Writing Systems of the World 10, Exercise 2: the toy decipherment

I ran the chapter's code with this sentence:

"a boy and a girl sat on a hill and the sun set on the hill"

Output from the run (letters outside the ten-letter key, such as b, y, g, l
and u, are left unchanged by the code):

△ b☆y △◇✦ △ g◁♢l ▽△□ ☆◇ △ ▷◁ll △◇✦ □▷○ ▽u◇ ▽○□ ☆◇ □▷○ ▷◁ll
[('△', 3), ('△◇✦', 2), ('☆◇', 2)]

The most frequent word is the single symbol △ (3 times), which is "a". Then
"and" (△◇✦) and "on" (☆◇), twice each. So yes, the most frequent words turn
out to be common English function words, which is typical of English text
(my general point, not from the article).

What would change if you did not know the language: you would still see that
△ is a very frequent one-symbol word and that △◇✦ is a repeated three-symbol
group, but you could not say what they mean. That is the Gelb and Whiting
"known writing, unknown language" situation the Decipherment article calls
extremely difficult to resolve from internal information alone.