the general "it's not Norse so it can't be right" argument - ok, so there's no certain Celtic source for "howe", but there's even less of a Norse source for "maes-".

Anyway - I would be delighted to hear anybody's thoughts on any of this, either here or by email.

Lian

PS If any of the experts I contaced happen to read this, please accept my apologies for posting without your permission. I have removed any clues to your identites.

This just received from another academic:

"Tradition has it that Mae was the personal name of a giant – this of course is the usual way of explaining unknown elements in Norse and other placenames.

From my limited knowledge of such matters, I would say that your suggestion is not an unreasonable one, given that the pre Norse ?Pictish language is now believed by some leading scholars to have been an early Celtic tongue akin to a primitive form of Welsh; rather than to Gaelic. This might help with the objections to the idea of your preferred derivation on the grounds of it not being a widespread one throughout Scotland, raised by the person who so far as I know first published the idea of this derivation, in the 1860s in the Proceedings of the Society of Antiquaries of Scotland.

Pre-Norse name survivals in the Northern Isles are scarce and all are much debated – for example some island names (such as Unst, Yell) don't fit obvious known Norse derivations but are plainly very ancient. But it seems to me that there would not be anything inherently unacceptable about a very prominent ancient landmark acquiring a placename which mixed an old local name with an incoming Norse element – the Howe bit is, of course, from Old Norse haugr, a mound (especially a burial mound). The main problem we have is that the name Maes Howe in any form does not appear in documents until very late, so we have no idea of any earlier name forms – if only we knew how it was spelled around 1000 AD, it would be much more likely that your idea could be supported or denied conclusively.

If you would like to discuss this idea with someone much more learned than <his organisation> on these matters, I would suggest you contact Dr <name and email address removed>. I suspect he has thought more deeply about this matter than anyone else currently living. Please feel free to say that I recommended you to him."

I checked the 1865 Proceedings of the Society of Antiquaries of Scotland and found the document he references. It's by Ralph Carr, and his main problem with this derivation was that the use of "maes" is not spread across the Scottish mainland. He suggests the Irish (Erse) source "mais", since several Orcadian promontories are designated by the Erse term "moul". He then goes on to translate and discuss the runes found at the site.

I will be following up by contacting the individual identified to me above.

While not immediately relevant to the "Maes-" question, I thought this site might interest many of you:

http://www.s-gabriel.org/names/tangwystyl/pictnames/