The Rebus Principle

I took two years of Latin in high school and our teacher Mrs. Albritton was able to instill in me a respect and admiration with how precise the language is.  A noun in Latin, for example, can have a number of meanings just by changing its suffix.  The noun manus for "hand" can mean "with or by hands" by changing it manibus.  The -bus suffix is quite common and also appears when res for "thing" becomes rebus meaning "with or by things".

We see an illustration of this in what's called a rebus puzzle. For example, the solution to the drawings of an eye plus a can plus a sea plus a female sheep would be "I can see you".  There was even a TV quiz show (Concentration) where contestants had to solve rebus puzzles to win prizes. 

When I first encountered "rebus" used to refer to a crossword puzzle that had multiple letters in a single grid square, I thought that was wrong.  If you're going to use a word from Latin for that, it should be litteris (by way of or with letters) or verbis (by way of or with words).  But I thought "No big deal" so I wasn't motivate to speak up about the error.

That changed in 2020 when I watched the PBS Nova show "A to Z: the First Alphabet" in which "The Rebus Principle" was used to describe how writing went from pictographs and hieroglyphics to modern alphabets.  Here's the relevant excerpt from the show's script:  

[begin quote]  NARRATOR: Such pictograms would be the basic building blocks of the first writing systems. And thousands of tablets like this one suggest that the reason for moving beyond a purely oral culture was something utterly prosaic: the need to keep ledgers. As far as we can tell from the evidence, for several centuries, the use of pictograms was limited to primitive accountancy. But then, sometime around 3000 B.C. there was the crucial conceptual leap.

IRVING FINKEL: The giant leap came when somebody conceived of this matter: that you could draw a picture which represented something that someone could recognize, but at the same time that sign could be used just for the sound of the thing it looked like. So, on this tablet here, there is an ear of barley. Now the word for barley in Sumerian is, is pronounced like “sheh.” So your Sumerian sees this and says “ah, “sheh,” “barley.” But at the same time, this scribe or a fellow scribe, in writing a totally different kind of document, could use this sign not to mean barley, but just to write the sound of “sheh.” And this giant leap is something rather simple, and it’s something which could have occurred to a child, but nevertheless it is of great lasting significance.

NARRATOR: Using a picture to represent a sound in this way is called the “rebus principle,” and it allows pictures to spell out words.

IRVING FINKEL: To give a really clear example. There’s a word “shega” in Sumerian, which means “beautiful” or “pretty” or “nice” or something like that. And so a scribe would write it syllabically, “she” “ga.” So, he would use this sign, the barley sign, for the “she” bit, and then he’d have to write “ga” for the second bit. As it happens, “ga” means milk. So, he would draw the picture which represented milk. And barley and milk together would spell “shega,” which had nothing to do with either barley or milk. So, this is a kind of rebus writing. Rebus is a smart word for it. It is really a pun in some sense. It is a kind of pun that you get another meaning out of the sign. [end quote]

When I came to this part of the program I had an "aha!" experience, sort of an epiphany lite.  And my attitude about the misuse of "rebus" in crosswords changed.  It was no longer a minor matter not really worth the effort to try to correct but now a much more serious one because "rebus" is presently used by learned scholars---historians, linguists, philologists, Egyptologists and the like---in a very different way, much closer to its Latin definition, "by way of things", than to how it is used in crosswords, "with multiple letters in a single puzzle grid square".  

So my question is can we continue to bring discredit to the crossword community by using a Latin word in a very different way than it is being used by contemporary scholars?  And these are people who crossword world should hold in high esteem and respect and not use one of their core terms in a blatantly incorrect way.

By the time we get to letters we have gone past the Rebus Principle.  Multiple letters in a single grid square do not a rebus make.  Non rebus sed litteris, not with things but with letters.  If the letters form an actual word then it would be Non rebus sed verbis, not with things but with words.  

Limited Space Information Display Matrix

I think a crossword puzzle grid is a limited-space information display matrix and that the ideal puzzle is one that maximizes the quantity and quality of the information on display within that matrix.

I  believe there are two criteria that are central to judging the quantity and quality of information in a grid.

