In Read Aloud of Japanese text, -的 is sometimes not properly pronounced as -TEKI, but as -MATO. For example, in ノエマは対象と本質的に異ならない: 超越論的還元が世界を排除するものとして誤解されると the first 的 is properly pronounced as TEKI in 本質的 (hon sitsu TEKI), but wrongly as MATO in 超越論的還元 (cho etsu ron MATO gengan). This happens in most of the -的 words, where the root noun such as 超越論 is modified with -的 to make an adjective, and the entire formulation, such as 超越論的, is not in the dictionary.  In this case, it appear that the first 本質的 is in the word list dictionary, whereas 超越論的 is treated as a sequence of nouns and 的 is not recognized as a suffix (TEKI), but a separate noun (MATO).

As something-的 construct is quite frequent in Japanese text, this severely impacts the read-alound quality in negative way.  Can someone take a look at this issue?

BTW, ChatGPT does not make any mistake in the pronunciation of something-的 at all.

​In Read Aloud of Japanese text, -的 is sometimes not properly pronounced as -TEKI, but as -MATO. For example, in ノエマは対象と本質的に異ならない: 超越論的還元が世界を排除するものとして誤解されると the first 的 is properly pronounced as TEKI in 本質的 (hon sitsu TEKI), but wrongly as MATO in 超越論的還元 (cho etsu ron MATO gengan). This happens in most of the -的 words, where the root noun such as 超越論 is modified with -的 to make an adjective, and the entire formulation, such as 超越論的, is not in the dictionary.  In this case, it appear that the first 本質的 is in the word list dictionary, whereas 超越論的 is treated as a sequence of nouns and 的 is not recognized as a suffix (TEKI), but a separate noun (MATO).                                                                                     As something-的 construct is quite frequent in Japanese text, this severely impacts the read-alound quality in negative way.  Can someone take a look at this issue? BTW, ChatGPT does not make any mistake in the pronunciation of something-的 at all.  Read More

​