Most students of Arabic are at one point introduced to and annoyed by what is known as deflected and strict agreement. This means that in VSO-type sentences, the verb only agrees with the subject in gender (1&2); in SVO-type sentences, the verb agrees with the subject in both number and gender (3&4):
1. ʾaman-a l-muʾminūn
believed-M.SG. the-believers, “The believers believed”2. kataba-t al-bināt
wrote.-f.sg. the-girls.F.PL, “The girls wrote” .3. ʾinna l-muʾminīna ʾaman-ū
indeed the.believers.PL believed-PL.M, “Indeed, the believers believed”4. ʾinna l-bināt katab-na
indeed the-girls wrote-PL.F. “Indeed the girls wrote”.
What’s more, in Classical Arabic there is a pretty strict distinction between animate and inanimate nouns. When something needs to agree with a noun (essentially verbs and adjectives), they can and sometimes have to (see above) take agreement in both number and gender. When a noun is inanimate, in the plural it only takes feminine agreement. The following phrase from al-Masʿūdī serves as a good example:
fa-lamma qutila yaḥyā ǧazaʿa-t ʿalayhi nufūs al-nās
and-when killed Yaḥyā gathered-F. to-him souls the-people
“and when Yaḥyā was killed the souls of the people gathered around him”
Although this is one of the most typical features of Arabic, it’s interestingly not present in pre-Classical Arabic. In the Qurʾān, for example, inanimate plurals take feminine plural agreement. Consider the following verse (Q2:184)
ʾayyāman maʿdūd-ātin fa-man kāna min-kum marīḍān
days limited-f.pl and who amongst you is sick […] “For a limited number of days. Those amongst you who are sick […]”
5. li-layla-tin ḫala-t mina l-rağab
night.SG. pass-F.SG. from Rajab, “One night having passed of Rajab”6. li-layla-tayn ḫala-tā mina l-rağab
to-night-DL. pass-F.DL from Rajab, “Two nights passed of Rajab”7. li-ṯalāṯin layālin ḫalaw-na mina l-rağab
three night.F.PL pass-F.PL from Rajab, “Three nights passed of Rajab’
So I checked the early Islamic sources to see if this system is maintained everywhere, and it turns out that yes, this is pretty much as it is everywhere. For example, in the tenth part of Ibn Kathīr’s Kitāb al-Bidāya wa-l-Nihāya, we find the following line:
tawallā l-muʾminūna l-ḫilafata fī l-muḥarram li-ḫamsin baqī-na min-hū
entrust the-believers the-caliphate in Muḥarram five remain.F.PL of-it
‘The believers were entrusted with the Caliphate on the 25th of Muḥarram’
And similarly, in the same environment:
yawm al-ʾarbāʿ li-ṯamān baqī-na min hāḏihī l-sina
‘Wednesday’ eight remain.F.PL of this the-year
‘On Wednesday, with eight [days] left in the year’
What’s interesting about this is that Ibn Kathir lived in the 14th century, centuries after the standardizations of Classical Arabic. For some reason, this old system of counting days and nights was maintained, with full agreement of subject and verb up until the re-standardization of Arabic during the Nahda.
al-Sarxāsī 