The trouble with "natural language" is that far too many people do not speak Natural.
I mean, I would argue that the trouble with "natural language" is that the way people actually use language naturally, very little (if any) concern is paid to "do my meanings match up consistently with others?" That's why we have things like dictionaries, and style guides, and writing classes, and all sorts of other things. Because humans are naturally quite bad at consistency and rigor.
Natural language
gets by because we don't really need consistency or rigor in everyday speech. We don't need precise terms for what "far" means, or whether "a couple" is precisely two (which is what the narrow definition requires) or "a small number" (usually between 2 and 4), or for narrowly specifying particular hues/shades/tints of "blue". (Heck, in Japanese, "aoi" means both blue
and green, simultaneously, just like how "blue" in English covers both sky blue and navy blue, but other languages have specific words; or how English separates "brown" from "orange" even though most browns are just dark orange.)
So, most people "speak Natural", but "Natural" isn't any specific thing. It's a huge nebulous
cloud of things, and two people
might be in more or less the same part of that cloud. Or they might be on opposite sides, where one sees the sun above, and the other sees the earth below. The more you bake that nebulous-cloud-thing into something like a set of rules, the more it becomes dependent on the participants massively over-explaining themselves to ensure that people get pushed to the correct end of the cloud. But that massive over-explaining process itself is at least as unnatural as jargon, and often moreso--and also a lot more effort to boot!
Hence why I said above: it's like the writers realized halfway through that jargon exists in nearly every part of human life because it's
useful. It's not just professional disciplines that develop it, though those tend to be the most robust and considered. Games develop jargon, it's part of what makes most sports nearly impenetrable to outsiders; I still to this day don't
fully understand the rules of gridiron football, but my parents made a valiant effort to teach me the basics. Cooks develop jargon; for the life of me I cannot understand why we call the
rippling surface of a pan of oil "shimmering" to indicate that it's at proper temperature, since a "shimmer" to me means like...being more reflective, but it really just means that the surface isn't smooth, it's subtly wavy and shifting. I could go on, but I doubt anyone cares. Point being, jargon forms because it's useful. Pretending you can just completely axe jargon and rely totally on casual terms was, fundamentally, a mistake.
Now, I know
why they made that mistake, and it wasn't a bad notion. Between 3e and 4e, folks had felt fatigued with what they saw as
excessive jargon, as fiddly nonsense that, rather than enhancing the game, was holding it back. All the hype about "natural language", all the ways it was pitched, all the ways 5e fans sang its praises all the live-long day (can you tell I found this period tedious?), was of the mold of "the rules will get out of the way". Not needing to learn the lingo is freeing! But if the consequence is that you end up having debate after debate about what things "really mean" or why XYZ statement needs to only be read with natural meaning Q when it also could be read as natural meaning P...it can be hard not to see it as being "freed" from traffic laws and thus seeing a strange rise in deaths caused by vehicle collisions.
Sometimes, eliminating rules means the rules can get out of the way. And sometimes, eliminating rules throws everything into chaos. Knowing which applies to any given game-design situation is damned hard, an extremely difficult game design challenge, and thus not at all something the "D&D Next" playtest was willing, or even really
able, to tackle.
An actually middle-of-the-road path--where you restrict jargon only to where it's needed, and otherwise avoid it--is much more productive in the long run. And it seems pretty clear to me that 5.5e is trying to address it...without truly
changing it...which is very much at risk of "worst of both worlds".
This goes for the term "line of sight" which does not mean the same as "can see", in natural language
Your problem is, you are asserting that something has a
singular, agreed meaning in natural language. It doesn't. It means both. Sometimes it means both to the same person at different times. Sometimes it means both at the same time to different people. And that's
precisely why dependence on natural language is liable to get confusing: two people can both "speak Natural" and yet be saying
different things.
Given the primary purpose of rules is, in general, to make clear what should or should not happen, having a system
built on this might be a hindrance some of the time!