Here is additional info that I believe rather conclusively indicates that the code requires cable seals for multiconductor cables with a gas/vaportight continous sheath capable of transmitting vapors through the core at any div. 2 area termination (not just at explosionproof enclosures). After further scrutinizing the rest of 501.15(E), I noticed that in 501.15(E)(2) the code is saying that cables with a gas/vaportight continuous sheath and do not transmit gases or vapors through the core are not required to be sealed, except at a required explosionproof enclosure. However, the cable length has to be great enough for the cable to limit the gas or vapor flow through the core to the same flow rate permitted for seal fittings. In other words, the cable has to perform the same function as a seal in terms of allowing passage of gases or vapors.
Then, if you look at 501.15(E)(3) Cables Capable of Transmitting Gases or Vapors, which is the same category as our multiconductor, it states that these cables have to comply with 501.15(E)(1) Terminations but otherwise do not require a seal except if connected to process equipment or devices that may cause a pressure in excess of 6 in. of water (the seal performance limit). (This condition is addressed in 501.17 Process Sealing.)
So, if you take the stance that 501.15(E)(1) is saying cable seals are only required at explosionproof enclosures and the sentence about multiconductors being sealed in the div. 2 area is only telling how to seal at the explosionproof enclosure, hence no seal is required at any other termination, then apparently the code is saying:
1) A cable that cannot transmit gases through the core has to perform the same function as a seal to limit the passage of gases or vapors;
2) A cable that can transmit gases through the core, nope, no seal required, we don't really care if gases or vapors are transmitted from a Div. 2 area to an unclassified one.
As Spock would likely say, this seems highly illogical.
My only logical conclusion is that 501.15(E)(1) is explaining the seal requirement at terminations for two different conditions: 1) Explosionproof Enclosures, and 2) All terminations of a multiconductor with gas/vaportight continous sheath capable of transmitting vapors through the core.
However, I do find it puzzling that not in any of the previously mentioned code sections does it mention the condition of the other end of the cable terminating in an unclassified area. Unless this condition is present, I would assume there would be no requirement for any sealing (except to complete explosionproof integrity of an enclosure, of course). Hopefully, the wording will be improved in future code cycles to at least minimize confusion. If my interpretation is correct, I am pretty sure that there are many cables not sealed that should be.