The asterisk seems to mean "start phrase" and the tilde "end phrase". The carets may be character or word separators. This is how often the blocks show up:
11 Count 1
12 Count 5
13 Count 2
14 Count 4
15 Count 6
18 Count 2
19 Count 1
20 Count 7
21 Count 4
23 Count 2
25 Count 1
125 Count 1
161 Count 1
199 Count 3
205 Count 1
231 Count 1
235 Count 2
239 Count 1
1435 Count 1
1656 Count 1
1851 Count 1
1895 Count 1
2085 Count 1
2089 Count 2
19395 Count 1
12g1 Count 1
13b51 Count 1
14|6 Count 1
144j Count 1
144y1 Count 1
15i213 Count 1
15n2531 Count 1
18u Count 1
18z Count 1
19|1 Count 2
1954d|8 Count 1
195h Count 1
195o Count 1
19f Count 1
19k9 Count 1
19l Count 1
19t6 Count 1
2054q Count 1
2085v Count 1
20m Count 1
21|1 Count 1
24s2 Count 1
24x Count 1
25c1 Count 1
25e Count 1
25r Count 1
25w Count 1
9a1 Count 1
9p Count 1
Grand Count 83
There are 83 code blocks and 54 different blocks.
The blocks that are simple short numbers are much more common than the more elaborate ones lower on the list.
Five blocks have the | symbol, which is the only non-alphanumeric that seems to be part of the character codes. In all cases but one (19|) the code has a single character after the |. I'd intuit that means those blocks, if they are words instead of characters, are contractions and the | is an apostrophe. 19| may be a typo for 19|1.
Its possible the encryptor used two different blocks to represent each character. This would throw off attempts at counting the blocks and trying the most common letter in the most common block, etc. The extra 2 block types may just be typos, and we should actually have 52, or else they may represent special characters like "oo" or "ee".
If not, notice that we have one set of doubles. This, in English, is likely cc (accord), dd (add), ee (feet), ff (afford), gg (aggregate), ll (allow), mm (commend), nn (annotate), oo (boot), pp (appoint), rr (arrest), ss (pass), or tt (attend). That applies only if each block represents a character.
One thing that bothers me is that the second phrase doesn't have a tilde to end it. So maybe the asterisk means a capital letter and you're meant to infer from the translated text where the sentences are. The second phrase is pretty short. So maybe there are only two phrases and there is a capital letter near the end of the first. Here are the divisions based on asterisks and tildes. The double 12 is bolded.
*9a1^13b51^199^12^25c1^13^21^1954d|8^15^1656^21^12^12^25e^2089^19f^239^12^12g1^13^21^195h^25^15^21|1^12^19^15i213^15^14|6^1895^19|1^144j^19395^1435~
*^2089^19k9^19l^161^18^20m^20^23^15n2531^21^195o
*9p^231^14^2054q^20^235^14^20^25r^199^24s2^1851^11^19t6^15^18u^2085v^20^235^14^20^25w^199^24x^125^20^205^18^19|1^144y1^14^15^2085^18z^20^23^15~