In a certain code language, 'speak nicely to all' is coded as “ka cu ma he”…

In sentence-based code-language ('Chinese coding') puzzles, several sentences are given along with their coded versions. Since different sentences use different sets of words, the SAME code repeating across two sentences must stand for the ONE word those two sentences have in common. So the method is: pick a pair of sentences sharing exactly one not-yet-known word, take the codes common to both, and — after excluding codes already matched to other identified words — assign the single remaining common code to that shared word.

1. List the four coded sentences for reference: (1) 'speak nicely to all' → “ka cu ma he”; (2) 'all are like us' → “si fo he to”; (3) 'teach us lesson nicely' → “po ma fo re”; (4) 'lesson like all humans' → “he re gu si”.

2. Sentences (1) and (2) share only the word 'all'; their common code is 'he', so all → he.

3. Sentences (1) and (3) share only the word 'nicely'; their common code is 'ma', so nicely → ma.

4. Sentences (2) and (4) share TWO words — 'all' and 'like'. Their common codes are 'he' and 'si'; since all → he is already known from the previous step, the remaining common code 'si' must correspond to 'like', so like → si.

5. Sentences (2) and (3) share only the word 'us'; their common code is 'fo', so us → fo.

6. Sentences (3) and (4) share the word 'lesson'. Removing the already-identified codes ('ma', 'fo' from sentence 3; 'he', 'si' from sentence 4) leaves 're' as the only code common to both, so lesson → re.

Assigning the remaining codes confirms the mapping is fully consistent: sentence (1)'s leftover codes 'ka, cu' cover 'speak, to'; sentence (2)'s leftover code 'to' covers 'are'; sentence (3)'s leftover code 'po' covers 'teach'; sentence (4)'s leftover code 'gu' covers 'humans'. Every code across all 4 sentences is now accounted for with no clashes, confirming 'lesson' corresponds to 're'.

Therefore, the code for 'lesson' is 're'.