First would be the amount of  value and interest in each grid entry.  This is tricky to define in a purely objective way.  One approach would be to imagine a scale with "Commonplace & Simple" at the low end and "Unique & Complex" at the high end.

Consider, for example, two 8-letter slots in the grid, one filled with SINECURE and the other with HE ATE ONE:  Same amount of matrix space filled up but with very different ratings on the value/interest scale.  You could write an essay on the meaning of the first entry, while the second entry is as simple, commonplace and uninteresting as it gets.

Second, an ideal grid would have only core words or base words.  These are words as they first appear alphabetically in the dictionary.  They are words that cannot be reduced further (have any letters removed) without losing their meaning, having their meaning completely changed, or losing their status as a word, e.g., becoming an abbreviation.

BATTLE would be an example.  Remove any letters and the meaning is gone. Add letters, however, and it is no longer a base word.   BATTLES, BATTLED, BATTLER, BATTLERS or BATTLING all fill up more of the limited matrix space without adding much of value or interest. 

This is the hallmark of Letter Count Inflation (LCI, discussed elsewhere on this blog).  It's largely wasted space, wasted opportunity.  It's like padding an essay to meet a minimum word count requirement, kind of like adding non-nutritive filler, so to speak, to increase volume without increasing value.  The more LCI in a grid the farther away it gets from what I think is the ideal crossword puzzle.

Crossword construction constraints such as symmetry, no two-letter entries, black square maximums, theme entries, etc., may make the ideal grid fill unachievable.  So my judgment of the quality of a grid hinges on how close or far away it is from the ideal.  One that comes as close as possible, given the constraints of that grid, would get the highest rating.   It would be an optimum grid, one with the least amount of wasted space and the most interesting fill.

Puzzography: The Early Years

The first crossword puzzle I ever saw was in a pulp magazine about science that was written for grade school kids.  This was in 1956, so I hope you'll forgive me for not remembering the particulars of the puzzle, though I think---really going out on a limb here--- it had a sciencey-type theme.

Along with the puzzle was an invitation to the readers to send in their own crossword creations.  So I set to work immediately, put one together, and sent it right away to the editor.  Again, the intervening years have left the details of the puzzle in an impenetrable mist.  Too bad I didn't make a copy.

What I do remember clearly (and painfully) is that I never heard a peep from the publication.  Nothing.  Nada.  Silence.  As the weeks, then months, passed without any word [cue crickets chirping] my hopes sank.  My fledgling crossword ambitions crushed, I finally gave up hope.

I decided I needed to start my own publication at my school so I would have a place for my puzzles.  I called the paper "The Weekly Blab".  It had some gossip about who was sweet on who, a serial adventure story featuring Donald Duck, a treasure hunt location clue, and much, much more.  And it only cost a nickel!

I did save a couple of the early issues.  The master copy was hand-printed with a toxic smelling purple ink and then laid on top of a flat gel surface that absorbed some of the ink.  After the master was removed, a few copies could be made by placing blank sheets of paper on the gel surface where they would reabsorb some of the ink.   Each copy would be slightly lighter than the previous one, until the ink, rather quickly as I remember, faded completely away.

So that was my first published crossword puzzle,  March 4, 1957, with eleven words and six black squares in a 6X6 grid.  No symmetry.  Here are the clues:

     Across
          1. Might; power
          2. Musical drama
          3. Change
          4. Foe
          5. Withered
          6. Lock of hair

     Down
          1. Woods
          2. Device for opening [needed some editing here!]
          3. Hold in great respect
          4. Violations of the law
          5. Not difficult

The filled grid looked something like this, where X represents a black square/block:

          F O R C E X
          O P E R A X
          R E V  I  S E
          E N E M Y X
          S E R E X X
          T R E S S X

Yeah, even some crosswordese with SERE at 5-Across, and my very first POC (Plural Of Convenience) with CRIMES at 4-Down.

I did a couple more crosswords in the next issues, but none ever reached the soaring heights of that first one.  Readership, never high to begin with,  began to dwindle and then the school year ended.  And with that, so did the brief but unremarkable life and times of "The Weekly Blab".

