UTF-8 was deliberately designed to be "backwards compatible" (for want of a better phrase) with ASCII.
Absent a BOM mark, there is no way to tell the difference between a piece of ASCII text on the one hand and, on the other, a piece of UTF-8 text that uses no code point above U+007f. How do they differ semantically?
Say you read the text into a program. You can load the text into an AsciiString variable or a UnicodeString variable, and it's valid in either of those cases. They have the same binary representation...but they're semantically different. They each have properties that the other one doesn't (ASCII has random access, Unicode is linear access, either one could have an ASCII string appended to it, but only the Unicode one could have a Unicode string appended, etc).
The difference isn't the representation; the difference is the interpretation that we attach to the representation.
When you deal with ASCII, that's also the semantic level on which you operate. Much simpler than Unicode (I don't think there are many people who really understand Unicode).
When you have a Unicode string that uses only characters that are also present in ASCII, that doesn't make it an ASCII string. It has more meaning.
Btw that has nothing to do with BOM. BOM is itself not part of ASCII, and not encoded as a single byte, and it doesn't affect single-byte sequences.
(I guess technically you can't tell the difference between an ASCII string and a string of random 8-bit numbers in the range 0 <= x <= 127, but if when represented as ASCII characters those numbers are meaningful -- comprehensible English, for example -- the odds are good it's the former not the latter.)
Or are you maybe talking at a software level, where you can make some simplifications and assumptions about a string when it is guaranteed to be ASCII?
Because otherwise I can't see how there could be any semantic or meaningful difference between ASCII and Unicode when the text is the same. An ASCII 'a' has exactly the same meaning as a Unicode 'Latin Small Letter A'.
Btw. I have no clue why I am constantly downvoted. Whoever did this, I don't think there is an incentive to downvote my two ancestor comments. It's demotivating.
In what SPECIFIC way is an only-ascii string not a UTF-8 string for ALL intends and purposes?
You throw some words like "semantics" but no actual arguments.