ja: read exponents in Japanese

[yasumorishima]

Second of the small PRs from #715, and independent of #720 (different rules).

How Japanese reads an exponent

山口雄仁・川根深・澤崎陽彦「日本語による数式読み上げ法の基本構成について」日本数学教育学会誌 78(9), 239–247 (1996), item (6)

• x² is 「x の 2 乗」 — base, の, exponent, 乗. 乗 is the closing marker, and the reference says it is spoken as a matter of course for exponents.
• There is no ordinal form: 2 stays 2. Japanese does not say "the second power" the way English does, so ToOrdinal has nothing to build here, and the English -th pronunciation hint does not apply either.
• When the exponent is complex, the reference switches to reading it as a superscript instead: 「の上付き … 上付き終了」.

That second tier solves the problem the English comments in these files worry about. Because 乗 already closes a simple exponent, a nested exponent would otherwise end 「… 乗 乗」 — the exact analogue of "...power power". Using 上付き for the nested case avoids it.

Changes

was	now
「スクエア」 (katakana "square")	「の 2 乗」
「キューブド」 (katakana "cubed")	「の 3 乗」
「パワー」 after a simple exponent	「乗」
「パワーの」 ("power of")	reordered — see below
「に上げられた」 ("was raised")	「の上付き」
「エンド指数」	「上付き終了」
「緊急事態に育つ」	「の上付き」
「力に育つ」	「の」 (with 「乗」 added to close)

The last two are worth spelling out: raised to the exponent had become 緊急事態に育つ, "to grow in an emergency", and raised to the power had become 力に育つ, "to grow into strength". Neither is about mathematics.

function-power needed reordering rather than a word swap. English puts the exponent first ("the fourth power of sine"); Japanese puts the base first — sin⁴ is 「サイン の 4 乗」 — so the rule now emits *[1], 「の」, *[2], 「乗」.

ToOrdinal and the pronounce: [text: "-th", ...] hint are removed from the power rules for the reason above. In ClearSpeak this also means $ClearSpeak_Exponents = 'Ordinal' no longer suppresses the closing 「乗」: that preference selects English's short ordinal form ("x squared" vs "x to the second power"), and Japanese has only the one form, so dropping 乗 there would leave 「x の 5」, which does not say that 5 is an exponent.

Tests

squared updated, and now checks SimpleSpeak as well as ClearSpeak. Added cubed, integer_exponent (x⁵), and variable_exponent (xⁿ), each in both styles, plus complex_exponent (x^(y+1) → 「x の上付き y プラス 1 上付き終了」) so the superscript frame is pinned too.

Not covered here: negative exponents. x⁻² currently comes out as 「x の 負の 2 乗」, and 「マイナス 2 乗」 is what the reference's worked example implies (it reads −b as 「マイナス b」). But 負の/正の is the prefix reading of negative numbers generally, and 正の is correct where it appears in general.yaml for "the set of positive integers", so that is a separate change rather than something to slip in here.