Research
Which contractions do the work?
Over a whole novel, contracted braille is 28.6% shorter than uncontracted braille. Sixteen contractions do half of that, and 24 of the 182 are never written at all.
A list of Grade 2 contractions is usually set out in the rulebook’s order: wordsigns, groupsigns, shortforms. That order says nothing about which ones a reader meets most, or how much each is worth.
Both can be counted. This site’s translator wrote the whole of Alice’s Adventures in Wonderland twice, once in each grade, and reported which contraction it used for every word, where the word stood. The count below is that report, and it adds up to the cell.
The method
- Question
- Which Grade 2 contractions account for the cells that contracted braille saves, and how unevenly is the work shared?
- Compared against
- Nothing — this is a count of what this site’s translator wrote. How far that translator can be trusted is measured separately.
- Over
- The body text of Alice’s Adventures in Wonderland, 26,439 words — the same file the page-count study lays out.
- Counted as
- A contraction’s saving is the letters it stands for, less the cells it takes. The sign for “the” saves two cells each time, “ing” two, “er” one, and “little”, written l-l, four. The contractions are the 182 that section 2.2 of the Rules of Unified English Braille lists, in its nine families. Five of them — and, for, of, the and with — are used both as a word and inside longer words; each is one contraction here, with both jobs added together.
- Checked by
- The savings of all 158 contractions used come to 34,447 cells. Grade 2 also spends 18 cells that Grade 1 does not, on the indicator in front of a letter standing alone. 34,447 less 18 is 34,429, which is exactly the difference between the two transcriptions: 120,501 cells and 86,072. No cell is unaccounted for.
- Published as
- /data/alice-contractions.tsv — every contraction with its family, uses, saving and commonest words — and the text itself, /data/alice-body.txt. The counting is done by the translator this site serves:
BR.explainText(text)in /tools/engine.js returns, for each word, the signs it was written with.
Sixteen contractions do half of it
The sign for “the” is 12.1% of everything Grade 2 saves in this book: 1,642 times as the word and 435 times inside words such as “they” and “then”, two cells each time. With “ing”, “and” and “er”, four contractions have done a quarter of the work (26.1%). Half is passed at the 16th, three quarters at the 39th and nine tenths at the 71st. The last 87 contractions used share the remaining tenth.
| Contraction | Cells | Family | Times used | Cells saved | Running share |
|---|---|---|---|---|---|
| 1. the | ⠮ | strong contraction | 2,077 | 4,154 | 12.07% |
| 2. ing | ⠬ | strong groupsign | 983 | 1,966 | 17.78% |
| 3. and | ⠯ | strong contraction | 950 | 1,900 | 23.29% |
| 4. er | ⠻ | strong groupsign | 966 | 966 | 26.10% |
| 5. said | ⠎⠙ | shortform | 460 | 920 | 28.77% |
| 6. ed | ⠫ | strong groupsign | 919 | 919 | 31.44% |
| 7. that | ⠞ | alphabetic wordsign | 300 | 900 | 34.06% |
| 8. sh | ⠩ | strong groupsign | 793 | 793 | 36.36% |
| 9. you | ⠽ | alphabetic wordsign | 396 | 792 | 38.66% |
| 10. en | ⠢ | lower groupsign | 726 | 726 | 40.77% |
| 11. ar | ⠜ | strong groupsign | 679 | 679 | 42.74% |
| 12. was | ⠴ | lower wordsign | 329 | 658 | 44.65% |
| 13. with | ⠾ | strong contraction | 209 | 627 | 46.47% |
| 14. of | ⠷ | strong contraction | 619 | 619 | 48.27% |
| 15. it | ⠭ | alphabetic wordsign | 592 | 592 | 49.99% |
| 16. in | ⠔ | lower groupsign | 562 | 562 | 51.62% |
| 17. th | ⠹ | strong groupsign | 506 | 506 | 53.09% |
| 18. little | ⠇⠇ | shortform | 126 | 504 | 54.56% |
| 19. ou | ⠳ | strong groupsign | 441 | 441 | 55.84% |
| 20. ow | ⠪ | strong groupsign | 438 | 438 | 57.11% |
| 21. wh | ⠱ | strong groupsign | 435 | 435 | 58.37% |
| 22. ea | ⠂ | lower groupsign | 423 | 423 | 59.60% |
| 23. this | ⠹ | strong wordsign | 131 | 393 | 60.74% |
| 24. st | ⠌ | strong groupsign | 391 | 391 | 61.88% |
| 25. very | ⠧ | alphabetic wordsign | 130 | 390 | 63.01% |
A learner who knows the first 40 contractions of this ranking has 75.7% of the saving in this book. The full ranking is in the data file.
