What¹s this?

[pne]

Occasionally, I'll come across text which uses a superscript 1 as an apostrophe; these also tend to use superscript 3 and 2 as opening and closing quotation marks, respectively. Here's an example message in this encoding which I found by googling "can¹t"; it also features the use of "‹" as a dash of some kind (invisible on the page itself, where it's rendered as character U+008B (the control character "PARTIAL LINE FORWARD"), but visible in the Google "snippet" on the SERP as "‹" (i.e., Google interpreted the byte 8B as Windows-1252, while the page shows it as Latin-1).

Does anyone have any idea where this particular variety of mojibake comes from?

I think that such texts come from Macs, but I'm not sure how; my first instinct was to have a look at the MacRoman charset, but there, 0xB9 is not an apostrophe but rather the Greek letter π, and 0xB3 and 0xB2 are not smart quotes but rather the mathematical symbols ≥ and ≤. (And 0x8B is ã, rather than a dash.)

Anyone else?

Comments

[sophie]

I came across the superscript-1-as-apostrophe too sometimes, when I was working at TNAUK, where my job required me to know quite a lot about character sets. However, it didn't do the superscript 2 and 3 as quotation marks.

My conclusion, after checking every character set my installation of Perl had access to, was that it was some sort of bad OCR combined with bad editing, as it didn't at all fit with any character set, and I already knew the character set it was *normally* done in (and which the rest of the document was).

I've also come across text where the capital ligature AE (Æ) was used as an apostrophe. From what I remember, this was a problem caused by double encoding, though I can't remember now.