review: the referee DOES hold a word-final ⟨c⟩ — my reason was wrong, the conclusion was not

I wrote that af.wiktionary-af.tsv contains no word-final ⟨c⟩, and that its C→sɪə entry reaches output
through the initialism path. Both false. The entry IS a word-final ⟨c⟩, eval.ts feeds af through
phonemizeWord directly, and the initialism normalizer only fires on runs of two or more letters — so
that entry traverses this exact branch, going s→k with this change.

The conclusion survives: it misses both ways against sɪə, which is why the folded backbone is unmoved
at 1658/2220. But the stated reason would have told the next reader that nothing exercises the branch,
which is the opposite of true. My original grep was case-sensitive and lowercase-only; I then saw the C
entry in later output and explained it away with an assumption I never checked.

Filed the underlying gap as #761: a bare single letter is read as a sound rather than spelled as its
name (B→p too), because the initialism normalizer needs ≥2 letters. Also moved the new test into the
g2p describe block, where it belongs.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_012Lc3WnUgogC7okV7n53vjr

Author: christopherthompson81

src/languages/afrikaans/afrikaans.ts

@@ -87,8 +87,9 @@ function phonemizeMorpheme(word: string): string {
             // ⟨c⟩ with nothing after it fell into the soft branch. franc→frans, arc→ars. #756 preserved
             // that verbatim (a data move must not change output) and #757 decided it: the rule's own
             // statement is "before a front vowel", and word-final has no following vowel at all, so the
-            // accident inverted the rule rather than extending it. The af referee has no word-final ⟨c⟩,
-            // so this rests on the rule being self-consistent, not on a measurement.
+            // accident inverted the rule rather than extending it.
+            // ⚠ NOT REFEREE-DECIDED: the corpus's only word-final ⟨c⟩ is the letter name C→sɪə, which this
+            // branch does reach and misses BOTH ways, so the score is unmoved either way (#761).
             out += C_SOFT.has(w[i + 1] ?? "") ? "s" : "k";
             i += 1;
             continue;

test/afrikaans.test.ts

@@ -46,6 +46,21 @@ describe("Afrikaans canonical IPA — greedy g2p + open/closed vowel length (Sta
     test("text: words + clause punctuation (stress deferred)", () => {
         expect(createAfrikaans().text("Die man loop huis toe.")).toBe("di man luəp ɦœys tu .");
     });
+    // ⚠ WORD-FINAL ⟨c⟩ IS [k] — issue #757. The rule is "soft [s] BEFORE a front vowel", and word-finally
+    // there is no following vowel, so the soft branch cannot apply. It used to, because the test was
+    // `"eiyêéè".includes(w[i + 1] ?? "")` and `includes("")` is TRUE: franc→frans, arc→ars.
+    // ⚠ THE REFEREE CANNOT ADJUDICATE THIS, though not for the reason first written here. af.wiktionary-af
+    // .tsv DOES hold one word-final ⟨c⟩ — the letter name `C → sɪə` — and eval.ts feeds af through
+    // phonemizeWord directly, so that entry does reach this branch. It simply misses BOTH ways (s, then k,
+    // against sɪə), which is why the folded backbone is unmoved at 1658/2220. Single letters are not
+    // spelled out by the g2p at all — `B → p` likewise — see issue #761.
+    test("a word-final ⟨c⟩ is [k], not the soft [s]", () => {
+        expect(phonemizeWord("franc")).toBe("frank");
+        expect(phonemizeWord("arc")).toBe("ark");
+        expect(phonemizeWord("bloc")).toBe("blɔk");
+        expect(phonemizeWord("cent")).toBe("sɛnt"); // …and ⟨c⟩ BEFORE a front vowel is still soft
+    });
+
 });
 
 // TEXT NORMALIZATION (src/languages/afrikaans/normalize.ts) — the pre-tokenizer pass. The
@@ -194,16 +209,5 @@ describe("Afrikaans text normalization", () => {
         expect(MANIFEST.morphology.prefixUnstressed.length).toBeGreaterThan(0); // and the scan found something
     });
 
-    // ⚠ WORD-FINAL ⟨c⟩ IS [k] — issue #757. The rule is "soft [s] BEFORE a front vowel", and word-finally
-    // there is no following vowel, so the soft branch cannot apply. It used to, because the test was
-    // `"eiyêéè".includes(w[i + 1] ?? "")` and `includes("")` is TRUE: franc→frans, arc→ars.
-    // ⚠ NOT REFEREE-CONFIRMED — af.wiktionary-af.tsv contains no word-final ⟨c⟩ at all (its only ⟨c⟩
-    // entries are surnames and the letter name C→sɪə). These pin the rule being self-consistent.
-    test("a word-final ⟨c⟩ is [k], not the soft [s]", () => {
-        expect(phonemizeWord("franc")).toBe("frank");
-        expect(phonemizeWord("arc")).toBe("ark");
-        expect(phonemizeWord("bloc")).toBe("blɔk");
-        expect(phonemizeWord("cent")).toBe("sɛnt"); // …and ⟨c⟩ BEFORE a front vowel is still soft
-    });
 
 });