In b. Sanh. 4b, there reads:
Has it not been taught: ‘letotafoth [frontlets] occurs…in the Torah…four in all, to indicate [that four sections are to be inserted in the phylacteries]…R. Akiba maintains…the word totafoth itself implies four, [it being composed of] tot which means two in Katpi and foth which means two in Afriki (טט בכתפי שתים פת באפריקי שתים).
There’s been some discussion on Classics-L as to what validity, if any, there is to this. The most recent post suggests “This is an example of a d:rash daHuq (proof by play on words, which, in my opinion, should not be taken literally or very seriously).”
I somewhat agree that this “should not be taken literally or very seriously.” I find כתפי curious, though. It has been suggested that this means ‘Coptic’. There is, of course, the כפתרים of Gen 10.13 – descendants of Mizraim. If we were to accept that this is what was being referred to, we would have to assume both metathesis and loss of resh, though (however, ‘Caphtor’ appears as Keftiu elsewhere). Also, the Targums often have, for Caphtor, an assimilation to Cappadocia: קפוטקיא.
There really is nothing like טט meaning ‘two’, however, in Coptic. I guess the closest thing would be tōt, ‘to be joined’ – but this isn’t really close at all. (There is also tou, ‘five’ in Coptic; but this has no relevance.) In Ugaritic, ṯiṯṯ is the numeral for six; and in Qatabanian, ‘one’ is ṭd. [Late addendum: apparently there’s a Fula(ni) word for two, didi.]
In regard to פת: there’s really nothing like this for ‘two’ in the אפריקי language(s) – although there is ftoou for ‘four’ in Coptic (Černý, 266 – see Crum, 625). This is seen also in several branches of Chadic languages.