I didn't forget the bitter disappointment of that first puzzle submission that was never answered yea or nay by the editor.  I think I developed some newbie constructor PTSD.  Thus the lengthy hiatus, 50 years or so, until my second submission attempt, to The Chronicle of Higher Education.  This time I got a timely reply---a couple of weeks as I recall--- from the editor, Patrick Berry, and with his patient and expert guidance (and believe me, I needed a lot of it!) I got my second publication with a puzzle titled "Think Again" in the Chronicle of Higher Education, September 26, 2008.

Letter Count Inflation (LCI)

When constructing a puzzle, I sometimes see a word that will fit beautifully with most of the letters of the crossing words in that section of the grid, except for one problem.  It has fewer letters than the number of squares in that position.  There are some options that can help me resolve this issue, i.e., increase the letter count of the base word* (usually the word in its first alphabetical appearance in the dictionary) to match the number of squares I'm trying to fill.

I call these devices Letter Count Inflation, partly because that's what's going on---the number of letters in a base word or phrase is being boosted---and partly because it makes for a spiffy initialism, LCI.  In the vast majority of cases the LCI is for expediency or convenience.  It's gratuitous.  It's there only because it ups the letter count and makes it easier to fill the grid, not because it adds value or interest to the puzzle.

One of the most frequently encountered methods to inflate the letter count is to take a base word, typically a noun or verb, and add an S, an ES, or drop a Y and add IES.  The resulting longer word is what I call a Plural Of Convenience, or POC, and I have posted about POCs elsewhere on this blog.

Another LCI option is to take a base word that is a verb and change it into a noun.  Consider the verb ASPIRE.  By adding an R, I can take that six-letter word and make it fit a seven-letter slot.  An additional letter-count boost to eight can be had by adding on a POC.   We would go from ASPIRE to ASPIRER to ASPIRERS.  A clue starting out along the lines of "One who...", or "Those who..." is often a tell-tale sign that a noun-to-verb LCI has gone down.

There are two more ways verbs can be used for LCIs.   The first is shifting the tense, usually from the present to the past, and the second is shifting from the base form of the verb to a gerund/present participle.  The typical tense shift will amp up the letter-count by two, e.g., CAMP to CAMPED (a 50% letter-count increase), while going to the gerund can net a three-letter boost, e.g., CAMP to CAMPING (a 75% letter-count increase).

Adjectives can provide a rich lode for mining LCIs, as when one takes on an adverb's clothing and gets a two-letter uptick, for example, from APT to APTLY.  For some advanced LCI, you can combine the verb-to-gerund and the adjective-to-adverb ploys to give some amazing letter-count boosting.  Consider the base word FIT.  An LCI to FITTING gives a 133% letter-count increase, while doubling up on LCIs to FITTINGLY gives a whopping 200% letter-count and subsequent grid fill increase.  Is there an equivalent 200% increase in value/meaning/interest?  Not for me, not even close.

Another LCI that I see regularly is shifting to the comparative or superlative form of an adjective. This can net a two-, three-, or even four- or more letter bump.   A good example would be FAR, FARTHER/FURTHER, FARTHEST/FURTHEST , with the superlative giving a 166% grid-space-filling increase.  The champion in this department, methinks, would be FITTINGEST, coming in with a 233% increase in grid-filling power.   

Using prefixes is another way to increase letter count, the most common probably being the oft encountered RE-, followed closely by UN- & IN-, with CON- et al. seen occasionally.  Just today I solved a top tier puzzle with both REROOT & UNNEATLY (!) in the grid.

The common thread for all LCI's is increasing the amount of grid space that gets filled without adding a commensurate amount of value or interest to the puzzle.  LCI's just make it easier to fill the grid.   They make the "artful arrangement of words crossing one another" less artful.

I think LCI's are like abbreviations, partial phrases, foreign words, random Roman numerals, crosswordese, and the like.  Any of these used judiciously to facilitate filling the grid of an otherwise excellent puzzle would be unremarkable and above reproach.   It's when they are used excessively that, for me, they become intrusive and degrade the overall quality and integrity of the puzzle, and diminish my enjoyment of the solving experience.

Be on the lookout for LCIs in a puzzle near you.

