Shivasar BeTammuz -- A Mistake?

I grew up hearing it referred to as Shivasar BeTammuz, rather that the more correct "Shiv'a Asar BeTammuz."

Indeed, a Google search for Shivasar B'Tammuz yields only three hits. Meanwhile, a Google search for Shiva Asar B'Tammuz yields 568 hits. It is possible that what I thought to be the common term is actually not that common. Or it could be that when writing web pages about that day, people tend to first look it up and make sure they have the "proper," grammatical spelling. I would lean towards the latter.

I would guess that phonologically speaking, since we don't pronounce the ayin in Shiv'a even as a glottal stop so as to get a new syllable, we get Shiva Asar. (And the ayin in asar is relaxed as well.) Faced with an "a" ending on the first word and an "a" start on the second word, the tendency would be to collapse them together.

Is such practice "correct?" It depends if you let it be. We see a similar phenomenon in Aramaic, where the teens become שיתסר and שבסר. The latter is likely shevsar (with the chirik becoming a segol), just as we have shev in Shev Shmaysa. The ayins disappear, and it becomes one word.

Thus, perhaps in Yiddish and in Judeo-English, we have a similar disappearance of ayins and coalescence into a single word, though preserving the kametz. And such is perfectly "correct" is we define it so.