Interesting! Thanks for the extra context. To your original question, no, as of today we don't have anything that would skip updates on certain fields while applying them to others. There are probably creative ways to accomplish this, but the way I've personally done something similar in the past was to change the load to what I call "append only" so that prior versions of records are never replaced, and the target table is only appended with new records.
This approach requires that you dedupe in a tool like dbt, calculating the latest version of a field when you want the latest data and the earliest/original version of a field from the first version of the record.
Does this sound like it could be a fit for your use case?