*By base or core word I mean the form of the word that cannot be reduced further without losing the meaning of the word. Remove any of its letters and it becomes meaningless or its meaning is completely changed, or it loses its standing as a word and becomes, for example, an abbreviation.

Addendum:  Another letter count, grid fill boosting method is adding an article in front of a noun in the grid .  In standard English discourse single, countable nouns must be preceded by a definite or indefinite article.  In crossword puzzles, however, the convention has been to use just the noun alone.  If a noun entry in a puzzle grid is arbitrarily preceded by an article while all other nouns in the grid stand alone then it has all the earmarks of an LCI, here an article of convenience.

P.S.  I first called this stratagem Letter Count Manipulation or LCM.  That's why LCM rather than LCI appears in some comments below.

Puzzography: The later years

Sometime in late 2007/early 2008 I was sitting at the kitchen table after finishing that morning's crossword puzzle when I heard myself saying "I could do better than that".  Turns out that's a whole bunch easier said than done, but with a lot of help from many sources and a few zillion hours work (I kid!), I've managed to get a handful of puzzles published.  Here they are, in chronological order:

9/26/08  "Think Again", Chronicle of Higher Education (CHE)

11/21/08  "One for the Ladies", USA Today (USA)

6/26/09  "Weaponyms", CHE

8/11/09  "Snake in the Grass", USA

8/15/09  "Giddyup", Universal Syndicate

9/23/09  "Minor Defects", USA

12/7/09  Untitled, Los Angeles Times (LAT)

3/22/10  Untitled, LAT

6/15/10  Untitled, LAT

10/11/10  Untitled, New York Times (NYT)

3/8/11  Untitled, LAT

7/18/11  Untitled, LAT

9/26/11  Untitled, LAT

7/20/12  "Natural Misunderstandings", CHE

11/8/12  Untitled, LAT

3/11/13  Untitled, NYT

2/19/15 Untitled, LAT

10/1/17  "That's One Way To Put It", NYT

4/23/19  Untitled, LAT

9/24/19 Untitled LAT

11/10/20 Untitled NYT

POC Levels

In crossword constructing, a POC (Plural Of Convenience) can be, well, convenient.  I can take a 5-letter word, say, and instantly make it into a  6- or 7-letter one by just adding an S or an ES.  How easy is that?  I can vouch that it is temptingly so.  As is often the case with temptations, however, there is a price to be paid for giving in.

The price for succumbing to the POC's siren song is a reduction in the puzzle's overall quality.  After all, the constructor just used a short cut to make it easier to complete the grid.  There was an increase in filled grid space without a commensurate increase in value.

The degree to which a puzzle's quality is impacted depends on what kind of POCs are used and how many times they appear in the puzzle (too many and it can become POC marked!).

A POC can be categorized into one of four levels, going from the hardly noticeable (Level 1) to a serious threat to the puzzle's overall integrity (Level 4).

Level 1 POC is the common-as-dirt kind, found in almost all puzzles.  POCs at this level involve only a single, relatively short entry being pluralized to up its letter count   One might see a Level 1 POC in a corner like this (the NE corner from my 11/8/12 LAT):


      J E E R
      A X L E
     W A L L
F L I  M S Y
      N
      G

There 'tis, ELLS.  Hardly noticeable, but a POC nonetheless.

Level 2 gives the constructor two POCs in one.  Two entries share an S at the end.  This is more of a threat to the puzzle's quality because that "S" is the equivalent of what's called a "cheater square", or, more charitably, a "helper square". This is a black square that doesn't change the word count, it just makes it easier to fill in the grid.  In other words, that shared final S in a Level 2 POC could be changed to a black square and then just clue the words as singular rather than plural.  Have I been guilty of this more serious POC?  Yup, sad to say.   From the same puzzle, SE corner:

       A Y E
O I  L E R
U V U L A
T E M PS

That's just too easy and it's why I think a Level 2 POC is a more serious threat to a puzzle's quality.  I was able to rework that corner to eliminate that two-for-one POC in about six or seven minutes.  Wish I had done that before submitting it for publication.

