Setting Things Right: The Manifezto Pt.4

Hi all, welcome back to Setting Things Right, aka The Manifezto. The usual caveat: these are just my ideas and opinions – please do add your own thoughts and disagreements in the comments!

Previous posts:

• Part 1: Say What You Mean
• Part 2: The Duck-Rabbit
• Part 3: Types of Rabbit

Part 4: Anatomy of the Common Rabbit

Last time I took a look at different types of clue. But I didn’t cover the most common clue type: definition and wordplay. So, this time we'll be looking at how to make anagrams and acrostics and deletions and Spoonerisms and ... won't we? Hmm, not quite yet. This is a different sort of blog, where I don't want just a 'how to' guide but a 'why to', so we really need to dig down into how the definition and wordplay clue works at a basic level, before dealing with the niceties of specific devices.

And before even getting into that, a brief recap to set the scene!

Clue types

A cryptic definition provides a pure illusion – the whole clue can be seen as one thing, the surface Duck, but looked at another way, also provides an unmistakable Rabbit.

Multiple definitions provide a slightly different illusion – the different parts of the clue are each a representation, so there’s also an illusion within the clue: the various definitions all depicting the same thing in different ways, plus an overall surface that itself makes sense as a Duck. A key point here is that all those various bits are meaningful representations, for example:

• Is able to record in advance (3)

Here we have a meaningful (if rather bland!) surface Duck, plus two distinct representations of the word CAN: “is able to”, and “record in advance” (as in, say, a canned laughter track on a sitcom). Those definitions actually mean "can", they aren't just representing the otherwise meaningless set of letters C-A-N.

But a hidden brings in the idea of words not actually ‘meaning’ anything, instead being seen as just sets of individual letters to play with. I think it’s this idea (of words not necessarily being representative of things, but simply sets of letter ‘objects’) that is perhaps the biggest leap for a new solver. I’ve said previously that clues are presented in a more-or-less ‘straightforward’ way, but they do require development of this rather odd way of thinking. When you see “cat” in a clue, the setter might be asking you to think of an animal, or a whip perhaps, with “cat” being a representation of that concept. Or they might be asking you to simply think of “the set of letters C-A-T”.

When a word is used just to provide a set of letters for the solver to play with we can think of this as otherwise meaningless fodder. So (as we’re thinking of hiddens right now), you might have:

• Animal in horrific attack (3)

And those last two words are simply fodder – you’re looking for a word for an animal (i.e., something for which the given definition might be “animal”) that’s hidden “in” the otherwise meaningless fodder, “the set of letters HORRIFIC ATTACK”.

Building blocks: Definitions, tokens and fodder

Charlie Methven (Methuselah / Chameleon) has suggested the concept of a “token” to describe 'meaningful' representations, as opposed to meaningless fodder. The words actually represent some thing or concept, most often as a synonym but possibly a descriptive phrase or maybe an abbreviation, say: in any case, the words carry meaning. (Note: all Charlie’s cruciverbal resources and thoughts on clues are well worth reading at Crosswords by Charlie Methven. And apologies to Charlie if I've used his "token" concept in a way other than intended!)

The ideas of definitions, tokens and fodder are central to the anatomy of a ‘common Rabbit’ – a definition and wordplay clue.

What things do you need to draw if you want to show a whole Rabbit? You need to show the eyes, ears, fur and so on. So, your clue has to somehow include all these elements for the solver to put together. You need a 'definition' too – a caption to your illustration, or maybe something like the picture on the front of a jigsaw puzzle box; it clearly shows the solver what they're aiming for. And for each bit you want the solver to draw, or jigsaw piece you want them to play with, you need either a ‘token’ – a representative element - or some ‘fodder’ – a set of letters from which the required letters can be generated.

So your cryptic reading, perhaps rather than just simply being “Rabbit” (as it might be in a non-cryptic puzzle), could be thought of as something like:

• Rabbit: Furry little creature with long ears and a twitchy nose

To ask the solver to draw the Rabbit’s ears, you might want to use a token such as “listeners” or “pieces of corn”, or some fodder such as GET ACROSS, PEARSON or ARSE (sorry!) from which the required bits can be extracted (in these cases, by alternation, hiding or 'cycling' the letters, respectively). The fodder could even be just the word "ears" (well, if taking the analogy at face value that'd be a bit same-sidey, but for purposes of illustration ...) – that is, the set of letters EARS which just gives the solver those required letters in plain sight.

