Listen With Others

I’m not a harpist myself, but I’ve always enjoyed listening to harp music, and the subject of this puzzle — Ravel’s Introduction and Allegro for Harp, Flute, Clarinet and String Quartet — has long been a favourite of mine.

There are many different forms of harp, but they can essentially be divided into two main types: “lever” harps, which use a lever system to change the note produce by a string by one semitone — small “Celtic” harps (clàrsach in Gaelic, or cláirseach in Irish) use this system; and “pedal” harps, which use a series of seven pedals, one for each note A to G, to raise each string by one or two semitones, and are thus fully chromatic. The concert harp is the best-known example of this type.

The innovative pedal system of the double-action pedal harp was first patented in London in the early 1800s by the French musical instrument maker Sebastien Érard. In 1905, following a commission from the Érard company, Maurice Ravel completed his Introduction and Allegro for Harp, Flute, Clarinet and String Quartet, which was intended to show off the capabilities of the Érard double-action pedal harp. This was in response to a similar commissioning of Claude Debussy by the rival Pleyel company for a piece to showcase their own new chromatic harp.

Hearing Ravel’s piece on the radio one day, it suddenly occurred to me that it might be possible to create a puzzle based on it, with a picture of a harp in the grid as the final denouement. Initially I played with some different grid sizes and harp designs, and found that a reasonable harp shape (though it really looks more like a Celtic harp than a concert harp) could be obtained with a 12×14 grid. My idea was to create the harp shape in the grid using contiguous cells spelling the name “The Double-action Pedal Harp”: this, however, has only 24 letters, and the harp design required 29 cells, so I initially decided that the remaining five cells could be filled with letters of the name Érard.

I also hoped to make the design a bit more realistic by including strings, represented by lines drawn between pairs of cells within the harp outline. The design allowed for eight strings — an octave — and I hoped to indicate to solvers where the strings must be drawn by including pairs of “notes” — C-C, D-D, etc., up to C-C again, inside the outline.

The “forced” cells arising from the letters creating the harp design were quite restrictive, and my early attempts to create a suitable grid (without the use of grid construction aids other than pencil, rubber, paper and dictionary, as is usual with me) were not successful. Eventually I dispensed with “Érard” and inserted “Ravel” instead, and this was a bit more successful: I managed to create a symmetrical grid, with a (just about) good enough average entry length. However, it required two entries forming the main stem of the harp to be jumbled, and also included some other grid deficiencies: so, after some thought, I decided to reject this effort and start again.

I noticed that the main stem of the harp could largely be formed from the “double-act(ion)” part of the name, and this led eventually to the final grid arrangement and other features, such as the name of the puzzle (with the entry “double act” at 19 down being unclued), and the form for the clues. In addition to the harp design, I was pleased also to be able to squeeze the name Érard into the grid (also to be highlighted by solvers). The grid I eventually settled for was not without its faults — the average entry length is slightly low, the entries include two abbreviations, and there is one potential ambiguity in creating the harp outline which required the instruction “including the whole of 19 down” to be included in the preamble. However, I decided at this point to go with what I had.

The “double-act(ion)” theme suggested the use of double clues. I wanted solvers to create the name of the thematic piece in some way from the clues, and eventually decided on using an extra word to separate the two parts of each clue, whose first and last letters would spell out the required name (or most of it — the “harp” element is omitted, occurring instead as the final grid design). Again, this gimmick is not without its problems: it is often fairly straightforward to guess which is the extra word, especially as more and more of the message is revealed — one solver mentioned that they worked out the entire message before solving a single clue! And I also had to deal with a few awkward letter-pairs, including GQ, which I did by using the term GHQ as the extra “word” — but, as one solver rightly noted, this isn’t actually a “word”, so I should really have expressed the instruction slightly differently in the preamble.

But the main point of these puzzles is to create something people enjoy solving, so I hope 19 Down managed to achieve that end: I’ve certainly received kind comments from a number of solvers, both directly and on-line, for which many thanks — such encouragement is very much appreciated. And, as ever, thanks to the Listener editors Roger and Shane, and to John, the statistician, for all their help and support, both before and after publication.

Rob. (Hedge-sparrow)