The objective of The Setting Room is to provide a knowledge base for setters and aspiring setters. If you have any observations, corrections, questions, or suggestions for topics to be covered, please leave a comment below.
all about cryptic crossword clues
The Setting Room is a welcome addition to the site.
There is a point I’d like to raise about indirectness in wordplay.
This clue of mine for KITTEN – “Youngster tailed relatives cycling around tent” (KI[n] + TENT cycled to TTEN) – was recently panned by a contributor to 225. Part of the criticism was that the cycling had to be applied to a word actually appearing in the clue, rather than a word indicated by the clue, thereby making the answer “obvious”. The reason I had done that was that the cycle element of the wordplay applied only to part of the answer, and I did not want to be accused of making things too complicated.
The question this raised for me is whether cycling clues that don’t actually specify exactly what is to be cycled are, in fact, fair. If the solver has to indentify a word which is a synonym of, or defined by, an element of the clue and then cycle its letters into the appropriate order, that is a two-stage process – so the answer is not obtained directly from the fodder.
Indirect anagrams have, for good reason, been excluded from reputable cryptics for years, but other instances of indirectness (by which I mean an element of wordplay that requires two or more operations to be performed) seem to be condoned, or, at any rate, tolerated.
Do you have any views on this?
Thanks, Richard
No-one has ever accused me of not having a view!
First of all, let me say that any criticism of your clue based on the easiness of solving it is unjustified. I initially intended to go into more detail, but I don’t think there’s any need.
Your question about indirectness revolves around where the line is drawn between the acceptable and the unacceptable, which is not necessarily the same as the line separating the fair and the unfair.
Consider four clues:
1. Criminal takes birds (5)
2. Capital of France staggers couples (5)
3. Topless model returning without English youth (3)
4. Shock two US states (4)
The answer (or, at least, one answer) to (1) is TERNS, an anagram of RENTS (Chambers gives ‘take’ = ‘rent’). This is an indirect anagram of the sort that would not now be published in any UK crossword.
The answer to (2) is PAIRS, an anagram of PARIS. This is an indirect anagram, but I can only think of two possible 5-letter words answering to ‘Capital of France’, EUROS amd PARIS. However, this clue would be tarred with the same brush as (1), and would not be allowed.
The answer to (3) is LAD, with IDEAL (‘model’) losing its first letter, being reversed, and shedding E. There are four stages involved in getting from ‘model’ to LAD, but they are all ‘licensed manoeuvres’, and there are no fundamental grounds that I can see for rejecting this clue.
I don’t think that indirectness per se is an issue, but if lines are to be drawn then realistically they can only be drawn based on general principles, so indirect anagrams are outlawed, while reductions, deletions and reversals of the sort in (3) are allowed. A word which has to be deduced by the solver can lose its last letter and be cycled, but only text in plain view can be anagrammed or have its first and last letters selected – in effect, a certain degree of integrity has to be retained following each manipulation.
In this respect, clue-writing is a bit like ice skating – you can perform a sequence of manoeuvres, but all of them have to be legal. Should you have wished, you could have replaced ‘tent’ with an indication thereof in your KITTEN clue.
The answer to (4) is MANE, MA + NE. A simple charade, but with 50 states to choose from, there are well over 2,000 permutations from which to choose. Would you say that this was a fairer clue than (2)? The answer is largely immaterial in the context of the foregoing, because this clue doesn’t breach any general principles. That’s where the editor should be the judge of fairness. Taking into account the solvers of that particular puzzle, they would have to decide whether to reject this clue, just as they would have to decide whether “Linking five notes (8)” for RELATING was a bridge too far.
Thank you very much, as ever, for your considered and helpful response.
From what you say, I assume that, by extension, the requirement to select (or ignore) letters appearing in regular positions should be imposed only in relation to text in plain sight.
But where do we stand on abbreviations? I was under the impression that single letter abbreviations had to be indicated directly, but have twice recently seen “monarch” used in Times puzzles to indicate R, and it is not unusual to see “direction” used for N, E, S or W.
