Thread by @PhDniX

Most modern Arabic fonts treats the right leg of the lām-ʾalif as a lām, and the left one as the ʾalif. This can be seen in al-ʾarḍ where the hamzah is placed on the left leg. But this was not always the case!
A thread on the history of the lām-ʾalif ->

The earliest vocalised Quranic manuscripts didn't have a dedicated sign for hamzah, instead they simply placed a red dot on top of the ʾalif for /ʾa/, below for /ʾi/ and to the left on the baseline for /ʾu/, e.g. ʾawlā, ʾilāhin and ʾulāʾika in the Quran of Amajur (262 AH/876CE)

In the case of a lām-ʾalif, it is the right stroke that receives the red dot of the /ʾa/, and /ʾi/ and /ʾu/ (which in the latter 2 cases crosses left at the bottom); This clearly suggests the right leg was considered the ʾalif, e.g. al-ʾanǧīlu (green) al-ʾingīlu (red)

This understanding of the lām-ʾalif remains remains in use for a long time. Compare for example the Ibn Shādhān Quran (361AH/972CE) for li-ʾādama where the left-leaning leg receive the karah, and the right-leaning leg the green dagger ʾalif to mark the /ʾā/.

The same interpretation remains current, even in later (approximately 12th century) Maghrebi Qurans, where defective /lā/ sequences are written with a lām with a red oblique ʾalif, imitating the shape of the true lām ʾalif, e.g. ʾulāʾika; salāmun; wa-lākinna

A memory of this practice is found even today in the Maghrebi Warš print Qurans, e.g. ʾulāʾika; aḍ-ḍalālata; wa-lākin with a tiny (black) miniature ʾalif crossing through the lām.

(Notice also just how closely the print Warš Qurans immitate Maghrebi Qurans in vocalisation)

So the right-leaning leg is the ʾalif in early Arabic writing. But why? For this we need to hop back to Pre-Islamic times. In the Nabataean Aramaic script -- the ancestor of the Arabic script -- the ʾalif looked somewhat like a 6; See, for example šmwʾl /samawʾal/

Already in pre-Islamic times, in the transitional Nabataeo Arabic script the ʾalif loses its loop, check out, for example the inscription UJadh 109 dated 455-6 CE which has ʾal-malik "the king", with a still fairly oblique ʾalif but otherwise a recognisable shape.

The lām-ʾalif however came into being *before* the loop of the ʾalif opened up, and the shape was simply followed from lām (a simple straight line) into the 6 shape of the ʾalif. See the inscription UJadh 367 which reads ʿbd ʾlʾḥšn = عبدالاحسن (note the straight ʾalif).

So where does the reinterpretation of the left leg as the alif come from? Perhaps it developed when in the Ruqʿah and Nastalīq styles formed a new shape for the lām-ʾalif which was more specifically a clear lām shape with an ʾalif resting on top.

Perhaps I am reversing causality? Perhaps the reinterpretation gave rise to possibility of the lām-ʾalif to take on the shape as we find it in this style. Already in the Naskh style the connected lām-ʾalif has taken on such a shape.

I am very happy to hear your thoughts! <end>

REFS:
The lām-ʾalif coming from the ancient ʾalif shape by @safaitic
academia.edu/35239575/Al-Ja…

Also shows the first lām-ʾalif is already found in the Namarah Inscription (328CE):
islamic-awareness.org/history/islam/…

Development of Nabataean script by Laïla Nehmé: academia.edu/2106858/_A_gli…

And I completely forgot to add: Also in the early 'Yazīd the King' inscription there is a "proto-hamza" above the right leg of the lām-ʾalif of the word al-ʾilāh, clearly marking the hamzah.