If you didn’t know, I’m currently in the midst of a month-long “mini crossword residency” at Puzzmo.com. It’s been a great opportunity for me to experiment with some fun puzzle ideas, so check it out!
In today’s April 20 puzzle, I have committed one of the gravest sins of all: Using 2-letter words. (May the world have mercy on me.)
But seriously, it’s a strange feeling: 2-letter words (which I’ll refer to as 2LWs from now on) are one of the biggest pitfalls/no-nos within grids of nascent constructors — so much so that many puzzle venues include an explicit “No two-letter words” rule within their spec sheets. Most construction software nowadays will scream at you if you try to use a 2LW, though to experienced constructors these spots often stick out as “wrong-looking” anyways.
And yet, my recent puzzle’s intentional inclusion of 2LWs has caused me to question this crossword convention. Which is why I hope to discuss: Does this rule really cultivate “better” puzzles? Or does it stifle them?
Spoilers (click for solving link):
Puzzmo: “Great Texture” on April 20, 2026 by Henry Lin-David
The Argument for 2LWs: A Solver Perspective
By now, you’ve either solved today’s Puzzmo mini or aren’t going to, so here’s a solution grid for reference:
As I mentioned in my constructor notes, the whole grid was built around QQ, which refers to the chewy textural description of mochi and QQ noodles.
One of the reasons I was excited to include QQ (beyond its significance to me and my Taiwanese roots) is its novelty. On one hand, it’s a cool entry with its double Q (20 Scrabble points, provided you steal a scoring Q from another set). But more importantly in my mind, it gets to “ask a new question” to the solver — which, regardless of whether they know it or don’t know it, contains some inherent “breaking new ground” excitement.
In a hypothetical world where 2LWs roam free, this novelty would be somewhat short-lived. There’s a mathematically limited pool of these words compared to longer entries. But I think the basic benefits still remain and continue serving the puzzle. If a large appeal of crosswords is how they encourage the solver to draw from a diverse knowledge pool, to me it seems better to maximize these possible sources.
Obviously, not all 2LWs are as distinct and unique as QQ; in fact many of them have basically already appeared in crosswords. In the puzzle above I include ET (i.e. extraterrestrial). No, it wouldn’t appear in a conventional puzzle — but what has appeared, many times, is ETS in the plural form.
Thus many 2LWs, to me, are at the very least equivalent to already-accepted plural 3-letter words. From a solver perspective, it’s difficult for me to imagine a meaningful difference between close-encountering [Out-of-this-world being] and [Out-of-this-world beings]. The length of the word is different, but you’re doing the same type of thinking to come up with (essentially) the same answer.
(As a sidenote observation, ET arguably has more cluing potential than ETS, through angles like the movie E.T. and Latin phrases like “et alia.”)
Setting aside these somewhat net-zero scenarios like ET vs. ETS, I think there are two major benefits to 2LWs. First, as previously discussed, the untapped potential of entries like QQ allows for a richer solving world that draws from more aspects of life. And second, these words allow constructors to do some very interesting things…
The Argument for 2LWs: A Constructor Perspective
Let’s just say the quiet part out loud: 2LWs make it easier to construct crosswords.
However, I don’t think this fact alone legitimizes 2LWs. But what I care more about is how these words might allow otherwise-impossible puzzles to exist.
As a big reason, grid art benefits greatly from the inclusion of 2LWs. Including these words gives the constructor more freedom when it comes to grid aesthetics, and allows them to create more striking designs. If 2LWs make it possible to create a cute ghost grid like in Alex Eaton-Salner’s October 31, 2025 Puzzmo midi crossword, one wonders what sorts of designs might be possible in a full-sized 15×15 puzzle.
