|>> || |
It should be noted that there was some trade-related borrowings between Akkadian, an old cousin of Arabic, and the PIE dialects. Notably, *septḿ is better reconstructed as something like *septh3óm when including early splits Tocharian and Luwian (whence Hittite); either form violates PIE phonotactics (two stops beside each other in a stem with a *s at the other end of the stem). A curiously close form akin to *sebohetem (I lost my source) existed at the time in Akkadian, related to Arabic sab?a (where ? is a glottal stop). Some people even advocate Akkadian's inferred "laryngeal collapse" (gathered from inconsistent spellings later on; KEs becoming KUs or QUs and what not) triggered an areal wave of mutation (like Chinese turning Thai, Lao, Hmong, Vietic, and Tibetan tonal) reducing laryngeals in Aeolian languages, neighboring semitic languages and language isolates, Caucasian Languages, and PIE itself.
The process is comparable to Uralic, exemplified by Finnish: yksi, kaksi, kolme, neljä, viisi, kuusi, and then seitsemän (reasonable as another borrowed seven, through PG *septum > **setpum > PFE *settym > PF *seitsem (with the an analogical to following forms), kahdeksan (still today transparently two-not-*deksan, transparently IE *dekm) yhdeksän (one-not-deksan), kymmenen; linguists can't reconstruct 7, 8, or 9 for Finno-Ugric, though only 2 and 5 seem to be cognate including Samoyedic. (they all could count beyond 6, it was just in base six, so 7 was one and six until seitsemän).
So they were in contact, though certainly not related.