Pronouncing Michael Robartes and the gyre

I am asked with some regularity about two pronunciation questions that seem to nag at readers of later Yeats, neither of them to do with Irish names or recondite terms such as 'congeries', all of which can be checked relatively easily. (Apparently it's 'con-jérry-ez', by the way, with elements of variation.)

The questions are: How do you pronounce the name Robartes? And is the 'g' of  'gyre' hard or soft?

Though it has been a while since I learnt that the surname Robartes is generally pronounced as two syllables with the stress on the first, being a variant spelling of Roberts or Robards, I still find it hard to shake a preference for the pronunciation that I first adopted of three syllables with the stress on the second (a little like Pilates as opposed to Pontius Pilate). I don't think that there are any recordings of Yeats reading a poem that names him (U Penn has, I think, as full an archive as is possible from the few extant recordings), nor have I come across any record of what he said, but I accept that robarts is probably the best pronunciation.

The problem is a little different with 'gyre'. Here, the standard accepted pronunciation for the word is with a soft 'g', and it is given by the dictionary as jīr (I'm using Chambers). Many Irish accents, like most Scottish and US accents, would normally pronounce a retroflex 'r' at the end — /dʒaɪɹ/ or /dʒaɪɻ /— while most English and some New England accents would have an open ending— /dʒaɪə/ .

However, Richard Ellmann states that the whirling symbol 'was a spiral, which Yeats preferred to call a gyre (and pronounced with a hard "g")' (Yeats: The Man and the Masks, p. 231). Although he gives no reason for this statement, he had probably heard it from George Yeats's own lips.* (Apparently Lewis Carroll also expected a hard 'g' for 'gyre' in 'The Jabberwocky', as the basis for this word, 'gyroscope', was pronounced with a hard 'g' in his day.)

Further confirmation comes from a typescript that Yeats dictated, where the word is misspelt as 'guyers', which indicates that the typist heard a hard 'g'. The word has been corrected in later occurrences.

NLI MS 36,272/24, p. 7

 (A similar instance elsewhere has the seventeenth-century theosophist Jakob Boehme typed up as 'Burmah'.)

Listening to Yeats's own readings of poetry, for instance 'The Lake Isle of Inisfree', the final 'r' sounds of 'there' and 'core' are strong, so that if one wants to follow what Yeats himself said, it would probably be /gaɪɻ / .

But do we adopt Yeats's pronunciation? Even if it goes against our own accent or region?  

How should the falcon and falconer of 'The Second Coming' be pronounced—long or short 'a'? with or without an 'l' sound? (fawk'n is a traditional pronunciation, as is fol-kon, but făl-kon has become general nowadays).
Does the philosopher George Berkeley echo London's Berkeley Square (barkly) or California's Berkeley University (birkly)? 
How close did Yeats intend a rhyme such as 'work' and 'clerk' to be in 'At Galway Races'? ('clerk' rhymes with bark in most British accents, but the rhyme implies clurk).

Do we keep an early-twentieth-century pronunciation in the twenty-first century? We know that Wordsworth and Byron rhymed water and matter, while Pope rhymed line and join. Do we keep the rhyme in order to preserve the internal sound patterns and music—and risk sounding ridiculous—or do we accept that time and place may erode some of these sound elements?  

As neither 'Robartes' nor 'gyre' is in any danger of being misunderstood if we pronounce it either way, we may prefer accepted standard usage or to heed what Yeats himself said, but there is nothing to stop our Robartes being trisyllabic if we choose and our gyres from being soft.

 

*Note (added May 2023): Curtis Bradford, in his article 'George Yeats: Poet's Wife', confirms that she would have been the source: 'Her verbal memory was astonishing and, in spite of the difference of sex and hence of vocal quality, I found her by far the most convincing imitator of Yeats talking or reading poetry.... She could always suggest the right Yeatsian stress: 'My fànatic hèart"; recollect Yeatsian peculiarities of pronunciation: "Bỳ zàn ti um", "gyres" with a hard "g", "gyrate" with a soft, "demesne" as though spelled domain, and so on.' (written in 1955; published in The Sewanee Review, 77: 3 (Summer, 1969), 385-404, at 398; https://www.jstor.org/stable/27541741 ).