By family
| Family | Used | Times used | Cells saved | Share |
|---|---|---|---|---|
| Strong groupsigns | 12 of 12 | 6,969 | 7,952 | 23.1% |
| Strong contractions | 5 of 5 | 4,044 | 7,678 | 22.3% |
| Alphabetic wordsigns | 23 of 23 | 2,777 | 5,422 | 15.7% |
| Shortforms | 53 of 75 | 1,745 | 4,841 | 14.1% |
| Initial-letter contractions | 31 of 33 | 1,402 | 2,929 | 8.5% |
| Lower groupsigns | 10 of 10 | 2,056 | 2,140 | 6.2% |
| Lower wordsigns | 6 of 6 | 1,008 | 1,639 | 4.8% |
| Strong wordsigns | 6 of 6 | 346 | 1,019 | 3.0% |
| Final-letter groupsigns | 12 of 12 | 490 | 827 | 2.4% |
The 12 strong groupsigns are the largest share, 23.1%: they are parts of words, and “ing” alone turns up in 282 different words here. The 5 strong contractions are close behind at 22.3%. The 75 shortforms are the longest list to memorise and save 14.1%; 53 of them appear.
Often is not the same as valuable
“er” is written 966 times and saves 966 cells. “said” is written 460 times and saves 920. A contraction that replaces a whole word is worth several cells each time it appears, so a few rarer ones sit high in the ranking: “little” is used 126 times and saves 504.
| Contraction | Cells | Family | Saved per use | Times used |
|---|---|---|---|---|
| altogether | ⠁⠇⠞ | shortform | 7 | 5 |
| question | ⠐⠟ | initial-letter contraction | 6 | 21 |
| children | ⠡⠝ | shortform | 6 | 10 |
| rather | ⠗ | alphabetic wordsign | 5 | 25 |
| because | ⠆⠉ | shortform | 5 | 15 |
| through | ⠐⠹ | initial-letter contraction | 5 | 14 |
| people | ⠏ | alphabetic wordsign | 5 | 13 |
| enough | ⠢ | lower wordsign | 5 | 11 |
| yourself | ⠽⠗⠋ | shortform | 5 | 10 |
| together | ⠞⠛⠗ | shortform | 5 | 9 |
What the book never uses
24 of the 182 contractions are never written, and in every case it is because the word is not in the book. Alice’s Adventures in Wonderland does not contain the word “mother”.
Initial-letter contractions: lord, mother. Shortforms: according, afternoon, although, beneath, beyond, blind, braille, conceive, conceiving, deceive, deceiving, declaring, necessary, paid, perceive, perceiving, receiving, rejoice, rejoicing, thyself, tonight, yourselves.
That is a fact about one children’s story of 1865 as much as about the code. Another text would leave a different set unused. Of the contractions that are written here, 37 appear fewer than ten times.
What Grade 2 costs
Contracted braille is not all saving. A letter standing alone would be read as its wordsign — a lone m as “more” — so it takes a grade 1 indicator that uncontracted braille has no need of. The book has 18 of them: “C”, “D”, “W”, “V”, “M” (3 times), “m”, “X”, “Hm”, “e” (6 times), “p”, “T”. The e is the Mock Turtle singing “e—e—evening”; “Hm” takes one because h-m is the shortform for “him”.
What this cannot tell you
- One book. These are the counts for one text. Which contractions lead in a textbook, a recipe or a modern novel is not measured here.
- Italics. The source marks 220 words and phrases as italic with underscores, and the translator writes the underscores as they stand. A word between them is not standing alone, so it loses its wordsign: “_very_” is v-er-y. Without the underscores the saving would be 34,623 cells rather than 34,429.
- A translator, not a transcriber. This counts what software wrote. Where a certified transcriber would write a word differently, the count differs with it; the accuracy study lists the known cases.
- Cells, not effort. A contraction that saves many cells is not thereby easy to learn or quick to read. This measures length only.