Listen With Others

eXternal:

When looking for a theme, I very often read through Brewer’s Dictionary of Phrase and Fable. I was reading the entry for King Midas, which featured the myth about Apollo giving the king ass’s ears as a consequence of judging against him in a musical contest. I had seen an animation based on the story of a king with ass’s ears, the secret with the barber and the subsequent betrayal of the secret by musical instruments, but I’d never linked it to King Midas (I think it may have been based on the Portzmach story, which is under the same entry). I then looked at on-line accounts of the myth to see whether I could get further detail to establish a thematic idea for a puzzle and various other versions mentioned donkey’s ears rather than ass’s ears. DONKEY’S YEARS suddenly came to mind as a potential way of hiding the donkey’s ears and noticing the accounts of them being covered by a Phrygian Cap with Midas being a king of Phrygia developed the thematic idea further. I noted that the second Y would need removing to reveal the ears and, as luck would have it, PHRYGIAN CAP contained a Y. This meant that I could have two large thematic entries in the grid with the ears being hidden in plain sight, rather than being hidden over bars through various entries or on a diagonal, for instance. That was immediately appealing as something a little different from the norm and potentially amusing for solvers, thus the endgame was born.

I decided to just check in Qxw that I could get a skeleton grid where the two long entries crossed and PHRYGIAN CAP could be removed leaving real words across the gap (a necessity, in my view). This is quite simple using Qxw’s free-light facility and I found it could easily be done with very few constraints on the grid-fill. It was quite obvious to me that the better-known myth regarding Midas should form the core of the puzzle, leading to the identification of the person. OR and AU are cryptic stalwarts and transformations are often used in barred puzzles, so the next question was how to implement this. I favoured the transformation being done in answers rather than clues, as OR and AU would be easy to see in clues once one or two had been found, although it could have been done the other way round with letters needing to be changed to gold. It seemed the way to work towards the endgame would be for the letters of MIDAS to be linked to the AU/OR transformations, so I did some checking in Qat to look at the possibilities and it confirmed that there were very many candidates for various positions in entries.

I have collaborated with Serpent on various puzzles (Two Listeners… is our twelfth published puzzle) over a few years and we pitch ideas at various stages of development. Sometimes, I will approach Serpent when I have an idea which is beyond my technical expertise, but at other times, it can simply be that we haven’t done a puzzle together for a while or that we don’t have many in the various barred-series’ pipelines. It must have been time for a collaboration, so I shared the idea and he was quite interested in working on it with me. He was a bit dubious as to whether we’d be able to get the five transformations for MIDAS in the puzzle, given the constraint of the real words remaining after the removal of PHRYGIAN CAP, so suggested using the gimmick in clues rather than the grid. He was quite busy at the time, so we decided I should try to produce a grid, firstly using the idea of the gimmick in answers.

The first thing to decide on was grid dimensions — DONKEYS YEARS is 12 letters, so I tried 12×12 grids, looking mainly at the two long entries which would be placed in symmetrical opposition to the ears and cap. With 12×12, the partner to DONKEYS YEARS would go through the C of PHRYGIAN CAP and the best I could come up with was WAPPENS(C)HAWS, which seemed very obscure. Changing to a 12×13 grid meant the long partner to the ears would go through the N of the cap, giving SOUTHER(N)MOST as an option as a more solver-friendly candidate and giving SEA ELEPHANT as a nice option for the partner to the cap, always better than some dull and obscure chemical compound.

I proceeded with the 12×13 arrangement and experimented with various options for the AU/OR words, making sure these were always in checked cells. I think I decided to have the ‘untouched/untransformed’ words in the grid, so that there weren’t so many AU/OR combinations within it. I realised that I could achieve a good grid with MIDAS being produced in clue order with the aforementioned constraints of cap removal, so decided to try for KING MIDAS, which would provide a more satisfying challenge for the solver and mean that we didn’t have to consider producing KING in some other way, such as from the clues . I managed to get a fill with 5.7 average word length, no entries smaller than four letters, only a pair of fully-checked fives and few obscurities. There would be various other combinations to check, but I was happy with this as a start and sent it to Serpent for further discussion/work.

Serpent:

The credit for this puzzle’s concept must definitely go to eXternal. (It is an understatement to say I was “quite interested in working on it”: I thought it was a really cute idea and was delighted to be able to contribute something to the puzzle.)

By the time I got properly involved in the development of the puzzle, eXternal had already established that it was possible to find a grid-fill satisfying the thematic requirements and having a respectable average entry-length. It remained to finalise the grid-fill, preferably using as few plurals, inflected forms and obscure words as possible (something we both set considerable store by). I set to work on the skeleton grid containing the four long entries, using Qxw and a customised dictionary, and fairly soon we had a grid we were both happy with.

All that remained was to write the clues. eXtent puzzles have 50% eXternal and 50% Serpent clues, written independently, then checked and commented on by each other. (The reader might like to speculate as to which of us wrote the across clues and which the downs.)

We then sent the puzzle to our estimable test-solvers, David Thomas and Norman Lusted. Their feedback was encouraging but it did prompt us to require the solver to demonstrate understanding of the theme by writing KING MIDAS under the grid.

As always, we received comprehensive feedback from the editors, which helped to improve a number of the clues and the preamble, for which we are very grateful. (I was pleased to see they retained “untouched” in the preamble and the title, both suggestions of mine.) We’re equally grateful to all solvers who provided such nice feedback via John Green.