Yes, I think the ‘rules’ preclude a regular selection from a derived word. Although something like “Silence former PM Margaret every now and then (4)” for TACE [T(h)A(t)C(h)E(r)] is accessible, I can think of other possibilities which would be very tough indeed.
When it comes to abbreviations, there’s a lot of ‘custom and practice’ involved. My view would be:
1. Explicit use of the word or words abbreviated. Always acceptable, eg ‘time’ for T and ‘Territorial Army’ for TA. Although R is an abbreviation for ‘River’ and not ‘river’, either is acceptable. Similarly, ‘Stokes’ is seen as acceptable for S (even when it doesn’t come at the start of a sentence), even though S is an abbreviation for ‘stokes’ and not ‘Stokes’.
2. Indications of abbreviations which have a life of their own. By this I mean abbreviations like LP or OTT, which can be used as standard parts of speech, as in “I used to have that LP” or “That’s a little OTT”. Normal definition rules should apply here, so ‘record’ or ‘album’ for LP is fine by me, as is ‘extreme’ or ‘too much’ for OTT.
3. Descriptive indications for abbreviations which directly identify a unique entity (arguably these are part of the preceding category ). I’m thinking here about something like ‘e’, the base of Napierian logarithms, or ‘c’, the speed of light, and I’m inclined to give setters a bit of latitude. I think ‘irrational’ can reasonably indicate E, and ‘top speed’ is ok for C.
4. English translations of abbreviations for foreign words or phrases, such as Latin ones. I would suggest that when these abbreviations are encountered in real life they would normally be pre-processed into English, so any relevant translation based on normal English usage is valid. Hence ‘namely’ or ‘specifically’ -> ‘scilicet’ -> SC, and ‘for instance’ or ‘say’ -> ‘exempli gratia’ -> EG. Could this justify ‘monarch’ -> ‘rex’ (or ‘regina’) -> R? No, in English the abbreviation is used specifically after the name of a king or queen, and has precisely that meaning, so ‘monarch’ or ‘ruler’ doesn’t cut it for me. Similarly, I wouldn’t accept anything other than ‘as above’ for ‘ut supra’ -> US.
5. Indications of abbreviations which don’t fall into categories 2, 3 or 4 but are part of crossword lore. When we see ‘artist’ we think RA, likewise when we see ‘gunners’. They are what they are – I don’t use them myself, but that’s largely because they are so hackneyed.
6. Indications of abbreviations which are not part of the language but are all around us. The points of the compass would fall into this category. I wouldn’t use ‘direction’ for E or SW myself, but if ‘pole’ is ok for N/S, I would find it hard to argue with ‘cardinal point’ for N/E/S/W. A marginal call, I would say. Best not to ask what I think about ‘Kent’ for SE.
7. Anything else. ‘Monarch’ for R, ‘Collection of verses’ for C, ‘Tory’ for C. No. Just no.
This is something of a rapid brain dump, and may not be very well expressed, but I hope that it gives an idea of roughly where I stand on the topic.
Thanks, Dr C, for this comprehensive reply. I think we are in pretty close agreement.
The misleading use of single letter abbreviations in blocked puzzles is something I find hard to understand, particularly in the Times and Telegraph, where setters are supposed to be bound by editorially-approved lists. For example, it is not at all uncommon to see “liberal” for L, “conservative” for C, “independent” for I, and “nationalist” for N. Yesterday, in the Times quick cryptic, we had an instance of “pounds” for L.
As far as the use of “cardinal point” for any of N, E, S and W is concerned, that is covered by the definition in Chambers, isn’t it?
Someone on 225 said yesterday, in a post offering help to new solvers, that “force” is often used for G. I don’t myself recall any instances, although I note that it’s given in Bradford’s so it must have some currency. A G-force is a force, but that doesn’t justify the use of “force”, without more, to indicate G.
