A song of middles and suffixes

🕑 3 min • 👤 Thomas Graf • 📆 May 21, 2019 in Discussions

Am I the only one who’s worn out by the total lack of fun and playfulness in all public matters? Everything is serious business, everything is one word away from a shit storm, everybody has to be proper and professional all the time, no fun allowed. It’s the decade of buzzkills, killjoys, and sourpusses. Linguistics is very much in line with this unfortunate trend. Gone are the days of Norbert Hornstein dressing up as Nim Chimpsky. It is unthinkable to publish a paper under the pseudonym Quang Phuc Dong (that’s at least a micro-aggression, if not worse). Even a tongue-in-cheek post on Faculty of Language is immediately under suspicion of being dismissive. Should have added that /s tag to spell it out.

Compared to other fields, linguistics has never been all that playful, perhaps because we’re already afraid of not being taken seriously by other fields. But we’ve had one proud torch bearer in this respect: Geoff Pullum. His Topic… comment column should be mandatory grad school reading. Formal Linguistics Meets the Boojum is a classic for the ages (and did, of course, get a very proper and professional response). My personal favorite is his poetic take on the Halting problem. So I figured instead of complaining I’d lead by example and inject some fun of my own. To be honest, I’m probably better at complaining, but here we go…

Round 1

We’ll investigate a set of strings
such is the passion of mathlings
it could encode a lot of things
DNA and words and sounds, methinks

We want to test locality
in the strict sense with formality
we’ll rely on math’s regality
to find the truth in totality

We’ll play a substitution game
where distinct strings should be the same
and if it fails we can proclaim
it’s not SL, oh what a shame

Pick strings \(s\), \(t\) from this set
and cut them up like a sextet
prefix, suffix can1 differ, yet
middle identity must be met

Now switch the suffixes around
and check if the result is sound
you may ask why, I shall expound
and so we start the second round:

Round 2

A recognizer for SL-\(k\)
is just too easy to lead astray
from left to right on its merry way
it collects all substrings of size \(k\)

If some \(k\)-grams are forbidden
a \(k\)-gram scanner won’t admit’em
and the whole string plainly didn’t
follow the rules as they are written

Now \(s\) and \(t\) are both just fine
their \(k\)-grams all do toe the line
the \(k\)-gram scanner should not whine
if \(s\) and \(t\) do intertwine

If the middle is long enough
we’ve got the perfect SL-bluff
suffix switching adds no stuff
but is the outcome up to snuff?

The \(k\)-grams are the same, you see,
the \(k\)-gram scanner must decree:
“Admit the string for its esprit!”
and that’s the trap! On to round 3:

Round 3

If the string is not permitted
while SL-scanners do admit it
then the proof has been submitted:
the string set’s not SL. You get it?

If after all this neat exposure
you still need the full disclosure:
An SL language’s composure
suffix substitution closure!


This didn’t actually take too long to write, I whipped it up in two hours while basking in the sun (in b4 someone says it shows). The meter roughly matches parts of this lovely Animaniacs song, so we can totally sing this at the next subregular workshop.

And if you have something linguistics- or academia-related that you did just for fun, perhaps even shits and giggles, I’d love to see it posted on the Outdex. Jokes, songs, limericks, anecdotes, parodies, or a short proof that Klingon is not TSL, whatever. Let’s add a dash of light-hearted, tongue-in-check, silly, goofy fun.

  1. This was should originally. Thanks to Jeff for pointing out how misleading that choice of words would have been.

Comments powered by Talkyard. Having trouble? Check the FAQ!