The letters that spread their elbows

An original exposition about an invented sign-design exercise.

To make a sign fit a narrow card, begin by counting the letters. That was my first suggestion, and it was not entirely useless. It was simply a poorer measurement than I thought. On our test sheet, the word minimum contained seven letters. Wide contained four. In the typeface we had chosen, the shorter word took up surprisingly generous space.

The letters had bodies. An i stood thinly between its neighbours; a w spread its elbows. Counting them as equal occupants was like counting passengers without asking what they were carrying. This did not make a long word short, but it made the number of letters an uncertain guide to the room it needed.

We tried squeezing the letters closer together. The word fitted, although reading it became less comfortable. Spaces that had seemed empty were doing work: they kept one shape from becoming part of another. Removing them all was a saving made at the reader's expense.

Then we changed the typeface. The same word acquired different shoulders and a different rhythm of dark strokes and gaps. A line that had looked crowded became calmer, but another label now seemed strangely sparse. There was no single adjustment that improved every card equally.

Our eventual sheet of signs was full of small decisions. Some phrases were rewritten. Some cards were wider. We kept enough consistency for the labels to belong together without pretending that every word occupied the same shape.

Letter counting remained useful when a limit really concerned characters, such as the number allowed in a form. For a physical sign, however, it could not do the eye's whole job. A word is something we recognise by meaning, but it is also something we have to see. The narrow card had made us attend to both.

How does the writer explain the limitations of counting letters when designing a physical sign? Support your answer with references to the passage.