But I’d even argue that 2LWs could even benefit more conventional themed puzzles. There are a lot of ways these entries might ease up the grid, but one specific use case might be dealing with words of notoriously unwieldy lengths. For instance, 12- to 14-letter words can be difficult to deal with in a typical 15×15 puzzle, especially as the final themed revealer. While other length entries often can go into the 3rd-to-bottom row, the 12s-to-14s often have to go in the 4th-to-bottom row in order to avoid 2LWs. This tends to scrunch up the grid towards the middle, moving the themers closer to each other and generally making the grid more restrained. On the other hand, allowing 2LWs could better allow these bulky revealers to go in the 3rd-to-bottom row, allowing for better spacing and generally an easier grid to fill.
Below: The left/1st grid is a possible arrangement of 5 themed entries following a 3-letter-word minimum – there are multiple points where you will have to find words that cross between 3 or more inflexible entries. By allowing 2LWs in the righthand/2nd grid, you can space the grid out and have more flexibility to place black squares that avoid constraining vertical entries that need to cross 3 or more theme entries.
Put plainly, 2LWs offer more gridding flexibility and make ambitious puzzle concepts more easily achievable. I experienced this firsthand when making my QQ mini. Originally I tried to create a grid without any 2LWs besides QQ, and it was a huge struggle. Allowing myself to use these words helped “grease the wheels,” so to speak, and ultimately allowed me to achieve my goal for the puzzle. I’ll admit, part of me would have liked to have all my non-QQ words be 3-letters and up (which would have really made QQ stand out). But at the same time, those misgivings are easily outweighed by the fact that I actually made the freaking puzzle.
And Yet…
There’s still a part of me, after all of these arguments, that’s still resistant to the idea of 2LWs.
I think it comes down to an aesthetic component. Bigger words mean a visually more “open” grid. In the side-by-side grids above, the right grid’s 2LW space underneath the black squares definitely feels more “scrunched” to me. I do think that I feel this way about 3-letter word sections compared to 4-letter word sections as well; though maybe the 2LWs stick out a little more due to the no-2LWs conventions I’ve been using for many years.
More to that point, I think the initial impact of the empty grid is important. Even if all of the 2LWs are legitimate and fun entries, the solver can’t know that when they first look at the puzzle. All they see are some very short entries – sort of the bare minimum of what a “crossword word” could physically be – which might feel like somewhat of a disappointment.
I stand by the fact that during the act of solving, there’s often not much of a value difference between a 2LW and a 3-letter word. (If the words ETS and YDS cross at the bottom right corner, I’d argue that the pluralized S might as well be a black square – it essentially doesn’t change the solver experience at all.) But I do think the initial possibilities of a more open grid are beneficial for setting the scene for solver expectations and excitement, which 2LWs may inhibit.
Final Thoughts
Look: The revolutionary in me wants to tear down the crossword walls and invite the masses of misunderstood 2LWs into puzzles across the land.
But I think even a moderate approach could be worth exploring. As a constructor, 2LWs sometimes slip into my grid without my noticing. To me, this signals that 2LWs might be possible to include in crosswords while mitigating some of the aesthetic drawbacks.
With the 12-letter-revealer example I mentioned earlier, there are other ways to work around that situation (e.g. adding black squares in that 2LW space to eliminate what would otherwise be 2LWs, or changing the grid size). Yet these solutions also come with their own potential drawbacks (e.g. more black squares might lead to feelings of being “cheated” out of solving squares, some venues literally will not publish non-15×15 puzzles).
I view 2LWs along the same lines as these other solutions. They’re a tool for constructors with their own benefits and drawbacks, and it’s up to each constructor to decide whether their incorporation is “worth it.”
At the very least, I’d like to entertain the possibility of moving past the blanket statement of “2LWs = Bad” – at least to a sentiment of “2LWs = Sometimes Reasonable in Order to Achieve my Amazing Ambitious Puzzle.”
That’s all for now! Tune into Puzzmo for the remainder of this month for a new mini crossword by yours truly!
Do you have additional ideas? Things I missed? Feel free to comment your thoughts!