How do you pick out those elements? Let’s consider clueing the word NINCOMPOOP. Alongside a definition, we need to break down the word into suitable tokens and fodder, to find the possible building blocks we can play with to construct our Rabbit.

It might break down into lots of little pieces, like N / IN / CO / MP / OO / P. All of those could provide tokens – “name” maybe as an abbreviation, “amongst” as a synonym, and some more abbreviations: “company”, “politician”, “pair of spectacles” (from cricket, two scores of zero), and “softly” (musically, piano). And maybe you can find some way to put those together neatly, but with six separate pieces – plus your definition still to go in the mix – it’s pretty unlikely you’ll come up with a coherent surface reading incorporating all of those elements.

What else could you use? There’s most of the word INCOM[E] in there. With a definition of “fool”, maybe “money” as a token for INCOM[E] ... is there a surface in there about a fool and his money? There’s a P for “Parted”, maybe (and the middle of [S]OO[N], too) but there’d still be a lot of engineering involved. Whilst it's a promising start, it looks like it might lead to an Otter – a clue that needs too many working parts to enable a satisfyingly smooth and snappy surafce; that is, it's 'OTT', over the top. I'm personally very fond of lengthy clues (by which I mean, say, 10+ words ... from a style point of view, others might think 'lengthy' is 8+ or even 6+), but they do have to be justified by a smooth and readable surface, rather than a complicated collection of 'bits'. You want to draw a Rabbit, not an Otter.

But we might have something with that ending POOP ... a fool wasting his “income” on “crap”? (Oh, apologies again – I really hadn't intended to use mildly naughty words, but sometimes that's where the words lead you!)

• Fool N most of wages on crap (10)

That looks promising – we just need something that’ll give us that N and also make sense in the surface, something like “wasted” or “spending”. But the obvious ‘tokens’ available for just a single letter N aren’t really helpful – number, name, north, and so on. So perhaps we need to generate the N with some fodder instead. We need a set of letters from which the N can be easily extracted – perhaps as a start, middle or end letter – that also could be a reasonable word or phrase for the surface. How about SQUANDERS?

• Fool essentially squanders most of wages on crap (10)

Yes, I think that works: we have the definition, “fool”, and three further elements: the fodder SQUANDERS, just a set of otherwise meaningless letters, and the tokens “wages” for INCOME and “crap” for POOP. And we have a Duck too: the surface makes sense. The Rabbit, the cryptic reading, can be seen as:

• Something meaning “fool”: the middle of the set of letters SQUANDERS, most of something that means “wages”, and something that means “crap” (oh, and as a bit of further help – it’s got 10 letters).

Alongside our definition, tokens and fodder (in bold above), we’ve also introduced, almost without thinking, a few other elements here. Those bits that are not in bold aren’t our building blocks, but they tell us what we need to do with those basic elements. And an absolutely crucial point here is that every word in the clue must perform some function – as part of a definition, token or fodder, or as some sort of indicator that provides useful information to help the solver. Otherwise, rather than a Duck-Rabbit, we'll have a Duck-RabbitTriangleSock.

Indicators: Modifiers, constructors, qualifiers and links

In the NINCOMPOOP example, the indicator “essentially” applies to the fodder (the set of letters SQUANDERS), asking for just the central letter from that set. (It’s quite whimsical really, but I think accurately conveys the idea of selecting the most central part, the essence.)

Another indicator, “most of”, applies not to some fodder but to a token: most of “something that means wages”; that is, most of INCOME. But surely, now, that INCOME is being treated just like fodder? You’re being asked to select “most of” the letters from the set INCOME. It’s indirect, isn't it? But that’s fine, we can use a token to represent something that then acts as fodder. In fact, more than that, a token always represents fodder! When you have the token "animal" that represents CAT, you're not asking solvers to find a meaningful word "cat", but just an otherwise meaningless set of letters, C-A-T, that help to make up the answer SCATTERBRAINED, say. Indeed, if that CAT does carry any 'meaning' within the answer – as part of ALLEYCAT, say – then that's a sure sign of same-sideyness. That's what distinguishes a definition from a wordplay token: a definition must resolve to a meaningful representation of the thing it represents; the 'thing' represented by a token should not usually carry any meaning (in the cryptic reading, that is) beyond its resolution as a set of letters. Fodder gives you the letters you need directly; a token gives you the letters you need but indirectly.

