We recently had an issue with handling Hebrew statuses that were retrieved via Twitter’s standard search API.
According to your documentation, the language filter is based on the ISO-639-1 standard. For Hebrew, the standard code is ‘he’, which replaced the old code, ‘iw’, in 1989. Accordingly, if you search for ‘Hanukkah lang:he’ you get the same results as when you search for ‘Hanukkah lang:iw’.
Yet, in the API, the retrieved statuses contain the old code (‘iw’), rather the new one, which causes mismatches when you wish to match the language from the retrieved status with the language in common ISO-639-1 tables.
Basically, we can treat the issue separately, but I wondered if by any chance it’s possible to update the API to work with the common standard.