Australian Crosswords

It’s amazing out of the thousands of possible words that can fit into a grid, how often words could end up being related in some way. It’s not uncommon to have two words that are synonymous or a few words that have a medical bent etc. Of course, it’s up to the setter what words are chosen, but still, being able to create a mini-theme without really trying must be statistically difficult but it often doesn’t seem that way. It also happens sometimes that words are included that seem harmless in isolation, but when combined with others in the grid, it may appear that the setter is making some kind of statement. In a recent Australian Financial Review crossword for Christmas, I finished the grid (complete with a colourful Christmas tree inset), filled in the words and wrote the clues. In some regards doing this is routine for me, or I may be working on a number of puzzles at once, and I can forget about context. Such was the case here. I exported the finished grid, and while eying it over, I noticed I had included a five-letter religion right in the middle of the crossword at the heart of the Christmas tree. It’s just a puzzle and including that religion doesn’t mean anything, but its positioning might make people think otherwise. I quickly rearranged a few words, clued them and sent it off. No need to cause an issue if it can be avoided.

I was alerted by the rec.puzzles.crosswords newsgroup to an editing mistake in a recent Times crossword, and thought I’d talk about it briefly as I think it highlights a couple of interesting things. Here’s the faulty clue: Disturbance in field scattered old fragments in two minutes (8,5)*. The mistake is that “scattered” should be “scattering”, possibly not a problem if it was an anagram indicator, but as “scattering” was actually part of the anagram fodder, the change made the clue incorrect from a wordplay perspective.
I’ve always believed the UK Times puzzles were test-solved, and if they are, then it’s clear that the editing process happens after the test-solving, as there’s no way such an error would get through. If there’s no test-solving, then maybe there should be, and if there is then maybe it should happen after the editor has done an initial rounds of changes, which leads me to the second point: the editor obviously doesn’t “cold-solve” the submitted crosswords, that is, solve them as if they were an actual solver – it’s more likely that they go through it with the answers at hand, which is a dangerous practice. Cold-solving takes more time but will force the editor to understand the workings of clues which will help them better realise the impact of making editorial changes. (BTW, the UK puzzles I’ve written have included explanations at the request of the editor – do they look at them, or only look at them when they are stumped?) Lastly, it’s clear the editor doesn’t feed back the changes to the original setter, if it had happened in this case the setter would have realised the editor’s mistake. I know from my dealings with editors they reserve the power to make changes without consultation or approval – which is their right – but a courtesy email listing changes would have been beneficial in this case.
So how did the mistake happen? It’s not a simple case of an “improvement” gone wrong, although the editor no doubt believed that’s what they were doing. There’s been an increased use over the years of descriptive language in clues that isn’t common in everyday speech. “AA boxes B” becomes “ABA” and “ABC changes in DD” becomes “DBACD” etc. No-one I know describes or defines things this way and I believe it only works if the same sort of language continues throughout the whole clue. “ABA” (definition) = “AA boxes B” (wordplay) just doesn’t work for me. However, many setters use this style of clue and that’s what’s happening in the faulty clue: “fragments” is used as an anagram indicator and should act on “scattering” (plus “o” for “old”), however, the editor missed “fragments” and saw “scattering” as the anagram indicator and decided to improve it by changing its form to “scattered”.
It’s rare that a puzzle like the Times contains such a mistake as it uses the best setters and the best people in the puzzle production process, but this mistake perhaps highlights a deficiency in the process that could possibly be eliminated with a tweak or two.

[*Disturbance in field scattered old fragments in two minutes (8,5) should be Disturbance in field scattering old fragments in two minutes (8,5) – MAGNETIC STORM = anagram of SCATTERING + O(ld) inside MM]

I think a cryptic clue is very much like an algebraic expression, where the wordplay is represented by the various symbols on one side of the equation that should equal an answer on the other side of the equation. It makes sense, then, that a cryptic clue should be completely verifiable, and able to be declared as unquestionable. However there doesn’t seem to be anything set down that allows solvers to consistently evaluate and grade cryptic clues. That is, there isn’t a formula that can be applied that will quantify a clue and give it a rating of some kind. Unlike a mathematical equation that can clearly be defined as valid and universally accepted as so, the same cryptic clue can elicit vastly different responses depending on who solves it, no matter what kind of construction is used in the wordplay. The only definitive aspect of a clue that I believe can be assessed and accepted is whether it contains a mistake – all other aspects come down to opinion. For example, a clue that may be technically accepted by all may have a suspect surface reading, making it for some a deficient clue; another may contain a definition that some don’t accept as good enough and others are happy with; yet another may be sound, read well and be perfectly acceptable to most but be considered “boring” by an experienced solver – so is such a clue a good one or not? I believe most aspects of cryptic clues are judged on a solver by solver basis, and no one clue will be considered the same way by all solvers. Note the setter doesn’t really have any control here – no matter their reasoning or justification for clue components, the solvers are the ultimate judges. A setter may be able to point to a particular definition of a word that justifies its use as an anagram indicator, but if a solver doesn’t like it, then the clue from the solver’s perspective (the one that matters) isn’t a good one.
The clue quality of “fair” that gets talked about regularly is a subjective view held by a solver and can’t be generally attached to a clue, as one man’s “fair” is another’s “rubbish”. Too often a clue is designated fair if it is solveable, which of course isn’t true, as all clues are solveable even those ones that are badly written, contain mistakes, or don’t have accurate definitions. The nature of a crossword provides cross-letters that can point to an answer regardless of the clue, making even the worst clue solveable.