Letter pairs and word recognition

[Mikey]

Goal: fit CPDV into a readable single volume on Amazon, (7.8 x 9.8 x 850) and on Snowfall Press (6 x 9 x 1280) page limit. 

Known: Ligatures in words to reduce letter spacing. More words can fit if more ligatures are composed. 

Problem: Ligatures reduce word recognition in some cases. 

Hypothesis: Visual word recognition in English and other Latin script languages depends on recognizing vowel sounds, then consonant sounds, then syllables, then words. Ligatures that span across a boundary between a vowel sound and a consonant sound break word recognition and require mental computation instead of recognition. Ligatures that remain within a phoneme and especially ligatures that tie an entire phoneme together improve recognition.

Example:

Given the letters KITTEN,

 we first recognize 2 vowel sounds,

 3 consonant sounds (K T N),

 2 syllables (kit en)

 1 word. 

• forming a ki ligature spans a consonant/vowel boundary and will slow reading and comprehension. 
• forming an it ligature spans a vowel/consonant boundary and will slow reading and comprehension
• forming a tt ligature ties the phenome together which helps recognition, speeds reading and comprehension.
• forming a te ligature  spans a vowel/consonant boundary and will slow reading and comprehension. 
• forming an en ligature  spans a vowel/consonant boundary and will slow reading and comprehension. 
  
  THEREFORE
  
  
• Kitten should only be compressed with a tt ligature. This is a beneficial ligature. 

Under this hypothesis. only certain letter pair ligatures make sense. If you look at current ligatures for the letter 'f' available: "ff ffl fft", the ff ligature is beneficial, but the ffl and fft frequently or universally span 2 consonant sounds. The ff ligature is beneficial, but the ffl and fft ligatures, while reducing space required for the word, also reduce readability. 

Consonant Pair Frequency in The Riverside New Testament.

In our sample collection, the most common consonant/consonant pairing is t+h.  This is very frequently the entire consonant phenome "th", but there are certain cases where this pairing does not form an 'f' sound but remain separate t and h consonants, for example 'nighthawk.'  So, universally forming ligatures where this letter pair occurs does fit more words, but it won't always improve or even maintain readability. Automation might be possible to pre-insert invisible separators in cases where common pairings have both single and multiple phenomes in the language. 

But with this caution about spanning multiple consonant sounds with a single ligature, the following small sample can make the biggest improvement on space required: 16 double consonant ligatures should reduce page requirement by about 10% (total guess without ligatures designed and tested.) 

• th, nd, ng, ll, st, wh, ch, rs, rd, sh, gh, ss, ns, ld, pr, ft, hr, tr

Whether some of these combinations can form ligatures that are easily recognizable remains to tested.. ll for example has little horizontal features to tie together, and parallel lines cause optical illusion effects that make design very complex.  But examining each of these pairs for their usage and ligature forming is a good opportunity for reducing page count of a work that exceeds print on demand limits. 

[Mikey]

311TruthMovement

•responded on Reddit

If I’m understanding your goal, you want to save space, correct?

You just want a condensed or compressed typeface. You want a smaller typeface.

Teeny-tiny Bibles have a long history, most of them as novelties to some extent.

I would read as much as you can about Retina by Tobias Frere-Jones: https://www.fastcompany.com/3063912/how-a-micro-font-designed-for-stock-indexes-became-a-classic

The principles in there were originally designed for printed stock listings in the Wall Street Journal, but those principles could be put to use in a wide range of ways.

Does it feel very "Biblical?" Not at all to me, but perhaps you could test Genesis or Jeremiah (two very long books) and see how much it reduces the length.

Spec by Ryan Bugden also came to mind: https://ryanbugden.com/Spec

Printing any long text in that is a bit bananas, but it does push the limits of "how small can you go?"

The pt size is going to make a lot more difference than this strategy of ligatures because you're fitting in more lines of type, not just reducing horizontal width by a small amount here and there.

I've had other projects on my backburner for a long time that take into account letter frequency and combos in English (and perhaps other Roman alphabet using languages), just for very different reasons with very different goals.

If we're sticking to the goal really being about saving space and reducing page count (and I’m understanding it right), then going as small as you can and as condensed/compressed as you can seems like the only answer, as well as reducing the leading as small as you can.

Will this be a pleasant reading experience? No. But something like Retina probably mitigates that as much as is possible.