I’m conscious of having taken up too much of your time on my concerns about indirectness, but shall no doubt be back before long about something else. It is so refreshing to be able to discuss points of interest to setters and to receive a well-argued response here. I find this site enormously helpful (despite the fact that I have no interest in barred cryptics), and make very frequent reference to it when compiling puzzles. Thank you so much for all your hard work.
Noting also that the listed “Special Clues” section does not exist as such.
Thanks for that – I’d written the section, but failed to include it. Now fixed, I hope.
Hello, A minor quibble I have with barred puzzles is the way multi-word or hyphenated clues are enumerated, specifically the lack of a break down of the enumerations. Having just perused the topic in the setting room, I wonder if an explanation as to how the differences came about is in order.
As for the setting room itself, many thanks for what must have been quite an effort to compile. I look forward to delving further into its content. J.
Thanks, John.
What I can say with reasonable confidence:
The differences appear to be a result of the evolution of the two types of puzzle. When blocked puzzles first appeared, no enumerations were given, with the solver simply having to count the number of cells available. This continued in the Times puzzle at least until 1933, but by 1936 that puzzle included enumerations much as they are found in blocked puzzles today; the accompanying note that “Figures in parentheses denote the number of letters in the words required” suggests that the enumerations were a relatively recent addition.
When Torquemada ‘invented’ barred puzzles in the 1920s, he did not provide answer lengths in his clues, and neither did the Listener puzzle when it started off in 1930. I believe that Ximenes was responsible for introducing them in barred puzzles during the 1940s, but I don’t have any Listener puzzles from that era; he simply gave the letter count and the number of words, treating hyphenated words as a single word, and Azed has always done the same. In the slip for competition 410, Azed wrote:
“… it’s suggested that I indicate hyphenated words either by simply showing that they are hyphenated or by stating where the hyphens come. My inclination has been and still is to follow X’s tradition of leaving the extra hard work to you.”
I think that Ximenes was probably influenced by the high proportion of checked letters in his early barred puzzles, quite a few entries (even some of 6 letters) being fully checked. As far as I know, all the major UK barred cryptics follow the same practice, with the exception of the Spectator puzzle, where the enumerations are as one would expect in a blocked crossword. A logical reason for this might be the fact that this puzzle often contains a number of unclued entries, but I think the reality is that it started out as a blocked puzzle, and by the 1950s (probably well before that) it was using the same enumerations as the other blocked puzzles. When it became a barred puzzle in July 1971, the previous enumeration style was retained, and it has not changed since. The crossword also continues to include clues that would not be acceptable in Ximenean circles, making it something of a hybrid.
I hope that is helpful, but I would be delighted to hear from anyone who has any further – or contradictory – information, so that I can add something that is as comprehensive as possible to the Setting Room.
It seems that a device I employ is unsound. (“[punctuation] must not itself mislead the solver when it comes to the cryptic reading”). Drat. I shall not rewrite affected clues, but I shall keep the admonition in mind for future constructions.
Hi Dr Daniel
Ultimately, it’s all about what the audience – or the editor – is happy to accept; my aim is to document practices that I feel could not reasonably be rejected by any editor or solver. I don’t think that the example “Male guards, one leading (4)” for MAIN (MAN around I) is fair, because it breaks Afrit’s rule that “I need not mean what I say, but I must say what I mean” – with the comma where it is, the wordplay simply cannot mean “A word for ‘man’ protects the Roman numeral for ‘one'”, however you read it. Change ‘guards’ to ‘guarding’, and while the wordplay might not be elegant it isn’t unfair either.
Forgive me, please, for obtuseness: how does “guarding” instead of “guards”, with the comma in place, render the construction acceptable?
A very fair question. What I meant was that in the first version, the position of the comma precludes the required interpretation, while in the alternative, it is undesirable, but it would not be entirely unreasonable to ask the solver to infer ‘[with] [a] Male guarding [it], one’. Another example on those lines would be “Group of animals outside, one learnt” for HEARD [A in HERD], where the wordplay works far better without the comma, but the surface reading demands it, and it is possible to infer ‘[with] [a] group of animals outside [it], one’.
My concerns are thus assuaged; thank you.