Support for grammatical order inversion.

[baha-bouali]

In languages like Arabic, grammatical rules often necessitate that the words' order in a sentence is inverted. This inversion is not a matter of style or reducing awkwardness but a fundamental grammatical requirement that alters the entire structure of the sentence.

As a native Arabic speaker, I can attest that Arabic is rich with cases where words' order in the sentence is changed due to grammatical factors, such as Definiteness.

In Arabic, the standard sentence structure places the subject before the predicate. However, when the subject is indefinite, the order is obligatorily inverted to maintain grammatical correctness. For example:

• Standard order (e.g., definite subject): (على الطاولة2️⃣) (الكتاب1️⃣) -> "(The book1️⃣) (is on the table2️⃣)".

• Standard order (e.g., Subject Pronoun) : (سعيدة2️⃣) (هي1️⃣) -> (She 1️⃣) (is happy2️⃣)

• Inverted order (indefinite subject): (كتاب1️⃣) (على الطاولة2️⃣) ->"(On the table 2️⃣) (a book 1️⃣)".

Here is a visual explanation to further illustrate this concept (note: Arabic is read from right to left)

This issue here is not about formatting a word or a segment in a sentence but rather about a grammatical property that changes the entire shape of the sentence.

I think this is similar to #14 regarding adjective order in English. I’m sure that a similar approach could be applied to handle order inversion in such cases.

I am currently working on this issue #87, and understanding how Unicode handles these cases.

I'm happy to provide additional examples and collaborate on this issue.
Thank you @grhoten for your time and consideration.

[grhoten]

Thanks for your interest. Unfortunately, this is a non-goal within this project. It is not capable of rewording sentences like that. It's too open ended.

For the 3 previous frameworks mentioned in the videos mentioned in the readme here is how each syntax would handle such a scenario.

In the first video about Automatic Grammar Agreement, the syntax makes it impossible to reword sentences like that. The syntax is easier for translators to handle at the expense of flexibility. If there is an unknown word, it can use an inflection alternative, which is the only time that an implied conditional is supported. For example, if "your" is gendered in a language, and you don't know the gender of the audience, you could have a sentence like, "Here are your messages" have the following alternate message provided instead, "Here are the messages that were requested". The alternate message is using a less desirable passive voice, but it's also grammatically correct.

In the second and third videos with the Siri syntax, you can detect if the object being referenced is indefinite, definite or unknown. Then you can switch between the 2-3 templates that you have, like an if statement. This puts a burden on translators to understand how conditionals work, but it's possible to make such sentences work. Translators are typically not programmers. So simplifying the syntax for common use cases is preferred.

As I recall, MF2 does have conditional formatting with the equivalent of switch statements. Conditionless formatting is easier for translators, but the scenario that you provided should probably be best handled as a conditional.

When there is reordering of words, such scenarios are usually handled in a limited way. For example, quantities in Hebrew do change the order of one, and other, and it will omit the number for two. Arabic will omit the number if it's one or two. So there is some minimal support for changing the order of words, or omitting words, but this library should not try to directly handle the open ended concept that was provided. Other frameworks should handle sentence rephrasing.

[baha-bouali]

Thank you @grhoten for your detailed response. I’ve been trying to figure out this issue since I started working on the project, and your response helped clarify a lot of things.
I completely agree that conditionals are the best way to handle such cases.
After experimenting with MF2, I implemented a solution using conditionals. Since MF2 doesn’t infer definiteness automatically, the only way to manage it is by passing multiple parameters and implementing Multiple selectors as shown in this example .

While this approach works, it becomes increasingly complex as the sentence structure evolves.

Interestingly, I found the process to be more straightforward with Apple Foundation largely due to its enhanced functionalities, particularly its ability to automatically detect definiteness.

Thank you once again for your valuable insights. This will help me focus my efforts on the relevant parts of the project moving forward.