September 26, 2004

Learn your grammar, Becky

A large amount of work went into the preparation of the recent spam message from "Becky Miranda" about tentative scheduling of a meeting that was sent to a random UCSC address and blind-copied to me (and doubtless hundreds of others). The body did nothing but display an icon which would take the viewer to a web site if clicked on. A significant amount of random Angloid text with English-type letter transition frequencies ("align fatbikini esquire granularhemorrhage applicable augerdominic chalet aggressivebarbudo wherefore verbsomewhat germane israelballroom toefl refrainnoetherian committal typewritethickish...") had been added to try and defeat spam detection algorithms which look for an excess of HTML over plain text. And work had been done to forward it through a trail of relay machines. But the imaginary Becky let herself down with the Subject line:

From: "Becky Miranda" <>
To: <>
Subject: tentative meeting on the 2th
Date: Fri, 24 Sep 2004 01:59:20 -0300

The suffix for numerically abbreviated ordinal numerals isn't always th in English, Becky. It's st those that end in 1 but not in 11; it's nd for those that end in 2 but not in 12; it's rd for those that end in 3 but not in 13; and otherwise it's th. (I have to admit to you that on page 1718 of The Cambridge Grammar of the English Language this is only implicit; it's carefully described for the spelled-out words, but not for the numerical abbreviations.) That little detail of the lexical structure of English number names (that we don't have a "2th" of any month) gave you away, and would have revealed you as a foreign spammer even if the incongruity of someone at a Lithuanian address inviting me to a meeting had not. You see how important grammar is?

Posted by Geoffrey K. Pullum at September 26, 2004 03:30 PM