OK, so tokens always resolve to fodder which might then be further manipulated, and they are by their nature indirect. So ... why can’t we use an indirect anagram, then? Well, in general, and hopefully without getting too confusing: if the fodder you create from a token is only played with in a simple way (losing a single letter, say), then it’s not really asking too much of the solver to make that leap; if, however, you’re asking for something that requires a bit more work on the fodder (only selecting initial or alternate letters, say, or mixing the whole thing up) then you’re starting to get into unfair territory. What’s the rule for where you draw the line with this? Ah, did you forget ... there are no rules! It's a case of using your – and importantly, your testers' – judgement as to whether it's too much of a leap to expect solvers to first correctly interpret your token, then use the resulting word(s) as fodder for further manipulation. See Charlie Methven's tips for some helpful examples. (Note that your judgement here might be influenced by your intended audience – are you setting a beginner-level 'quiptic' or an advanced 'barred thematic', maybe?)

Then we have one more word, that doesn't indicate how to modify one or more of your building blocks, but instead describes how they fit together: "on". (Preferably, this'll be a Down clue so that "on" is easily interpreted; in an Across, the "on" often means before, although I don't think you could argue that using it to mean after is either inaccurate or unfair – it's a question of style that different setters and editors will have different opinions on.) To make the distinction between indicators that modify your building blocks, and those that show how to put them together, maybe we could introduce the terms modifiers and constructors?

One further element that you might need to apply to your tokens, or to your definition, is a qualifier, indicating that the 'thing' you've chosen needs a bit more information in order to make it either accurate or fair, or both. This may be to do with things being examples of what you mean: if you want to clue a DOG you could use a synonym like "animal" or "canine", but you might prefer a "setter" as a possible example – so you'll need to qualify your token "setter" as "setter, perhaps" or similar. Or, you might want to clue STAKE using the soundalike STEAK, a cut of meat – your token "cut" only gets you STEAK, so you need to qualify that with its being a homophone, maybe "reportedly cut", so that it actually resolves to STAKE as intended. Or, you might want to clue LABOR, but your token "work" would resolve to LABOUR for a UK audience, so you need to qualify it as "work for Trump", say. And so on. (So: a qualifier tells you some way that you need to regard your token in order to accurately give you the required fodder; a modifier tells you what you need to do with the fodder that is generated.)

And a final possible element: you might need a link to join your definition to the wordplay. In the NINCOMPOOP example above, there's no link word. When the clue was broken down into its constituent parts, the link between the definition and the wordplay was simply shown by a colon; in the clue itself, that can reasonably be hidden so the definiton just sits alongside the wordplay. But sometimes, to enable a smooth surface reading (and in some cases, to enable a smooth cryptic reading too) you'll need to make that link explicit. Link words might explain that the definition somehow "comes from" the wordplay, or that the wordplay "leads to" the definition; or, they may demonstrate that the definition is equivalent to the wordplay – the definition "is" the wordplay, say; or, they might just show that the two representations, definition and wordplay, are shown together – wordplay "and" definition, perhaps. What's important here is that the link needs to do its job in a way that ensures the cryptic grammar of the clue as a whole remains intact.

Cryptic grammar

I mentioned earlier that the idea of identifying 'tokens' and 'fodder', of words sometimes carrying meaning and sometimes being just sets of letters, may be the big leap for a new solver. For a new setter, I think it's a grasp of the concept of cryptic grammar that is the big leap. I've been setting for a fair few years now, and I still find there's plenty I need to learn! It's relatively easy to find definitions for words, and to break words down into constituent parts; it's not too hard to find ways to join these together and move them around. And with practice you can start to find just the right definitions and building blocks and indicators that might all sit together nicely with potential to produce a clear, smooth surface reading. Where this then becomes rather more tricky – and where I'd argue the real art of setting comes in – is in ensuring that all of those various elements, presented as a lovely Duck, are combined in a way that produces a single, unified clue that successfully stands alone as a complete and correct representation of a Rabbit. That's where cryptic grammar comes in, and that's what I'll concentrate on next time, when thinking about the actual words you might use to produce your anagrams and acrostics and deletions and Spoonerisms ... and to link them all together.

