In order to address the question, I suppose we have to temporarily suspend disbelief and "admit" some initial (highly debatable) assumptions. Namely that:
stage hypnotism is a scientifically-backed practice and...
people who are hypnotised are recollecting actual past memories and/or exhibiting behaviours tied to actual past events, not acting out randomly.
With these two fairly strong (euphemism) assumptions, your question then becomes about the existence of fetal memory. Studies suggest foetuses do have short-term and possibly long-term memory functions (in agreement with what is known of early brain development).
However, there is no scientific evidence that this memory extends beyond an order of days (and studies on infants and young children suggest they only slowly reach an order of years, in the years after birth). On the contrary, studies suggest that memory of particular events starts at a much later age of development.
speaking a language he never knew.
The term for this is glossalia. It isn't really language, but just gibberish. There are several scientific explanations for it that do not require pre-natal memory.
Saṃsāra, as defined by Wikipedia, is a theological concept, not a scientific hypothesis. As a model, it doesn't make any testable predictions, and is unfalsifiable. Therefore, it doesn't count as an "explanation" in the context of skeptical investigation.
For our purpose, the genetic make-up of an individual is solely dependent on its parents (and on a very limited range of conditions during a comparatively negligible interval of time around fertilisation) and has absolutely nothing to do with potential pre-natal conditions, let alone some hypothetical transfer of information.
In conclusion, strictly speaking, "pre-birth hypnotism" is neuro-biologically plausible, though unbacked by current scientific studies, if you overlook the questionability of the two essential assumptions above.