Level 3 is when a long, non-theme entry is a POC.  Say, for example, HEX BOLT fits beautifully with some fantastic crossing fill, but it's a letter short for that slot.  No problem, just add an S.  But now a much more integral part of the puzzle has been compromised than happens in Levels 1 & 2.

Level 4 POCs so compromise a puzzle's quality that I think that they should be deal killers.  This is where a theme entry is a POC.  The theme is the heart and soul of the puzzle and shouldn't have to rely on an S or ES tacked on to one of the theme entries to make the letter counts match.  It's kinda like adding a gratuitous syllable in one of the lines of a Haiku poem to get the 5-7-5 syllable counts.

Have I been guilty of using a Level 3 or 4 POC in my puzzles?  Nope, glad to say.

The other factor that can impact puzzle quality is how many POCs appear in the puzzle.  Obviously, the more POCs, the greater the impact.  Since the relative frequency of the letter S in standard English text is around 6%, I think any frequency significantly above this in a grid, say 12% or higher, shouts "POC Assisted" or even "POC Marked".  (Only 4% of the 100 tiles in a Scrabble set are Ss.)

Combining POC frequency with POC Levels offers a way to give each puzzle a POC Score.   The POC Score could then be used along with more commonly discussed  factors, like theme consistency, excess abbreviations, long partials, etc., to rate the puzzle's overall quality.

POC doc.

I became aware of what I like to call a "plural of convenience" (POC) after I tried my hand at crossword constructing.  I remember an early attempt where I had only one pesky corner remaining to be filled.  The solution turned out to be simply a matter of adding an "s" to one of the words in that corner and, kaboom, the grid was complete.  This is a big milestone for a fledgling constructor, believe me.

Imagine that you have invested hours on a puzzle and have filled in all of the grid except the last corner.  There, one possible solution would work except that a word that fits four out of the five required squares, say FANG, is, unfortunately, a letter short of the number of squares in that slot.  And then you see that by simply adding an "S" to FANG, it all falls into place.  Woo-hoo!

Who could resist?    I couldn't, and it appears other constructors are in the same boat.  I've never seen a puzzle that didn't have a POC or two, or sometimes many more, even in top-tier ones like the NYT or LAT.

That first experience got me hooked on POCs, but I still had a tinge of regret that I had to resort to what is essentially a short cut in constructing my puzzle.  I had filled up additional grid space without adding anything of substance.  FANGS takes up 25% more grid space than FANG, but adds little if any additional value.

Because a POC makes it easier to fill the grid, I think it lowers what might be called a puzzle's degree of difficulty. It's common to rate some games and competitions by not only how well they are performed but also by their degree of difficulty. Of two competitive dives that are performed equally well, for example, the one with a higher degree of difficulty will get the higher score or ranking. I think the same is true of crosswords. Since POCs make it easier to fill the grid, they diminish the puzzle's overall rating or score.

 
I should mention that POC is not a grammatical term.   It's a crossword term that means adding an "S", "ES", or dropping a "Y" and adding "IES", (whether the word be a noun or verb*) in order to boost the word's letter count and give it more grid space filling power. POCs make it easier to fill the grid.

Not all POCs are the same.  I'm seeing four levels, from the common-as-dirt to the deal-killer.

For more, check out my post,  POC levels.

*Often times the word getting an "S" or "ES" can be clued as either a noun or verb.  LOVES would be an example.  Clueing it as a verb rather than a noun doesn't change the fact that a POC has been used to boost the letter count and---I think it's worth repeating---make it easier to fill the grid.

Post script:  After becoming aware of how much adding an "S" or "ES" to a word or phrase can help in filling grid space, I realized that a POC isn't the only letter-count-boosting device available to the constructor to make it easier to fill the grid.  I have added another post on this blog, Letter Count Inflation (LCI), to discuss these methods.

Post post script; Most of the time a POC increases a word/phrase letter count so it will fill a bigger slot or space in the grid. There is, however, a rarely seen POC where the pluralized form is shorter than its non-pluralized one. It lowers the letter count rather than raises it.  An example would be using ESOPHAGI rather than ESOPHAGUS.  One POC feature remains the same, though. The POC is there for convenience (not because it adds anything of interest or value to the puzzle); it makes it easier to fill the grid.