For now, just a quick taste of the sort of things that'll come up:

• Time for a general election, principally (3)

Here, the parse is:

• Something meaning "time" for the letter A, the set of letters GENERAL, the set of letters ELECTION, principally

Which gives AGE. Now the 'scope' of that "principally" is two wordplay elements, but that's fine as it can apply to a list of things: x, y, and z principally. And the wordplay elements aren't explicitly linked together by any constructors – but we don't need those, as the parts just sit together in order to give us an accurate representation (no descriptions or instructions are needed). The problem is the linking for – it's the wrong way round for a directional link.

• I am in jail with a murderer (4)

With the parse:

• The letter I am in the letters in a word that means "jail" with something meaning "a murderer"

Giving you CAIN. Does that "with" work as a link? Yes, I'd say, if you read it as perhaps "Here's the wordplay bringing you the definition" – similar to, "Here's Alan Partridge with the sports news". But many might only accept a linking "with" the other way round, "Here's the definition by means of the wordplay", say (using a verbatim Chambers definition of "with") ... and plenty frown on "with" as a link whichever way it's used. But anyway, the concern here isn't really the arguable link, it's that am that's definitely wrong. "The letter I am in jail" is just plain wrong grammatically, needing is instead. (Note that the usual way around this particular mismatch is something like "I must be put in jail ...", although in this case that then makes the linking "with" rather jarring as the wordplay becomes a descriptive, instructional phrase. More on such nuances next time.)

• The bad doctor went for a swim? (6)

Parsing as:

• The set of letters THE BAD doctor: something meaning "went for a swim?"

That's BATHED. The definition probably doesn't need a question mark, but it does no harm ... that doctor might've gone for a bath rather than a swim, I suppose. And having "the" appear as is in both the anagram fodder and the answer isn't ideal, stylistically. But the real problem lies in the anagram indicator doctor. Look up a list of anagrinds (as they're sometimes abbreviated) and you're likely to find "doctor" there. But it needs to be used as an imperative verb, an instruction to do something: to doctor the evidence, maybe. It makes no sense at all, in the cryptic reading, if it appears after the fodder.

Remember, in your clue you are trying to make a Rabbit – a completely coherent representation, description or instruction, that works as a whole with correct and consistent grammar putting all the pieces togther. And it will only really work in that way if you pay attention to accurate cryptic grammar.

Summary

Your definition and wordplay clue involves various elements that all play a part in the clue:

• Definition: an accurate representation of the answer
• Building blocks: the various bits that you're going to use to construct the letters that make up the answer – you might think of these as Lego blocks or jigsaw pieces, though perhaps as something more malleable like clay, as we might be moulding them and breaking them up in all sorts of ways

These building blocks in turn can be either direct fodder or indirect tokens:

• Fodder: a set of letters from which the necessary letters for (part of) the answer can be generated
• Token: a representation of some 'thing' or 'concept', which when resolved can then be treated as fodder, a set of letters to help construct the answer

Note that a token or a definition might need some further sort of qualifier to ensure it remains fair and accurate.

Then there are elements that show the solver what they need to do to their building blocks, so that they produce the answer:

• Modifiers: indicators of how the letters from your building blocks might need to be treated – by just using part of them, by mixing them around, by reversing them, and so on
• Constructors: indicators of how your various elements are put together – showing their relative positions in the answer, putting them side by side, or one before or inside the other perhaps

And you might also want to explicitly join your definition to your wordplay using a link, usually to help the surface reading but sometimes to improve the cryptic reading too.

Now the skill – or is it the art? – is to not only come up with suitable elements that can contribute to a convincing surface reading, but to ensure that they are all put togther in a way that maintains a completely correct and consistent cryptic reading: a proper Rabbit, not a Rabtib. And that's what'll always be at the back of our minds as we think about the particular sorts of wordplay, the specific devices, we can employ ... which is what we'll look at next time.

Til then, cheers!

Fez